Citation
The ecological response of step-pool streams following the 2012 Waldo Canyon fire of Colorado, USA

Material Information

Title:
The ecological response of step-pool streams following the 2012 Waldo Canyon fire of Colorado, USA
Creator:
Solverson, Anna Parker ( author )
Language:
English
Physical Description:
1 electronic file (172 pages) : ;

Subjects

Subjects / Keywords:
Ecology -- Colorado ( lcsh )
Wildfires -- Colorado ( lcsh )
Environmental sciences -- Ecology ( lcsh )
Hydrology -- Ecology ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Abstract:
Step-pool mountain streams are increasingly subjected to a variety of natural and human-induced disturbances such as climate change, urban encroachment, and wildfires. In the western USA, wildfires are increasing in frequency and magnitude due to warming climates. Because step-pool sequences provide critical hydraulic and ecological functions in the river system, understanding the impact of wildfires on step-pool streams is crucial for effective management of aquatic resources. This study investigated the ecological response of step-pool streams to the 2012 Waldo Canyon Fire of Colorado, using benthic macroinvertebrates as indicators of changes in the ecological condition of stream channels. The analysis tracked changes in benthic macroinvertebrates immediately following the fire (2012), one year post-fire (2013), and two years post-fire (2014). Four categories of metrics represented characteristics of macroinvertebrate communities: overall richness and composition, community tolerance, functional feeding groups, and habit types. Non-parametric statistical analysis tested for differences in the responses of macroinvertebrate communities as a function of (1) presence of burn; (2) severity of burn; (3) habitat type (step or pool); and (4) time. Additionally, ordination allowed exploration of the ecological responses together with the changing physical characteristics of the step-pool streams. ( ,, )
Abstract:
This study produced the following main results. First, the ecological condition of channels burned by wildfire was significantly poorer than that of unburned channels. Second, channels burned with high severity generally reflected poorer condition compared to channels burned with low severity and unburned channels. Channels burned with low severity only differed from unburned channels in functional feeding groups. Third, pool habitats showed the negative effects of wildfire more than were step habitats. Fourth, although channels burned with high severity showed evidence of impact at each of the time steps examined, some channels exhibited signs of recovery within two years following the fire. Finally, the extent of ecological impact and recovery apparently correlates with the presence of the step-pool structure. Even when severely burned, channels able to retain the step-pool structure experienced minimal ecological impact and recovered more quickly.
Thesis:
Thesis (M.S.) - University of Colorado Denver
Bibliography:
Includes bibliographics references
System Details:
System requirements: Adobe Reader.
General Note:
Department of Geography and Environmental Sciences
Statement of Responsibility:
by Anna Parker Solverson.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
946536115 ( OCLC )
ocn946536115
Classification:
LD1193.L547 2015m S65 ( lcc )

Downloads

This item has the following downloads:


Full Text
THE ECOLOGICAL RESPONSE OF STEP-POOL STREAMS FOLLOWING THE 2012
WALDO CANYON FIRE OF COLORADO, USA
by
ANNA PARKER SOLVERSON
B.S., College of William & Mary, 2007
A thesis submitted to the
Faculty of the Graduate School of the
University of Colorado in partial fulfillment
of the requirements for the degree of
Master of Science
Environmental Sciences
2015


This thesis for the Master of Science degree by
Anna Parker Solverson
has been approved for the
Environmental Sciences Program
by
Anne Chin, Chair
Peter Anthamatten
Alison ODowd
November 29, 2015


Solverson, Anna Parker (M.S., Environmental Sciences)
The Ecological Response of Step-Pool Streams Following the 2012 Waldo Canyon Fire of
Colorado, USA
Thesis directed by Professor Anne Chin
ABSTRACT
Step-pool mountain streams are increasingly subjected to a variety of natural and
human-induced disturbances such as climate change, urban encroachment, and wildfires. In
the western USA, wildfires are increasing in frequency and magnitude due to warming
climates. Because step-pool sequences provide critical hydraulic and ecological functions in
the river system, understanding the impact of wildfires on step-pool streams is crucial for
effective management of aquatic resources. This study investigated the ecological response
of step-pool streams to the 2012 Waldo Canyon Fire of Colorado, using benthic
macroinvertebrates as indicators of changes in the ecological condition of stream channels.
The analysis tracked changes in benthic macroinvertebrates immediately following the fire
(2012), one year post-fire (2013), and two years post-fire (2014). Four categories of metrics
represented characteristics of macroinvertebrate communities: overall richness and
composition, community tolerance, functional feeding groups, and habit types. Non-
parametric statistical analysis tested for differences in the responses of macroinvertebrate
communities as a function of (1) presence of burn; (2) severity of burn; (3) habitat type (step
or pool); and (4) time. Additionally, ordination allowed exploration of the ecological
responses together with the changing physical characteristics of the step-pool streams.
This study produced the following main results. First, the ecological condition of
m


channels burned by wildfire was significantly poorer than that of unburned channels. Second,
channels burned with high severity generally reflected poorer condition compared to
channels burned with low severity and unburned channels. Channels burned with low
severity only differed from unbumed channels in functional feeding groups. Third, pool
habitats showed the negative effects of wildfire more than were step habitats. Fourth,
although channels burned with high severity showed evidence of impact at each of the time
steps examined, some channels exhibited signs of recovery within two years following the
fire. Finally, the extent of ecological impact and recovery apparently correlates with the
presence of the step-pool structure. Even when severely burned, channels able to retain the
step-pool structure experienced minimal ecological impact and recovered more quickly.
The form and content of this abstract are approved. I recommend its publication.
Approved: Anne Chin
IV


ACKNOWLEDGMENTS
The completion of this thesis would not have been possible without the support and
guidance of my advisor, Dr. Anne Chin. She committed boundless time and energy to my
development as a writer and researcher. Her dedication, ambition, and attention to detail have
been crucial in shaping my graduate career. I am also grateful to my committee members, Dr.
Alison ODowd and Dr. Peter Anthamatten, for their expertise and advice throughout the
research process, particular regarding the analytic methods used. I would also like to thank
Rhonda Barton, Lauren Tyner, Rachel Gidley, Alex Key, Sam Epperly, Corine Roberts-
Niemann, and Dan Ben-Horin for their assistance in the field and in the lab. Several sponsors
provided financial and in-kind support (to Dr. Chin) for this project: the National Science
Foundation (EAR 1254989), the University of Colorado Denver, U.S. Forest Service, U.S.
Geological Survey, and The Navigators. A monetary grant through the 2015 M. Gordon
Reds Wolman Graduate Student Research Award from the Geomorphology Specialty
Group of the Association of American Geographers also enabled me to complete the analysis
and preparation of this thesis document. I would like to thank my parents for instilling in me
a love of learning and the drive to pursue my ambitions. Finally, I am eternally grateful to my
wonderful husband, Keith, for his patience and encouragement through all of my highs and
lows over the last three years. Thank you for always being there for me.
v


TABLE OF CONTENTS
CHAPTER
I. INTRODUCTION AND OBJECTIVES.................................... 1
II. BACKGROUND.....................................................5
Ecology of Step-Pool Mountain Streams.....................5
Ecological Disturbances in Fluvial Systems................7
Wildfire as a Disturbance in Stream Ecosystems............10
Research Questions and Hypotheses.........................15
III. STUDY AREA.....................................................17
Environmental Characteristics.............................17
Waldo Canyon Fire of 2012.................................20
Study Reaches.............................................22
IV. METHODS........................................................28
Data Collection: Field....................................28
Laboratory Methods........................................ 32
Data Analysis: Statistical and Computational..............33
V. RESULTS: RESPONSE OF BENTHIC MACROINVERTEBRATES................44
Description of Data.......................................44
Research Question (1): Presence of Bum....................47
Research Question (2): Severity of Burn...................55
Research Question (3): Stream Habitat Type................68
Research Question (4): Change over Time...................96
Discussion.................................................117
VI. EXPLORING THE ECOLOGICAL RESPONSE WITHIN A CHANGING STEP-
POOL MORPHOLOGY ................................................ 123
Integrating Bio-physical Characteristics...................123
Results of Ordination Analysis............................ 127
A Changing Step-Pool Morphology........................... 133
VII. SUMMARY AND CONCLUSIONS.........................................138
vi


Summary of Major Findings.........................................138
Limitations of Study............................................. 140
Significance and Implications for Management......................143
Future Work...................................................... 144
REFERENCES.................................................................... 147
APPENDICES
A. Abbreviations...........................................................157
B. Summary Statistics of Study Reaches.....................................158
vii


CHAPTERI
INTRODUCTION AND OBJECTIVES
Step-pool sequences are important features of stream channels that have received
increasing scientific attention over the last several decades. A step-pool system develops
where the size of substrate particles is large in relation to the size of the channel (Figure 1.1,
Chin 1989). These stream features form during low-frequency, high-magnitude flood events
that are capable of mobilizing large instream particles. Over time, large events reorganize
channel substrate into an alternating pattern of steps comprising large particles and pools
made up of finer materials. Step-pool sequences promote geomorphic stability, and typically
remain in place for 50 years or more (Chin and Wohl 2005). The main function of step-pools
is to provide hydraulic resistance in high-gradient river channels. As flow tumbles over a step
into the pool below, energy is dissipated that would otherwise be translated downstream and
cause excessive erosion and other potentially hazardous consequences (Chin and Wohl
2005).
Figure 1.1. Schematic diagram of the profile of a step-pool sequence.
1


In addition to hydraulic function, step-pool sequences are important ecological
features that provide high-quality habitat for aquatic life. The range of substrate and flow
types found within a step-pool system create a broad diversity of habitats capable of
sustaining aquatic organisms with widely variable habitat requirements (Chin et al. 2009a).
Recent research has focused on the ecological role of step-pools in maintaining abundance
and diversity of benthic macroinvertebrates in mountain streams (e.g. Chung et al. 2012;
Milner and Gilvear 2012). These studies extend traditional studies on step-pool channels that
emphasized geomorphological processes. Nevertheless, little is known about the ecological
response of step-pool streams to disturbances, discrete events that alter the condition of
aquatic ecosystems. Disturbances in small headwater streams are particularly important
because changes in headwaters affect the entire fluvial network downstream.
This study investigates the ecological response of step-pool streams to wildfire
disturbance, emphasizing changes in the assemblages of benthic macroinvertebrates in steps
and pools. Benthic macroinvertebrates serve as bioindicators of stream quality (Cairns and
Pratt 1993). Changes in macroinvertebrate communities therefore provide insight into the
ecological response of step-pool streams after a disturbance such as wildfire. Although
research on the impact of wildfire on aquatic ecosystems has recently accelerated in
recognition of their increasing occurrences, the role of step-pool systems in facilitating the
ecological response of streams to wildfire has not been investigated. A conceptual framework
for describing the impact of wildfire on stream channels links a fire disturbance to an
ecological response through interacting biophysical processes of post-fire hydrology,
sediment dynamics, and channel morphology (Figure 1.2). The 2012 Waldo Canyon Fire in
2


Colorado provided a case study for investigating the ecological response of step-pool streams
to wildfire. This study also explores this response in the context of physical changes to the
step-pool morphology of stream channels following the fire.
Figure 1.2. Conceptual framework displaying the interacting biophysical processes within
stream channels following a wildfire. This study focuses on the ecological response, followed
by an exploration of the morphological linkage between fire and ecological response.
This study was conducted within the context of ongoing research into the response of
step-pool streams to the Waldo Canyon Fire, sponsored by the National Science Foundation
(EAR 1254989) and the University of Colorado Denver and directed by Dr. Anne Chin of the
Department of Geography and Environmental Sciences at the University of Colorado Denver.
The sponsored project focused on establishing baseline data for the channel morphology and
ecological character in Fall 2012, and on initial changes following the first flood season of
summer 2013, including the use of terrestrial LiDAR technology to quantify
geomorphological changes. This thesis research expanded the ecological components
3


including a second year of data collection during the summer of 2014 and subsequent
analysis and synthesis.
This thesis is organized into seven chapters. Chapter II outlines the background
literature relevant to this study, followed by specific research questions and hypotheses.
Chapter III describes the environmental characteristics of the study area and the background
to the 2012 Waldo Canyon Fire. Chapter IV outlines the methods of the study, including field
and laboratory protocols, as well as analytical procedures. Chapter V presents the results to
the research questions. Chapter VI investigates the morphological response of stream
channels after the fire as a possible linkage between wildfire disturbance and the ecological
response, using ordination to explore interactions. Chapter VII summarizes and interprets the
major findings of this study. It discusses the limitations of this research, suggests
implications for managing aquatic resources in areas disturbed by wildfire, and proposes
future lines of inquiry.
4


CHAPTER II
BACKGROUND
Ecology of Step-Pool Mountain Streams
Small headwater step-pool streams have historically generated comparatively low
interest in ecological research due to their lack of fish (Cole et al. 2003) and absence on
topographic maps. These streams are typically covered by riparian vegetation canopies
(Moore and Richardson 2003). Before the 1980s, consistent terminology did not exist to
distinguish upland step-pool systems with coarse, rocky substrata from lowland riffle-pool
systems with fine substrata (Logan and Brooker 1983). Logan and Brooker (1983) compiled
several studies of the occurrence of macroinvertebrates in riffles and pools of rocky upland
streams in the United States, Canada, and the United Kingdom. Although macroinvertebrate
communities appeared similar between the two habitat types, they found higher densities of
invertebrates in riffles, higher proportions of sensitive Ephemeroptera taxa, and lower
proportions of more tolerant Diptera taxa. The authors proposed that, overall, the literature
indicated that the effects of pollution may be greater in pool habitats than in riffles. The
riffle-pool structures described in this study displayed similar characteristics to what are
today differentiated as step-pool systems, including steep slopes, rocky substrate, and fast-
flowing water (Logan and Brooker 1983).
It wasnt until the late 1980s and 1990s that the scientific community adopted a
clearer distinction between step-pools and riffle-pools (Chin 1989). Since this time, literature
regarding the role of step-pool sequences in the fluvial system has proliferated. The vast
majority of studies concerning these streams have emphasized geomorphological aspects
5


including their importance as hydraulic and hydrologic features (e.g., Chin and Wohl 2005;
Church and Zimmermann 2007; Waters and Curran 2012). Nevertheless, interest is growing
in the ecological significance of step-pool channels as unique and diverse habitats for an
array aquatic species. These species range from benthic macroinvertebrates and fishes to the
Rocky Mountain tailed frog (Dupuis and Friele 2006).
Benthic macroinvertebrates have long served as bioindicators of stream quality, with
references to biological monitoring appearing as early as 1908 (Cairns and Pratt 1993). By
the 1970s, researchers began to connect the distributions of benthic macroinvertebrates to
physical environmental factors such as stream structure, flow velocity, and substratum
composition (e.g., Minshall and Minshall 1976; Winterbourn 1978; Reice 1980). Many
studies provided strong evidence of an inverse relationship between stream size and the
density and diversity of benthic macroinvertebrates (e.g., Danehy et al. 1999), whereby larger
streams are less diverse and less dense in benthic macroinvertebrates. This correlation is
shown through higher habitat heterogeneity in step-pool sequences, which are typically found
in small headwater streams (Clausen and Biggs 1997). Studies have also clearly established
the relationships between sediment and benthic communities, with excess sedimentation
decreasing the abundance and richness of macroinvertebrates (Wang et al. 2013). This
relationship is especially strong with very fine particles (Wood and Armitage 1997; Cole et
al. 2003; Kaller and Hartman 2004).
Studies attempting to use macroinvertebrates as bioindicators of overall habitat
quality have found step-pool sequences to provide exceptional habitats due to their inherent
stability and habitat diversity (Sullivan et al. 2004; Harvey et al. 2008; Milner and Gilvear
6


2012). Wang et al. (2009) suggested that the step-pool system is the best ecologically-sound
riverbed pattern in mountain streams. They found the diversity of benthic
macroinvertebrates to be several hundred times higher in step-pools than that of nearby
streams.
Because step-pool streams are increasingly vulnerable to a range of disturbances that
include climate change, urban encroachment upon mountain fronts, and the growing
frequency and magnitude of wildfires in mountain areas (Isaak et al. 2012; Holsinger et al.
2014; Simmoneaux et al. 2015), greater understanding of how step-pool streams respond to
disturbances is urgently needed. Although interest in the ecological response of streams to
disturbance is growing, literature regarding the response of step-pool systems to such
disturbances is notably lacking. The following sections of this review outline current
knowledge about stream response to disturbance. Many of the studies discussed investigated
disturbances in small headwater streams, but they did not reference specifically the observed
responses in relation to the role of steps and pools.
Ecological Disturbances in Fluvial Systems
According to Resh et al. (1988), disturbance in a lotic environment is any relatively
discrete event in time that disrupts ecosystem, community, or population structure, and that
changes resources, availability of substratum, or the physical environment. Disturbance is
unpredictable in terms of frequency and intensity. Stream disturbances may occur through
floods, drought, fire, logging, and many other natural and anthropogenic phenomena.
Disturbances regulate the form and function of streams by altering flow regime, rearranging
substrata, and introducing biotic and abiotic stressors to aquatic life (Resh et al. 1988).
7


Hydrologic processes such as runoff rates, peak streamflow, sediment dynamics, and bank
erosion are major determinants of stream habitat quality. A wide variety of disturbances may
profoundly impact these processes (Gresswell 1999). For example, unusual periods of
dryness can cause flow to cease over steps, eliminating the aquatic habitat heterogeneity
provided by step-pool systems. Stagnation of flow in pools during droughts may also lead to
algal proliferation and depletion of dissolved oxygen, subsequently degrading habitat
condition for benthic macroinvertebrates and fishes (Whitehead et al. 2009).
In the 1970s, a theory of dynamic equilibrium emerged relating the occurrence of
stream disturbance to the condition of macroinvertebrate communities (Connell 1978, Huston
1979). This theory suggested that the traditional concept of equilibrium exists in the absence
of disturbance. A community at equilibrium in this traditional sense is shaped by competitive
exclusion, dominated by successfully competitive species. When the ecosystem is at
equilibrium, these species out-compete inferior competitors, and the community experiences
decreased species diversity. However, ecosystem disturbances may reduce or extirpate
species that dominate in the absence of disturbance, allowing the emergence of less
competitive, disturbance-adapted species (Huston 1979; OBryan et al. 2009; Biswas and
Mallik 2011). These colonizing species remain dominant when disturbances are frequent
enough to regulate the growth of competitive species. The idea of dynamic equilibrium
entails an intermediate disturbance regime in which the ecosystem fluctuates between
equilibrium and non-equilibrium (post-disturbance) conditions. Fluctuating stream conditions
limit competitive exclusion and maximize species diversity. This process enables a dynamic
flux in the growth rate and dominance of successful and inferior competitors (Huston 1979;
8


Resh et al. 1988; Tremolieres 2004).
Large floods are capable of altering the highly stable structure of established step-
pool systems (Chin 1989). Such alteration to stream geomorphology is highly influential in
shaping the response of ecosystems to floods. The natural stability and habitat diversity of
step-pool systems make them more resilient to large events. As a result, they may exhibit
fewer losses to biotic abundance and diversity by providing physical complexity and
permanent features (at least through individual events) that act as refugia for stream
organisms (Pearsons et al. 1992; Sedell et al. 1990). When large flood events upset the
stability of step-pool channels, intermediate flows in subsequent years are generally
successful in reorganizing these streams to their natural structure (Roghair et al. 2002).
The hydrologic regime has great impact on the community composition of benthic
macroinvertebrates (Boulton 2003), periphyton (Clausen and Biggs 1997), and fishes
(Pearsons et al. 1992; Roghair et al. 2002). Floods have the ability to disrupt stream
equilibrium, remove steps and natural dams to create unstable channels, alter sediment
dynamics, and severely erode streambeds (Fuller et al. 2011). Such changes to stream
characteristics can significantly alter habitat quality and availability for aquatic life.
Clausen and Biggs (1997) explored hydrological properties across a range of stream
sizes as they relate to biotic characteristics. They recognized the role of step-pool systems in
headwater streams in creating channel stability, and thus stable flows, and related this
property of step-pool streams to periphyton biomass under normal flow conditions. No
connection was made, however, between the stability of step-pool channels and the response
of ecological communities to changes in flow regime.
9


Wildfire as a Disturbance to Stream Ecosystems
Research has accelerated on the effects of wildfire on stream ecosystems, even though
these studies have not addressed responses in step-pool or riffle-pool mountain channels
explicitly. The growing focus on wildfire has occurred as warming climates are increasingly
acknowledged as a driver of a changing fire regime (Westerling et al. 2006; Schoennagel et
al. 2007; Adams 2013; Lippitt et al. 2013). Literature typically categorizes the temporal
trajectory of stream response to wildfire into three phases: short-term (immediate to one or
two years post-fire), midterm (two to ten years), and long-term (decades to centuries) (e.g.,
Oliver et al. 2012; Malison and Baxter 2010a; Legleiter et al. 2003; Minshall et al. 1998).
The vast majority of research on the effects of wildfire focuses on the short-term timeframe,
typically two to three years following fire. The short-term response is also the focus of this
research. Short-term responses, as they relate to benthic macroinvertebrates, may be
classified into direct and indirect effects of fire (Figure 2.1) (Minshall 2003).
Direct effects of wildfire on benthic macroinvertebrates stem from characteristics of
the fire itself, such as intense heat and extended exposure to dense smoke (Minshall 2003).
Indirect effects result from secondary disturbances, such as major precipitation events
occurring after wildfire that induce increased runoff and sediment transport. Changes
resulting from the direct effects of burning, such as increased soil hydrophobicity and loss of
riparian vegetation cover, amplify the impacts of these secondary disturbances. For example,
soil hydrophobicity is a direct effect of fire that itself has little, if any, impact on aquatic life.
Soil hydrophobicity in burned areas decreases infiltration, however, leading to increased
runoff into streams. Greater runoff may, in turn, disrupt flow regimes, input excess fine
10


sediment into aquatic habitats, and alter channel morphology and particle size distribution
(Gresswell 1999; Benda et al. 2003; Legleiter et al. 2003; Minshall 2003; Ryan et al. 2011;
Moody et al. 2013).
Direct
Effects
Indirect
Effects
Figure 2.1. The succession of direct and indirect effects of wildfire on benthic
macroinvertebrate communities (based on Minshall 2003).
Although the impacts of direct effects of wildfire on stream ecology are variable, the
combination of stream size with fire intensity and severity play an apparent role. Gresswell
(1999) and Minshall (2003) reported minor effects in the absence of secondary disturbances
such as increased runoff, sediment influx, and channel alteration. Minshall (2003) suggested,
however, that direct effects may be more evident in small streams with high-intensity fires.
Consistent with this suggestion, Oliver et al. (2012) found decreased macroinvertebrate
11


densities and percentages of sensitive taxa post-fire in a first-order stream with and without
accompanying flooding and scouring. Short-term changes within ecosystems following
wildfire are most likely a result of both direct and indirect effects dictated by numerous
factors such as bum severity, precipitation regime, and watershed characteristics.
Studies focusing on indirect effects in the short term indicate decimated (Rinne 1996)
or highly altered (Minshall et al. 1997) populations of fishes and macroinvertebrates
following wildfire. While the decrease in the abundance of macroinvertebrates resulting from
increased runoff post-fire is widely undisputed, effects on taxa richness and diversity are less
well understood. Vieira et al. (2004) witnessed a reduction of abundance and taxa richness to
near-zero following a large post-fire flood event. This study found that benthic
macroinvertebrate abundance was quick to recover, but richness took several years to return
to pre-fire levels. Hall and Lombardozzi (2008) also found reduced richness one year after
the 2002 Hayman Fire in the Colorado Front Range, but saw recovery to pre-fire levels the
following year. Elsewhere in a small stream in the Lake Tahoe Basin, California, Oliver et al.
(2012) found a decrease in abundance, as well as a decrease in percentage sensitive taxa in
the first two years following a fire, but found no consistent results regarding taxa richness
and diversity.
Although research has shown inconclusive results regarding the short-term effects of
wildfire on the taxa richness of benthic macroinvertebrates, studies of community
composition have yielded more consistent findings. Loss of riparian vegetation cover is a
direct effect of fire that leads to a shift in the major energy resources available to aquatic
biota from allochthonous to autochthonous sources. Allocthonous sources are those
12


originating outside of the aquatic habitat, such as course particulate organic matter (CPOM)
arriving from the riparian zone. In contrast, autochthonous sources are those produced within
the aquatic habitat, such as periphyton growing in streams (Moss 2010). Benthic
macroinvertebrates are largely unable to survive on burned particulate matter, and thus are
forced to turn to autochthonous periphyton, which proliferate with the increased insolation
associated with the loss of riparian cover (Mihuc and Minshall 2005). This shift from
allochthonous to autochthonous sources of organic matter favors generalist over specialist
taxa. Burned streams are therefore often re-colonized by collector-gatherers, which are
generalist feeders that feed on the fine particulate organic matter (FPOM) associated with
post-fire sediment influx. These generalists displace more specialized shredders (Mihuc and
Minshall 1995; Vieira et al. 2004; Mihuc and Minshall 2005; Oliver et al 2012). Some of the
dominant taxa following wildfire in the short-term are Baetidae (Mihuc and Minshall 1995;
Vieira et al. 2004), Nemouridae (Mihuc and Minshall 1995), Simuliidae, and Chironomidae
(Vieira et al. 2004).
Studies of the effects of wildfire at a midterm time frame are also variable. The time
frame of ecological recovery depends upon many environmental factors that influence the
rate of recolonization by macroinvertebrates. Such factors include topographic barriers,
availability of refugia, and habitat stability (Sedell et al. 1990; Richards and Minshall 1992;
Vieira et al. 2004). Habitat stability after fire depends largely on the weakening of soil
hydrophobicity over time, and subsequent attenuation of the impact of hydrologic events.
Studies of wildfire effects on soil hydrophobicity indicate that soil water repellency declines
with time and with repeated wetting of the soil (Doerr et al. 2009). The timeframe for this
13


decline is highly variable, and processes that dictate the persistence of soil hydrophobicity
remain undefined. Possible factors include bum severity, vegetation cover and regrowth, soil
type, slope aspect, and precipitation regime. Studies in the Colorado Front Range have seen
infiltration return to pre-fire rates in the span of one to two years (Huffman et al. 2001;
MacDonald and Huffman 2004), or have seen rates increase without reaching pre-fire rates
after two years (Doerr et al. 2006).
Studies on macroinvertebrate response in the mid-term following wildfire have found
that the magnitude of response dampens over time, with initially reduced benthic
macroinvertebrate abundance and diversity approaching, but not surpassing pre-fire levels
within ten years (Minshall et al. 2001; Vieira et al. 2004; Mihuc and Minshall 2005). Some
studies, however, yield strikingly different results from short-term studies regarding biotic
response at severely burned sites. Higher abundances of macroinvertebrates and emergence
rates of adult aquatic insects have been recorded at high-severity burn locations, and are
associated with increased occurrence of bats and spiders in riparian zones (Malison and
Baxter 2010a). This response suggests a shift in direction of the flux of resources following
wildfires. Movement of nutrients, organic materials, and terrestrial organisms (Nakano et al.
1999) from land to water immediately following fires is elevated with higher runoff rates.
Thus, in the midterm, aquatic productivity is bolstered, and the flux of energy shifts to
movement from water to land (Malison and Baxter 2010a). Little research has focused on
long-term effects of wildfire on stream ecology, and thus few datasets exist to demonstrate
responses on this timescale (Moody et al. 2013).
14


Research Questions and Hypotheses
The review of the literature suggests the following fruitful research questions deriving
from the Waldo Canyon Fire.
1. How do benthic macroinvertebrate communities in step-pools mountain streams
respond to wildfire disturbance?
2. How does the response of macroinvertebrate communities to wildfire vary with the
severity of bum?
3. How does the response of macroinvertebrate communities to wildfire vary as a
function of habitat type (step or pool)?
4. How does the response of macroinvertebrate communities to wildfire vary over time?
Literature review and initial field observations suggest the following hypotheses.
First, wildfire will disturb benthic macroinvertebrate communities. Communities in burned
watersheds will exhibit lower taxa richness, lower proportion of sensitive taxa, and altered
composition compared to communities in unburned streams. These responses presumably
result from the indirect effects of fire, such as changes in channel morphology and sediment
dynamics following post-fire floods, and decreased coarse organic particulate input and
increased periphyton growth (Figure 2.1).
Second, ecological responses will vary between streams in watersheds burned with
high and low severity. This is because increasing severity of bum is expected to magnify the
impacts of wildfire on all of the response variables (Figure 2.1), including increased soil
hydrophobicity, decreased infiltration, and increased runoff. Magnification of these responses
is expected to generate a greater alteration in macroinvertebrate communities.
15


Third, within individual step-pool habitats in burned channels, steps will retain better
ecological quality compared to pools. This is because steps are morphologically stable stream
features, whereas pools comprise finer, less stable substrata. Pools are therefore more
susceptible to scouring during the expected post-fire floods. The morphological instability of
pools is expected to result in lower ecological quality compared to steps. The enhanced
ecological quality of steps is expected to be reflected in higher taxa richness and higher
proportion of sensitive macroinvertebrate taxa compared to steps.
Fourth, macroinvertebrate communities in burned sites are expected to reflect
significant differences in richness, proportion of community comprising sensitive taxa, and
composition from year to year after wildfire. Stream ecological quality of burned sites is
expected to degrade from immediately after the fire in 2012 to one year post-fire in 2013, as
a result of post-fire effects that include flooding and inputs of fine sediment (Figure 2.1).
This degradation will be reflected in decreased taxa richness, decreased proportion of the
community comprising sensitive taxa, and altered composition of macroinvertebrate
communities. Ecological condition is expected to show initial signs of recovery by two years
after the fire disturbance. These trends are expected to initiate the projected trajectory of
recovery toward pre-fire community composition over the next decade.
The next chapter describes the study area, including environmental characteristics of
the watersheds of interest and details of the Waldo Canyon Fire.
16


CHAPTER III
STUDY AREA
Environmental Characteristics
The study area lies within the Pike National Forest along the eastern slope of the
Colorado Front Range (Figure 3.1). This chapter briefly describes the geology, vegetation,
climate, and hydrology of the area, followed by a description of the specific study reaches
selected for analysis.
Geology
The Pikes Peak batholith dominates the geology of the study area. The Pikes Peak
batholith is a granitic structure spanning much of the central Colorado Front Range. Since its
formation during the Middle Proterozoic Eon approximately one billion years ago,
weathering and tectonic activity have exposed and uplifted the batholith. This process
resulted in large areas of exposed bedrock on the steep slopes of the region (Stoeser 2005).
The coarse grain of the Pikes Peak granite is conducive to weathering. Weathered granite
often yields a shallow layer of coarse scree overlaying the bedrock (.Fountain Creek
Watershed Study 2009). Other rock types found in the area include Ogallala shale, sandstone,
gravels and alluviums, Williams Canyon limestone, and gneiss (Stoeser et al. 2005). Despite
the coarse texture and rapid drainage of these types of soils, the shallow depth of
impermeable bedrock places the study area largely in USGS Hydrologic Soil Group D
(Natural Resources Conservation Service).
17


Vegetation
Ponderosa forest, mixed conifer, mixed montane shrubland, and lower montane-
foothills shrubland characterize the vegetation types in the study area (Colorado Natural
Heritage Program 2011). Major tree species include conifers dominated by ponderosa pine
(Pinusponderosa), as well as lodgepole pine (Pinus contorta), pinyon pine (Pinus edulis),
and western hardwoods such as Gambel oak (Quercus gambelii) (USDA Forest Service and
U.S. Geological Survey 2002). Ponderosa pines are fire-resistant due to their open crowns,
the thickness of their bark, and high moisture content of their leaves (Howard 2003). The thin
bark of lodgepole pines makes them more susceptible to fire damage, but they thrive in the
aftermath of wildfire due to their serotinous cones which require high temperatures to open
for seed dispersal (Anderson 2003). Pinyon pines are not well-adapted to wildfire, with thin
bark and flammable foliage (Anderson 2002). Gambel oak is fairly resistant to low-severity
fires and reestablishes quickly (Simonin 2000). Understory grasses and forbs in the area
include: grama (Bouteloua sp.), western wheatgrass (Pascopyrum smithii), and needle-and-
thread (Slipa comata) (Fountain Creek Watershed Study 2009).
Climate
The area receives an average of 500 to 650 mm (approximately 20 to 25 inches) of
precipitation annually, following a gradient from dry to wet as elevation increases. A notable
drying trend has been documented over the last several decades (Natural Resources
Conservation Service 2013). The widely variable elevation of the Front Range yields
changing climate and weather patterns, however, exhibiting large spatial and temporal
temperature swings, as well as unpredictable precipitation. Summer thunderstorms generate
18


much of the precipitation received on the eastern slopes (i.e., the Front Range), typically
peaking in July or August with monthly precipitation totals averaging 100 mm (four inches).
Individual storms may produce between 25 and 75 mm of rain (one to three inches; National
Water Information System). Pikes Peak, located southwest of the study area, is especially apt
to produce summer thunderstorms in the Pike National Forest. At an elevation of 4,302 m
14,114 feet), nearly 3,000 m (10,000 feet) higher than the neighboring eastern plains, Pikes
Peak creates orographic effects that strongly influence precipitation patterns of the
surrounding area (Doesken et al. 2003).
Annual average high temperatures range from approximately 10 degrees Celsius (50
F at Eagle Reach) to 15 degrees Celsius (59 F at Gage Reach). Annual average lows range
from approximately -5 to 0 degrees Celsius (low 20s to low 30s F). Summer temperatures
range from average highs approaching 20 degrees Celsius (68 F) at higher elevations, with
overnight lows approaching freezing, to highs averaging around 30 degrees Celsius (86 F) at
lower elevations. Winter highs decrease with increasing elevation from approximately 5 to 0
degrees Celsius (41 F to 32 F), and lows from -10 to -15 degrees Celsius (5 F to 14 F;
Colorado Climate Center 2010).
Hydrology
Spring snowmelt and intense summer thunderstorms generate most of the streamflow
in the study channels, although thunderstorms are the primary cause of flash floods in the
area. Discharge typically rises sharply in late spring as perennial snowpack melts, then
decreases gradually before peaking again with precipitation events (National Water
Information System). U.S. Geological Survey (USGS) weather stations provide precipitation
19


and discharge data at various locations within and around the study area. Additionally, the
University of Colorado Denver (UCD) and the USGS jointly installed four rain gages in the
study area from 2013 to 2014 for the purpose of this project (see Figure 3.3 for locations).
Watersheds selected for this study occur within the Fountain Creek Basin (2,401 km2;
927 mi2), which eventually drains into the Arkansas River (Figure 3.1). The Fountain Creek
Basin includes a portion of the city of Colorado Springs, which is situated at the base of the
mountainous northwestern region of the watershed. Fountain Creek provides 15% of the
citys drinking water (Fountain Creek Watershed 2015). As the population of this metropolis
has burgeoned over the last several decades, development has increased the percentage of
impervious area. This process has accelerated the hydrologic response to precipitation events.
This development, combined with the steep slopes of the upstream area and highly variable
precipitation regime put the watershed at risk for flash flooding after fire (Stogner 2000).
Waldo Canyon Fire of 2012
The Waldo Canyon Fire burned for 17 days within Pike National Forest along the
Front Range of Colorado. It was ignited on Saturday, 23 June 2012 at approximately 12:00
pm, roughly 15 km (nine miles) northwest of Colorado Springs, CO (Parker 2012, Figure
3.1). The U.S. Forest Service determined the fire to be human-caused, but the intent remains
unknown (Steiner 2013).
Originating in the Headwaters Fountain Creek basin, the fire spread into other major
watersheds, including Cascade Creek-Fountain Creek, Garden of the Gods, West Monument
Creek, and Lower Monument Creek before containment on 10 July 2012. Of the 74 km2
(18,247 acres) burned, 41% were classified as low severity bum, 40% moderate severity, and
20


19% high severity (Inciweb 2012). Areas burned with moderate and high severity were
characterized by long-term soil damage that increased soil water repellency and the risk of
erosion. These areas exhibited substantial losses of surface vegetation and ground litter, as
well as destruction of roots up to 10 cm (four inches) below the soil surface (Young and Rust
2012) (Figure 3.2).
Figure 3.1. Location of the Waldo Canyon Fire northwest of Colorado Springs, Colorado.
Until the Black Forest Fire in the following summer (June 2013), the Waldo Canyon
Fire was cited as the most destructive in Colorados history. Initial insurance damages totaled
an estimated $453 million. The fire also directly destroyed 346 homes, forced evacuation of
over 32,000 people, and killed two people (Coalition for the Upper South Platte, Associated
21


Press 2013). Chin et al. (2015) provide additional details of human-environment interactions
following the fire.
Study Reaches
Two canyons within the burned area were selected for analysis: Williams Canyon and
the Camp Creek watershed (Queens Canyon) (Figure 3.3). Williams Canyon is oriented
approximately north to south and flows toward Manitou Springs. Camp Creek flows
southwest toward Colorado Springs. The Camp Creek watershed (20.9 km2, 8.1 mi2) is larger
than Williams Canyon (6.2 km2, 2.4 mi2) (Rosgen et al. 2013). Channel gradients, hill slopes,
and drainage areas of tributaries to Camp Creek are thus more comparable to channel reaches
in Williams Canyon than are points along the main channel of Camp Creek.
22


I
' "ill
USAFAr;
XrS'
- ..;4'sW
Upper Queens ^ ^
_ _ Canyon >= \,*T " .
i: Sb/r -.'' A
7 V '^,A""iC F
^Meadow ^
- -
UCD-USGS-2 .
> *

rV


cv
Hmi .;
_ V ,-\ a, r \ yas
's£v. '
N
+
T Km i
Willis Upper Williams ''i
, i Canyon
,-'M ^
>1' ^ ..................
^ Academy \ Af ^eSt
(Data sources: U.S. Forest Service Burned Area Emergency Response;
Pirmin Kalberer & Mathias Walker, Sourcepole, Switzerland)
Burn Severity
____ Very Low/Low
[] Moderate
H High
Study Reaches
A Unbumed
Low Severity
H High Severity
{) UCD Rain Gage
^ Exiting Rain Gage
Figure 3.3. Locations of study reaches and rain
gages within the burn area, showing variations in
burn severity.
£ Springs
"1 Km*1
C 2M 0 NAVTEQ C A> C 1S M
23


This study examines five study reaches in the two burned watersheds: Meadow
Reach, Eagle Reach, and Aussie Reach within the Camp Creek watershed, and Willis Reach
and Tributary Reach within Williams Canyon (Figure 3.3). Similarities in characteristics such
as drainage area, elevation, channel slope, and vegetation guided the selection of the study
sites (Table 3.1). All reaches contain well-defined step-pool sequences within a span of 30 to
50 meters in length (Figure 3.4).
Table 3.1. Characteristics of study and reference reaches.
Reach Watershed Burn Severity Upstream Drainage Area (km2) % High Severity Burn % Moderate Severity Burn % Low Severity Burn Elevation (m) Channel Slope*
Willis Williams Canyon High 3.37 19.13 57.08 20.75 2,250 0.047
Tributary Williams Canyon High 1.09 43.95 53.62 2.40 2,250 0.119
Aussie Camp Creek High 2.15 16.99 58.12 22.47 2,500 0.111
Eagle Camp Creek Low 1.30 6.91 37.60 47.32 2,775 0.083
Meadow Camp Creek Low 1.54 1.11 33.55 50.36 2,650 0.044
Gage Bear Creek Unbumed 17.8 0 0 0 2,000 0.053
Hunter Bear Creek Unbumed 2.70 0 0 0 2,050 0.107
Academy West Monument Creek Unbumed 0.70 0 0 0 2,175 0.217
*slopes calculated from 2012 surveys
Additionally, the five study reaches were selected to represent sites of high and low
burn severity. Bum severity of each site was determined by the areal percentage of the
upstream watershed classified as low, moderate, and high burn severity. Reaches of low
severity bum have more than 50% of the upstream watershed classified as low burn severity
or unburned (Table 3.1). Reaches classified as high severity bum, for the purposes of this
study, have more than 50% of the upstream watershed classified as moderate or high burn
severity.
24


Figure 3.4. Step-pool stream channel within the burned area. Photo by Derek
Strickler.
The two reaches in Williams Canyon (Willis Reach and Tributary Reach) provide
examples of high severity bum sites. Willis Reach lies on the main channel at an elevation of
2,250 m (U.S. Geological Survey 2013). The average channel slope along this channel reach
is 0.047. The upstream drainage area of Willis Reach encompasses 3.37 km2. Of this area,
77.83% was burned with high or moderate severity.
Tributary Reach lies on a tributary of the main channel in Williams Canyon upstream
of Willis Reach. The tributary of interest flows from the west. The elevation of Tributary
Reach is approximately 2,250 m, and the average channel slope is 0.119. Of the 1.09 km2 that
drains into this channel reach, 97.57% of this area was burned with high or moderate
severity.
The three study reaches within the Camp Creek watershed (Queens Canyon) are
25


tributaries. Aussie Reach lies on a tributary that flows into the main channel from the north.
This reach represents a third site of high burn severity, with 75.11% of the upstream area
burned with high or moderate severity. Aussie Reach sits at an elevation of 2,500 m. The
drainage area of this channel reach is 2.15 km2, and the average channel slope is 0.111.
Eagle Reach in the Camp Creek watershed represents a site of low burn severity. This
channel reach is part of a tributary near the headwaters of the canyon at an elevation of 2,775
m. The tributary flows into the main channel from the north. Of the upstream drainage area
of Eagle Reach (1.30 km2), 44.51% was burned with high or moderate severity. The average
channel slope of this reach is 0.083.
Meadow Reach is a portion of a tributary that flows into Camp Creek from the west.
This channel reach lies downstream of Eagle Reach at an elevation of approximately 2,650
m. The drainage area of Meadow Reach is 1.54 km2, of which 34.66% was burned with high
or moderate severity. The average channel slope is 0.044.
In addition to the five study reaches burned by the Waldo Canyon Fire, three
unburned reference reaches were selected for comparison. These reaches lie within two
different watersheds: the Bear Creek watershed the West Monument Creek watershed (see
Figure 3.3). Bear Creek flows toward Colorado Springs from the southwest and drains into
Fountain Creek. This watershed lies south of the study area. West Monument Creek flows
toward Colorado Springs from the northwest. This watershed is north of the burned area.
Gage Reach lies on the main channel of Bear Creek at an elevation of 2,000 m. The
drainage area of this channel reach is 17.8 km2, and the average channel slope is 0.053.
Hunter Reach is a portion of a tributary of Bear Creek flowing from the west. This reference
26


reach lies upstream of Gage Reach at an elevation of 2,050 m. The upstream drainage area of
this reach is 2.70 km2, and the average channel slope is 0.107.
The third reference reach, Academy Reach, is located north of the burned area on a
tributary of West Monument Creek. This reach sits at an elevation of 2,175 m. The upstream
drainage area is 0.70 km2, and the average channel slope is 0.217.
The next chapter outlines the methods employed in this study. It begins by describing
methods of field data collection and laboratory processing, followed by a description of the
statistical and computational treatment for each research question.
27


CHAPTER IV
METHODS
The goal of this project was to investigate post-fire changes in the ecological quality
of step-pool streams. Benthic macroinvertebrate samples portrayed the ecological character
of stream channels and provided an indication of habitat quality. Ecological condition was
monitored a few months after the fire (2012), one year after fire (2013), and two years after
fire (2014). These data enabled comparison of characteristics among study sites at each time
step (year) and within each study site over time. The following sections detail the methods of
collection of ecological data, the laboratory procedures to process these data, and the
statistical and computational procedures utilized in analysis.
Data Collection: Field
Within each reach of approximately 50 m, two cross sections were established at
well-developed step-pool features. Benthic macroinvertebrate samples were collected from
the pool and the step at each designated cross section. This procedure provided two replicates
of steps and two of pools for each reach. Overall, 20 samples were collected from burned
reaches and 12 from unburned reaches at each time step.
Standard protocols in stream ecology guided the collection of benthic
macroinvertebrate samples in this study (Barbour et al. 1999, Merritt et al. 2008). Sampling
took place in a systematic manner, starting at downstream cross sections and working
upstream. A 500 micron D-frame net collected samples from the channel bed in pools. Pool
sampling began with substrate disturbance at the downstream margin of the pool.
Maintaining constant flow into the net, the D-frame gradually moved upstream to sample the
28


entire area of the pool in one minute. A flexible frame net, devised for collecting samples
from steps, enabled a seal between the net and the rocks of steps (Chin et al. 2009b). Rubbing
the rock surfaces and shifting rocks and debris helped to dislodge macroinvertebrates
clinging to the substrate. Sampling of the step proceeded for one minute. Samples were
transferred from the D-net to a plastic tub filled with stream water following collection from
the stream. The net underwent visual inspection for remaining macroinvertebrates. Large
rocks and organics in the sample were rinsed, inspected, and returned to the stream. The
remaining sample passed through a 500 pm sieve before transfer to a 32-oz plastic Nalgene
for storage in 95% ethanol.
From October through December 2012, benthic macroinvertebrate samples were
collected from steps and pools in the following study reaches: Willis Reach in William
Canyon; Aussie Reach, Meadow Reach, and Eagle Reach in Camp Creek; and reference
reaches Gage Reach and Hunter Reach in the Bear Creek watershed, and Academy Reach in
the tributary to West Monument Creek (Table 4.1). Samples could not be collected from
Tributary Reach in 2012 due to lack of flow at time of sampling.
Table 4.1. Timing of macroinvertebrate sample collection at each time step.
Watershed Reach 2012 Year 2013 2014
Williams Canyon Willis November August September
Tributary - September August
Camp Creek Aussie December November October
Eagle December November September
Meadow December *May November
Bear Creek Hunter December December December
Gage December October December
West Monument Creek Academy December December December
*Samples were collected from Meadow Reach in May 2014 instead of Fall 2013
29


Sampling efforts in 2013 and 2014 followed major precipitation events that occurred
during the summer and fall seasons (Table 4.2). These storms induced post-fire floods as a
result of the augmented runoff from hydrophobic soils in the burned watersheds. Post-fire
flood events of 2013 highly altered stream structure in many reaches. Several original cross
sections no longer encompassed step-pool sequences, but rather spanned across widened,
smoothed channel segments comprising shallow riffles. In these cases, samples were
collected at the nearest step-pool feature. This approach ensured consistent representation of
the ecological quality of step-pool features. Sample collection from each stream feature type
enabled comparison of the resilience of steps and pools, providing insight into the specific
role of step-pools in stream recovery after wildfire disturbance. Data collected between late
August and December 2013 included seven of the eight study reaches: Willis Reach and
Tributary Reach within Williams Canyon; Aussie Reach and Eagle Reach in Camp Creek;
and the three reference reaches, Gage Reach, Hunter Reach, and Academy Reach. In 2014,
samples from all study reaches were collected between August and December (Table 4.1).
Although data were not collected from Meadow Reach in Camp Creek in 2013 (a
reach burned with low severity), this reach was sampled in May 2014 before the first major
precipitation event of the summer (Table 4.1). Data collected at this time served to represent
the ecological condition of Meadow Reach in 2013. The research design minimized seasonal
differences in macroinvertebrate communities by calling for sample collection during spring
and fall. Communities exhibit the greatest seasonal differences during summer months
(Hilsenhoff 1988). The greatest physical and ecological changes over time after fire were
expected to occur as a result of secondary hydrological disturbances. Data collected before
30


o < CD "I C/3 c o
p £:
Table 4.2. Major precipitation events in 2013 and 2014 following the Waldo Canyon Fire. Data available from o C/3 £
existing USGS weather stations and rain gauges installed jointly by UC Denver and the USGS (see Figure 3.3 for p P 3* p
locations). o r+ CD p o CD
1 July 2013 10 July 2013 9 August 2013 11-12 September 2013 O >-b 5'
5-min 5-min 5-min 5-min Definitions and Units =r cd N> O
Station Dur Dep Avgl peak 1 R.I. Dur Dep Avgl peak I R.l Dur Dep Avgl peak I R.l. Dur Dep Avg I peak I R.1 Dur: Duration (minutes) C/3 r-K
Upper Queens 10 4.8 29.0 45.7 <2 35 10.9 18.7 36.6 <2 90 7.1 4.7 24.4 <2 2 days 170.9 3.6 70.1 100- 200 Dep: Depth (mm) CD P
Avg I: Average intensity 3 cd
Upper 10 15.0 89.9 131.1 5-10 25 13.2 31.7 73.2 <2 120 39.9 19.9 131.1 5-10 2 184.9 3.9 94.5 100- (rnm/hr) 5"
Williams days 200 5-min peak I: 5 minute to o =r cd
UCD- 15 8.1 33.5 76.2 <2 25 1.0 2.4 6.1 <2 120 21.8 11.0 67.1 <2 2 184.9 3.6 70.1 100- peak intensity (mm/hr) UJ cd
USGS-2 31 13.7 26.0 39.6 <2 days 200 R.I.: Recurrence Interval o
cd
(years) p
Bear 55 12.2 13.3 51.8 <2 Negligible rainfall 75 8.4 6.7 24.4 <2 2 244.6 5.1 51.8 500- X/l
Creek days 1000 3
110 10.2 5.5 42.7 <2 3 cd

16 July 2014 24 July 2014 29 July 2014 14 August 2014 9 October 2014 O a" cd p
5-min 5-min 5-min 5-min 5-min
Station Dur Dep Avgl peak I R.l. Dur Dep Avgl peak I R.I. Dur Dep Avgl peak I R.l Dur Dep Avg I peak I R.l Dur Dep Avgl peak 1 R.l. 3
Upper 40 9.9 14.9 48.8 <2 60 18.0 18.0 61.0 2 30 6.9 13.7 36.6 <2 24 5.8 14.0 24.3 <2 250 31.0 7.4 24.4 2-5 o
Queens o
20 9.1 27.4 61.0 2 35 14.0 24.0 61.0 2 125 22.1 10.6 61.0 2-5
Upper 15 16.0 64.0 97.6 5 15 8.9 35.6 48.8 <2 20 6.1 18.3 24.4 <2 Negligible rainfall Not available cr
Williams 25 15.0 36.0 131.1 2 115 6.9 3.6 12.2 <2 CD -t CD -t CD C/3
UCD- USGS-2 60 13.0 13.0 73.2 <2 15 6.4 25.4 45.7 <2 220 25.2 6.9 82.3 2 30 2.0 4.1 6.1 <2 525 38.6 4.4 21.3 2-5
15 7.4 29.5 45.7 <2 65 3.6 3.3 6.1 <2 125 5.1 2.4 6.1 <2 555 15.2 1.7 6.1 <2 CD P K
Bear 45 24.1 32.2 73.2 5 25 12.2 29.3 57.2 <2 15 2.3 9.1 12.2 <2 30 6.6 13.2 33.5 <2 500 55.9 6.7 24.4 10 £
Creek o
25 8.4 20.1 45.7 <2 195 25.4 7.8 64.1 <2 3 O
CD


Laboratory Methods
Macroinvertebrate samples collected in the field were returned to the laboratory at the
University of Colorado Denver for processing. Sorting was the first step in processing
samples in the laboratory (Barbour et al. 1999). The sample was first rinsed with water in a
500 pm sieve. Upon initial inspection, large organisms were separated from leaves, sediment,
and other debris. A portion of the sample was then transferred to a three-section
compartmentalized petri dish and covered with ethanol. Under a dissecting microscope, each
compartment of the petri dish was inspected over white and black backgrounds to separate
remaining macroinvertebrates from debris. This process was repeated until the entire sample
was inspected. 20 mL scintillation vials stored sorted macroinvertebrates in ethanol, and the
remaining sample was archived in its original Nalgene.
All benthic macroinvertebrate samples from 2012 were sorted in the laboratory and
identified by Aquatic Biology Associates, Inc. (Corvalis, Oregon). Sorting and identification
of samples from 2013 and 2014 occurred in the laboratory at the University of Colorado
Denver. A dissecting microscope with magnification power of up to 30x enabled these
macroinvertebrates to be identified. Insects were identified to the family level using an
illustrated identification guide (Merritt et al. 2008). Non-insects were identified to higher
taxonomic levels (subclass, class, or phylum). Identified organisms from each sample were
separated by taxon in 20 mL scintillation vials and labeled with taxon name, reach name,
location of collection (step or pool replicate), and date of collection. Identification of
macroinvertebrates provided the data for calculating a range of metrics for statistical
analysis. The following section elaborates on these metrics.
32


Data Analysis: Statistical and Computational
Variables and Constants
Dependent Variables
This study examined a range of metrics that described the composition and tolerance
of macroinvertebrate communities. These metrics provided insight into the ecological
condition of study streams. Table 4.3 outlines 18 candidate metrics with descriptions and
predicted responses to perturbation. Of the candidate metrics, 10 metrics provide key
indicators of ecological condition (Barbour et al. 1999). The metrics described in Table 4.3
fall into four general categories for characterizing macroinvertebrate communities in the
study streams. These categories are measures of (1) overall richness and composition, (2)
tolerance to perturbation, (3) composition of functional feeding groups, and (4) composition
of habit types (Barbour et al. 1999).
Overall richness and composition. Measures of overall community richness and
composition include four metrics that characterize the structure of macroinvertebrate
communities. The four metrics are (a) taxa richness; (b) percentage of organisms belonging
to the insect orders Ephemeroptera, Plecoptera, and Trichoptera (EPT); (c) percentage of
organisms belonging to the dominant taxon; and (d) percentage of organisms belonging to the
family Chironomidae. These compositional metrics reflect the relative contributions
(percentages) of certain populations to the total community of macroinvertebrates in a stream
ecosystem.
33


Table 4.3. Macroinvertebrate metrics calculated for all samples and their predicted
response to perturbation (adapted from Barbour et al. 1999, Merritt et al. 2008). This table
separates metrics into four categories of characteristics: overall richness and composition,
tolerance, functional feeding groups, and habit types.
Metric Description Predicted Response to Perturbation
Taxa Richness* Total number of taxa Decrease
*3 s S .2 % EPT* Percentage of individuals belonging to the orders Ephemeroptera, Plecoptera, and Trichoptera Decrease
o 3 <3 % Dominant Taxon* % Chironomidae Percentage of individuals belonging to the single most dominant taxon Percentage of individuals belonging to the family Chironomidae Increase Increase
1) Family Biotic Index* Calculated for a community based on weighted values of tolerance to pollution Increase
2 1) % Intolerant Taxa* Percentage of individuals with tolerance values < 2 Decrease
H % Tolerant Taxa* Percentage of individuals with tolerance values > 8 Increase
m rv % Scraper* Percentage of individuals that scrape or graze on periphyton attached to substrate Decrease
S3 2 O % Shredder* Percentage of individuals that shred coarse particulate organic matter (CPOM) Decrease
1) 1) Ph % Collector-Gatherer Percentage of individuals that gather fine particulate organic matter (FPOM) deposited in streams Variable
§ _o % Collector-Filterer* Percentage of individuals that filter FPOM from the water column Variable
=3 Ph % Predator Percentage of individuals that feed by predation on live animals Variable
% Burrower Percentage of individuals with adaptations for digging in sand or silt Variable
% Climber Percentage of individuals with adaptations for climbing plants or debris Variable
in 1) Qh % Clinger* Percentage of individuals that have adaptations for attachment to surfaces in streamflow Decrease
1 K % Skater Percentage of individuals with adaptations for moving on the waters surface Variable
% Sprawler Percentage of individuals with adaptations for remaining on top of fine substrate Variable
% Swimmer Percentage of individuals with adaptations for clinging to submerged objects between bursts of swimming Variable
*Key metrics according to Barbour et al. (1999)
34


Of the four metrics outlined above, three metrics are considered key indicators of
stream quality: taxa richness, percentage EPT, and percentage dominant taxon (Barbour et al.
1999). Taxa richness, or the number of different taxa present in a sample, represents the
diversity within a macroinvertebrate community. High taxa richness suggests a stable
community and a variety of available niches within the sampled habitat. In order for
measures of taxa richness to be comparable among samples from variable areas, samples
were subsampled electronically to a count of 300 organisms using a subsample program
provided by the Western Center for Monitoring & Assessment of Freshwater Ecosystems of
Utah State University fhttp://qcnr.usu.edu/: subsample.exe created by Dr. Dave Roberts).
Percentage EPT indicates the proportion of organisms in a sample that belong to the
orders Ephemeroptera, Plecoptera, and Trichoptera. Taxa in these orders are generally
considered to be sensitive to pollution (although this is not always true; Merritt et al. 2008).
Therefore, high percentage EPT generally suggests a healthy stream (Table 4.3).
Percentage dominant taxon is another measure of diversity in a community.
Communities that predominately comprise a single taxon typically indicate the dominance of
a tolerant taxon that can outcompete other taxa in poor ecological conditions (Barbour et al.
1999). Therefore, a high percentage dominant taxon usually suggests poor stream quality.
Chironomidae is a pollution-tolerant family of the order Diptera that frequently
appears as the dominant taxon in stream samples. This family is especially prevalent
following a disturbance to the aquatic environment (Mihuc and Minshall 2005). Therefore, a
high percentage of Chironomidae typically indicates poor ecological stream condition.
35


Tolerance to perturbation. Measures of the tolerance of macroinvertebrate
communities to perturbation include (a) the Family Biotic Index; (b) percentage intolerant
taxa; and (c) percentage tolerant taxa (Table 4.3). These metrics derive from the
Environmental Protection Agencys Regional Tolerance Values assigned to macroinvertebrate
taxa (Barbour et al. 1999), supplemented by values made available by the California Aquatic
Bioassessment Workshop (CABW; 2003). These tolerance values represent the sensitivity of
macroinvertebrate taxa to pollution, ranging from zero (intolerant) to ten (tolerant).
The three measures of community tolerance to perturbation are considered key
metrics of ecological condition (Barbour et al. 1999). The Family Biotic Index (FBI)
provides weighted tolerance values for macroinvertebrate communities based on percentages
of each taxon in a sample. To determine the FBI of a sample, the tolerance value assigned to
a taxon is multiplied by the number of organisms of that taxon in the sample. This product is
calculated for each taxon present in the sample. The total of these products is divided by the
total number of macroinvertebrates to yield the weighted tolerance value that reflects the FBI
for that sample. A low FBI indicates low macroinvertebrate community tolerance and high
ecological quality. Percentage intolerant taxa reflects the proportion of organisms belonging
to taxa that are assigned tolerance values of two or lower. Percentage tolerant taxa refers to
the proportion of organisms belonging to taxa with tolerance values of eight or higher.
Composition of functional feeding groups. Functional feeding groups classify
macroinvertebrates by their morphological and behavioral adaptations for acquiring food
(Merritt et al. 2008). The five major functional feeding groups are (a) scrapers; (b);
shredders; (c) collector-gatherers; (d) collector-filterers; and (e) predators. Community
36


composition based on functional feeding groups provides an indication of available food
sources, which often change following a disturbance to the stream system.
Of the five major functional feeding groups, three are considered key indicators of
disturbance: percentage scraper, percentage shredder, and percentage collector-filterer (Table
4.3). Scrapers and shredders exhibit specialized feeding mechanisms that obligate these
organisms to certain habitat conditions. Scrapers rely on periphyton attached to substrata,
particularly cobble surfaces (Allan and Castillo 2007). Shredders have specialized
mouthparts for shredding coarse particulate organic matter (CPOM) from allochthonous
sources (e.g., leaves of riparian vegetation) (Merritt et al. 2008). Environmental disturbance
typically leads to a decrease in the proportions of these two functional feeding groups in
macroinvertebrate communities.
Collector-gatherers and collector-filterers are generalists that feed on a wide variety
of fine particulate organic matter (FPOM). These organisms are more resilient to disturbance
because they do not rely on specialized food sources (Barbour et al. 1999). Percentage
collector-filterer, however, is considered a key metric because this feeding group is an
indicator of the type and availability of food sources from upstream. These organisms filter
FPOM directly from the water column (drift FPOM), taking advantage of the movement of
water to acquire food while expending little energy. As a result, this feeding group can be
found in high proportions relative to other feeding groups when the transport of FPOM
through a stream habitat increases (Wallace and Webster 1996).
Composition of habit types. Macroinvertebrate community composition can also be
characterized by macroinvertebrate habit type. Measures of habit type indicate the
37


mechanisms employed by organisms for locomotion or stasis within the aquatic environment.
Often, when the environment is disturbed, the altered habitat favors particular types of
locomotion, leading to a change in the type of organisms found. The six metrics of habit type
included in this analysis are: (a) percentage burrower; (b) percentage climber; (c) percentage
clinger; (d) percentage skater; (e) percentage sprawler; and (f) percentage swimmer.
Close associations exist between habit types and functional feeding groups.
Swimmers, for example, are generalist collector-gatherers that acquire food by swimming.
Clingers maintain position in flowing water by clinging to the substrate. This habit type
correlates with collector-filterers that adhere to surfaces and filter FPOM from the water
column. Clingers also include scrapers that attach to the substrate to scrape periphyton from
the surface. Clingers typically occupy exposed positions in swift streamflow, placing them at
high risk for displacement (Merritt et al. 2008). Percentage clinger is a key metric in
characterizing ecological condition that is expected to decrease in response to perturbation
(Barbour et al. 1999).
Independent Variables
Values of the dependent variables were expected to vary with the following
independent variables: presence of burn in the upstream catchment, severity of burn, stream
habitat, and time since fire event. Presence of bum was a categorical variable with values of
burned (present) and unburned (absent). Burn severity was a categorical variable assuming
values of high severity, low severity, and unburned. In this study, the term bum category is
an inclusive term that refers to presence of burn and severity of burn together. Stream habitat
(also called stream feature) was a nominal variable that may be defined as step or pool.
38


Categorization of stream habitat types indicates differences in habitat structure and quality
(representing a morphological parameter). Time since fire event was treated as a categorical
variable, and expressed as immediate (2012), one year post-fire (2013), and two years post-
fire (2014).
In addition to these independent variables that are directly related to wildfire, step-
pool stability is an independent variable that is indirectly affected by wildfire. This variable is
expected to depend upon burn severity and post-fire hydrologic disturbance. Changes in this
variable may help to explain the variation seen in macroinvertebrate communities as a
function of burn severity (Minshall et al. 1998, Legleiter et al. 2003). The geomorphological
response to fire is driven by changes in hydrologic regimes to yield ecological response (see
Figure 1.2).
Statistical and Analytical Procedures
Data analysis relied on statistical procedures to test for differences in the ecological
conditions among the study sites and over time. R data analysis provided the software to
analyze the data derived from laboratory analysis. This statistical software package is freely
available from http://www.r-project.org. Initial exploration of the data indicated non-normal
distributions; nonparametric statistical tests were therefore used. Statistical tests included the
Wilcoxon Rank Sum test, the Kruskal-Wallis test, the Wilcoxon Signed Rank test, and the
Friedman test (Corder and Foreman 2009). These statistical procedures addressed the
research questions in turn.
39


Research Question 1: Presence of Bum
First, how do benthic macroinvertebrate communities in step-pool streams respond to
wildfire disturbance? Specifically, are the median values of macroinvertebrate metrics
significantly different in burned versus unburned reaches immediately after wildfire, one year
post-fire, and two years post-fire? The Wilcoxon Rank Sum test addressed this research
question. The Wilcoxon Rank Sum test is a non-parametric statistical procedure that
compares two independent samples by rank ordering the values from each sample and
identifying whether samples cluster at opposite ends of the rank ordering (Corder and
Foreman 2009). Medians are analyzed, rather than means, because this test considers the
ranks of the values from two groups, rather than the values themselves. The Wilcoxon Rank
Sum test was an appropriate test to address this question because each component
investigated differences in macroinvertebrate metrics (Table 4.3) between two unrelated
groups of reaches. The test compared values of the macroinvertebrate metrics of all samples
collected in each year from burned reaches to the values of the metrics of samples from
reference reaches in order to distinguish differences at each time step.
Research Question 2: Bum Severity
Second, how does the response of macroinvertebrate communities to wildfire vary
with severity of burn? In other words, are the medians of the values of macroinvertebrate
metrics significantly different among channels with high severity burn, low severity burn,
and unbumed reaches at each time step? The second research question used the Kruskal-
Wallis Rank Sum test to compare macroinvertebrate metrics in study reaches of varying burn
severity. The Kruskal-Wallis test is a statistical procedure that compares more than two
40


independent samples (Corder and Foreman 2009). This test was appropriate for addressing
this question because the question investigated differences among three groups of study
channel reaches that were spatially and temporally unrelated. This comparison was done
within each sample collection year in order to determine at each stage of recovery whether
burn severity helped to explain the magnitude of ecological response.
Research Question 3: Stream Habitat Type
The third research question incorporated habitat type as a possible explanatory
variable for differences in macroinvertebrate metrics. The research question was: How does
the response of macroinvertebrate communities vary as a function of habitat type (step or
pool)? In other words, do the medians of the values of macroinvertebrate metrics differ
significantly between steps and pools within each burn category? Additionally, are the
medians of the values of macroinvertebrate metrics of each habitat type significantly
different across burn categories? For example, do the values of the metrics of steps in burned
sites differ from the values of the metrics of steps in unburned sites?
Steps versus pools. The first component of this question required the pairwise
Wilcoxon Signed Rank test to compare metrics of samples collected in steps versus pools.
This test was appropriate because it accounted for non-independence of groups by pairing
steps with adjacent pools. The Wilcoxon Signed Rank test compared steps and pools within
burned study reaches at each time step. This test also compared steps and pools within sites
of each burn severity, as well as within unbumed study reaches. Results of this test indicated
whether the specific stream feature (i.e., step or pool) influenced ecological response to
wildfire.
41


Habitat type across burn categories. The second component of this question used
the Wilcoxon Rank Sum test to compare values of the metrics from each habitat type across
burn categories. In other words, this test compared metrics from steps of burned reaches with
those from steps of unburned reaches and metrics from pools of burned reaches with those
from pools of unbumed reaches. This component also used the Kruskal-Wallis Rank Sum test
to compare samples from each habitat type across three burn severities. The results of this
analysis indicated how macroinvertebrate communities within each stream feature (i.e., step
or pool) responded to wildfire.
Research Question 4: Change over Time
The fourth research question was: How does the response of benthic
macroinvertebrates to wildfire vary over time? In other words, did the median values of the
macroinvertebrate metrics change significantly in the study reaches from 2012 to 2013, 2013
to 2014, or from 2012 to 2014? The Friedman Rank Sum test provided the means to compare
the values of macroinvertebrate metrics at each time step within each reach. The Friedman
test is a statistical procedure that compares more than two dependent samples (Corder and
Foreman 2009). The Friedman test was appropriate here because the question investigated
how the values of the macroinvertebrate metrics changed with repeated measures. In other
words, the three groups (years) in question were not independent because each group derived
from the same location (reach). Where differences over time emerged, the Friedman multiple
comparisons test further identified in which years they occurred. Analyzing each reach
separately effectively eliminated spatial variation as a confounding factor. From the results of
this test, patterns of change among reaches were examined and plotted to provide further
42


insight into the response of the dependent variables at each reach over time.
The next chapter provides an overview of the data for all study reaches to characterize
the central tendency and variability of the characteristics of response of macroinvertebrate
communities to wildfire. It then presents the results of quantitative analysis to reveal the
statistical differences and significance in the data. The statistical analyses will provide
answers to the specific research questions.
43


CHAPTER V
RESULTS: RESPONSE OF MACROINVERTEBRATES
This chapter presents the results of statistical analysis. The first section describes the
data generated from processing the macroinvertebrate samples. The subsequent sections
provide answers to the four research questions posed:
1. How do benthic macroinvertebrate communities in step-pools mountain streams
respond to wildfire disturbance?
2. How does the response of macroinvertebrate communities to wildfire vary with the
severity of bum?
3. How does the response of macroinvertebrate communities to wildfire vary as a
function of habitat type (step or pool)?
4. How does the response of macroinvertebrate communities to wildfire vary over
time?
The analysis focused primarily on the ten macroinvertebrate metrics the dependent
variables identified as key indicators of ecological condition (see Table 4.3).
Description of Data
Table 5.1 displays the means and standard deviations of the ten key macroinvertebrate
metrics (Table 4.3) for each study reach and at each time step (year). Appendix B provides
additional summary statistics (median, range, and interquartile range) for these metrics at
each reach.
44


Table 5.1. Means standard deviations of the ten key macroinvertebrate metrics for each
study reach at each time step (year). Refer to Figure 3.3 for locations of study reaches.
Category Richness and Composition Community Tolerance Functional Feeding Group Habit Type
Reach 'C ts s Year Taxa Richness H Ph W ox % Dominant Taxon Family Biotic Index % Intolerant Taxa % Tolerant Taxa 1) & 2 o GO vP 0s- % Shredder % Collector- Filterer S-H D U vP 0s-
Willis 2012 7.3 2.1 0.8 0.8 89.6 4.2 5.8 0.1 2.1 2.5 1.4 1.6 0.0 0.0 3.4 1.5 0.6 0.5 0.6 0.5
2013 8.3 1.5 0.7 1.4 30.7 4.3 6.4 0.6 0.0 0.0 10.8 7.0 0.0 0.0 2.0 4.1 4.7 9.5 8.3 9.8
e 2014 10 1.4 28.3 14.5 56.4 15.0 5.4 0.3 1.1 0.3 0.03 0.0 0.0 0.0 2.3 1.0 35.6 38.5 35.2 38.5
> CJ TJl s 5 Tributary 2013 3.5 0.6 0.0 0.0 49.5 19.5 6.4 1.1 0.0 0.0 12.5 25.0 4.2 8.3 0.0 0.0 0.0 0.0 0.0 0.0
2014 7.8 2.5 2.6 5.3 36.3 1.8 6.1 0.4 0.0 0.0 8.8 7.5 0.0 0.0 3.2 2.9 20.0 17.8 22.5 21.3
X a Aussie 2012 9.3 2.2 6.4 5.4 78.1 18.1 5.8 0.2 1.9 1.2 1.0 1.7 0.0 0.0 3.4 1.8 14.7 25.6 15.0 25.5
2013 12.5 2.6 8.8 7.3 66.3 27.1 5.7 0.1 3.4 2.6 2.1 2.1 0.2 0.3 5.3 2.0 16.9 20.0 17.3 20.5
2014 14.3 2.1 43.6 12.3 45.9 15.6 4.4 0.7 25.4 17.7 2.2 1.7 0.2 0.4 26.9 20.6 23.5 32.0 26.8 32.7
Eagle 2012 11.5 1.0 25.8 18.7 51.0 12.8 5.3 1.2 20.3 18.1 23.9 28.0 1.9 1.6 21.4 13.7 0.2 0.3 5.4 4.5
>% c 2013 14.3 4.6 51.4 29.3 58.3 11.3 2.5 1.0 49.9 29.7 11.4 7.6 0.03 0.06 47.8 27.5 8.0 6.7 2.7 1.7
i cn c 2014 18 2.7 23.2 25.1 28.6 9.2 4.0 1.1 16.4 12.9 14.0 8.5 0.2 0.3 27.2 8.1 18.1 4.1 14.2 9.9
- 5 a Meadow 2012 10.3 3.1 20.4 14.0 50.3 17.0 5.3 1.0 20.0 14.7 20.5 19.7 0.0 0.0 23.6 19.7 0.0 0.0 3.4 2.8
£ o hJ 2013 12.5 2.9 17.6 18.8 73.3 22.6 5.3 0.8 14.2 17.3 2.4 0.9 0.1 0.2 15.7 17.4 1.8 2.0 3.6 3.1
2014 16 1.8 54.2 19.0 38.6 9.4 4.3 0.7 18.5 15.8 2.9 3.0 0.1 0.1 30.3 10.5 10.3 7.7 11.5 8.7
Hunter 2012 12.3 1.7 8.8 4.0 51.9 20.1 5.2 0.3 8.6 4.0 6.4 8.3 0.4 0.4 14.2 11.4 0.5 0.5 6.3 5.8
2013 10.5 1.3 4.0 1.3 90.4 3.4 5.9 0.03 1.6 0.6 0.2 0.2 0.2 0.2 2.7 0.5 3.7 3.3 4.1 3.3
2014 13 2.2 23.8 8.2 57.6 14.0 4.9 0.5 16.9 4.2 4.6 4.2 2.1 2.2 18.7 2.4 14.5 13.8 12.7 15.6
-o Gage 2012 14.3 1.7 60.5 23.0 33.5 14.3 3.8 1.0 32.4 20.5 11.0 7.2 4.5 4.0 20.4 12.8 22.4 20.9 35.3 13.8
g 5 -Q 2013 11.5 5.2 54.8 4.4 34.2 9.7 4.1 0.4 20.9 11.7 7.8 2.1 2.9 3.3 6.4 7.8 7.4 9.9 37.1 17.1
p 2014 16.5 2.4 62.1 18.0 39.8 12.5 4.1 0.6 26.1 14.2 5.4 1.2 1.7 1.7 22.4 12.7 23.9 17.5 25.5 17.1
Academy 2012 15.5 2.6 19.0 9.4 51.3 10.2 5.3 0.4 7.2 3.3 9.4 8.7 0.2 0.3 7.3 2.2 7.9 5.6 14.6 9.5
2013 15.5 6.5 16.4 13.6 68.1 11.4 5.6 0.4 4.3 3.6 3.2 2.5 0.1 0.1 4.8 4.2 39.3 42.4 42.7 40.2
2014 16.5 1.3 32.7 15.5 49.9 16.1 5.2 0.6 12.5 11.1 11.1 0.1 0.3 0.4 12.9 11.1 11.3 4.5 3.4 2.1
45


Initial qualitative assessment of these statistics suggested several trends related to the
severity of burn. First, a distinct difference was apparent in the values of many metrics for
sites of high burn severity compared to sites of low burn severity and unbumed sites. This
distinction was particularly evident for Willis Reach and Tributary Reach, the two reaches of
high burn severity in Williams Canyon. The mean values of taxa richness, percentage EPT,
and percentage intolerant taxa were generally lower for these sites compared to other reaches.
For example, the mean percentages of EPT organisms for Willis Reach and Tributary Reach
in 2012 were 0.8% and 0.0%, respectively (Table 5.1). The percentage for the third site of
high burn severity (Aussie Reach) was 6.4% in 2012. In contrast, the mean percentages were
25.8% and 20.4% in reaches of low severity burn for 2012, and 8.8%, 60.5%, and 19.0% for
unburned reaches. Low values of these metrics typically indicate poor ecological condition.
Other metrics suggested a similar differential impact according to burn severity.
Values of the Family Biotic Index for sites of high burn severity were generally higher
compared to other sites. For example, mean values of the Family Biotic Index for high
severity bum sites in 2012 were 5.8, 6.4, and 5.8 (Table 5.1). These values for other sites
were 5.3, 5.3, 5.2, 3.8, and 5.3. A high Family Biotic Index reflects overall poor ecological
condition. Additionally, proportions of scrapers and shredders, two specialized functional
feeding groups, were somewhat low at sites burned with high severity. For example, scrapers
were entirely absent in samples from Willis Reach for each time step.
Qualitative review also suggested temporal trends in the values of the ten key
macroinvertebrate metrics for ecological condition. Changes over time in these metrics were
more evident in burned sites compared to unburned sites. Taxa richness, percentage EPT, and
46


percentage intolerant taxa appeared, overall, to increase over time in reaches of burned
watersheds, while percentage dominant taxon and the Family Biotic Index generally
appeared to decrease following the fire. For example, the mean percentage of intolerant taxa
at Aussie Reach increased from 1.9% in 2012 and 3.4% in 2013 to 25.4% in 2014 (Table
5.1). At Willis Reach, the trend was a decrease in ecological condition from 2012 to 2013,
followed by an increase in 2014. For example, the mean value of the Family Biotic Index at
Willis Reach was 5.8 in 2012, 6.4 in 2013, and 5.4 in 2014. These trends suggested that the
study channels that were affected by the Waldo Canyon Fire began to recover ecologically
within two years after the disturbance.
Research Question (1): Presence of Burn
The first research question concerned the effects of wildfire on benthic
macroinvertebrate communities at each time step (year). Were the medians of
macroinvertebrate metrics significantly different in burned versus unburned reaches
immediately after fire, one year post-fire, and two years post-fire? Statistical analysis
identified significant differences between values of the metrics of communities in step-pool
streams in burned and unbumed watersheds. This section is divided into three subsections,
each addressing one time step.
Immediately after Burn: 2012
Results of the Wilcoxon Rank Sum test showed that the median values of two key
metrics of overall richness and composition were different between burned and unburned
sites in 2012: taxa richness and percentage dominant taxon (Table 5.2). Taxa richness was
statistically lower in burned sites (median = 10) than in unbumed sites (median = 14) (p =
47


0.00027). This difference indicated that the habitat provided by step-pool features of burned
sites supported a less diverse community of macroinvertebrates. Percentage dominant taxon
was greater in burned sites (median = 71.4%) than in unburned sites (median = 48.3%) (p =
0.017). This difference indicated that the communities in burned sites were more
homogenous than those in unburned sites, with greater proportions of the benthic
communities belonging to a single taxon.
Table 5.2. Key macroinvertebrate metrics of samples collected from burned and
unburned sites in 2012, with results of the Wilcoxon Rank Sum test.
Median
Category Metric Burned Unburned p-value
Richness and Taxa Richness 10 14 0.00027 ***
Composition % EPT 7.8 19.5 0.059
% Dominant Taxon 71.4 48.3 0.017 *
Tolerance Family Biotic Index 5.8 5.1 0.0031 **
% Intolerant Taxa 4.2 8.6 0.090
% Tolerant Taxa 2.7 6.1 0.16
Functional % Scraper 0.0 0.6 0.066
Feeding Groups % Shredder 5.2 8.7 0.30
% Collector-Filterer 0.1 5.1 0.0050 **
Habit Type % Clinger 2.0 15.2 0.0032 **
*p<0.05 **p<0.01 ***p<0.001
Of the key metrics of community tolerance, only the Family Biotic Index reflected a
significant difference between burned and unbumed sites (Table 5.2). Values of the Family
Biotic Index were higher in burned sites (median = 5.8) than in unbumed sites (median = 5.1)
(p = 0.0031). This index is based on the tolerance values assigned to individual taxa (ranging
from one to ten), and the percentages of those taxa in a macroinvertebrate community
(Barbour et al. 1999; CABW 2003). The results suggested that communities in burned sites
comprised greater proportions of organisms belonging to taxa with high tolerance values and
lower proportions of organisms belonging to taxa with low tolerance values. This indicated
48


higher overall tolerance of the communities in burned sites compared to unburned sites.
Of the key metrics focused on functional feeding groups, the data for percentage
collector-filterer was significantly lower in burned sites than in unburned sites (Figure 5.1).
For burned sites, the median percentage of collector-filterers was 0.1, compared to 5.1 in
unburned sites (p = 0.0050). Lower percentage collector-filterer in burned sites indicated that
these communities comprised lower proportions of macroinvertebrates that feed on fine
particulate organic matter suspended in the water column (drift FPOM). Low availability of
drift FPOM suggests poor allochthonous food sources.
80
m 60
CS
"H
d
a
£ 40
*3
1)
£ 20
0
Scr
Functional Feeding Groups Habit Types
Shr C-G C-F Pred Bu Cb Cn Sk
Burned
Unbumed
Sp Sw
*denotes significant difference
Figure 5.1. Median proportions of macroinvertebrate functional feeding groups and
habit types in burned and unburned sites in 2012.
Scr: Scraper Bu: Burrower
Shr: Shredder Cb: Climber
C-G: Collector-Gatherer Cn: Clinger
C-F: Collector-Filterer Sk: Skater
Pred: Predator Sp: Sprawler
Sw: Swimmer
For habit type, the percentage of dingers was significantly greater in unburned sites
(median = 15.2%) than in burned sites (median = 2.0%) (p = 0.0032; Figure 5.1). This
suggested that unburned sites supported communities in which greater proportions of
49


macroinvertebrates had adaptations for clinging to the substrate in flowing water. Figure 5.1
displays the percentages of organisms of all functional feeding groups and habit types in
burned and unburned sites for 2012, immediately after the bum.
Additional (non-key) metrics reflecting significant differences between burned and
unburned sites included higher percentage Chironomidae in burned sites (71.4%) than in
unburned sites (36.7%) (p < 0.05). Chironomidae are organisms that are adapted to
disturbance. Therefore this difference was consistent with the expected response to
perturbation. The median percentage of climbers was higher in unburned sites (4.4%) than in
burned sites (1.1%) (p < 0.05). These macroinvertebrates have adaptations for vertical
movement on the surfaces of submerged plant material (e.g., stems, roots, and overhanging
branches) (Merritt et al. 2008). Lower percentages of climbers in burned study reaches
suggests, therefore, that wildfire disrupted the plant material present in these habitats.
One Year Post-Fire: 2013
Of the key metrics in all categories for the year 2013, only percentage clinger was
significantly different between burned and unburned sites (Table 5.3). Median values of the
key measures of overall richness and composition (taxa richness, percentage EPT and
percentage dominant taxon) were higher at reference sites, though the differences were not
statistically significant. The three measures of community tolerance showed no difference
between burned and unburned sites. Median values of these metrics were similar between
burned and unburned sites. The characteristics of macroinvertebrate communities based on
functional feeding groups also did not indicate any significant differences between burned
and unbumed sites. The key metric of macroinvertebrate habit type, percentage clinger, was
50


significantly lower in burned sites (median = 1.1%) than in unburned sites (median = 17.9%)
(p = 0.0022). Higher percentage clinger in unburned sites means that communities in
undisturbed sites had greater proportions of organisms that cling to the substrate in flowing
water. Figure 5.2 shows the proportions of functional feeding groups and habit types in
burned and unburned sites in 2013, one year after fire.
Table 5.3. Key macroinvertebrate metrics of samples collected from burned and unburned
sites in 2013, with results of the Wilcoxon Rank Sum test.________________________
Median
Category Metric Burned Unburned p-value
Richness and Taxa Richness 10 11.5 0.27
Composition % EPT 3.3 12.6 0.059
% Dominant Taxon 50.1 71.1 0.39
Tolerance Family Biotic Index 5.8 5.7 0.40
% Intolerant Taxa 1.8 3.9 0.36
% Tolerant Taxa 3.2 3.1 0.55
Functional % Scraper 0.0 0.1 0.057
Feeding Groups % Shredder 4.4 2.6 0.64
% Collector-Filterer 0.6 3.6 0.15
Habit Type % Clinger 1.1 17.9 0.0022 **
*p<0.05 **p<0.01 ***p<0.001
Other metrics that indicated significant differences between burned and unbumed
sites included two measures of the composition of habit types, in addition to the percentage
of dingers. These differences were reflected in the median percentage sprawler (p < 0.005)
and percentage swimmer (p < 0.005). Values of the percentage of sprawlers were lower
unburned sites (median = 1.9%) than in burned sites (median = 15.3%). Median values of
percentage swimmer were greater in unburned sites (9.5%) than in burned sites (0.3%)
(Figure 5.2).
51


Functional Feeding Groups Habit Types
80 50
Scr Shr C-G C-F Pred Bu Cb Cn Sk Sp Sw
*denotes significant difference
Figure 5.2. Median proportions of macroinvertebrate functional feeding groups and
habit types in burned and unburned sites in 2013.
Scr: Scraper Bu: Burrower
Shr: Shredder Cb: Climber
C-G: Collector-Gatherer Cn: Clinger
C-F: Collector-Filterer Sk: Skater
Pred: Predator Sp: Sprawler
Sw: Swimmer
Two Years Post-Fire: 2014
In 2014 (two years after fire), metrics of community richness and composition did not
exhibit significant differences in the values of burned and unburned sites (Table 5.4). The
median values of taxa richness, percentage EPT, and percentage dominant taxon were similar
between burned sites and reference sites.
Key metrics of community tolerance differed significantly between burned and
unburned sites in percentage intolerant taxa. Median percentage intolerant taxa was higher in
unburned sites (15.8%) than in burned sites (6.7%) (p = 0.045). This suggested that
undisturbed habitats supported greater proportions of organisms that are sensitive to
pollution.
52


Table 5.4. Key macroinvertebrate metrics of samples collected from burned and unburned
sites in 2014, with results of the Wilcoxon Rank Sum test.
Median
Category Metric Burned Unburned p-value
Richness and Taxa Richness 14 15 0.16
Composition % EPT 36.1 37.0 0.57
% Dominant Taxon 39.2 48.5 0.31
Tolerance Family Biotic Index 5.0 4.8 0.71
% Intolerant Taxa 6.7 15.8 0.045 *
% Tolerant Taxa 3.7 6.2 0.11
Functional % Scraper 0.0 0.8 0.000026 ***
Feeding Groups % Shredder 13.7 15.3 0.72
% Collector-Filterer 15.5 11.3 1.0
Habit Type % Clinger 10.4 6.4 0.39
*p<0.05 **p<0.01 ***p<0.001
Of the key functional feeding groups, the percentage scraper characteristic exhibited a
significant difference between burned and unbumed sites in 2014 (Figure 5.3). The median
percentage scraper was higher in samples collected from unburned sites (0.8%) compared to
burned sites (0.0%) (p = 0.000026). This means that undisturbed sites supported greater
proportions of organisms that acquire food by scraping periphyton from the surface of the
substrate.
Analysis of macroinvertebrate habit types suggested that the median percentage of
clingers did not differ between burned and unburned sites. Figure 5.3 displays the median
percentages of all functional feeding groups and habit types in burned and unburned sites for
2014.
Several additional metrics of functional feeding groups exhibited significant
differences between the macroinvertebrate communities of burned and unburned sites for
2014. Collector-gatherer was the dominant feeding group in burned and unburned sites, but
the median percentage was significantly higher in unbumed sites (65.4%) than in burned sites
53


(46.4%) (p < 0.05). Median percentage predator was greater in burned sites (9.4%) compared
to unburned sites (1.5%) (p < 0.0005; Figure 5.3).
Functional Feeding Groups Habit Types
Scr Shr C-G C-F Pred Bu Cb Cn Sk Sp Sw
*denotes significant difference
Figure 5.3. Median proportions of macroinvertebrate functional feeding groups and
habit types in burned and unburned sites in 2014.
Scr: Scraper Bu: Burrower
Shr: Shredder Cb: Climber
C-G: Collector-Gatherer Cn: Clinger
C-F: Collector-Filterer Sk: Skater
Pred: Predator Sp: Sprawler
Sw: Swimmer
Summarizing Key Results
These results suggest answers to the first research question: How do benthic
macroinvertebrate communities in step-pools mountain streams respond to wildfire
disturbance? Key findings from the analysis of macroinvertebrate metrics as a function of the
presence or absence of burn include the following:
In 2012, immediately after the Waldo Canyon Fire, the median value of taxa richness
was higher in unburned sites than in burned sites. Median values of percentage
dominant taxon and the Family Biotic Index were higher in burned sites than in
unburned sites. The percentages of collector-filterers and dingers were higher in
54


unburned sites.
In 2013, the macroinvertebrate habit types of clingers and swimmers were more
prevalent in unburned sites; sprawlers were more prevalent in burned sites.
In 2014, the percentage of intolerant taxa was higher in unbumed sites than in burned
sites. The functional feeding groups of scrapers and collector-gatherers were more
prevalent in unburned sites than in burned sites. Predators were more prevalent in
burned sites than in unbumed sites.
Overall, these key results support the hypothesis that streams in burned watersheds
would be negatively affected by wildfire. In particular, these results suggest a significant
impact of the fire on many metrics of macroinvertebrate communities immediately after the
Waldo Canyon Fire. Indications of impact, however, were not seen as prominently in 2013
and 2014. Effects on functional feeding groups and habit types were variable, in that the
groups that exhibited differences between burned and unburned reaches were not consistent
from year to year.
Research Question (2): Severity of Burn
The second research question addressed ecological response as a function of
variations in burn severity. Were the median values of macroinvertebrate metrics significantly
different among study sites with high severity burn, low severity burn, and those unaffected
by bum at each time step? This part of the analysis separated the burned sites into categories
of high and low bum severity. This arrangement produced three levels of burn severity for
comparison: high severity bum, low severity bum, and unburned. The following subsections
present the results of these statistical comparisons at each time step (year).
55


Immediately after Burn: 2012
In 2012, macroinvertebrate communities in sites of varying burn severity exhibited
statistically significant differences in the median values of each of the ten key metrics (Table
5.5; Figure 5.4). All metrics of richness and composition showed differences between the
medians of high severity burn and unburned sites. For example, study channels burned with
high severity exhibited low richness values (median = 8.5) compared to channels with low
severity bum (median = 12) or no bum (median = 14). This difference was statistically
significant between channels of high bum severity and unburned channels (p = 0.00030).
Sites of high burn severity differed from sites of low bum severity as well as from unburned
sites in values of percentage EPT and percentage dominant taxon. Median percentage EPT
was significantly lower in sites burned with high severity (1.4%) than in sites of low bum
severity (22.9%) and unburned sites (19.5%) (p = 0.0021). Median percentage dominant
taxon was significantly higher in sites of high bum severity (89.2%) than in sites of low bum
severity (46.4%) and unburned sites (48.3%) (p = 0.00066).
The three key measures of community tolerance differed between high burn severity
and unbumed sites. Two metrics also differed between sites of high and low burn severity.
These metrics were percentage intolerant taxa and percentage tolerant taxa (Table 5.5; Figure
5.4). Communities in channels of high burn severity exhibited significantly higher values of
the Family Biotic Index (median = 5.8) than did unburned unbumed (median = 5.1) (p =
0.0076). Median percentage intolerant taxa was significantly lower in sites of high bum
severity (1.7%) than in sites of low bum severity (17.1%) and unburned site (8.6%) (p =
0.00063). Median percentage tolerant taxa was also significantly lower in sites of high bum
56


severity (0.2%) compared to sites of low bum severity (16.6%) and unburned sites (6.1%) (p
= 0.010).
Table 5.5. Key macroinvertebrate metrics of samples collected from high severity bum, low
severity burn, and unburned sites in 2012, with results of the Kruskal-Wallis test.
Median
Category Metric High Severity Low Severity Unburned p-value
Richness and Taxa Richness t8.5 t%2 14 0.00030 ***
Composition % EPT tl.4 22.9 19.5 0.0021 **
% Dominant Taxon 189.2 46.4 48.3 0.00066 ***
Tolerance Family Biotic Index 15.8 1^5.3 5.1 0.0076 **
% Intolerant Taxa U.7 17.1 8.6 0.00063 ***
% Tolerant Taxa t0.2 16.6 6.1 0.010 *
Functional % Scraper 0.0 0.0 0.6 0.048 *
Feeding Groups % Shredder 13.5 18.2 8.7 0.0013 **
% Collector-Filterer t^O.7 to.o 5.1 0.0013 **
Habit Type % Clinger 10.9 t%.7 15.2 0.0091 **
^Identical symbols identify groups that do not differ significantly from each other
*p<0.05 **p<0.01 ***p<0.001
Analysis of the key metrics representing functional feeding group indicated that
percentage scraper differed statistically among levels of bum severity (p = 0.048) (Table 5.5;
Figure 5.5). However, the Kmskal-Wallis multiple comparisons test was unable to identify
the bum categories between which a significant difference occurred. The median values were
0.0%, 0.0%, and 0.6% for sites of high burn severity, low burn severity, and no burn,
respectively. The median percentage shredder was significantly lower in sites of high burn
severity (3.5%) than in sites of low bum severity (18.2%) and unburned sites (8.7%) (p =
0.0013). Percentage collector-filterer differed statistically between low severity (median =
0.0%) and unburned sites (median = 5.1%) (p = 0.0013).
57


Taxa Richness
Percent EPT
Percent Dominant Taxon
low
unburned
high
Bum Severity
Percent Intolerant Taxa
high
low
unburned
Burn Severity
Percent Tolerant Taxa
high
low
unbumed
Bum Severity
Burn Severity
Identical symbols identify groups that do not differ significantly from each other
Figure 5.4. Metrics of community composition and tolerance in high burn severity, low
burn severity, and unburned sites in 2012, with results of the Kruskal-Wallis test. The boxes
encompass the interquartile range and the horizontal line within each box represents the
median value. The whiskers extend beyond the boxes to reflect the variability in the data
outside of the interquartile range. Open circles indicate outliers in the data.
A difference in macroinvertebrate habit types also occurred in the metric percentage
clinger (Table 5.5; Figure 5.5). The median value of the percentage of clingers was
significantly lower in sites burned with high severity (0.9%) than in unburned sites (15.2%).
Figure 5.5 shows graphically these results as well as additional metrics representing
functional feeding groups and habit types.
Several metrics additional to the key characteristics emphasized for ecological
condition (Table 4.3) exhibited significant differences among categories of burn severity. Of
the additional metrics of community composition, the values of percentage Chironomidae
58


were significantly higher in samples from sites of high burn severity (89.2%) compared to
sites of low burn severity (41.6%) and sites of no bum (36.7%) (p < 0.005).
Functional Feeding Groups
Habit "types
100
too
H High Severity
Low Severity
Unbumed
Scr Shr C-G C-F Pred
Bu Cb Cn Sk Sp Sw
*denotes significant difference
Figure 5.5. Median proportions of macroinvertebrate functional feeding groups and
habit types in high severity bum, low severity bum, and unburned sites in 2012.
Analysis of additional metrics of functional feeding groups suggested that the
percentage of collector-gatherers was significantly higher in high severity burn sites (median
= 93.9%) than in low severity burn (median = 59.6%) and unburned sites (66.6%) (p < 0.001.
The percentage of predators was significantly lower in high severity burn sites (median =
1.3%) than in low severity bum (median = 8.8%) and unbumed sites (median = 8.9%) (p <
0.005).
Additional differences in community composition based on macroinvertebrate habit
types included significantly lower percentage climber in sites burned with high severity
Scr: Scraper
Shr: Shredder
Bu: Burrower
Cb: Climber
Cn: Clinger
Sk: Skater
Sp: Sprawler
C-G: Collector-Gatherer
C-F: Collector-Filterer
Pred: Predator
Sw: Swimmer
59


(median = 0.3%) than in unburned sites (median = 4.4%) (p < 0.05). Burrowers, while
dominant in all burn categories, occurred in significantly higher proportion in sites of high
burn severity (median = 91.5%) than in low severity burn (median = 47.2%) and unbumed
sites (median = 51.2%) (p < 0.05). The metric of percentage sprawler was significantly
higher in sites of low bum severity (median = 12.7%) than in sites of high bum severity
(median = 1.4%) (p < 0.05), though neither level of bum severity differed from unburned
sites.
One Year Post-Fire: 2013
Analysis of metrics as a function of burn severity suggested that in 2013, two key
metrics of richness and composition exhibited significant differences among high burn
severity, low bum severity, and unburned sites. These metrics were taxa richness and
percentage EPT (Table 5.6; Figure 5.6). The values of taxa richness were significantly lower
in sites of high burn severity (median = 9) than in sites of low burn severity (median = 11.5)
(p = 0.024). Values of the percentage of EPT organisms were significantly lower in sites of
high severity burn (median = 0.0%) than in both low burn severity (median = 31.8%) and
unburned sites (median = 12.6%) (p = 0.00061).
Of the measures of community tolerance, the Family Biotic Index and percentage
intolerant taxa differed statistically among levels of bum severity. Values of the Family Biotic
Index were significantly lower in low severity burn sites (median = 4.1) than in high severity
burn sites (median = 6.0) (p = 0.003). The percentage of intolerant taxa was significantly
lower in sites of high bum severity (median = 0.0%) than in sites of low burn severity
(median = 26.6%) and no burn (median = 3.9%) (p = 0.00046). The values of all metrics of
60


composition and tolerance did not differ significantly between sites of low burn severity burn
and unbumed sites.
Table 5.6. Key macroinvertebrate metrics of samples collected from high severity bum, low
severity burn, and unburned sites in 2013, with results of the Kruskal-Wallis test.
Median
Category Metric High Severity Low Severity Unbumed p-value
Richness and Taxa Richness t9 12.5 t^ 11.5 0.024 *
Composition % EPT to.o %1.8 %2.6 0.00061 ***
% Dominant Taxon 42.4 65.4 71.1 0.14
Tolerance Family Biotic Index t6.0 M.l t^5.7 0.003 **
% Intolerant Taxa to.o ^26.6 %.9 0.00046 ***
% Tolerant Taxa 2.8 3.8 3.1 0.58
Functional % Scraper 0.0 0.0 0.1 0.15
Feeding Groups % Shredder to.o ^28.0 t2.6 0.0017 **
% Collector-Filterer 0.0 3.0 3.6 0.14
Habit Type % Clinger to.l t^2.8 T7.9 0.0071 **
^Identical symbols identify groups that do not differ significantly from each other
*p<0.05 **p<0.01 ***p<0.001
Analysis of the metrics representing functional feeding groups suggested that the
percentage of shredders was significantly higher in sites burned with low severity (median =
28.0%) than in either high severity bum (median = 0.0%) or unburned sites (median = 2.6%)
(p = 0.0017; Figure 5.7; Table 5.6). Median values of percentage scraper and percentage
collector-filterer showed no significant differences among levels of burn severity.
Of the macroinvertebrate habit types, the percentage of dingers differed statistically
between sites of high burn severity and no burn (Figure 5.7). The median value of percentage
dinger was significantly lower in high severity burn sites (0.1%) than in unburned sites
(17.9%) (p = 0.0071).
61


E-
u-
O
Si
O
-O
a
3
2;
Taxa Richness
to
1 1 1 1 Hn
1 1 1 high low unburned Bum Severity Family Biotic Index t
- O O r~i
high low unburned
Bum Severity
a *
Percent EPT
t
Percent Dominant Taxon
high
low
unbumed
Burn Severity
Percent Intolerant Taxa
Burn Severity
Bum Severity
Identical symbols identify groups that do not differ significantly from each other
Figure 5.6. Metrics of community composition and tolerance in high bum severity, low bum
severity, and unbumed sites in 2013, with results of the Kruskal-Wallis test.
Additional metrics (Table 4.3) that exhibited statistically significant differences
among high severity bum, low severity bum, and unburned sites included two habit types:
percentage sprawler and percentage swimmer. Values of the percentage of sprawlers were
significantly lower in unburned sites (median = 1.5%) than in low severity (median = 26.2%)
and high severity burn sites (median = 13.7%) (p < 0.005). The percentage of swimmers
differed only between channels burned with high severity and unburned channels (p < 0.005).
This habit type was more prevalent in samples from unburned channels (median = 9.5%
compared to 0.0%).
62


Scr Shr C-G C-F Pred Bu Cb Cn Sk Sp Sw
*denotes significant difference
Figure 5.7. Median proportions of macroinvertebrate functional feeding groups and habit
types in high severity burn, low severity burn, and unburned sites in 2013.
Scr: Scraper
Shr: Shredder
C-G: Collector-Gatherer
C-F: Collector-Filterer
Pred: Predator
Bu: Burrower
Cb: Climber
Cn: Clinger
Sk: Skater
Sp: Sprawler
Sw: Swimmer
Two Years Post-Fire: 2014
In 2014, statistical analysis of the key metrics of overall community richness and
composition identified a significant difference among levels of burn severity in the values of
taxa richness (Table 5.7; Figure 5.8). The values of taxa richness were significantly lower in
high severity burn sites (median = 11) than in sites of low burn severity (median = 17.5) and
unburned sites (median = 15) (p = 0.00053). The other key metrics of overall richness and
composition (the percentage of EPT organisms and the percentage of the dominant taxon)
exhibited no differences among levels of bum severity two years after the Waldo Canyon
Fire.
63


Table 5.7. Key macroinvertebrate metrics of samples collected from high severity bum, low
severity burn, and unburned sites in 2014.
Median
Category Metric High Severity Low Severity Unbumed p-value
Richness and Taxa Richness tit 17.5 15 0.00054 ***
Composition % EPT 22.0 25.2 36.7 0.085
% Dominant Taxon 40.4 33.7 48.5 0.093
Tolerance Family Biotic Index t5.4 4.3 1"4.8 0.015 *
% Intolerant Taxa n.o 1"T3.4 15.8 0.018 *
% Tolerant Taxa 0.7 5.6 6.2 0.044 *
Functional % Scraper to.o to.o 0.8 0.00012 ***
Feeding Groups % Shredder t4.6 30.2 t^l5.3 0.0033 **
% Collector-Filterer 17.0 15.5 11.3 1.0
Habit Type % Clinger 19.4 9.7 6.4 0.64
^Identical symbols identify groups that do not differ significantly from each other
*p<0.05 **p<0.01 ***p<0.001
Taxa Richness Percent EPT Percent Dominant Taxon
* Identical symbols identify groups that do not differ significantly from each other
Figure 5.8. Key macroinvertebrate metrics of composition and tolerance in sites of high
burn severity, sites of low burn severity, and unburned sites in 2014.
64


Each of the three key measures of community tolerance exhibited differences as a
function of burn severity (Table 5.7; Figure 5.8). Values of the Family Biotic Index were
significantly lower in low severity burn sites (median = 4.3) than in high severity bum sites
(median = 5.4) (p = 0.015). The percentage of intolerant taxa was significantly lower in high
severity bum sites (median = 1.0%) than in unbumed sites (median = 15.8%) (p = 0.018).
The Kruskal-Wallis test detected a significant difference among burn categories in the value
of percentage tolerant taxa (p = 0.044), but the test failed to identify the levels of bum that
differed. The median values for percentage tolerant taxa were 0.7%, 5.6%, and 6.2% for sites
of high severity burn, low severity burn, and no burn, respectively.
Of the key metrics of functional feeding groups, percentage scraper and percentage
shredder exhibited significant differences as a function of bum severity (Table 5.7; Figure
5.9). The values of percentage scraper were higher in unburned sites (median = 0.8%) than in
low (median = 0.0%) and high severity bum sites (median = 0.0%) (p = 0.00012). The values
of percentage shredder were significantly higher in low severity bum sites (median = 30.2%)
than in high severity burn sites (median = 4.6%) (p = 0.0033). For macroinvertebrate habit
types, values of the percentage of dingers in a community did not differ significantly among
levels of burn severity.
Sites of varying burn severities differed in one additional metric representing
functional feeding groups for 2014: the percentage of predators (Figure 5.9). Values of
percentage predator were significantly higher in high severity burn (median = 10.4%) and
low severity burn (median = 8.5%) sites compared to unburned sites (median = 1.5%) (p <
0.001). The only significant difference in macroinvertebrate habit types among levels of bum
65


severity appeared in the metric of percentage climber, which was higher in channels of low
severity bum (median = 12.0%) compared to channels burned with high severity (median =
1.9%) (p< 0.05).
*denotes significant difference
Figure 5.9. Median proportions of macroinvertebrate functional feeding groups and
habit types in high severity burn, low severity burn, and unburned sites in 2014.
Scr: Scraper
Shr: Shredder
C-G: Collector-Gatherer
C-F: Collector-Filterer
Pred: Predator
Bu: Burrower
Cb: Climber
Cn: Clinger
Sk: Skater
Sp: Sprawler
Sw: Swimmer
Summarizing Key Results
In 2012, one year following the Waldo Canyon Fire, channels burned with high severity
differed from unbumed channels with lower taxa richness, percentage EPT, percentage
intolerant taxa, percentage tolerant taxa, percentage shredder, and percentage clinger.
These channels also exhibited higher percentage dominant taxon and the Family Biotic
Index compared to unburned channels. High severity burn sites differed from low
severity bum sites in many of the same metrics, except taxa richness, the Family Biotic
Index, and percentage clinger. Low severity burn sites differed from unbumed sites
66


only with a lower percentage of collector-filterers.
In 2013, sites of high burn severity differed from unbumed sites in many of the same
metrics as in 2012: percentage EPT, percentage intolerant taxa, and percentage clinger.
High severity burn sites differed from low severity bum sites with lower taxa richness,
percentage EPT, percentage intolerant taxa, and percentage shredder, and higher Family
Biotic Index. Low severity burn sites differed from unbumed sites only with a higher
percentage of shredders.
In 2014, high severity burn sites differed from unburned sites in fewer metrics: taxa
richness, percentage intolerant taxa, and percentage scraper were lower in sites of high
burn severity. Study reaches burned with high severity differed from reaches of low
severity bum with a higher Family Biotic Index and lower percentage shredder. Low
severity bum sites differed from unbumed sites with a lower percentage of scrapers.
Regarding Research Question 2 How does the response of macroinvertebrate
communities to wildfire vary with the severity of burn? the results supported the
hypothesis that the severity of burn influences the magnitude of ecological response. Overall,
the differences identified in this analysis suggested lower community diversity and higher
tolerance to pollution in high severity bum sites when compared with low severity burn and
unburned sites. Sites of high bum severity also exhibited altered proportions of functional
feeding groups and habit types compared to reference sites. Low severity burn sites differed
from unbumed sites only in metrics of functional feeding groups. These results support the
hypothesis posed for this research question.
67


Research Question (3): Stream Habitat Type
The third research question introduced habitat type steps and pools as an
independent variable contributing to variability in the response of macroinvertebrates
following wildfire. First, did the median values of macroinvertebrate metrics differ
significantly between steps and pools within each burn category? Second, were the values of
macroinvertebrate metrics for each habitat type significantly different across burn categories?
The first subsection of this section outlines the results of comparisons for samples collected
from pools with those from steps within each bum category. The second subsection considers
differences among samples from each bum category within each habitat type (e.g., pools in
burned sites compared with pools in unburned sites). Each subsection is further divided to
address differences at each time step (year). These subsections present the results of analysis
of steps and pools within all burn categories. These categories are identified as follows:
burned sites (referring to the inclusion of high and low severity burn sites), high burn severity
sites, low burn severity sites, and unburned sites. The analysis utilized median values because
the nonparametric tests were based on the ranks of the data values for each category of burn.
Steps versus Pools
This subsection addresses the following question: Did the values of macroinvertebrate
metrics differ significantly between steps and pools within each burn category? To answer
this question, pairwise comparison of the values of macroinvertebrate metrics by habitat type
identified differences between the communities of steps and pools. This analysis addressed
differences between habitat types within each category of burn: burned sites, sites burned
with high severity, sites burned with low severity, and unburned sites.
68


Immediately after Bum: 2012
For burned sites immediately after the fire, statistical analysis revealed significant
differences between steps and pools. These differences were reflected in the following
metrics: percentage EPT, percentage dominant taxon, percentage intolerant taxa, and
percentage shredder (Table 5.8). The median value of percentage EPT was higher in steps
(19.5%) than in pools (3.6%) (p = 0.016), as was the median value of percentage intolerant
taxa (12.8% compared to 1.9%) (p = 0.0078). The values of percentage dominant taxon were
lower in steps (median = 52.9%) than in pools (median = 81.4%) (p = 0.0078). Percentage
shredder was higher in steps (median = 12.7%) than in pools (median = 4.6%) (p = 0.039).
Table 5.8. Key macroinvertebrate metrics of samples collected from steps and pools of
burned sites in 2012, with results of the Wilcoxon Signed Rank test.
Median
Category Metric Pool Step p-value
Richness and Taxa Richness 10 10 0.35
Composition % EPT 3.6 19.5 0.016 *
% Dominant Taxon 81.4 52.9 0.0078 **
Tolerance Family Biotic Index 5.8 5.6 0.11
% Intolerant Taxa 1.9 12.8 0.0078 **
% Tolerant Taxa 1.9 2.7 0.55
Functional Feeding % Scraper 0.0 0.0 0.37
Groups % Shredder 4.6 12.7 0.039 *
% Collector-Filterer 0.0 0.5 0.11
Habit Type % Clinger 0.9 4.4 0.20
*p<0.05 **p<0.01 ***p<0.001
Additional significant differences the values of metrics for steps and pools in burned
sites included percentage Chironomidae (p < 0.05), percentage collector-gatherer (p < 0.05),
percentage burrower (p < 0.05), and percentage sprawler (p < 0.05). The median value of
percentage Chironomidae was higher in pools (81.4%) than in steps (43.9%). Percentages of
collector-gatherers were higher in pools (median = 87.7%) compared to steps (median =
69


59.5%), as were the percentages of burrowers (median = 84.1% in pools versus 53.2% in
steps). The median value of percentage sprawler was higher in steps (8.7%) than in pools
(1.6%).
Analysis of the metrics of steps and pools within sites of high and low burn severity
did not detect any significant differences in the composition of habit types in
macroinvertebrate communities for 2012. Tables 5.9 and 5.10 display the results of these
analyses for the ten key macroinvertebrate metrics.
Table 5.9. Key macroinvertebrate metrics of samples collected from steps and pools of high
severity burn sites in 2012, with results of the Wilcoxon Signed Rank test.
Median
Category Metric Pool Step p-value
Richness and Taxa Richness 6.5 10 0.25
Composition % EPT 0.7 5.6 0.25
% Dominant Taxon 92.4 80.5 0.13
Tolerance Family Biotic Index 5.8 5.7 0.25
% Intolerant Taxa 0.6 2.8 0.13
% Tolerant Taxa 0.2 1.3 1.0
Functional Feeding % Scraper 0.0 0.0 *n/a
Groups % Shredder 3.0 3.7 0.63
% Collector-Filterer 0.3 3.2 0.25
Habit Type % Clinger 0.5 3.5 0.25
*p<0.05 **p<0.01 ***p<0.001
^percentage scraper in all sites was zero
For unburned reaches in 2012, immediately after the fire, values for metrics
representing overall richness and composition, community tolerance, and habit type exhibited
no significant differences between steps and pools (Table 5.11). The only significant
difference identified in metrics of functional feeding groups was a greater median value of
percentage collector-filterer in steps than in pools (10.7% and 2.3%, respectively).
70


Table 5.10. Key macroinvertebrate metrics of samples collected from steps and pools of
low severity bum sites in 2012, with results of the Wilcoxon Signed Rank test.
Median
Category Metric Pool Step p-value
Richness and Taxa Richness 11 12 1.0
Composition % EPT 10.3 31.6 0.13
% Dominant Taxon 57.0 43.9 0.13
Tolerance Family Biotic Index 5.9 4.7 0.25
% Intolerant Taxa 7.7 31.6 0.13
% Tolerant Taxa 24.2 15.1 0.88
Functional Feeding % Scraper 0.0 1.6 0.37
Groups % Shredder 9.4 38.6 0.13
% Collector-Filterer 0.0 0.0 1.0
Habit Type % Clinger 3.7 5.3 0.63
*p<0.05 **p<0.01 ***p<0.001
Table 5.11. Key metrics of for steps and pools of unburned sites in 2012, with results of the
Wilcoxon Signed Rank test.
Median
Category Metric Pool Step p-value
Richness and Taxa Richness 14.5 13.5 1.0
Composition % EPT 20.8 19.5 0.84
% Dominant Taxon 48.3 47.2 0.69
Tolerance Family Biotic Index 5.0 5.1 0.56
% Intolerant Taxa 10.3 7.7 0.44
% Tolerant Taxa 6.2 5.0 0.69
Functional Feeding % Scraper 0.7 0.3 0.11
Groups % Shredder 8.7 11.1 0.84
% Collector-Filterer 2.3 10.7 0.031 *
Habit Type % Clinger 9.0 21.2 0.063
*p<0.05 **p<0.01 ***p<0.001
One Year Post-Fire: 2013
Analysis of samples collected from steps and pools of burned sites in 2013, one year
post-fire, yielded no significant difference in any macroinvertebrate metric (Table 5.12).
Similarly, no differences were detected in the characteristics of macroinvertebrate samples
from steps and pools in either high or low severity burn sites (Tables 5.13 and 5.14).
71


Table 5.12. Key metrics for steps and pools of burned sites in 2013, with results of the
Wilcoxon Signed Rank test.
Median
Category Metric Pool Step p-value
Richness and Taxa Richness 9.5 12 0.28
Composition % EPT 2.6 14.5 0.15
% Dominant Taxon 66.4 44.8 0.32
Tolerance Family Biotic Index 5.9 5.5 0.28
% Intolerant Taxa 1.4 4.9 0.21
% Tolerant Taxa 2.2 4.1 0.81
Functional Feeding % Scraper 0.0 0.0 0.36
Groups % Shredder 3.3 7.4 0.27
% Collector-Filterer 0.1 3.3 0.11
Habit Type % Clinger 0.8 4.6 0.080
*p<0.05 **p<0.01 ***p<0.001
Table 5.13. Key metrics for steps and pools of high severity burn sites in 2013, with results
of the Wilcoxon Signed Rank test.
Median
Category Metric Pool Step p-value
Richness and Taxa Richness 7.5 9 0.28
Composition % EPT 0.0 1.4 0.18
% Dominant Taxon 35.7 42.9 1.0
Tolerance Family Biotic Index 6.0 5.8 0.56
% Intolerant Taxa 0.0 0.0 0.37
% Tolerant Taxa 0.5 4.1 1.0
Functional Feeding % Scraper 0.0 0.0 0.37
Groups % Shredder 0.0 1.3 0.80
% Collector-Filterer 0.0 9.5 0.18
Habit Type % Clinger 0.1 9.5 0.20
*p<0.05 **p<0.01 ***p<0.001
For unburned study reaches, statistical analysis did not detect significant differences
in key metrics between steps and pools in 2013, one year after the fire (Table 5.15). The one
exception was the median percentage collector-filterer, a key metric representing functional
feeding groups. Steps contained higher percentage collector-filterer (14.8%) compared to
pools (1.2%) (p = 0.031). Medians of the metrics of richness and composition, community
72


tolerance, and macroinvertebrate habit types did not differ between steps and pools.
Unburned sites did, however, show differences in one additional metric for 2013. The
percentage of predators was higher in pools than in steps (median = 2.7% in pools compared
to 1.2% in steps; p < 0.05).
Table 5.14. Key metrics for steps and pools of low severity burn sites in 2013, with results
of the Wilcoxon Signed Rank test.
Median
Category Metric Pool Step p-value
Richness and Taxa Richness 11 14.5 0.88
Composition % EPT 6.8 51.1 0.38
% Dominant Taxon 80.1 50.1 0.13
Tolerance Family Biotic Index 4.9 3.3 0.25
% Intolerant Taxa 4.4 48.2 0.38
% Tolerant Taxa 3.8 4.7 1.0
Functional Feeding % Scraper 0.0 0.0 1.0
Groups % Shredder 7.2 44.1 0.38
% Collector-Filterer 2.2 41.0 0.88
Habit Type % Clinger 1.1 4.6 0.13
*p<0.05 **p<0.01 ***p<0.001
Table 5.15. Key metrics for steps and pools of unburned sites in 2013, with results of the
Wilcoxon Signed Rank test.
Median
Category Metric Pool Step p-value
Richness and Taxa Richness 11.5 11.5 0.69
Composition % EPT 26.5 6.4 0.84
% Dominant Taxon 60.7 75.5 0.84
Tolerance Family Biotic Index 5.3 5.8 0.14
% Intolerant Taxa 7.2 2.1 0.44
% Tolerant Taxa 5.4 1.1 0.31
Functional Feeding % Scraper 0.2 0.04 0.10
Groups % Shredder 4.2 2.5 0.84
% Collector-Filterer 1.2 14.8 0.031 *
Habit Type % Clinger 8.0 30.0 0.44
*p<0.05 **p<0.01 ***p<0.001
73


Two Years Post-Fire: 2014
Steps and pools of burned sites only differed statistically in two metrics in 2014, two
years after fire (Table 5.16). Of the key functional feeding groups, the percentage of
collector-filterers was significantly higher in steps (median = 29.0%) than in pools (median =
4.1%) (p = 0.014). The percentage of clingers was also significantly higher in steps (median
= 31.4%) compared to pools (median = 3.7%) (p = 0.0020).
Table 5.16. Key metrics for steps and pools of burned sites in 2014, with results of the
Wilcoxon Signed Rank test.
Median
Category Metric Pool Step p-value
Richness and Taxa Richness 14 13 0.84
Composition % EPT 36.1 38.7 0.15
% Dominant Taxon 37.6 40.9 0.23
Tolerance Family Biotic Index 5.0 4.9 1.0
% Intolerant Taxa 5.0 10.3 0.44
% Tolerant Taxa 5.3 0.8 0.32
Functional Feeding % Scraper 0.0 0.0 0.18
Groups % Shredder 27.2 9.3 0.084
% Collector-Filterer 4.1 29.0 0.014 *
Habit Type % Clinger 3.7 31.4 0.0020 **
*p<0.05 **p<0.01 ***p<0.001
Analysis of additional macroinvertebrate metrics (Table 4.3) indicated that two of
these metrics differed statistically between steps and pools of burned channels in 2014. Of
the metrics of overall community composition, percentage Chironomidae was significantly
higher in pools of burned reaches than in steps (median = 29.4% in pools versus 9.3% in
steps; p < 0.05). In addition to percentage clinger, macroinvertebrate habit types differed
between steps and pools in the percentage of burrowers. The median value of percentage
burrower was greater in pools of burned channels (48.7%) than in steps (18.4%) (p < 0.005).
For samples from high severity burn sites in 2014, statistical analysis detected
74


significant differences only in the values of percentage collector-filterer, percentage clinger,
and percentage sprawler which represent functional feeding groups and habit types (Table
5.17). The values of percentage collector-filterer and percentage clinger were significantly
higher in steps than in pools (median = 48.7% in steps and 1.1% in pools for percentage
collector-filterer, p = 0031; median = 53.8% in steps and 2.8% in pools for percentage
clinger, p = 0.031). The value of percentage sprawler was significantly higher in pools
(median = 37.8%) than in steps (median= 8.8%) (p < 0.05). For channels burned with low
severity, statistical analysis detected no significant differences between steps and pools in any
metric for 2014 (Table 5.18).
Table 5.17. Key metrics for steps and pools of high severity burn sites in 2014, with results
of the Wilcoxon Signed Rank test.
Median
Category Metric Pool Step p-value
Richness and Taxa Richness 10.5 11 1.0
Composition % EPT 40.7 15.9 0.11
% Dominant Taxon 39.4 51.5 0.31
Tolerance Family Biotic Index 5.1 5.6 0.059
% Intolerant Taxa 0.9 1.3 0.58
% Tolerant Taxa 2.1 0.5 0.56
Functional Feeding % Scraper 0.0 0.0 1.0
Groups % Shredder 4.6 4.0 0.094
% Collector-Filterer 1.1 48.7 0.031 *
Habit Type % Clinger 2.8 53.8 0.031 *
*p<0.05 **p<0.01 ***p<0.001
For unburned sites for 2014, statistical analysis identified differences in only two key
metrics (Table 5.19). Of the metrics of community tolerance, the median value of percentage
tolerant taxa was significantly greater in pools than in steps (8.2% versus 4.3%, respectively)
(p = 0.031). Macroinvertebrate habit types differed in median percentage clinger. Clingers
were more prevalent in steps than in pools of unburned sites (median = 21.7% in steps
75


compared to 4.3% in pools) (p = 0.031). Percentages of key functional feeding groups did not
differ between habitat types. Of the additional metrics of functional feeding groups, however,
percentage collector-gatherer was greater in pools (median = 71.4%) compared to steps
(median = 53.4%) (p < 0.05).
Table 5.18. Key metrics for steps and pools of low severity burn sites in 2014, with results
of the Wilcoxon Signed Rank test.
Median
Category Metric Pool Step p-value
Richness and Taxa Richness 17 17.5 0.88
Composition % EPT 32.8 62.7 0.13
% Dominant Taxon 29.4 38.0 0.66
Tolerance Family Biotic Index 4.7 3.6 0.13
% Intolerant Taxa 8.4 25.2 0.13
% Tolerant Taxa 12.9 2.6 0.13
Functional Feeding % Scraper 0.0 0.1 0.37
Groups % Shredder 31.1 25.5 0.88
% Collector-Filterer 12.6 15.5 0.88
Habit Type % Clinger 5.6 22.6 0.13
*p<0.05 **p<0.01 ***p<0.001
Table 5.19. Key metrics for steps and pools of unburned sites in 2014, with results of the
Wilcoxon Signed Rank test.
Median
Category Metric Pool Step p-value
Richness and Taxa Richness 15 16 0.83
Composition % EPT 34.7 37.0 0.69
% Dominant Taxon 58.3 39.9 0.44
Tolerance Family Biotic Index 5.1 4.6 1.0
% Intolerant Taxa 13.7 18.7 0.44
% Tolerant Taxa 8.2 4.3 0.031 *
Functional Feeding % Scraper 0.7 0.8 0.69
Groups % Shredder 13.5 18.3 0.44
% Collector-Filterer 8.9 15.7 0.22
Habit Type % Clinger 4.3 21.7 0.031 *
*p<0.05 **p<0.01 ***p<0.001
76


Habitat Type across Burn Categories
This subsection addresses the question: Were the characteristics of macroinvertebrate
communities for each habitat type (step and pool) significantly different across bum
categories? To answer this question, statistical analysis compared metrics (Table 4.3) within
each habitat type (step and pool) across categories of burn severity. Subsections address the
analysis at each time step (year). These subsections are further divided to present
comparisons between burned and unburned sites (presence of burn) and between high
severity, low severity, and unburned sites (bum severity).
Immediately after Bum: 2012
Presence of burn. For 2012, immediately after the Waldo Canyon Fire, statistical
analysis identified significant differences between samples from pools of burned and
unburned sites in the three key metrics of overall richness and composition: taxa richness,
percentage EPT, and percentage dominant taxon (Table 5.20). The median value of taxa
richness was higher in pools of unburned sites than in pools of burned sites (14.5 versus 10,
respectively) (p = 0.019). The percentage of EPT organisms was also higher in pools of
unburned sites than in pools of burned sites (median = 20.8% compared to 3.6%) (p = 0.043).
The median value of percentage dominant taxon was higher in pools of burned sites (81.4%)
than in pools of unburned sites (48.3%) (p = 0.043).
Two key metrics of community tolerance differed statistically between pools of
burned and unburned sites for 2012 (Table 5.20). The median value of the Family Biotic
Index was significantly lower in pools of unbumed channels than in pools of burned channels
(5.0 compared to 5.8, respectively) (p = 0.030). The percentage of intolerant taxa was
77


significantly higher in pools of unburned reaches than in pools of burned reaches (median =
10.3% compared to 1.9%, respectively) (p = 0.043).
Table 5.20. Key macroinvertebrate metrics in pools of burned and unbumed sites in 2012,
with results of the Wilcoxon Rank Sum test.
Median
Category Metric Burned Unburned p-value
Richness and Taxa Richness 10 14.5 0.019 *
Composition % EPT 3.6 20.8 0.043 *
% Dominant Taxon 81.4 48.3 0.043 *
Tolerance Family Biotic Index 5.8 5.0 0.030 *
% Intolerant Taxa 1.9 10.3 0.043 *
% Tolerant Taxa 1.9 6.2 0.33
Functional Feeding % Scraper 0.0 0.7 0.060
Groups % Shredder 4.6 8.7 0.18
% Collector-Filterer 0.0 2.3 0.022 *
Habit Type % Clinger 0.9 9.0 0.020 *
*p<0.05 **p<0.01 ***p<0.001
Of the metrics representing functional feeding groups and habit types, median values
of percentage collector-filterer and percentage clinger were greater in pools of unburned sites
compared with pools of burned sites (2.3% versus 0.0% for collector-filterers, p = 0.022;
9.0% versus 0.9% for dingers, p = 0.020). Additionally, pools of burned sites exhibited
higher median percentage Chironomidae (81.4%) compared to pools of unburned sites
(39.2%) (p < 0.05).
For 2012, immediately after fire, steps of burned and unbumed sites differed
statistically in three key metrics: taxa richness, percentage tolerant taxa, and percentage
clinger (Table 5.21). The median value of taxa richness was significantly higher in steps of
unburned reaches (13.5) compared to steps of burned reaches (10) (p = 0.0065). Percentage
tolerant taxa was also higher in steps of unbumed reaches than in steps of burned reaches
78


(median = 5.0% versus 2.7%, respectively; p = 0.0012), as was percentage clinger (median =
21.2% versus 4.4%, respectively; p = 0.043).
Table 5.21. Key macroinvertebrate metrics in steps of burned and unburned sites in 2012,
with results of the Wilcoxon Rank Sum test.
Median
Category Metric Burned Unburned p-value
Richness and Taxa Richness 10 13.5 0.0065 **
Composition % EPT 19.5 19.5 1.0
% Dominant Taxon 52.9 47.2 0.35
Tolerance Family Biotic Index 5.6 5.1 0.14
% Intolerant Taxa 12.8 7.7 1.0
% Tolerant Taxa 2.7 5.0 0.012 *
Functional Feeding % Scraper 0.0 0.3 0.50
Groups % Shredder 12.7 11.1 0.95
% Collector-Filterer 0.5 10.7 0.14
Habit Type % Clinger 4.4 21.2 0.043 *
*p<0.05 **p<0.01 ***p<0.001
Burn severity. For 2012, analysis of pools in high severity burn, low severity burn,
and unbumed sites identified the following differences in the three key metrics of overall
richness and composition: taxa richness, percentage EPT, and percentage dominant taxon
(Table 5.22; Figure 5.10). The median value of taxa richness was lower in pools of high
severity bum sites (6.5) than in pools of unbumed sites (14.5) (p = 0.023). The percentage of
EPT organisms was also significantly lower in pools of sites with high severity burn
compared to pools of unburned sites (median = 0.7% compared to 20.8%, respectively) (p =
0.015). The median value of percentage dominant taxon was higher in pools of high severity
burn sites (92.3%) than in pools of unburned sites (48.3%) (p = 0.015).
One key metric of community tolerance differed among pools of varying burn
severities for 2012 (Table 5.22; Figure 5.10). The median value of percentage intolerant taxa
79


was significantly lower in pools of channels burned with high severity (0.6%) than in pools
of unburned channels (10.3%) (p = 0.015).
Table 5.22. Key macroinvertebrate metrics in pools of high severity bum, low severity
burn, and unbumed sites in 2012, with results of the Kruskal-Wallis test.
Median
Category Metric High Severity Low Severity Unbumed p-value
Richness and Taxa Richness t6.5 tm %4.5 0.023 *
Composition % EPT t0.7 t"^ 10.3 ^20.8 0.015 *
% Dominant Taxon 192.3 1^57.0 ^48.3 0.015 *
Tolerance Family Biotic Index 5.8 5.9 5.0 0.090
% Intolerant Taxa 10.6 t7.7 T0.3 0.015 *
% Tolerant Taxa 0.2 24.2 6.2 0.073
Functional % Scraper 0.0 0.0 0.7 0.12
Feeding Groups % Shredder 3.0 9.4 8.7 0.076
% Collector-Filterer t^0.3 to.o ^2.3 0.028 *
Habit Type % Clinger t0.5 t%.7 ^9.0 0.021 *
^Identical symbols identify groups that do not differ significantly from each other
*p<0.05 **p<0.01 ***p<0.001
6
3
z
Taxa Richness
t fo o
high low unburned
Burn Severity
Percent EPT
Percent Dominant Taxon
high low unburned
Burn Severity
1 1 1 o 'O o 00
T- 1 -1- Percent 0 20 40 j i i 1=1 1 Percent 20 40 60 _i i i h
high low unburned
Burn Severity
Family Biotic Index
Percent Intolerant Taxa
Percent Tolerant Taxa
o
high low unburned high low unburned high low unburned
Burn Severity Burn Severity Burn Severity
^Identical symbols identify groups that do not differ significantly from each other
Figure 5.10. Key metrics of composition and tolerance of macroinvertebrate communities
in pools of sites of varying bum severities in 2012, with results of the Kruskal-Wallis test.
80


Of the key functional feeding groups, percentage collector-filterer was lower in pools
of low severity burn sites (median = 0.0%) than in pools of unburned sites (median = 2.3%)
(p = 0028). Median percentage clinger was lower in pools of high severity burn sites (0.5%)
compared to those of unburned sites (9.0%) (p = 0.021).
Additional metrics of community composition that exhibited significant differences
among pools of varying bum severities included the percentage of Chironomidae organisms
(Table 4.3). Values of percentage Chironomidae were higher in pools of high severity bum
sites than in pools of unburned sites (median = 92.3% versus 39.2%, respectively; p < 0.05).
Chironomidae are a family of macroinvertebrates that are relatively tolerant of pollution and
adapted to disturbance. For additional functional feeding groups, the values of percentage
predator were significantly higher in pools of low severity burn sites than in pools of high
severity bum sites (median = 12.3% versus 0.9%, respectively; p < 0.005). In addition to
percentage clinger, the percentage of burrowers also differed between high severity bum sites
and unbumed sites. Values of percentage burrower were significantly higher in pools of high
severity bum sites (median = 96.3% for sites of high bum severity versus 55.0% for
unburned sites; p < 0.05).
Statistical analysis for 2012 suggested similar significant differences among steps of
varying burn severities. These differences were reflected in the three key metrics of
community composition: taxa richness, percentage EPT, percentage dominant taxon (Table
5.23; Figure 5.11). Values of taxa richness were significantly higher in steps of unbumed
sites (median = 13.5) than in those of high severity burn sites (median = 10) (p = 0.013).
Values of the percentage of EPT organisms were higher in steps of sites burned with low
81


severity (median = 31.6%) than in steps of sites burned with high severity (median = 5.6%)
(p = 0.026). Although the percentage of the dominant taxon was found to differ among levels
of burn severity (p = 0.036), the multiple comparisons test failed to identify which levels
differed. The median value for percentage dominant taxon in steps was 80.5% for sites
burned with high severity, 43.9% for sites burned with low severity, and 47.2% for unburned
sites.
Table 5.23. Key macroinvertebrate metrics in steps of high severity burn, low severity burn,
and unburned sites in 2012, with results of the Kruskal-Wallis test.
Median
Category Metric High Severity Low Severity Unbumed p-value
Richness and Taxa Richness no 1"T2 %3.5 0.013 *
Composition % EPT t5.6 %1.6 t%9.5 0.026 *
% Dominant Taxon 80.5 43.9 47.2 0.036 *
Tolerance Family Biotic Index 5.7 4.7 5.1 0.072
% Intolerant Taxa t2.8 %1.6 1"7.7 0.0058 **
% Tolerant Taxa 1.3 15.1 5.0 0.10
Functional % Scraper 0.0 1.6 0.3 0.25
Feeding Groups % Shredder 13.7 %8.6 t%l.l 0.0075 **
% Collector-Filterer t%.2 to.o %0.7 0.037 *
Habit Type % Clinger 3.2 5.3 21.2 0.11
^Identical symbols identify groups that do not differ significantly from each other
*p<0.05 **p<0.01 ***p<0.001
Of the key metrics of community tolerance, percentages of intolerant taxa differed
statistically among steps of varying bum severities in 2012, immediately following the fire
(Table 5.23; Figure 5.11). Values of the percentage of intolerant taxa were significantly lower
in steps of high severity burn sites (median = 2.8%) compared to steps of low severity burn
sites (median = 31.6%) (p = 0.0058). Values of the Family Biotic Index and the percentages
of tolerant taxa showed no statistical differences among steps of varying burn severity in
2012.
82


Taxa Richness
Percent EPT
Percent Dominant Taxon
6
3
z
Percent Intolerant Taxa
Percent Tolerant Taxa
high low unburned high low unburned high low unburned
Burn Severity Burn Severity Burn Severity
^Identical symbols identify groups that do not differ significantly from each other
Figure 5.11. Key metrics of composition and tolerance of macroinvertebrate communities
in steps of varying burn severities in 2012, with results of the Kruskal-Wallis test.
Of the key metrics representing functional feeding groups, statistical analysis
identified significant differences among steps of varying bum severities in the percentages of
shredders and collector-filterers (Table 5.23). The median value of percentage shredder was
significantly higher in steps of low severity burn sites (38.6%) than in steps of high severity
burn sites (3.7%) (p = 0.0075). The median value of percentage collector-filterer was
significantly higher in steps of unbumed sites (10.7%) than in those of low severity bum sites
(0.0%) (p = 0.037). Macroinvertebrate habit types differed only in the percentages of
sprawlers. Values of percentage sprawler were significantly greater in steps of low severity
burn sites (median = 31.4%) than in those of either high severity burn or unbumed sites
(median = 2.2% and 2.6%, respectively) (p < 0.05).
83


One Year Post-Fire: 2013
Presence of burn. For 2013, the key metrics of richness and composition,
community tolerance, and functional feeding groups showed no statistically significant
differences between burned and unbumed sites in either pools or steps (Tables 5.24 and
5.25). In pools, statistical analysis identified significant differences between burned and
unburned sites in two metrics representing macroinvertebrate habit types: percentage clinger
and percentage swimmer. Pools of unbumed sites exhibited significantly higher values of the
percentage of clingers than pools of burned sites (median = 8.0% versus 0.9%, respectively)
(p = 0.011). Values of the percentage of swimmers were significantly higher in pools of
unburned sites than in pools of burned sites (median = 18.6% compared to 9.6%,
respectively) (p < 0.05).
Table 5.24. Key macroinvertebrate metrics in pools of burned and unbumed sites in 2013,
with results of the Wilcoxon Rank Sum test.
Median
Category Metric Burned Unburned p-value
Richness and Taxa Richness 9.5 11.5 0.21
Composition % EPT 2.6 26.5 0.056
% Dominant Taxon 66.4 60.7 0.71
Tolerance Family Biotic Index 5.9 5.3 0.17
% Intolerant Taxa 1.4 7.2 0.14
% Tolerant Taxa 2.2 5.4 0.79
Functional Feeding % Scraper 0.0 0.2 0.26
Groups % Shredder 3.3 4.2 0.87
% Collector-Filterer 0.1 1.2 0.40
Habit Type % Clinger 0.9 8.0 0.011 *
*p<0.05 **p<0.01 ***p<0.001
For steps, statistical analysis identified no significant differences between burned and
unburned sites in any of the ten key macroinvertebrate metrics (Table 5.25). Of all the
metrics (Table 4.3), analysis only suggested a significant difference between communities in
84


the steps of burned and unburned sites in one measure of macroinvertebrate habit types: the
percentage of sprawlers. The values of percentage sprawler were significantly higher in steps
of burned channels (median = 16.4%) than in steps of unbumed channels (median = 1.6%) (p
< 0.005).
Table 5.25. Key macroinvertebrate metrics in steps of burned and unburned sites in 2013,
with results of the Wilcoxon Rank Sum test.
Median
Category Metric Burned Unburned p-value
Richness and Taxa Richness 12 11.5 0.87
Composition % EPT 14.5 6.4 0.70
% Dominant Taxon 44.8 75.5 0.22
Tolerance Family Biotic Index 5.5 5.8 0.79
% Intolerant Taxa 4.9 2.1 0.87
% Tolerant Taxa 4.1 1.1 0.17
Functional Feeding % Scraper 0.0 0.04 0.12
Groups % Shredder 7.4 2.5 0.48
% Collector-Filterer 3.3 14.8 0.11
Habit Type % Clinger 4.6 30.0 0.057
*p<0.05 **p<0.01 ***p<0.001
Burn severity. For 2013, one year after the fire, key measures of richness and
composition in pools of varying burn severities differed only in percentage EPT (Table 5.26;
Figure 5.12). The median value of percentage EPT was significantly greater in pools of
unburned sites (26.5%) than in pools of high severity bum sites (0.0%). Of the measures of
community tolerance, statistical analysis identified significant differences in the values of the
Family Biotic Index (p = 0.044) and percentage intolerant taxa (p = 0.024) of pools across
burn severities. The multiple comparisons test failed to identify between which categories
these differences occurred. For the Family Biotic Index, the median values in pools were 6.0,
4.9, and 5.3 for sites of high severity bum, low severity bum, and no burn, respectively. For
85


percentage intolerant taxa, median values in pools were 0.0%, 4.4%, and 7.2% for sites of
high severity burn, low severity burn, and no burn, respectively.
Table 5.26. Key macroinvertebrate metrics in pools of high severity bum, low severity
burn, and unbumed sites in 2013, with results of the Kruskal-Wallis test.________
Median
Category Metric High Severity Low Severity Unbumed p-value
Richness and Taxa Richness 7.5 11 11.5 0.10
Composition % EPT to.o 1^6.8 ^26.5 0.012 *
% Dominant Taxon 35.7 80.1 60.7 0.26
Tolerance Family Biotic Index 6.0 4.9 5.3 0.044 *
% Intolerant Taxa 0.0 4.4 7.2 0.024 *
% Tolerant Taxa 0.5 3.8 5.4 0.59
Functional % Scraper 0.0 0.0 0.2 0.37
Feeding Groups % Shredder 0.0 7.2 4.2 0.12
% Collector-Filterer 0.0 2.2 1.2 0.15
Habit Type % Clinger to.l tn.t %.o 0.019 *
^Identical symbols identify groups that do not differ significantly from each other
*p<0.05 **p<0.01 ***p<0.001
Taxa Richness
Percent EPT
Percent Dominant Taxon
6
3
z
high low unburned
Burn Severity
high low unburned
Burn Severity
Family Biotic Index
Percent Intolerant Taxa
Percent Tolerant Taxa
^Identical symbols identity groups that do not differ significantly from each other
Figure 5.12. Key metrics of composition and tolerance of macroinvertebrate communities
in pools in sites of varying burn severities in 2013, with results of the Kruskal-Wallis test.
86


Metrics representing functional feeding groups in pools did not exhibit significant
differences among levels of bum severity in 2013 (Table 5.26). Macroinvertebrate habit types
differed in pools of high bum severity and unburned sites in percentage clinger (p = 0.019)
and percentage swimmer (p < 0.05). The median percentages of these habit types were
greater in pools of unbumed reference sites than in pools of high severity burn sites (8.0%
versus 0.1% for percentage clinger; 18.6% versus 0.0% for percentage swimmer).
Analysis of macroinvertebrate communities in steps for 2013 identified statistically
significant differences in one metric of overall richness and composition (Table 5.27; Figure
5.13). Values of the percentage of EPT organisms were significantly higher in steps of low
severity bum sites (median = 51.1%) compared to steps of high severity burn sites (median =
1.4%) (p = 0.026).
Table 5.27. Key macroinvertebrate metrics in steps of high severity burn, low severity burn,
and unburned sites in 2013, with results of the Kruskal-Wallis test.
Median
Category Metric High Severity Low Severity Unbumed p-value
Richness and Taxa Richness 9 14.5 11.5 0.21
Composition % EPT 1-1.4 51.1 1^6.4 0.026 *
% Dominant Taxon 42.9 50.1 75.5 0.31
Tolerance Family Biotic Index 15.8 3.3 1^5.8 0.026 *
% Intolerant Taxa to.o 48.2 1^2.1 0.0098 **
% Tolerant Taxa 4.1 4.7 1.1 0.36
Functional % Scraper 0.0 0.0 0.04 0.17
Feeding Groups % Shredder tl.3 44.1 1^2.5 0.017
% Collector-Filterer 9.5 3.3 14.8 0.25
Habit Type % Clinger 9.5 4.6 30.0 0.14
^Identical symbols identify groups that do not differ significantly from each other
*p<0.05 **p<0.01 ***p<0.001
87


Taxa Richness
Percent EPT
Percent Dominant Taxon
high low unburned
Burn Severity
high low unburned
Burn Severity
Family Biotic Index
Percent Intolerant Taxa
Percent Tolerant Taxa
high low unburned high low unburned high low unburned
Burn Severity Burn Severity Burn Severity
^Identical symbols identify groups that do not differ significantly from each other
Figure 5.13. Key metrics of composition and tolerance in macroinvertebrate communities
in steps of varying burn severities in 2013, with results of the Kruskal-Wallis test.
Of the measures of community tolerance, statistical analysis identified differences in
the Family Biotic Index and the percentage of intolerant taxa in steps as a function of burn
severity for 2013 (Table 5.27; Figure 5.13). Values of the Family Biotic Index were
significantly lower in steps of low severity burn sites (median = 3.3) than in those of high
severity bum sites (median = 5.8) (p = 0.026). Values of the percentage of intolerant taxa
were significantly greater in steps of low severity bum (median = 48.2%) than in steps of
high severity burn (median = 0.0%) (p = 0.0098).
Functional feeding groups differed among the steps of different levels of burn
severity only in the percentage of shredders (Table 5.27). The median values of percentage
shredder were significantly higher in steps of low burn severity (44.1%) than in steps of high
88


burn severity (1.3%) (p = 0.017). Habit types differed only in the percentage of sprawlers.
Median percentage sprawler was higher in low severity burn sites (47.8%) than in unburned
sites (1.6%) (p < 0.01).
Two Years Post-Fire: 2014
Presence of burn. For 2014, two years after the Waldo Canyon Fire, statistical
analysis of pools in burned and unburned sites did not identify significant differences in any
of the three key metrics of overall richness and composition (Table 5.28). Similarly, the three
measures of community tolerance showed no differences between the pools of burned and
unburned sites.
Table 5.28. Key metrics in pools of burned and unburned sites in 2014, with results of the
Wilcoxon Rank Sum test.
Median
Category Metric Burned Unburned p-value
Richness and Taxa Richness 14 15 0.27
Composition % EPT 36.1 34.7 0.64
% Dominant Taxon 37.6 58.3 0.093
Tolerance Family Biotic Index 5.0 5.1 0.87
% Intolerant Taxa 5.0 13.7 0.12
% Tolerant Taxa 5.3 8.2 0.25
Functional Feeding % Scraper 0.0 0.7 0.0012 **
Groups % Shredder 27.2 13.5 0.79
% Collector-Filterer 4.1 8.9 0.17
Habit Type % Clinger 3.7 4.3 1.0
*p<0.05 **p<0.01 ***p<0.001
Of the key functional feeding groups, values of the percentage of scrapers were
higher in pools of unbumed sites for 2014 (median = 0.7%) than in pools of burned sites
(median = 0.0%) (p = 0.0012; Table 5.28). Other key measures of functional feeding groups
(percentage shredder and percentage collector-filterer) did not differ between pools of burned
and unbumed study reaches. Additional metrics exhibiting differences in pools included only
89


the percentage of predators. Median percentage predator was significantly lower in pools of
unburned sites (1.7%) compared to pools of burned sites (9.4%) (p < 0.001). Proportions of
habit types did not differ significantly between pools of burned and unburned sites.
In steps for 2014, metrics of overall richness and composition did not differ
significantly between burned and unbumed sites (Table 5.29). Metrics of community
tolerance also showed no significant differences between burned and unbumed sites. Of the
key functional feeding groups, median percentage scraper was higher in steps of unburned
sites (0.8%) than in steps of burned sites (0.0%) (p = 0.0039). Additionally, the median
percentage of predators was higher in steps of burned channels (7.7%) than in those of
unburned channels (1.2%) (p < 0.05). Analysis of macroinvertebrate habit types identified no
statistically significant differences between steps of burned and unburned sites.
Table 5.29. Key metrics in steps of burned and unburned sites in 2014, with results of the
Wilcoxon Rank Sum test.
Median
Category Metric Burned Unburned p-value
Richness and Taxa Richness 13 16 0.41
Composition % EPT 38.7 37.0 1.0
% Dominant Taxon 40.9 39.9 0.79
Tolerance Family Biotic Index 4.9 4.6 0.96
% Intolerant Taxa 10.3 18.7 0.17
% Tolerant Taxa 0.8 4.3 0.22
Functional Feeding % Scraper 0.0 0.8 0.0039 **
Groups % Shredder 9.3 18.3 0.26
% Collector-Filterer 29.0 25.7 0.37
Habit Type % Clinger 31.4 21.7 0.22
*p<0.05 **p<0.01 ***p<0.001
Burn severity. In pools, statistical analysis of metrics for 2014 by burn severity
indicated that values of taxa richness were significantly higher in channels burned with low
severity (median = 17) than in those burned with high severity (median = 10.5) (p = 0.013;
90


Table 5.30; Figure 5.14). Other key metrics of community composition did not differ
significantly among levels of burn severity. Similarly, measures of community tolerance
showed no statistically significant differences in pools of varying bum severities.
Table 5.30. Key macroinvertebrate metrics in pools of high severity bum, low severity
burn, and unbumed sites in 2014, with results of the Kruskal-Wallis test.
Median
Category Metric High Severity Low Severity Unbumed p-value
Richness and Taxa Richness f 10.5 I? 1"T5 0.013 **
Composition % EPT 40.7 32.8 34.7 0.83
% Dominant Taxon 39.4 29.4 58.3 0.11
Tolerance Family Biotic Index 5.1 4.7 5.1 0.61
% Intolerant Taxa 0.9 8.4 13.7 0.24
% Tolerant Taxa 2.1 12.9 8.2 0.073
Functional % Scraper to.o tAi.o A).7 0.0044 **
Feeding Groups % Shredder 4.6 31.1 13.5 0.32
% Collector-Filterer 1.1 12.6 8.9 0.033 *
Habit Type % Clinger 2.8 5.6 4.3 0.48
^Identical symbols identify groups that do not differ significantly from each other *p<0.05 **p<0.01 ***p<0.001
Key functional feeding groups in pools for 2014 differed in the percentage of scrapers
and the percentage of collector-filterers (Table 5.30). Values of percentage scraper were
higher in pools of unbumed sites (median = 0.7%) than in those of high severity burn sites
(median = 0.0%) (p = 0.0044). Although a difference among burn categories was identified in
percentage collector-filterer (p = 0.033), the difference could not be assigned to specific
levels of burn severity. Median values of percentage collector-filterer were 1.1%, 12.6%, and
8.9% for sites of high severity burn, low severity burn, and no burn, respectively.
Additionally, the median percentage of predators was higher in pools of sites of high severity
burn (10.4%) than in unburned sites (1.7%) (p < 0.01).
91


Percent EPT
Percent Dominant Taxon
Taxa Richness
t
high low unburned
Burn Severity
Family Biotic Index
Percent Intolerant Taxa
Percent Tolerant Taxa
Burn Severity Burn Severity Burn Severity
^Identical symbols identify groups that do not differ significantly from each other
Figure 5.14. Key metrics of composition and tolerance of macroinvertebrate communities
in pools of varying burn severities in 2014, with results of the Kruskal-Wallis test.
For macroinvertebrate habit types of pools for 2014, percentage clinger was not
shown to differ among levels of burn severity (Table 5.30). Habit types did, however, exhibit
differences in percentages of climbers and sprawlers. Values of percentage climber were
higher in pools of low burn severity sites (median = 13.7%) than in pools of either unburned
(median = 4.1%) or high severity bum sites (median = 4.1%) (p < 0.05). Median percentage
sprawler was higher in pools of high burn severity sites (37.7%) than in those of low severity
burn sites (4.1%) (p < 0.05).
In steps for 2014, statistical analysis across bum severities identified significant
differences in two metrics of overall richness and composition: taxa richness and percentage
EPT (Table 5.31; Figure 5.15). Although the Kmskal-Wallis test detected a difference in taxa
92


richness (p = 0.44), it failed to identify which levels of burn severity differed. The median
values of taxa richness were 11 for sites of high bum severity, 17.5 for sites of low bum
severity, and 16 for unburned sites. Median percentage EPT was higher in steps of low
severity bum sites (62.7%) than in those of high severity bum sites (15.9%) (p = 0.012).
Table 5.31. Key macroinvertebrate metrics in steps of high severity burn, low severity
burn, and unbumed sites in 2014, with results of the Kruskal-Wallis test.
Median
Category Metric High Severity Low Severity Unbumed p-value
Richness and Taxa Richness 11 17.5 16 0.044 *
Composition % EPT U5.9 %2.7 1^37.0 0.012 *
% Dominant Taxon 51.5 38.0 39.9 0.36
Tolerance Family Biotic Index 15.6 %.6 1M.6 0.0041 **
% Intolerant Taxa U.3 ^25.2 T8.7 0.0059 **
% Tolerant Taxa 0.5 2.6 4.3 0.31
Functional % Scraper to.o tV).l V).8 0.012 *
Feeding Groups % Shredder 14.0 ^25.5 %8.3 0.0049 **
% Collector-Filterer 48.7 15.5 25.7 0.036 *
Habit Type % Clinger 53.8 22.6 21.7 0.032 *
^Identical symbols identity groups that do not differ significantly from each other
*p<0.05 **p<0.01 ***p<0.001
Of the metrics representing community tolerance, two exhibited differences among
steps in sites of varying bum severity: the Family Biotic Index and the percentage of
intolerant taxa (Table 5.31). The values of the Family Biotic Index were lower in steps of low
severity bum sites (median = 3.6) than in steps of high severity bum sites (median = 5.6) (p =
0.0041). Values of the percentage of intolerant taxa were higher in steps of low severity burn
(median = 25.2%) and unburned sites (median = 18.7%) than in steps of high severity bum
sites (median = 1.3%) (p = 0.0041).
93


Full Text

PAGE 1

THE ECOLOGICAL RESPONSE OF STEP-POOL STREAMS FOLLOWING THE 2012 WALDO CANYON FIRE OF COLORADO, USA by ANNA PARKER SOLVERSON B.S., College of William & Mary, 2007 A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Master of Science Environmental Sciences 2015

PAGE 2

This thesis for the Master of Science degree by Anna Parker Solverson has been approved for the Environmental Sciences Program by Anne Chin, Chair Peter Anthamatten Alison O'Dowd November 29, 2015 ii

PAGE 3

Solverson, Anna Parker (M.S., Environmental Sciences) The Ecological Response of Step-Pool Streams Following the 2012 Waldo Canyon Fire of Colorado, USA Thesis directed by Professor Anne Chin ABSTRACT Step-pool mountain streams are increasingly subjected to a variety of natural and human-induced disturbances such as climate change, urban encroachment, and wildfires. In the western USA, wildfires are increasing in frequency and magnitude due to warming climates. Because step-pool sequences provide critical hydraulic and ecological functions in the river system, understanding the impact of wildfires on step-pool streams is crucial for effective management of aquatic resources. This study investigated the ecological response of step-pool streams to the 2012 Waldo Canyon Fire of Colorado, using benthic macroinvertebrates as indicators of changes in the ecological condition of stream channels. The analysis tracked changes in benthic macroinvertebrates immediately following the fire (2012), one year post-fire (2013), and two years post-fire (2014). Four categories of metrics represented characteristics of macroinvertebrate communities: overall richness and composition, community tolerance, functional feeding groups, and habit types. Nonparametric statistical analysis tested for differences in the responses of macroinvertebrate communities as a function of (1) presence of burn; (2) severity of burn; (3) habitat type (step or pool); and (4) time. Additionally, ordination allowed exploration of the ecological responses together with the changing physical characteristics of the step-pool streams. This study produced the following main results. First, the ecological condition of iii

PAGE 4

channels burned by wildfire was significantly poorer than that of unburned channels. Second, channels burned with high severity generally reflected poorer condition compared to channels burned with low severity and unburned channels. Channels burned with low severity only differed from unburned channels in functional feeding groups. Third, pool habitats showed the negative effects of wildfire more than were step habitats. Fourth, although channels burned with high severity showed evidence of impact at each of the time steps examined, some channels exhibited signs of recovery within two years following the fire. Finally, the extent of ecological impact and recovery apparently correlates with the presence of the step-pool structure. Even when severely burned, channels able to retain the step-pool structure experienced minimal ecological impact and recovered more quickly. The form and content of this abstract are approved. I recommend its publication. Approved: Anne Chin iv

PAGE 5

ACKNOWLEDGMENTS The completion of this thesis would not have been possible without the support and guidance of my advisor, Dr. Anne Chin. She committed boundless time and energy to my development as a writer and researcher. Her dedication, ambition, and attention to detail have been crucial in shaping my graduate career. I am also grateful to my committee members, Dr. Alison O'Dowd and Dr. Peter Anthamatten, for their expertise and advice throughout the research process, particular regarding the analytic methods used. I would also like to thank Rhonda Barton, Lauren Tyner, Rachel Gidley, Alex Key, Sam Epperly, Corine RobertsNiemann, and Dan Ben-Horin for their assistance in the field and in the lab. Several sponsors provided financial and in-kind support (to Dr. Chin) for this project: the National Science Foundation (EAR 1254989), the University of Colorado Denver, U.S. Forest Service, U.S. Geological Survey, and The Navigators. A monetary grant through the 2015 M. Gordon "Reds" Wolman Graduate Student Research Award from the Geomorphology Specialty Group of the Association of American Geographers also enabled me to complete the analysis and preparation of this thesis document. I would like to thank my parents for instilling in me a love of learning and the drive to pursue my ambitions. Finally, I am eternally grateful to my wonderful husband, Keith, for his patience and encouragement through all of my highs and lows over the last three years. Thank you for always being there for me. v

PAGE 6

TABLE OF CONTENTS CHAPTER I. INTRODUCTION AND OBJECTIVES . . . . . . . . 1 II. BACKGROUND . . . . . . . . . . . . 5 Ecology of Step-Pool Mountain Streams . . . . . . 5 Ecological Disturbances in Fluvial Systems . . . . . . 7 Wildfire as a Disturbance in Stream Ecosystems . . . . . 10 Research Questions and Hypotheses . . . . . . . 15 III. STUDY AREA . . . . . . . . . . . . . 17 Environmental Characteristics . . . . . . . . 17 Waldo Canyon Fire of 2012 . . . . . . . . . 20 Study Reaches . . . . . . . . . . . 22 IV. METHODS . . . . . . . . . . . . . 28 Data Collection: Field . . . . . . . . . . 28 Laboratory Methods . . . . . . . . . . 32 Data Analysis: Statistical and Computational . . . . . 33 V. RESULTS: RESPONSE OF BENTHIC MACROINVERTEBRATES . . 44 Description of Data . . . . . . . . . . 44 Research Question (1): Presence of Burn . . . . . . 47 Research Question (2): Severity of Burn . . . . . . 55 Research Question (3): Stream Habitat Type . . . . . . 68 Research Question (4): Change over Time . . . . . . 96 Discussion . . . . . . . . . . . . 117 VI. EXPLORING THE ECOLOGICAL RESPONSE WITHIN A CHANGING STEPPOOL MORPHOLOGY . . . . . . . . . . . 123 Integrating Bio-physical Characteristics . . . . . . 123 Results of Ordination Analysis . . . . . . . . 127 A Changing Step-Pool Morphology . . . . . . . 133 VII. SUMMARY AND CONCLUSIONS . . . . . . . . . 138 vi

PAGE 7

Summary of Major Findings . . . . . . . . . 138 Limitations of Study . . . . . . . . . . 140 Significance and Implications for Management . . . . . 143 Future Work . . . . . . . . . . . . 144 REFERENCES . . . . . . . . . . . . . . 147 APPENDICES A. Abbreviations . . . . . . . . . . . . . 157 B. Summary Statistics of Study Reaches . . . . . . . . 158 vii

PAGE 8

CHAPTER I INTRODUCTION AND OBJECTIVES Step-pool sequences are important features of stream channels that have received increasing scientific attention over the last several decades. A step-pool system develops where the size of substrate particles is large in relation to the size of the channel (Figure 1.1, Chin 1989). These stream features form during low-frequency, high-magnitude flood events that are capable of mobilizing large instream particles. Over time, large events reorganize channel substrate into an alternating pattern of steps comprising large particles and pools made up of finer materials. Step-pool sequences promote geomorphic stability, and typically remain in place for 50 years or more (Chin and Wohl 2005). The main function of step-pools is to provide hydraulic resistance in high-gradient river channels. As flow tumbles over a step into the pool below, energy is dissipated that would otherwise be translated downstream and cause excessive erosion and other potentially hazardous consequences (Chin and Wohl 2005). Figure 1.1. Schematic diagram of the profile of a step-pool sequence. Distance Elevation Pool Step 1

PAGE 9

In addition to hydraulic function, step-pool sequences are important ecological features that provide high-quality habitat for aquatic life. The range of substrate and flow types found within a step-pool system create a broad diversity of habitats capable of sustaining aquatic organisms with widely variable habitat requirements (Chin et al. 2009a). Recent research has focused on the ecological role of step-pools in maintaining abundance and diversity of benthic macroinvertebrates in mountain streams (e.g. Chung et al. 2012; Milner and Gilvear 2012) These studies extend traditional studies on step-pool channels that emphasized geomorphological processes. Nevertheless, little is known about the ecological response of step-pool streams to disturbances, discrete events that alter the condition of aquatic ecosystems. Disturbances in small headwater streams are particularly important because changes in headwaters affect the entire fluvial network downstream. This study investigates the ecological response of step-pool streams to wildfire disturbance, emphasizing changes in the assemblages of benthic macroinvertebrates in steps and pools. Benthic macroinvertebrates serve as bioindicators of stream quality (Cairns and Pratt 1993). Changes in macroinvertebrate communities therefore provide insight into the ecological response of step-pool streams after a disturbance such as wildfire. Although research on the impact of wildfire on aquatic ecosystems has recently accelerated in recognition of their increasing occurrences, the role of step-pool systems in facilitating the ecological response of streams to wildfire has not been investigated. A conceptual framework for describing the impact of wildfire on stream channels links a fire disturbance to an ecological response through interacting biophysical processes of post-fire hydrology, sediment dynamics, and channel morphology (Figure 1.2). The 2012 Waldo Canyon Fire in 2

PAGE 10

Colorado provided a case study for investigating the ecological response of step-pool streams to wildfire. This study also explores this response in the context of physical changes to the step-pool morphology of stream channels following the fire. This study was conducted within the context of ongoing research into the response of step-pool streams to the Waldo Canyon Fire, sponsored by the National Science Foundation (EAR 1254989) and the University of Colorado Denver and directed by Dr. Anne Chin of the Department of Geography and Environmental Sciences at the University of Colorado Denver. The sponsored project focused on establishing baseline data for the channel morphology and ecological character in Fall 2012, and on initial changes following the first flood season of summer 2013, including the use of terrestrial LiDAR technology to quantify geomorphological changes. This thesis research expanded the ecological components Figure 1.2. Conceptual framework displaying the interacting biophysical processes within stream channels following a wildfire. This study focuses on the ecological response, followed by an exploration of the morphological linkage between fire and ecological response. Ecological Response FIRE Sedimentologic Response Hydrologic Response Morphological Response 3

PAGE 11

including a second year of data collection during the summer of 2014 and subsequent analysis and synthesis This thesis is organized into seven chapters. Chapter II outlines the background literature relevant to this study, followed by specific research questions and hypotheses. Chapter III describes the environmental characteristics of the study area and the background to the 2012 Waldo Canyon Fire. Chapter IV outlines the methods of the study, including field and laboratory protocols, as well as analytical procedures. Chapter V presents the results to the research questions. Chapter VI investigates the morphological response of stream channels after the fire as a possible linkage between wildfire disturbance and the ecological response, using ordination to explore interactions. Chapter VII summarizes and interprets the major findings of this study. It discusses the limitations of this research, suggests implications for managing aquatic resources in areas disturbed by wildfire, and proposes future lines of inquiry. 4

PAGE 12

CHAPTER II BACKGROUND Ecology of Step-Pool Mountain Streams Small headwater step-pool streams have historically generated comparatively low interest in ecological research due to their lack of fish (Cole et al. 2003) and absence on topographic maps. These streams are typically covered by riparian vegetation canopies (Moore and Richardson 2003). Before the 1980s, consistent terminology did not exist to distinguish upland step-pool systems with coarse, rocky substrata from lowland riffle-pool systems with fine substrata (Logan and Brooker 1983). Logan and Brooker (1983) compiled several studies of the occurrence of macroinvertebrates in riffles and pools of rocky upland streams in the United States, Canada, and the United Kingdom. Although macroinvertebrate communities appeared similar between the two habitat types, they found higher densities of invertebrates in riffles, higher proportions of sensitive Ephemeroptera taxa, and lower proportions of more tolerant Diptera taxa. The authors proposed that, overall, the literature indicated that the effects of pollution may be greater in pool habitats than in riffles. The riffle-pool structures described in this study displayed similar characteristics to what are today differentiated as step-pool systems, including steep slopes, rocky substrate, and fastflowing water (Logan and Brooker 1983). It wasn't until the late 1980s and 1990s that the scientific community adopted a clearer distinction between step-pools and riffle-pools (Chin 1989). Since this time, literature regarding the role of step-pool sequences in the fluvial system has proliferated. The vast majority of studies concerning these streams have emphasized geomorphological aspects 5

PAGE 13

including their importance as hydraulic and hydrologic features (e.g., Chin and Wohl 2005; Church and Zimmermann 2007; Waters and Curran 2012). Nevertheless, interest is growing in the ecological significance of step-pool channels as unique and diverse habitats for an array aquatic species. These species range from benthic macroinvertebrates and fishes to the Rocky Mountain tailed frog (Dupuis and Friele 2006). Benthic macroinvertebrates have long served as bioindicators of stream quality, with references to biological monitoring appearing as early as 1908 (Cairns and Pratt 1993). By the 1970s, researchers began to connect the distributions of benthic macroinvertebrates to physical environmental factors such as stream structure, flow velocity, and substratum composition (e.g., Minshall and Minshall 1976; Winterbourn 1978; Reice 1980). Many studies provided strong evidence of an inverse relationship between stream size and the density and diversity of benthic macroinvertebrates (e.g., Danehy et al. 1999), whereby larger streams are less diverse and less dense in benthic macroinvertebrates. This correlation is shown through higher habitat heterogeneity in step-pool sequences, which are typically found in small headwater streams (Clausen and Biggs 1997). Studies have also clearly established the relationships between sediment and benthic communities, with excess sedimentation decreasing the abundance and richness of macroinvertebrates (Wang et al. 2013). This relationship is especially strong with very fine particles (Wood and Armitage 1997; Cole et al. 2003; Kaller and Hartman 2004). Studies attempting to use macroinvertebrates as bioindicators of overall habitat quality have found step-pool sequences to provide exceptional habitats due to their inherent stability and habitat diversity (Sullivan et al. 2004; Harvey et al. 2008; Milner and Gilvear 6

PAGE 14

2012). Wang et al. (2009) suggested that the step-pool system is the "best ecologically-sound riverbed pattern in mountain streams." They found the diversity of benthic macroinvertebrates to be several hundred times higher in step-pools than that of nearby streams. Because step-pool streams are increasingly vulnerable to a range of disturbances that include climate change, urban encroachment upon mountain fronts, and the growing frequency and magnitude of wildfires in mountain areas (Isaak et al. 2012; Holsinger et al. 2014; Simmoneaux et al. 2015), greater understanding of how step-pool streams respond to disturbances is urgently needed. Although interest in the ecological response of streams to disturbance is growing, literature regarding the response of step-pool systems to such disturbances is notably lacking. The following sections of this review outline current knowledge about stream response to disturbance. Many of the studies discussed investigated disturbances in small headwater streams, but they did not reference specifically the observed responses in relation to the role of steps and pools. Ecological Disturbances in Fluvial Systems According to Resh et al. (1988), disturbance in a lotic environment is "any relatively discrete event in time that disrupts ecosystem, community, or population structure, and that changes resources, availability of substratum, or the physical environment." Disturbance is unpredictable in terms of frequency and intensity. Stream disturbances may occur through floods, drought, fire, logging, and many other natural and anthropogenic phenomena. Disturbances regulate the form and function of streams by altering flow regime, rearranging substrata, and introducing biotic and abiotic stressors to aquatic life (Resh et al. 1988). 7

PAGE 15

Hydrologic processes such as runoff rates, peak streamflow, sediment dynamics, and bank erosion are major determinants of stream habitat quality. A wide variety of disturbances may profoundly impact these processes (Gresswell 1999). For example, unusual periods of dryness can cause flow to cease over steps, eliminating the aquatic habitat heterogeneity provided by step-pool systems. Stagnation of flow in pools during droughts may also lead to algal proliferation and depletion of dissolved oxygen, subsequently degrading habitat condition for benthic macroinvertebrates and fishes (Whitehead et al. 2009). In the 1970s, a theory of dynamic equilibrium emerged relating the occurrence of stream disturbance to the condition of macroinvertebrate communities (Connell 1978, Huston 1979). This theory suggested that the traditional concept of equilibrium exists in the absence of disturbance. A community at equilibrium in this traditional sense is shaped by competitive exclusion, dominated by successfully competitive species. When the ecosystem is at equilibrium, these species out-compete inferior competitors, and the community experiences decreased species diversity. However, ecosystem disturbances may reduce or extirpate species that dominate in the absence of disturbance, allowing the emergence of less competitive, disturbance-adapted species (Huston 1979; O'Bryan et al. 2009; Biswas and Mallik 2011). These colonizing species remain dominant when disturbances are frequent enough to regulate the growth of competitive species. The idea of dynamic equilibrium entails an intermediate disturbance regime in which the ecosystem fluctuates between equilibrium and non-equilibrium (post-disturbance) conditions. Fluctuating stream conditions limit competitive exclusion and maximize species diversity. This process enables a dynamic flux in the growth rate and dominance of successful and inferior competitors (Huston 1979; 8

PAGE 16

Resh et al. 1988; TrŽmolires 2004 ). Large floods are capable of altering the highly stable structure of established steppool systems (Chin 1989). Such alteration to stream geomorphology is highly influential in shaping the response of ecosystems to floods. The natural stability and habitat diversity of step-pool systems make them more resilient to large events. As a result, they may exhibit fewer losses to biotic abundance and diversity by providing physical complexity and "permanent features" (at least through individual events) that act as refugia for stream organisms (Pearsons et al. 1992; Sedell et al. 1990). When large flood events upset the stability of step-pool channels, intermediate flows in subsequent years are generally successful in reorganizing these streams to their natural structure (Roghair et al. 2002). The hydrologic regime has great impact on the community composition of benthic macroinvertebrates (Boulton 2003), periphyton (Clausen and Biggs 1997), and fishes (Pearsons et al. 1992; Roghair et al. 2002). Floods have the ability to disrupt stream equilibrium, remove steps and natural dams to create unstable channels, alter sediment dynamics, and severely erode streambeds (Fuller et al. 2011). Such changes to stream characteristics can significantly alter habitat quality and availability for aquatic life. Clausen and Biggs (1997) explored hydrological properties across a range of stream sizes as they relate to biotic characteristics. They recognized the role of step-pool systems in headwater streams in creating channel stability, and thus stable flows, and related this property of step-pool streams to periphyton biomass under normal flow conditions. No connection was made, however, between the stability of step-pool channels and the response of ecological communities to changes in flow regime. 9

PAGE 17

Wildfire as a Disturbance to Stream Ecosystems Research has accelerated on the effects of wildfire on stream ecosystems, even though these studies have not addressed responses in step-pool or riffle-pool mountain channels explicitly. The growing focus on wildfire has occurred as warming climates are increasingly acknowledged as a driver of a changing fire regime (Westerling et al. 2006; Schoennagel et al. 2007; Adams 2013; Lippitt et al. 2013). Literature typically categorizes the temporal trajectory of stream response to wildfire into three phases: short-term (immediate to one or two years post-fire), midterm (two to ten years), and long-term (decades to centuries) (e.g., Oliver et al. 2012; Malison and Baxter 2010a; Legleiter et al. 2003; Minshall et al. 1998). The vast majority of research on the effects of wildfire focuses on the short-term timeframe, typically two to three years following fire. The short-term response is also the focus of this research. Short-term responses, as they relate to benthic macroinvertebrates, may be classified into direct and indirect effects of fire (Figure 2.1) (Minshall 2003). Direct effects of wildfire on benthic macroinvertebrates stem from characteristics of the fire itself, such as intense heat and extended exposure to dense smoke (Minshall 2003). Indirect effects result from secondary disturbances, such as major precipitation events occurring after wildfire that induce increased runoff and sediment transport. Changes resulting from the direct effects of burning, such as increased soil hydrophobicity and loss of riparian vegetation cover, amplify the impacts of these secondary disturbances. For example, soil hydrophobicity is a direct effect of fire that itself has little, if any, impact on aquatic life. Soil hydrophobicity in burned areas decreases infiltration, however, leading to increased runoff into streams. Greater runoff may, in turn, disrupt flow regimes, input excess fine 10

PAGE 18

sediment into aquatic habitats, and alter channel morphology and particle size distribution (Gresswell 1999; Benda et al. 2003; Legleiter et al. 2003; Minshall 2003; Ryan et al. 2011; Moody et al. 2013). Although the impacts of direct effects of wildfire on stream ecology are variable, the combination of stream size with fire intensity and severity play an apparent role. Gresswell (1999) and Minshall (2003) reported minor effects in the absence of secondary disturbances such as increased runoff, sediment influx, and channel alteration. Minshall (2003) suggested, however, that direct effects may be more evident in small streams with high-intensity fires. Consistent with this suggestion, Oliver et al. (2012) found decreased macroinvertebrate Direct Effects Intense heat Extended exposure to smoke Soil hydrophobicity Decreased riparian vegetation cover " Indirect Effects Macroinvertebrate mortality Macroinvertebrate mortality Decreased infiltration Decreased coarse organic particulate matter input Increased insolation Increased runoff volume Altered macroinvertebrate communities Increased periphyton growth Increased sediment transport Altered channel morphology Altered macroinvertebrate communities Decreased average stream substrate size Altered macroinvertebrate communities Altered macroinvertebrate communities Figure 2.1. The succession of direct and indirect effects of wildfire on benthic macroinvertebrate communities (based on Minshall 2003). 11

PAGE 19

densities and percentages of sensitive taxa post-fire in a first-order stream with and without accompanying flooding and scouring. Short-term changes within ecosystems following wildfire are most likely a result of both direct and indirect effects dictated by numerous factors such as burn severity, precipitation regime, and watershed characteristics. Studies focusing on indirect effects in the short term indicate decimated (Rinne 1996) or highly altered (Minshall et al. 1997) populations of fishes and macroinvertebrates following wildfire. While the decrease in the abundance of macroinvertebrates resulting from increased runoff post-fire is widely undisputed, effects on taxa richness and diversity are less well understood. Vieira et al. (2004) witnessed a reduction of abundance and taxa richness to near-zero following a large post-fire flood event. This study found that benthic macroinvertebrate abundance was quick to recover, but richness took several years to return to pre-fire levels. Hall and Lombardozzi (2008) also found reduced richness one year after the 2002 Hayman Fire in the Colorado Front Range, but saw recovery to pre-fire levels the following year. Elsewhere in a small stream in the Lake Tahoe Basin, California, Oliver et al. (2012) found a decrease in abundance, as well as a decrease in percentage sensitive taxa in the first two years following a fire, but found no consistent results regarding taxa richness and diversity. Although research has shown inconclusive results regarding the short-term effects of wildfire on the taxa richness of benthic macroinvertebrates, studies of community composition have yielded more consistent findings. Loss of riparian vegetation cover is a direct effect of fire that leads to a shift in the major energy resources available to aquatic biota from allochthonous to autochthonous sources. Allocthonous sources are those 12

PAGE 20

originating outside of the aquatic habitat, such as course particulate organic matter (CPOM) arriving from the riparian zone. In contrast, autochthonous sources are those produced within the aquatic habitat, such as periphyton growing in streams (Moss 2010). Benthic macroinvertebrates are largely unable to survive on burned particulate matter, and thus are forced to turn to autochthonous periphyton, which proliferate with the increased insolation associated with the loss of riparian cover (Mihuc and Minshall 2005). This shift from allochthonous to autochthonous sources of organic matter favors generalist over specialist taxa. Burned streams are therefore often re-colonized by collector-gatherers, which are generalist feeders that feed on the fine particulate organic matter (FPOM) associated with post-fire sediment influx. These generalists displace more specialized shredders (Mihuc and Minshall 1995; Vieira et al. 2004; Mihuc and Minshall 2005; Oliver et al 2012). Some of the dominant taxa following wildfire in the short-term are Baetidae (Mihuc and Minshall 1995; Vieira et al. 2004), Nemouridae (Mihuc and Minshall 1995), Simuliidae, and Chironomidae (Vieira et al. 2004). Studies of the effects of wildfire at a midterm time frame are also variable. The time frame of ecological recovery depends upon many environmental factors that influence the rate of recolonization by macroinvertebrates. Such factors include topographic barriers, availability of refugia, and habitat stability (Sedell et al. 1990; Richards and Minshall 1992; Vieira et al. 2004). Habitat stability after fire depends largely on the weakening of soil hydrophobicity over time, and subsequent attenuation of the impact of hydrologic events. Studies of wildfire effects on soil hydrophobicity indicate that soil water repellency declines with time and with repeated wetting of the soil (Doerr et al. 2009). The timeframe for this 13

PAGE 21

decline is highly variable, and processes that dictate the persistence of soil hydrophobicity remain undefined. Possible factors include burn severity, vegetation cover and regrowth, soil type, slope aspect, and precipitation regime. Studies in the Colorado Front Range have seen infiltration return to pre-fire rates in the span of one to two years (Huffman et al. 2001; MacDonald and Huffman 2004), or have seen rates increase without reaching pre-fire rates after two years (Doerr et al. 2006). Studies on macroinvertebrate response in the mid-term following wildfire have found that the magnitude of response dampens over time, with initially reduced benthic macroinvertebrate abundance and diversity approaching, but not surpassing pre-fire levels within ten years (Minshall et al. 2001; Vieira et al. 2004; Mihuc and Minshall 2005). Some studies, however, yield strikingly different results from short-term studies regarding biotic response at severely burned sites. Higher abundances of macroinvertebrates and emergence rates of adult aquatic insects have been recorded at high-severity burn locations, and are associated with increased occurrence of bats and spiders in riparian zones (Malison and Baxter 2010a). This response suggests a shift in direction of the flux of resources following wildfires. Movement of nutrients, organic materials, and terrestrial organisms (Nakano et al. 1999) from land to water immediately following fires is elevated with higher runoff rates. Thus, in the midterm, aquatic productivity is bolstered, and the flux of energy shifts to movement from water to land (Malison and Baxter 2010a). Little research has focused on long-term effects of wildfire on stream ecology, and thus few datasets exist to demonstrate responses on this timescale (Moody et al. 2013). 14

PAGE 22

Research Questions and Hypotheses The review of the literature suggests the following fruitful research questions deriving from the Waldo Canyon Fire. 1. How do benthic macroinvertebrate communities in step-pools mountain streams respond to wildfire disturbance? 2. How does the response of macroinvertebrate communities to wildfire vary with the severity of burn ? 3. How does the response of macroinvertebrate communities to wildfire vary as a function of habitat type (step or pool)? 4. How does the response of macroinvertebrate communities to wildfire vary over time? Literature review and initial field observations suggest the following hypotheses. First, wildfire will disturb benthic macroinvertebrate communities Communities in burned watersheds will exhibit lower taxa richness, lower proportion of sensitive taxa, and altered composition compared to communities in unburned streams. These responses presumably result from the indirect effects of fire, such as changes in channel morphology and sediment dynamics following post-fire floods, and decreased coarse organic particulate input and increased periphyton growth (Figure 2.1). Second, ecological responses will vary between streams in watersheds burned with high and low severity. This is because increasing severity of burn is expected to magnify the impacts of wildfire on all of the response variables (Figure 2.1), including increased soil hydrophobicity, decreased infiltration, and increased runoff. Magnification of these responses is expected to generate a greater alteration in macroinvertebrate communities. 15

PAGE 23

Third, within individual step-pool habitats in burned channels, steps will retain better ecological quality compared to pools. This is because steps are morphologically stable stream features, whereas pools comprise finer, less stable substrata. Pools are therefore more susceptible to scouring during the expected post-fire floods. The morphological instability of pools is expected to result in lower ecological quality compared to steps. The enhanced ecological quality of steps is expected to be reflected in higher taxa richness and higher proportion of sensitive macroinvertebrate taxa compared to steps. Fourth, macroinvertebrate communities in burned sites are expected to reflect significant differences in richness, proportion of community comprising sensitive taxa, and composition from year to year after wildfire. Stream ecological quality of burned sites is expected to degrade from immediately after the fire in 2012 to one year post-fire in 2013, as a result of post-fire effects that include flooding and inputs of fine sediment (Figure 2.1). This degradation will be reflected in decreased taxa richness, decreased proportion of the community comprising sensitive taxa, and altered composition of macroinvertebrate communities. Ecological condition is expected to show initial signs of recovery by two years after the fire disturbance. These trends are expected to initiate the projected trajectory of recovery toward pre-fire community composition over the next decade. The next chapter describes the study area, including environmental characteristics of the watersheds of interest and details of the Waldo Canyon Fire. 16

PAGE 24

CHAPTER III STUDY AREA Environmental Characteristics The study area lies within the Pike National Forest along the eastern slope of the Colorado Front Range (Figure 3.1). This chapter briefly describes the geology, vegetation, climate, and hydrology of the area, followed by a description of the specific study reaches selected for analysis Geology The Pikes Peak batholith dominates the geology of the study area. The Pikes Peak batholith is a granitic structure spanning much of the central Colorado Front Range. Since its formation during the Middle Proterozoic Eon approximately one billion years ago, weathering and tectonic activity have exposed and uplifted the batholith. This process resulted in large areas of exposed bedrock on the steep slopes of the region (Stoeser 2005). The coarse grain of the Pikes Peak granite is conducive to weathering. Weathered granite often yields a shallow layer of coarse scree overlaying the bedrock ( Fountain Creek Watershed Study 2009). Other rock types found in the area include Ogallala shale, sandstone, gravels and alluviums, Williams Canyon limestone, and gneiss (Stoeser et al. 2005). Despite the coarse texture and rapid drainage of these types of soils, the shallow depth of impermeable bedrock places the study area largely in USGS Hydrologic Soil Group D (Natural Resources Conservation Service) 17

PAGE 25

Vegetation Ponderosa forest, mixed conifer, mixed montane shrubland, and lower montanefoothills shrubland characterize the vegetation types in the study area (Colorado Natural Heritage Program 2011). Major tree species include conifers dominated by ponderosa pine ( Pinus ponderosa ), as well as lodgepole pine ( Pinus contorta ), pinyon pine ( Pinus edulis ), and western hardwoods such as Gambel oak ( Quercus gambelii ) (USDA Forest Service and U.S. Geological Survey 2002). Ponderosa pines are fire-resistant due to their open crowns, the thickness of their bark, and high moisture content of their leaves (Howard 2003). The thin bark of lodgepole pines makes them more susceptible to fire damage, but they thrive in the aftermath of wildfire due to their serotinous cones which require high temperatures to open for seed dispersal (Anderson 2003). Pinyon pines are not well-adapted to wildfire, with thin bark and flammable foliage (Anderson 2002). Gambel oak is fairly resistant to low-severity fires and reestablishes quickly (Simonin 2000). Understory grasses and forbs in the area include: grama ( Bouteloua sp. ), western wheatgrass ( Pascopyrum smithii ), and needle-andthread ( Stipa comata ) (Fountain Creek Watershed Study 2009). Climate The area receives an average of 500 to 650 mm (approximately 20 to 25 inches) of precipitation annually, following a gradient from dry to wet as elevation increases. A notable drying trend has been documented over the last several decades (Natural Resources Conservation Service 2013). The widely variable elevation of the Front Range yields changing climate and weather patterns, however, exhibiting large spatial and temporal temperature swings, as well as unpredictable precipitation. Summer thunderstorms generate 18

PAGE 26

much of the precipitation received on the eastern slopes (i.e., the Front Range), typically peaking in July or August with monthly precipitation totals averaging 100 mm (four inches). Individual storms may produce between 25 and 75 mm of rain (one to three inches; National Water Information System). Pikes Peak, located southwest of the study area, is especially apt to produce summer thunderstorms in the Pike National Forest. At an elevation of 4,302 m 14,114 feet), nearly 3,000 m (10,000 feet) higher than the neighboring eastern plains, Pikes Peak creates orographic effects that strongly influence precipitation patterns of the surrounding area (Doesken et al. 2003). Annual average high temperatures range from approximately 10 degrees celsius (50 # F at Eagle Reach) to 15 degrees celsius (59 # F at Gage Reach). Annual average lows range from approximately -5 to 0 degrees celsius (low 20s to low 30s # F). Summer temperatures range from average highs approaching 20 degrees celsius (68 # F) at higher elevations, with overnight lows approaching freezing, to highs averaging around 30 degrees celsius (86 # F) at lower elevations. Winter highs decrease with increasing elevation from approximately 5 to 0 degrees celsius (41 # F to 32 # F), and lows from -10 to -15 degrees celsius (5 # F to 14 # F; Colorado Climate Center 2010). Hydrology Spring snowmelt and intense summer thunderstorms generate most of the streamflow in the study channels, although thunderstorms are the primary cause of flash floods in the area. Discharge typically rises sharply in late spring as perennial snowpack melts, then decreases gradually before peaking again with precipitation events (National Water Information System). U.S. Geological Survey (USGS) weather stations provide precipitation 19

PAGE 27

and discharge data at various locations within and around the study area. Additionally, the University of Colorado Denver (UCD) and the USGS jointly installed four rain gages in the study area from 2013 to 2014 for the purpose of this project (see Figure 3.3 for locations). Watersheds selected for this study occur within the Fountain Creek Basin (2,401 km 2 ; 927 mi 2 ), which eventually drains into the Arkansas River (Figure 3.1). The Fountain Creek Basin includes a portion of the city of Colorado Springs, which is situated at the base of the mountainous northwestern region of the watershed. Fountain Creek provides 15% of the city's drinking water (Fountain Creek Watershed 2015). As the population of this metropolis has burgeoned over the last several decades, development has increased the percentage of impervious area. This process has accelerated the hydrologic response to precipitation events. This development, combined with the steep slopes of the upstream area and highly variable precipitation regime put the watershed at risk for flash flooding after fire (Stogner 2000). Waldo Canyon Fire of 2012 The Waldo Canyon Fire burned for 17 days within Pike National Forest along the Front Range of Colorado. It was ignited on Saturday, 23 June 2012 at approximately 12:00 pm, roughly 15 km (nine miles) northwest of Colorado Springs, CO (Parker 2012, Figure 3.1). The U.S. Forest Service determined the fire to be human-caused, but the intent remains unknown (Steiner 2013). Originating in the Headwaters Fountain Creek basin, the fire spread into other major watersheds, including Cascade Creek-Fountain Creek, Garden of the Gods, West Monument Creek, and Lower Monument Creek before containment on 10 July 2012. Of the 74 km 2 (18,247 acres) burned, 41% were classified as low severity burn, 40% moderate severity, and 20

PAGE 28

19% high severity (Inciweb 2012). Areas burned with moderate and high severity were characterized by long-term soil damage that increased soil water repellency and the risk of erosion. These areas exhibited substantial losses of surface vegetation and ground litter, as well as destruction of roots up to 10 cm (four inches) below the soil surface (Young and Rust 2012) (Figure 3.2). Until the Black Forest Fire in the following summer (June 2013), the Waldo Canyon Fire was cited as the most destructive in Colorado's history. Initial insurance damages totaled an estimated $453 million. The fire also directly destroyed 346 homes, forced evacuation of over 32,000 people, and killed two people (Coalition for the Upper South Platte, Associated Fountain Arkansas River Creek Manitou Springs Colorado Springs Waldo Canyon Fire COLORADO Denver Waldo Canyon Fire Colorado Springs Figure 3.1. Location of the Waldo Canyon Fire northwest of Colorado Springs, Colorado. (Data sources: USGS Geospatial Multi-Agency Coordination: Wildland Fire Support; Pirmin Kalberer & Mathias Walker, Sourcepole, Switzerland ) 21

PAGE 29

Press 2013). Chin et al. (2015) provide additional details of human-environment interactions following the fire. Study Reaches Two canyons within the burned area were selected for analysis: Williams Canyon and the Camp Creek watershed (Queens Canyon) (Figure 3.3). Williams Canyon is oriented approximately north to south and flows toward Manitou Springs. Camp Creek flows southwest toward Colorado Springs. The Camp Creek watershed (20.9 km 2 8.1 mi 2 ) is larger than Williams Canyon (6.2 km 2 2.4 mi 2 ) (Rosgen et al. 2013). Channel gradients, hill slopes, and drainage areas of tributaries to Camp Creek are thus more comparable to channel reaches in Williams Canyon than are points along the main channel of Camp Creek. Figure 3.2. Fire-ravaged hillslope following the Waldo Canyon Fire. Photo by Alicia Kinoshita. 22

PAGE 30

! West Monument Creek Bear Creek Academy Hunter Gage Figure 3.3. Locations of study reaches and rain gages within the burn area, showing variations in burn severity. (Data sources: U.S. Forest Service Burned Area Emergency Response; Pirmin Kalberer & Mathias Walker, Sourcepole, Switzerland ) UCD Rain Gage Exiting Rain Gage Study Reaches Unburned Low Severity High Severity Eagle Meadow Aussie Tributary Willis Camp Creek Williams Canyon Burn Severity Very Low/Low Moderate High Upper Queens Canyon Upper Williams Canyon UCD-USGS-2 Bear Creek Colorado Springs N 23

PAGE 31

This study examines five study reaches in the two burned watersheds: Meadow Reach, Eagle Reach, and Aussie Reach within the Camp Creek watershed, and Willis Reach and Tributary Reach within Williams Canyon (Figure 3.3). Similarities in characteristics such as drainage area, elevation, channel slope, and vegetation guided the selection of the study sites (Table 3.1). All reaches contain well-defined step-pool sequences within a span of 30 to 50 meters in length (Figure 3.4). Additionally, the five study reaches were selected to represent sites of high and low burn severity. Burn severity of each site was determined by the areal percentage of the upstream watershed classified as low, moderate, and high burn severity. Reaches of low severity burn have more than 50% of the upstream watershed classified as low burn severity or unburned (Table 3.1). Reaches classified as high severity burn, for the purposes of this study, have more than 50% of the upstream watershed classified as moderate or high burn severity. Table 3.1. Characteristics of study and reference reaches. Reach Watershed Burn Severity Upstream Drainage Area (km 2 ) % High Severity Burn % Moderate Severity Burn % Low Severity Burn Elevation (m) Channel Slope* Willis Williams Canyon High 3.37 19.13 57.08 20.75 2,250 0.047 Tributary Williams Canyon High 1.09 43.95 53.62 2.40 2, 250 0.119 Aussie Camp Creek High 2.15 16.99 58.12 22.47 2,500 0.111 Eagle Camp Creek Low 1.30 6.91 37.60 47.32 2,775 0.083 Meadow Camp Creek Low 1.54 1.11 33.55 50.36 2,6 50 0.044 Gage Bear Creek Unburned 17.8 0 0 0 2,000 0.053 Hunter Bear Creek Unburned 2.70 0 0 0 2,050 0.107 Academy West Monument Creek Unburned 0.70 0 0 0 2, 175 0.217 *slopes calculated from 2012 surveys 24

PAGE 32

! The two reaches in Williams Canyon (Willis Reach and Tributary Reach) provide examples of high severity burn sites. Willis Reach lies on the main channel at an elevation of 2,250 m (U.S. Geological Survey 2013). The average channel slope along this channel reach is 0.047. The upstream drainage area of Willis Reach encompasses 3.37 km 2 Of this area, 77.83% was burned with high or moderate severity. Tributary Reach lies on a tributary of the main channel in Williams Canyon upstream of Willis Reach. The tributary of interest flows from the west. The elevation of Tributary Reach is approximately 2,250 m, and the average channel slope is 0.119. Of the 1.09 km 2 that drains into this channel reach, 97.57% of this area was burned with high or moderate severity. The three study reaches within the Camp Creek watershed (Queens Canyon) are Figure 3.4. Step-pool stream channel within the burned area. Photo by Derek Strickler. 25

PAGE 33

tributaries. Aussie Reach lies on a tributary that flows into the main channel from the north. This reach represents a third site of high burn severity, with 75.11% of the upstream area burned with high or moderate severity. Aussie Reach sits at an elevation of 2,500 m. The drainage area of this channel reach is 2.15 km 2 and the average channel slope is 0.111. Eagle Reach in the Camp Creek watershed represents a site of low burn severity. This channel reach is part of a tributary near the headwaters of the canyon at an elevation of 2,775 m. The tributary flows into the main channel from the north. Of the upstream drainage area of Eagle Reach (1.30 km 2 ), 44.51% was burned with high or moderate severity. The average channel slope of this reach is 0.083. Meadow Reach is a portion of a tributary that flows into Camp Creek from the west. This channel reach lies downstream of Eagle Reach at an elevation of approximately 2,650 m. The drainage area of Meadow Reach is 1.54 km 2 of which 34.66% was burned with high or moderate severity. The average channel slope is 0.044. In addition to the five study reaches burned by the Waldo Canyon Fire, three unburned reference reaches were selected for comparison. These reaches lie within two different watersheds: the Bear Creek watershed the West Monument Creek watershed (see Figure 3.3). Bear Creek flows toward Colorado Springs from the southwest and drains into Fountain Creek. This watershed lies south of the study area. West Monument Creek flows toward Colorado Springs from the northwest. This watershed is north of the burned area. Gage Reach lies on the main channel of Bear Creek at an elevation of 2,000 m. The drainage area of this channel reach is 17.8 km 2 and the average channel slope is 0.053. Hunter Reach is a portion of a tributary of Bear Creek flowing from the west. This reference 26

PAGE 34

reach lies upstream of Gage Reach at an elevation of 2,050 m. The upstream drainage area of this reach is 2.70 km 2 and the average channel slope is 0.107. The third reference reach, Academy Reach, is located north of the burned area on a tributary of West Monument Creek. This reach sits at an elevation of 2,175 m. The upstream drainage area is 0.70 km 2 and the average channel slope is 0.217. The next chapter outlines the methods employed in this study. It begins by describing methods of field data collection and laboratory processing, followed by a description of the statistical and computational treatment for each research question. 27

PAGE 35

CHAPTER IV METHODS The goal of this project was to investigate post-fire changes in the ecological quality of step-pool streams. Benthic macroinvertebrate samples portrayed the ecological character of stream channels and provided an indication of habitat quality. Ecological condition was monitored a few months after the fire (2012), one year after fire (2013), and two years after fire (2014). These data enabled comparison of characteristics among study sites at each time step (year) and within each study site over time. The following sections detail the methods of collection of ecological data, the laboratory procedures to process these data, and the statistical and computational procedures utilized in analysis. Data Collection: Field Within each reach of approximately 50 m, two cross sections were established at well-developed step-pool features. Benthic macroinvertebrate samples were collected from the pool and the step at each designated cross section. This procedure provided two replicates of steps and two of pools for each reach. Overall, 20 samples were collected from burned reaches and 12 from unburned reaches at each time step. Standard protocols in stream ecology guided the collection of benthic macroinvertebrate samples in this study (Barbour et al. 1999, Merritt et al. 2008). Sampling took place in a systematic manner, starting at downstream cross sections and working upstream. A 500 micron D-frame net collected samples from the channel bed in pools. Pool sampling began with substrate disturbance at the downstream margin of the pool. Maintaining constant flow into the net, the D-frame gradually moved upstream to sample the 28

PAGE 36

entire area of the pool in one minute. A flexible frame net, devised for collecting samples from steps, enabled a seal between the net and the rocks of steps (Chin et al. 2009b). Rubbing the rock surfaces and shifting rocks and debris helped to dislodge macroinvertebrates clinging to the substrate. Sampling of the step proceeded for one minute. Samples were transferred from the D-net to a plastic tub filled with stream water following collection from the stream. The net underwent visual inspection for remaining macroinvertebrates. Large rocks and organics in the sample were rinsed, inspected, and returned to the stream. The remaining sample passed through a 500 m sieve before transfer to a 32-oz plastic Nalgene for storage in 95% ethanol. From October through December 2012, benthic macroinvertebrate samples were collected from steps and pools in the following study reaches: Willis Reach in William Canyon; Aussie Reach, Meadow Reach, and Eagle Reach in Camp Creek; and reference reaches Gage Reach and Hunter Reach in the Bear Creek watershed, and Academy Reach in the tributary to West Monument Creek (Table 4.1). Samples could not be collected from Tributary Reach in 2012 due to lack of flow at time of sampling. Table 4.1. Timing of macroinvertebrate sample collection at each time step. Watershed Reach Year 2012 2013 2014 Williams Canyon Willis November August September Tributary September August Camp Creek Aussie December November October Eagle December November September Meadow December *May November Bear Creek Hunter December December December Gage December October December West Monument Creek Academy December December December *Samples were collected from Meadow Reach in May 2014 instead of Fall 2013 29

PAGE 37

Sampling efforts in 2013 and 2014 followed major precipitation events that occurred during the summer and fall seasons (Table 4.2). These storms induced post-fire floods as a result of the augmented runoff from hydrophobic soils in the burned watersheds. Post-fire flood events of 2013 highly altered stream structure in many reaches. Several original cross sections no longer encompassed step-pool sequences, but rather spanned across widened, smoothed channel segments comprising shallow riffles. In these cases, samples were collected at the nearest step-pool feature. This approach ensured consistent representation of the ecological quality of step-pool features. Sample collection from each stream feature type enabled comparison of the resilience of steps and pools, providing insight into the specific role of step-pools in stream recovery after wildfire disturbance. Data collected between late August and December 2013 included seven of the eight study reaches: Willis Reach and Tributary Reach within Williams Canyon; Aussie Reach and Eagle Reach in Camp Creek; and the three reference reaches, Gage Reach, Hunter Reach, and Academy Reach. In 2014, samples from all study reaches were collected between August and December (Table 4.1) Although data were not collected from Meadow Reach in Camp Creek in 2013 (a reach burned with low severity), this reach was sampled in May 2014 before the first major precipitation event of the summer (Table 4.1). Data collected at this time served to represent the ecological condition of Meadow Reach in 2013. The research design minimized seasonal differences in macroinvertebrate communities by calling for sample collection during spring and fall. Communities exhibit the greatest seasonal differences during summer months (Hilsenhoff 1988). The greatest physical and ecological changes over time after fire were expected to occur as a result of secondary hydrological disturbances. Data collected before 30

PAGE 38

such disturbances in 2014 were therefore assumed to be an acceptable representation of the overall character of the stream in 2013. Table 4.2. Major precipitation events in 2013 and 2014 following the Waldo Canyon Fire. Data available from existing USGS weather stations and rain gauges installed jointly by UC Denver and the USGS (see Figure 3.3 for locations). Definitions and Units Dur: Duration (minutes) Dep: Depth (mm) Avg I: Average intensity (mm/hr) 5-min peak I: 5 minute peak intensity (mm/hr) R.I.: Recurrence Interval (years) 31

PAGE 39

Laboratory Methods Macroinvertebrate samples collected in the field were returned to the laboratory at the University of Colorado Denver for processing. Sorting was the first step in processing samples in the laboratory (Barbour et al. 1999). The sample was first rinsed with water in a 500 m sieve. Upon initial inspection, large organisms were separated from leaves, sediment, and other debris. A portion of the sample was then transferred to a three-section compartmentalized petri dish and covered with ethanol. Under a dissecting microscope, each compartment of the petri dish was inspected over white and black backgrounds to separate remaining macroinvertebrates from debris. This process was repeated until the entire sample was inspected. 20 mL scintillation vials stored sorted macroinvertebrates in ethanol, and the remaining sample was archived in its original Nalgene. All benthic macroinvertebrate samples from 2012 were sorted in the laboratory and identified by Aquatic Biology Associates, Inc. (Corvalis, Oregon). Sorting and identification of samples from 2013 and 2014 occurred in the laboratory at the University of Colorado Denver. A dissecting microscope with magnification power of up to 30x enabled these macroinvertebrates to be identified. Insects were identified to the family level using an illustrated identification guide (Merritt et al. 2008). Non-insects were identified to higher taxonomic levels (subclass, class, or phylum). Identified organisms from each sample were separated by taxon in 20 mL scintillation vials and labeled with taxon name, reach name, location of collection (step or pool replicate), and date of collection. Identification of macroinvertebrates provided the data for calculating a range of metrics for statistical analysis. The following section elaborates on these metrics. 32

PAGE 40

Data Analysis: Statistical and Computational Variables and Constants Dependent Variables This study examined a range of metrics that described the composition and tolerance of macroinvertebrate communities. These metrics provided insight into the ecological condition of study streams. Table 4.3 outlines 18 candidate metrics with descriptions and predicted responses to perturbation. Of the candidate metrics, 10 metrics provide key indicators of ecological condition (Barbour et al. 1999). The metrics described in Table 4.3 fall into four general categories for characterizing macroinvertebrate communities in the study streams. These categories are measures of (1) overall richness and composition, (2) tolerance to perturbation, (3) composition of functional feeding groups, and (4) composition of habit types (Barbour et al. 1999). Overall richness and composition. Measures of overall community richness and composition include four metrics that characterize the structure of macroinvertebrate communities. The four metrics are (a) taxa richness; (b) percentage of organisms belonging to the insect orders Ephemeroptera, Plecoptera, and Trichoptera (EPT); (c) percentage of organisms belonging to the dominant taxon; and (d) percentage of organisms belonging to the family Chironomidae. These compositional metrics reflect the relative contributions (percentages) of certain populations to the total community of macroinvertebrates in a stream ecosystem. 33

PAGE 41

Table 4.3. Macroinvertebrate metrics calculated for all samples and their predicted response to perturbation (adapted from Barbour et al. 1999, Merritt et al. 2008). This table separates metrics into four categories of characteristics: overall richness and composition, tolerance, functional feeding groups, and habit types. Metric Description Predicted Response to Perturbation Taxa Richness* Total number of taxa Decrease % EPT* Percentage of individuals belonging to the orders Ephemeroptera, Plecoptera, and Trichoptera Decrease % Dominant Taxon* Percentage of individuals belonging to the single most dominant taxon Increase % Chironomidae Percentage of individuals belonging to the family Chironomidae Increase Family Biotic Index* Calculated for a community based on weighted values of tolerance to pollution Increase % Intolerant Taxa* Percentage of individuals with tolerance values $ 2 Decrease % Tolerant Taxa* Percentage of individuals with tolerance values % 8 Increase % Scraper* Percentage of individuals that scrape or graze on periphyton attached to substrate Decrease % Shredder* Percentage of individuals that shred coarse particulate organic matter (CPOM) Decrease % Collector-Gatherer Percentage of individuals that gather fine particulate organic matter (FPOM) deposited in streams Variable % Collector-Filterer* Percentage of individuals that filter FPOM from the water column Variable % Predator Percentage of individuals that feed by predation on live animals Variable % Burrower Percentage of individuals with adaptations for digging in sand or silt Variable % Climber Percentage of individuals with adaptations for climbing plants or debris Variable % Clinger* Percentage of individuals that have adaptations for attachment to surfaces in streamflow Decrease % Skater Percentage of individuals with adaptations for moving on the water's surface Variable % Sprawler Percentage of individuals with adaptations for remaining on top of fine substrate Variable % Swimmer Percentage of individuals with adaptations for clinging to submerged objects between bursts of swimming Variable *Key metrics according to Barbour et al. (1999) Habit Types Functional Feeding Groups Tolerance Richness and Composition 34

PAGE 42

Of the four metrics outlined above, three metrics are considered key indicators of stream quality: taxa richness, percentage EPT, and percentage dominant taxon (Barbour et al. 1999). Taxa richness, or the number of different taxa present in a sample, represents the diversity within a macroinvertebrate community. High taxa richness suggests a stable community and a variety of available niches within the sampled habitat. In order for measures of taxa richness to be comparable among samples from variable areas, samples were subsampled electronically to a count of 300 organisms using a subsample program provided by the Western Center for Monitoring & Assessment of Freshwater Ecosystems of Utah State University ( http://qcnr.usu.edu/ ; subsample.exe created by Dr. Dave Roberts). Percentage EPT indicates the proportion of organisms in a sample that belong to the orders Ephemeroptera, Plecoptera, and Trichoptera. Taxa in these orders are generally considered to be sensitive to pollution (although this is not always true; Merritt et al. 2008). Therefore, high percentage EPT generally suggests a healthy stream (Table 4.3). Percentage dominant taxon is another measure of diversity in a community. Communities that predominately comprise a single taxon typically indicate the dominance of a tolerant taxon that can outcompete other taxa in poor ecological conditions (Barbour et al. 1999). Therefore, a high percentage dominant taxon usually suggests poor stream quality. Chironomidae is a pollution-tolerant family of the order Diptera that frequently appears as the dominant taxon in stream samples. This family is especially prevalent following a disturbance to the aquatic environment (Mihuc and Minshall 2005). Therefore, a high percentage of Chironomidae typically indicates poor ecological stream condition. 35

PAGE 43

Tolerance to perturbation. Measures of the tolerance of macroinvertebrate communities to perturbation include (a) the Family Biotic Index; (b) percentage intolerant taxa; and (c) percentage tolerant taxa (Table 4.3). These metrics derive from the Environmental Protection Agency's Regional Tolerance Values assigned to macroinvertebrate taxa (Barbour et al. 1999), supplemented by values made available by the California Aquatic Bioassessment Workshop (CABW; 2003). These tolerance values represent the sensitivity of macroinvertebrate taxa to pollution, ranging from zero (intolerant) to ten (tolerant). The three measures of community tolerance to perturbation are considered key metrics of ecological condition (Barbour et al. 1999). The Family Biotic Index (FBI) provides weighted tolerance values for macroinvertebrate communities based on percentages of each taxon in a sample. To determine the FBI of a sample, the tolerance value assigned to a taxon is multiplied by the number of organisms of that taxon in the sample. This product is calculated for each taxon present in the sample. The total of these products is divided by the total number of macroinvertebrates to yield the weighted tolerance value that reflects the FBI for that sample. A low FBI indicates low macroinvertebrate community tolerance and high ecological quality. Percentage intolerant taxa reflects the proportion of organisms belonging to taxa that are assigned tolerance values of two or lower. Percentage tolerant taxa refers to the proportion of organisms belonging to taxa with tolerance values of eight or higher. Composition of functional feeding groups. Functional feeding groups classify macroinvertebrates by their morphological and behavioral adaptations for acquiring food (Merritt et al. 2008). The five major functional feeding groups are (a) scrapers; (b); shredders; (c) collector-gatherers; (d) collector-filterers; and (e) predators. Community 36

PAGE 44

composition based on functional feeding groups provides an indication of available food sources, which often change following a disturbance to the stream system. Of the five major functional feeding groups, three are considered key indicators of disturbance: percentage scraper, percentage shredder, and percentage collector-filterer (Table 4.3). Scrapers and shredders exhibit specialized feeding mechanisms that obligate these organisms to certain habitat conditions. Scrapers rely on periphyton attached to substrata, particularly cobble surfaces (Allan and Castillo 2007). Shredders have specialized mouthparts for shredding coarse particulate organic matter (CPOM) from allochthonous sources (e.g., leaves of riparian vegetation) (Merritt et al. 2008). Environmental disturbance typically leads to a decrease in the proportions of these two functional feeding groups in macroinvertebrate communities. Collector-gatherers and collector-filterers are generalists that feed on a wide variety of fine particulate organic matter (FPOM). These organisms are more resilient to disturbance because they do not rely on specialized food sources (Barbour et al. 1999). Percentage collector-filterer, however, is considered a key metric because this feeding group is an indicator of the type and availability of food sources from upstream. These organisms filter FPOM directly from the water column (drift FPOM), taking advantage of the movement of water to acquire food while expending little energy. As a result, this feeding group can be found in high proportions relative to other feeding groups when the transport of FPOM through a stream habitat increases (Wallace and Webster 1996). Composition of habit types. Macroinvertebrate community composition can also be characterized by macroinvertebrate habit type. Measures of habit type indicate the 37

PAGE 45

mechanisms employed by organisms for locomotion or stasis within the aquatic environment. Often, when the environment is disturbed, the altered habitat favors particular types of locomotion, leading to a change in the type of organisms found. The six metrics of habit type included in this analysis are: (a) percentage burrower; (b) percentage climber; (c) percentage clinger; (d) percentage skater; (e) percentage sprawler; and (f) percentage swimmer. Close associations exist between habit types and functional feeding groups. Swimmers, for example, are generalist collector-gatherers that acquire food by swimming. Clingers maintain position in flowing water by clinging to the substrate. This habit type correlates with collector-filterers that adhere to surfaces and filter FPOM from the water column. Clingers also include scrapers that attach to the substrate to scrape periphyton from the surface. Clingers typically occupy exposed positions in swift streamflow, placing them at high risk for displacement (Merritt et al. 2008). Percentage clinger is a key metric in characterizing ecological condition that is expected to decrease in response to perturbation (Barbour et al. 1999). Independent Variables Values of the dependent variables were expected to vary with the following independent variables: presence of burn in the upstream catchment, severity of burn, stream habitat, and time since fire event. Presence of burn was a categorical variable with values of burned (present) and unburned (absent). Burn severity was a categorical variable assuming values of high severity, low severity, and unburned. In this study, the term "burn category" is an inclusive term that refers to presence of burn and severity of burn together. Stream habitat (also called stream feature) was a nominal variable that may be defined as step or pool. 38

PAGE 46

Categorization of stream habitat types indicates differences in habitat structure and quality (representing a morphological parameter). Time since fire event was treated as a categorical variable, and expressed as immediate (2012), one year post-fire (2013), and two years postfire (2014). In addition to these independent variables that are directly related to wildfire, steppool stability is an independent variable that is indirectly affected by wildfire. This variable is expected to depend upon burn severity and post-fire hydrologic disturbance. Changes in this variable may help to explain the variation seen in macroinvertebrate communities as a function of burn severity (Minshall et al. 1998, Legleiter et al. 2003). The geomorphological response to fire is driven by changes in hydrologic regimes to yield ecological response (see Figure 1.2). Statistical and Analytical Procedures Data analysis relied on statistical procedures to test for differences in the ecological conditions among the study sites and over time. R data analysis provided the software to analyze the data derived from laboratory analysis. This statistical software package is freely available from http://www.r-project.org. Initial exploration of the data indicated non-normal distributions; nonparametric statistical tests were therefore used. Statistical tests included the Wilcoxon Rank Sum test, the Kruskal-Wallis test, the Wilcoxon Signed Rank test, and the Friedman test (Corder and Foreman 2009). These statistical procedures addressed the research questions in turn. 39

PAGE 47

Research Question 1: Presence of Burn First, how do benthic macroinvertebrate communities in step-pool streams respond to wildfire disturbance? Specifically, are the median values of macroinvertebrate metrics significantly different in burned versus unburned reaches immediately after wildfire, one year post-fire, and two years post-fire? The Wilcoxon Rank Sum test addressed this research question. The Wilcoxon Rank Sum test is a non-parametric statistical procedure that compares two independent samples by rank ordering the values from each sample and identifying whether samples cluster at opposite ends of the rank ordering (Corder and Foreman 2009). Medians are analyzed, rather than means, because this test considers the ranks of the values from two groups, rather than the values themselves. The Wilcoxon Rank Sum test was an appropriate test to address this question because each component investigated differences in macroinvertebrate metrics (Table 4.3) between two unrelated groups of reaches. The test compared values of the macroinvertebrate metrics of all samples collected in each year from burned reaches to the values of the metrics of samples from reference reaches in order to distinguish differences at each time step. Research Question 2: Burn Severity Second, how does the response of macroinvertebrate communities to wildfire vary with severity of burn? In other words, are the medians of the values of macroinvertebrate metrics significantly different among channels with high severity burn, low severity burn, and unburned reaches at each time step? The second research question used the KruskalWallis Rank Sum test to compare macroinvertebrate metrics in study reaches of varying burn severity. The Kruskal-Wallis test is a statistical procedure that compares more than two 40

PAGE 48

independent samples (Corder and Foreman 2009). This test was appropriate for addressing this question because the question investigated differences among three groups of study channel reaches that were spatially and temporally unrelated. This comparison was done within each sample collection year in order to determine at each stage of recovery whether burn severity helped to explain the magnitude of ecological response. Research Question 3: Stream Habitat Type The third research question incorporated habitat type as a possible explanatory variable for differences in macroinvertebrate metrics. The research question was: How does the response of macroinvertebrate communities vary as a function of habitat type (step or pool)? In other words, do the medians of the values of macroinvertebrate metrics differ significantly between steps and pools within each burn category? Additionally, are the medians of the values of macroinvertebrate metrics of each habitat type significantly different across burn categories? For example, do the values of the metrics of steps in burned sites differ from the values of the metrics of steps in unburned sites? Steps versus pools. The first component of this question required the pairwise Wilcoxon Signed Rank test to compare metrics of samples collected in steps versus pools. This test was appropriate because it accounted for non-independence of groups by pairing steps with adjacent pools. The Wilcoxon Signed Rank test compared steps and pools within burned study reaches at each time step. This test also compared steps and pools within sites of each burn severity, as well as within unburned study reaches. Results of this test indicated whether the specific stream feature (i.e., step or pool) influenced ecological response to wildfire. 41

PAGE 49

Habitat type across burn categories. The second component of this question used the Wilcoxon Rank Sum test to compare values of the metrics from each habitat type across burn categories. In other words, this test compared metrics from steps of burned reaches with those from steps of unburned reaches and metrics from pools of burned reaches with those from pools of unburned reaches. This component also used the Kruskal-Wallis Rank Sum test to compare samples from each habitat type across three burn severities. The results of this analysis indicated how macroinvertebrate communities within each stream feature (i.e., step or pool) responded to wildfire. Research Question 4: Change over Time The fourth research question was: How does the response of benthic macroinvertebrates to wildfire vary over time? In other words, did the median values of the macroinvertebrate metrics change significantly in the study reaches from 2012 to 2013, 2013 to 2014, or from 2012 to 2014? The Friedman Rank Sum test provided the means to compare the values of macroinvertebrate metrics at each time step within each reach. The Friedman test is a statistical procedure that compares more than two dependent samples (Corder and Foreman 2009). The Friedman test was appropriate here because the question investigated how the values of the macroinvertebrate metrics changed with repeated measures. In other words, the three groups (years) in question were not independent because each group derived from the same location (reach). Where differences over time emerged, the Friedman multiple comparisons test further identified in which years they occurred. Analyzing each reach separately effectively eliminated spatial variation as a confounding factor. From the results of this test, patterns of change among reaches were examined and plotted to provide further 42

PAGE 50

insight into the response of the dependent variables at each reach over time. The next chapter provides an overview of the data for all study reaches to characterize the central tendency and variability of the characteristics of response of macroinvertebrate communities to wildfire. It then presents the results of quantitative analysis to reveal the statistical differences and significance in the data. The statistical analyses will provide answers to the specific research questions. 43

PAGE 51

CHAPTER V RESULTS : RESPONSE OF MACROINVERTEBRATES This chapter presents the results of statistical analysis. The first section describes the data generated from processing the macroinvertebrate samples. The subsequent sections provide answers to the four research questions posed: 1. How do benthic macroinvertebrate communities in step-pools mountain streams respond to wildfire disturbance? 2. How does the response of macroinvertebrate communities to wildfire vary with the severity of burn ? 3. How does the response of macroinvertebrate communities to wildfire vary as a function of habitat type (step or pool)? 4. How does the response of macroinvertebrate communities to wildfire vary over time? The analysis focused primarily on the ten macroinvertebrate metrics the dependent variables identified as key indicators of ecological condition (see Table 4.3). Description of Data Table 5.1 displays the means and standard deviations of the ten key macroinvertebrate metrics (Table 4.3) for each study reach and at each time step (year). Appendix B provides additional summary statistics (median, range, and interquartile range) for these metrics at each reach. 44

PAGE 52

Table 5.1. Means standard deviations of the ten key macroinvertebrate metrics for each study reach at each time step (year). Refer to Figure 3.3 for locations of study reaches. Category Richness and Composition Community Tolerance Functional Feeding Group Habit Type Reach Year Willis 2012 7.3 2.1 0.8 0.8 89.6 4.2 5.8 0.1 2.1 2.5 1.4 1.6 0.0 0.0 3.4 1.5 0.6 0.5 0.6 0.5 2013 8.3 1.5 0.7 1.4 30.7 4.3 6.4 0.6 0.0 0.0 10.8 7.0 0.0 0.0 2.0 4.1 4.7 9.5 8.3 9.8 2014 10 1.4 28.3 14.5 56.4 15.0 5.4 0.3 1.1 0.3 0.03 0.0 0.0 0.0 2.3 1.0 35.6 38.5 35.2 38.5 Tributary 2013 3.5 0.6 0.0 0.0 49.5 19.5 6.4 1.1 0.0 0.0 12.5 25.0 4.2 8.3 0.0 0.0 0.0 0.0 0.0 0.0 2014 7.8 2.5 2.6 5.3 36.3 1.8 6.1 0.4 0.0 0.0 8.8 7.5 0.0 0.0 3.2 2.9 20.0 17.8 22.5 21.3 Aussie 2012 9.3 2.2 6.4 5.4 78.1 18.1 5.8 0.2 1.9 1.2 1.0 1.7 0.0 0.0 3.4 1.8 14.7 25.6 15.0 25.5 2013 12.5 2.6 8.8 7.3 66.3 27.1 5.7 0.1 3.4 2.6 2.1 2.1 0.2 0.3 5.3 2.0 16.9 20.0 17.3 20.5 2014 14.3 2.1 43.6 12.3 45.9 15.6 4.4 0.7 25.4 17.7 2.2 1.7 0.2 0.4 26.9 20.6 23.5 32.0 26.8 32.7 Eagle 2012 11.5 1.0 25.8 18.7 51.0 12.8 5.3 1.2 20.3 18.1 23.9 28.0 1.9 1.6 21.4 13.7 0.2 0.3 5.4 4.5 2013 14.3 4.6 51.4 29.3 58.3 11.3 2.5 1.0 49.9 29.7 11.4 7.6 0.03 0.06 47.8 27.5 8.0 6.7 2.7 1.7 2014 18 2.7 23.2 25.1 28.6 9.2 4.0 1.1 16.4 12.9 14.0 8.5 0.2 0.3 27.2 8.1 18.1 4.1 14.2 9.9 Meadow 2012 10.3 3.1 20.4 14.0 50.3 17.0 5.3 1.0 20.0 14.7 20.5 19.7 0.0 0.0 23.6 19.7 0.0 0.0 3.4 2.8 2013 12.5 2.9 17.6 18.8 73.3 22.6 5.3 0.8 14.2 17.3 2.4 0.9 0.1 0.2 15.7 17.4 1.8 2.0 3.6 3.1 2014 16 1.8 54.2 19.0 38.6 9.4 4.3 0.7 18.5 15.8 2.9 3.0 0.1 0.1 30.3 10.5 10.3 7.7 11.5 8.7 Hunter 2012 12.3 1.7 8.8 4.0 51.9 20.1 5.2 0.3 8.6 4.0 6.4 8.3 0.4 0.4 14.2 11.4 0.5 0.5 6.3 5.8 2013 10.5 1.3 4.0 1.3 90.4 3.4 5.9 0.03 1.6 0.6 0.2 0.2 0.2 0.2 2.7 0.5 3.7 3.3 4.1 3.3 2014 13 2.2 23.8 8.2 57.6 14.0 4.9 0.5 16.9 4.2 4.6 4.2 2.1 2.2 18.7 2.4 14.5 13.8 12.7 15.6 Gage 2012 14.3 1.7 60.5 23.0 33.5 14.3 3.8 1.0 32.4 20.5 11.0 7.2 4.5 4.0 20.4 12.8 22.4 20.9 35.3 13.8 2013 11.5 5.2 54.8 4.4 34.2 9.7 4.1 0.4 20.9 11.7 7.8 2.1 2.9 3.3 6.4 7.8 7.4 9.9 37.1 17.1 2014 16.5 2.4 62.1 18.0 39.8 12.5 4.1 0.6 26.1 14.2 5.4 1.2 1.7 1.7 22.4 12.7 23.9 17.5 25.5 17.1 Academy 2012 15.5 2.6 19.0 9.4 51.3 10.2 5.3 0.4 7.2 3.3 9.4 8.7 0.2 0.3 7.3 2.2 7.9 5.6 14.6 9.5 2013 15.5 6.5 16.4 13.6 68.1 11.4 5.6 0.4 4.3 3.6 3.2 2.5 0.1 0.1 4.8 4.2 39.3 42.4 42.7 40.2 2014 16.5 1.3 32.7 15.5 49.9 16.1 5.2 0.6 12.5 11.1 11.1 0.1 0.3 0.4 12.9 11.1 11.3 4.5 3.4 2.1 % Dominant Taxon Taxa Richness High Burn Severity % Clinger % Shredder % Tolerant Taxa Unburned % EPT Family Biotic Index Metric Low Burn Severity % CollectorFilterer % Scraper % Intolerant Taxa 45

PAGE 53

Initial qualitative assessment of these statistics suggested several trends related to the severity of burn. First, a distinct difference was apparent in the values of many metrics for sites of high burn severity compared to sites of low burn severity and unburned sites. This distinction was particularly evident for Willis Reach and Tributary Reach, the two reaches of high burn severity in Williams Canyon. The mean values of taxa richness, percentage EPT, and percentage intolerant taxa were generally lower for these sites compared to other reaches. For example, the mean percentages of EPT organisms for Willis Reach and Tributary Reach in 2012 were 0.8% and 0.0%, respectively (Table 5.1). The percentage for the third site of high burn severity (Aussie Reach) was 6.4% in 2012. In contrast, the mean percentages were 25.8% and 20.4% in reaches of low severity burn for 2012, and 8.8%, 60.5%, and 19.0% for unburned reaches. Low values of these metrics typically indicate poor ecological condition. Other metrics suggested a similar differential impact according to burn severity. Values of the Family Biotic Index for sites of high burn severity were generally higher compared to other sites. For example, mean values of the Family Biotic Index for high severity burn sites in 2012 were 5.8, 6.4, and 5.8 (Table 5.1). These values for other sites were 5.3, 5.3, 5.2, 3.8, and 5.3. A high Family Biotic Index reflects overall poor ecological condition. Additionally, proportions of scrapers and shredders, two specialized functional feeding groups, were somewhat low at sites burned with high severity. For example, scrapers were entirely absent in samples from Willis Reach for each time step. Qualitative review also suggested temporal trends in the values of the ten key macroinvertebrate metrics for ecological condition. Changes over time in these metrics were more evident in burned sites compared to unburned sites. Taxa richness, percentage EPT, and 46

PAGE 54

percentage intolerant taxa appeared, overall, to increase over time in reaches of burned watersheds, while percentage dominant taxon and the Family Biotic Index generally appeared to decrease following the fire. For example, the mean percentage of intolerant taxa at Aussie Reach increased from 1.9% in 2012 and 3.4% in 2013 to 25.4% in 2014 (Table 5.1). At Willis Reach, the trend was a decrease in ecological condition from 2012 to 2013, followed by an increase in 2014. For example, the mean value of the Family Biotic Index at Willis Reach was 5.8 in 2012, 6.4 in 2013, and 5.4 in 2014. These trends suggested that the study channels that were affected by the Waldo Canyon Fire began to recover ecologically within two years after the disturbance. Research Question (1): Presence of Burn The first research question concerned the effects of wildfire on benthic macroinvertebrate communities at each time step (year). Were the medians of macroinvertebrate metrics significantly different in burned versus unburned reaches immediately after fire, one year post-fire, and two years post-fire? Statistical analysis identified significant differences between values of the metrics of communities in step-pool streams in burned and unburned watersheds. This section is divided into three subsections, each addressing one time step. Immediately after Burn: 2012 Results of the Wilcoxon Rank Sum test showed that the median values of two key metrics of overall richness and composition were different between burned and unburned sites in 2012: taxa richness and percentage dominant taxon (Table 5.2). Taxa richness was statistically lower in burned sites (median = 10) than in unburned sites (median = 14) (p = 47

PAGE 55

0.00027). This difference indicated that the habitat provided by step-pool features of burned sites supported a less diverse community of macroinvertebrates. Percentage dominant taxon was greater in burned sites (median = 71.4%) than in unburned sites (median = 48.3%) (p = 0.017). This difference indicated that the communities in burned sites were more homogenous than those in unburned sites, with greater proportions of the benthic communities belonging to a single taxon. Of the key metrics of community tolerance, only the Family Biotic Index reflected a significant difference between burned and unburned sites (Table 5.2). Values of the Family Biotic Index were higher in burned sites (median = 5.8) than in unburned sites (median = 5.1) (p = 0.0031). This index is based on the tolerance values assigned to individual taxa (ranging from one to ten), and the percentages of those taxa in a macroinvertebrate community (Barbour et al. 1999; CABW 2003). The results suggested that communities in burned sites comprised greater proportions of organisms belonging to taxa with high tolerance values and lower proportions of organisms belonging to taxa with low tolerance values. This indicated Table 5.2. Key macroinvertebrate metrics of samples collected from burned and unburned sites in 2012, with results of the Wilcoxon Rank Sum test Category Metric Median p-value Burned Unburned Richness and Composition Taxa Richness 10 14 0.00027 *** % EPT 7.8 19.5 0.059 % Dominant Taxon 71.4 48.3 0.017 Tolerance Family Biotic Index 5.8 5.1 0.0031 ** % Intolerant Taxa 4.2 8.6 0.090 % Tolerant Taxa 2.7 6.1 0.16 Functional Feeding Groups % Scraper 0.0 0.6 0.066 % Shredder 5.2 8.7 0.30 % Collector-Filterer 0.1 5.1 0.0050 ** Habit Type % Clinger 2.0 15.2 0.0032 ** *p<0.05 **p<0.01 ***p<0.001 48

PAGE 56

higher overall tolerance of the communities in burned sites compared to unburned sites. Of the key metrics focused on functional feeding groups, the data for percentage collector-filterer was significantly lower in burned sites than in unburned sites (Figure 5.1). For burned sites, the median percentage of collector-filterers was 0.1, compared to 5.1 in unburned sites (p = 0.0050). Lower percentage collector-filterer in burned sites indicated that these communities comprised lower proportions of macroinvertebrates that feed on fine particulate organic matter suspended in the water column (drift FPOM). Low availability of drift FPOM suggests poor allochthonous food sources. For habit type, the percentage of clingers was significantly greater in unburned sites (median = 15.2%) than in burned sites (median = 2.0%) (p = 0.0032; Figure 5.1). This suggested that unburned sites supported communities in which greater proportions of Functional Feeding Groups Median Percentage 0 20 40 60 80 Scr Shr C-G C-F Pred Habit Types 0 20 40 60 80 Bu Cb Cn Sk Sp Sw Burned Unburned *denotes significant difference Figure 5.1. Median proportions of macroinvertebrate functional feeding groups and habit types in burned and unburned sites in 2012. Scr: Scraper Bu: Burrower Shr: Shredder Cb: Climber C-G: Collector-Gatherer Cn: Clinger C-F: Collector-Filterer Sk: Skater Pred: Predator Sp: Sprawler Sw: Swimmer 49

PAGE 57

macroinvertebrates had adaptations for clinging to the substrate in flowing water. Figure 5.1 displays the percentages of organisms of all functional feeding groups and habit types in burned and unburned sites for 2012, immediately after the burn. Additional (non-key) metrics reflecting significant differences between burned and unburned sites included higher percentage Chironomidae in burned sites (71.4%) than in unburned sites (36.7%) (p < 0.05). Chironomidae are organisms that are adapted to disturbance. Therefore this difference was consistent with the expected response to perturbation. The median percentage of climbers was higher in unburned sites (4.4%) than in burned sites (1.1%) (p < 0.05) These macroinvertebrates have adaptations for vertical movement on the surfaces of submerged plant material (e.g., stems, roots, and overhanging branches) (Merritt et al. 2008). Lower percentages of climbers in burned study reaches suggests, therefore, that wildfire disrupted the plant material present in these habitats. One Year Post-Fire: 2013 Of the key metrics in all categories for the year 2013, only percentage clinger was significantly different between burned and unburned sites (Table 5.3). Median values of the key measures of overall richness and composition (taxa richness, percentage EPT and percentage dominant taxon) were higher at reference sites, though the differences were not statistically significant. The three measures of community tolerance showed no difference between burned and unburned sites. Median values of these metrics were similar between burned and unburned sites. The characteristics of macroinvertebrate communities based on functional feeding groups also did not indicate any significant differences between burned and unburned sites. The key metric of macroinvertebrate habit type, percentage clinger, was 50

PAGE 58

significantly lower in burned sites (median = 1.1%) than in unburned sites (median = 17.9%) (p = 0.0022). Higher percentage clinger in unburned sites means that communities in undisturbed sites had greater proportions of organisms that cling to the substrate in flowing water. Figure 5.2 shows the proportions of functional feeding groups and habit types in burned and unburned sites in 2013, one year after fire. Other metrics that indicated significant differences between burned and unburned sites included two measures of the composition of habit types, in addition to the percentage of clingers. These differences were reflected in the median percentage sprawler (p < 0.005) and percentage swimmer (p < 0.005). Values of the percentage of sprawlers were lower unburned sites (median = 1.9%) than in burned sites (median = 15.3%). Median values of percentage swimmer were greater in unburned sites (9.5%) than in burned sites (0.3%) (Figure 5.2). Table 5.3. Key macroinvertebrate metrics of samples collected from burned and unburned sites in 2013, with results of the Wilcoxon Rank Sum test Category Metric Median p-value Burned Unburned Richness and Composition Taxa Richness 10 11.5 0.27 % EPT 3.3 12.6 0.059 % Dominant Taxon 50.1 71.1 0.39 Tolerance Family Biotic Index 5.8 5.7 0.40 % Intolerant Taxa 1.8 3.9 0.36 % Tolerant Taxa 3.2 3.1 0.55 Functional Feeding Groups % Scraper 0.0 0.1 0.057 % Shredder 4.4 2.6 0.64 % Collector-Filterer 0.6 3.6 0.15 Habit Type % Clinger 1.1 17.9 0.0022 ** *p<0.05 **p<0.01 ***p<0.001 51

PAGE 59

! Two Years Post-Fire: 2 014 In 2014 (two years after fire), metrics of community richness and composition did not exhibit significant differences in the values of burned and unburned sites (Table 5.4). The median values of taxa richness, percentage EPT, and percentage dominant taxon were similar between burned sites and reference sites. Key metrics of community tolerance differed significantly between burned and unburned sites in percentage intolerant taxa. Median percentage intolerant taxa was higher in unburned sites (15.8%) than in burned sites (6.7%) (p = 0.045). This suggested that undisturbed habitats supported greater proportions of organisms that are sensitive to pollution. Habit Types 0 10 20 30 40 50 Bu Cb Cn Sk Sp Sw Burned Unburned Functional Feeding Groups Median Percentage 0 20 40 60 80 Scr Shr C-G C-F Pred *denotes significant difference Figure 5.2. Median proportions of macroinvertebrate functional feeding groups and habit types in burned and unburned sites in 2013. Scr: Scraper Bu: Burrower Shr: Shredder Cb: Climber C-G: Collector-Gatherer Cn: Clinger C-F: Collector-Filterer Sk: Skater Pred: Predator Sp: Sprawler Sw: Swimmer 52

PAGE 60

Of the key functional feeding groups, the percentage scraper characteristic exhibited a significant difference between burned and unburned sites in 2014 (Figure 5.3). The median percentage scraper was higher in samples collected from unburned sites (0.8%) compared to burned sites (0.0%) (p = 0.000026). This means that undisturbed sites supported greater proportions of organisms that acquire food by scraping periphyton from the surface of the substrate. Analysis of macroinvertebrate habit types suggested that the median percentage of clingers did not differ between burned and unburned sites. Figure 5.3 displays the median percentages of all functional feeding groups and habit types in burned and unburned sites for 2014. Several additional metrics of functional feeding groups exhibited significant differences between the macroinvertebrate communities of burned and unburned sites for 2014. Collector-gatherer was the dominant feeding group in burned and unburned sites, but the median percentage was significantly higher in unburned sites (65.4%) than in burned sites Table 5.4. Key macroinvertebrate metrics of samples collected from burned and unburned sites in 2014, with results of the Wilcoxon Rank Sum test Category Metric Median p-value Burned Unburned Richness and Composition Taxa Richness 14 15 0.16 % EPT 36.1 37.0 0.57 % Dominant Taxon 39.2 48.5 0.31 Tolerance Family Biotic Index 5.0 4.8 0.71 % Intolerant Taxa 6.7 15.8 0.045 % Tolerant Taxa 3.7 6.2 0.11 Functional Feeding Groups % Scraper 0.0 0.8 0.000026 *** % Shredder 13.7 15.3 0.72 % Collector-Filterer 15.5 11.3 1.0 Habit Type % Clinger 10.4 6.4 0.39 *p<0.05 **p<0.01 ***p<0.001 53

PAGE 61

(46.4%) (p < 0.05). Median percentage predator was greater in burned sites (9.4%) compared to unburned sites (1.5%) (p < 0.0005; Figure 5.3). Summarizing Key Results These results suggest answers to the first research question: How do benthic macroinvertebrate communities in step-pools mountain streams respond to wildfire disturbance? Key findings from the analysis of macroinvertebrate metrics as a function of the presence or absence of burn include the following: In 2012, immediately after the Waldo Canyon Fire, the median value of taxa richness was higher in unburned sites than in burned sites. Median values of percentage dominant taxon and the Family Biotic Index were higher in burned sites than in unburned sites. The percentages of collector-filterers and clingers were higher in Functional Feeding Groups Median Percentage 0 20 40 60 80 Scr Shr C-G C-F Pred Habit Types 0 10 20 30 40 50 Bu Cb Cn Sk Sp Sw Burned Unburned *denotes significant difference Figure 5.3. Median proportions of macroinvertebrate functional feeding groups and habit types in burned and unburned sites in 2014. Scr: Scraper Bu: Burrower Shr: Shredder Cb: Climber C-G: Collector-Gatherer Cn: Clinger C-F: Collector-Filterer Sk: Skater Pred: Predator Sp: Sprawler Sw: Swimmer 54

PAGE 62

unburned sites. In 2013, the macroinvertebrate habit types of clingers and swimmers were more prevalent in unburned sites; sprawlers were more prevalent in burned sites. In 2014, the percentage of intolerant taxa was higher in unburned sites than in burned sites. The functional feeding groups of scrapers and collector-gatherers were more prevalent in unburned sites than in burned sites. Predators were more prevalent in burned sites than in unburned sites. Overall, these key results support the hypothesis that streams in burned watersheds would be negatively affected by wildfire. In particular, these results suggest a significant impact of the fire on many metrics of macroinvertebrate communities immediately after the Waldo Canyon Fire. Indications of impact, however, were not seen as prominently in 2013 and 2014. Effects on functional feeding groups and habit types were variable, in that the groups that exhibited differences between burned and unburned reaches were not consistent from year to year. Research Question (2): Severity of Burn The second research question addressed ecological response as a function of variations in burn severity. Were the median values of macroinvertebrate metrics significantly different among study sites with high severity burn, low severity burn, and those unaffected by burn at each time step? This part of the analysis separated the burned sites into categories of high and low burn severity. This arrangement produced three levels of burn severity for comparison: high severity burn, low severity burn, and unburned. The following subsections present the results of these statistical comparisons at each time step (year). 55

PAGE 63

Immediately after Burn: 2012 In 2012, macroinvertebrate communities in sites of varying burn severity exhibited statistically significant differences in the median values of each of the ten key metrics (Table 5.5; Figure 5.4). All metrics of richness and composition showed differences between the medians of high severity burn and unburned sites. For example, study channels burned with high severity exhibited low richness values (median = 8.5) compared to channels with low severity burn (median = 12) or no burn (median = 14). This difference was statistically significant between channels of high burn severity and unburned channels (p = 0.00030). Sites of high burn severity differed from sites of low burn severity as well as from unburned sites in values of percentage EPT and percentage dominant taxon. Median percentage EPT was significantly lower in sites burned with high severity (1.4%) than in sites of low burn severity (22.9%) and unburned sites (19.5%) (p = 0.0021). Median percentage dominant taxon was significantly higher in sites of high burn severity (89.2%) than in sites of low burn severity (46.4%) and unburned sites (48.3%) (p = 0.00066). The three key measures of community tolerance differed between high burn severity and unburned sites. Two metrics also differed between sites of high and low burn severity. These metrics were percentage intolerant taxa and percentage tolerant taxa (Table 5.5; Figure 5.4). Communities in channels of high burn severity exhibited significantly higher values of the Family Biotic Index (median = 5.8) than did unburned unburned (median = 5.1) (p = 0.0076). Median percentage intolerant taxa was significantly lower in sites of high burn severity (1.7%) than in sites of low burn severity (17.1%) and unburned site (8.6%) (p = 0.00063). Median percentage tolerant taxa was also significantly lower in sites of high burn 56

PAGE 64

severity (0.2%) compared to sites of low burn severity (16.6%) and unburned sites (6.1%) (p = 0.010). Analysis of the key metrics representing functional feeding group indicated that percentage scraper differed statistically among levels of burn severity (p = 0.048) (Table 5.5; Figure 5.5). However, the Kruskal-Wallis multiple comparisons test was unable to identify the burn categories between which a significant difference occurred. The median values were 0.0%, 0.0%, and 0.6% for sites of high burn severity, low burn severity, and no burn, respectively. The median percentage shredder was significantly lower in sites of high burn severity (3.5%) than in sites of low burn severity (18.2%) and unburned sites (8.7%) (p = 0.0013). Percentage collector-filterer differed statistically between low severity (median = 0.0%) and unburned sites (median = 5.1%) (p = 0.0013). Table 5.5. Key macroinvertebrate metrics of samples collected from high severity burn, low severity burn, and unburned sites in 201 2, with results of the Kruskal-Wallis test. Category Metric Median p-value High Severity Low Severity Unburned Richness and Composition Taxa Richness   8.5   12 14 0.00030 *** % EPT   1.4 22.9 19.5 0.0021 ** % Dominant Taxon   89.2 46.4 48.3 0.00066 *** Tolerance Family Biotic Index   5.8   5.3 5.1 0.0076 ** % Intolerant Taxa   1.7 17.1 8.6 0.00063 *** % Tolerant Taxa   0.2 16.6 6.1 0.010 Functional Feeding Groups % Scraper 0.0 0.0 0.6 0.048 % Shredder   3.5 18.2 8.7 0.0013 ** % Collector-Filterer   0.7   0 .0 5.1 0.0013 ** Habit Type % Clinger   0.9   3.7 15.2 0.0091 **   Identical symbols identify groups that do not differ significantly from each other *p<0.05 **p<0.01 ***p<0.001 57

PAGE 65

! A difference in macroinvertebrate habit types also occurred in the metric percentage clinger (Table 5.5; Figure 5.5). The median value of the percentage of clingers was significantly lower in sites burned with high severity (0.9%) than in unburned sites (15.2%). Figure 5.5 shows graphically these results as well as additional metrics representing functional feeding groups and habit types. Several metrics additional to the key characteristics emphasized for ecological condition (Table 4.3) exhibited significant differences among categories of burn severity. Of the additional metrics of community composition, the values of percentage Chironomidae high low unburned 6 8 10 12 14 16 18 Taxa Richness Burn Severity Number of Taxa high low unburned 0 20 40 60 Percent EPT Burn Severity Percent high low unburned 20 40 60 80 Percent Dominant Taxon Burn Severity Percent high low unburned 3 4 5 6 Family Biotic Index Burn Severity Index high low unburned 0 10 20 30 40 50 Percent Intolerant Taxa Burn Severity Percent high low unburned 0 10 20 30 40 50 Percent Tolerant Taxa Burn Severity Percent   Identical symbols identify groups that do not differ significantly from each other Figure 5.4. Metrics of community composition and tolerance in high burn severity, low burn severity, and unburned sites in 2012, with results of the Kruskal-Wallis test. The boxes encompass the interquartile range and the horizontal line within each box represents the median value. The whiskers extend beyond the boxes to reflect the variability in the data outside of the interquartile range. Open circles indicate outliers in the data.                 58

PAGE 66

were significantly higher in samples from sites of high burn severity (89.2%) compared to sites of low burn severity (41.6%) and sites of no burn (36.7%) (p < 0.005). Analysis of additional metrics of functional feeding groups suggested that the percentage of collector-gatherers was significantly higher in high severity burn sites (median = 93.9%) than in low severity burn (median = 59.6%) and unburned sites (66.6%) (p < 0.001. The percentage of predators was significantly lower in high severity burn sites (median = 1.3%) than in low severity burn (median = 8.8%) and unburned sites (median = 8.9%) (p < 0.005). Additional differences in community composition based on macroinvertebrate habit types included significantly lower percentage climber in sites burned with high severity 0 25 50 75 100 Bu Cb Cn Sk Sp Sw High Severity Low Severity Unburned Habit Types * Median Percentage 0 25 50 75 100 Scr Shr C-G C-F Pred Functional Feeding Groups * *denotes significant difference Figure 5.5. Median proportions of macroinvertebrate functional feeding groups and habit types in high severity burn, low severity burn, and unburned sites in 2012. Scr: Scraper Bu: Burrower Shr: Shredder Cb: Climber C-G: Collector-Gatherer Cn: Clinger C-F: Collector-Filterer Sk: Skater Pred: Predator Sp: Sprawler Sw: Swimmer 59

PAGE 67

(median = 0.3%) than in unburned sites (median = 4.4%) (p < 0.05). Burrowers, while dominant in all burn categories, occurred in significantly higher proportion in sites of high burn severity (median = 91.5%) than in low severity burn (median = 47.2%) and unburned sites (median = 51.2%) (p < 0.05). The metric of percentage sprawler was significantly higher in sites of low burn severity (median = 12.7%) than in sites of high burn severity (median = 1.4%) (p < 0.05), though neither level of burn severity differed from unburned sites. One Year Post-Fire: 2013 Analysis of metrics as a function of burn severity suggested that in 2013, two key metrics of richness and composition exhibited significant differences among high burn severity, low burn severity, and unburned sites. These metrics were taxa richness and percentage EPT (Table 5.6; Figure 5.6). The values of taxa richness were significantly lower in sites of high burn severity (median = 9) than in sites of low burn severity (median = 11.5) (p = 0.024). Values of the percentage of EPT organisms were significantly lower in sites of high severity burn (median = 0.0%) than in both low burn severity (median = 31.8%) and unburned sites (median = 12.6%) (p = 0.00061). Of the measures of community tolerance, the Family Biotic Index and percentage intolerant taxa differed statistically among levels of burn severity. Values of the Family Biotic Index were significantly lower in low severity burn sites (median = 4.1) than in high severity burn sites (median = 6.0) (p = 0.003). The percentage of intolerant taxa was significantly lower in sites of high burn severity (median = 0.0%) than in sites of low burn severity (median = 26.6%) and no burn (median = 3.9%) (p = 0.00046). The values of all metrics of 60

PAGE 68

composition and tolerance did not differ significantly between sites of low burn severity burn and unburned sites. Analysis of the metrics representing functional feeding groups suggested that the percentage of shredders was significantly higher in sites burned with low severity (median = 28.0%) than in either high severity burn (median = 0.0%) or unburned sites (median = 2.6%) (p = 0.0017; Figure 5.7; Table 5.6). Median values of percentage scraper and percentage collector-filterer showed no significant differences among levels of burn severity. Of the macroinvertebrate habit types, the percentage of clingers differed statistically between sites of high burn severity and no burn (Figure 5.7). The median value of percentage clinger was significantly lower in high severity burn sites (0.1%) than in unburned sites (17.9%) (p = 0.0071). Table 5.6. Key macroinvertebrate metrics of samples collected from high severity burn, low severity burn, and unburned sites in 201 3, with results of the Kruskal-Wallis test. Category Metric Median p-value High Severity Low Severity Unburned Richness and Composition Taxa Richness   9 1 2.5   1 1.5 0.024 % EPT   0.0 31.8 1 2.6 0.00061 *** % Dominant Taxon 42.4 65.4 71.1 0.14 Tolerance Family Biotic Index   6 .0 4.1   5.7 0.003 ** % Intolerant Taxa   0 .0 26.6 3.9 0.00046 *** % Tolerant Taxa 2.8 3.8 3.1 0.58 Functional Feeding Groups % Scraper 0.0 0.0 0.1 0.15 % Shredder   0 .0 28 .0   2.6 0.0017 ** % Collector-Filterer 0.0 3.0 3.6 0.14 Habit Type % Clinger   0.1   2.8 17.9 0.0071 **   Identical symbols identify groups that do not differ significantly from each other *p<0.05 **p<0.01 ***p<0.001 61

PAGE 69

! Additional metrics (Table 4.3) that exhibited statistically significant differences among high severity burn, low severity burn, and unburned sites included two habit types: percentage sprawler and percentage swimmer. Values of the percentage of sprawlers were significantly lower in unburned sites (median = 1.5%) than in low severity (median = 26.2%) and high severity burn sites (median = 13.7%) (p < 0.005). The percentag e of swimmers differed only between channels burned with high severity and unburned channels (p < 0.005). This habit type was more prevalent in samples from unburned channels (median = 9.5% compared to 0.0%). high low unburned 5 10 15 20 Taxa Richness Burn Severity Number of Taxa high low unburned 0 20 40 60 80 Percent EPT Burn Severity Percent high low unburned 20 40 60 80 Percent Dominant Taxon Burn Severity Percent high low unburned 2 3 4 5 6 7 8 Family Biotic Index Burn Severity Index high low unburned 0 20 40 60 Percent Intolerant Taxa Burn Severity Percent high low unburned 0 10 20 30 40 50 Percent Tolerant Taxa Burn Severity Percent   Identical symbols identify groups that do not differ significantly from each other Figure 5.6. Metrics of community composition and tolerance in high burn severity, low burn severity, and unburned sites in 2013, with results of the Kruskal-Wallis test.             62

PAGE 70

! Two Years Post-Fire: 2014 In 2014, statistical analysis of the key metrics of overall community richness and composition identified a significant difference among levels of burn severity in the values of taxa richness (Table 5.7; Figure 5.8). The values of taxa richness were significantly lower in high severity burn sites (median = 11) than in sites of low burn severity (median = 17.5) and unburned sites (median = 15) (p = 0.00053). The other key metrics of overall richness and composition (the percentage of EPT organisms and the percentage of the dominant taxon) exhibited no differences among levels of burn severity two years after the Waldo Canyon Fire. Functional Feeding Groups Median Percentage 0 20 40 60 80 Scr Shr C-G C-F Pred Habit Types 0 10 20 30 40 50 Bu Cb Cn Sk Sp Sw High Severity Low Severity Unburned * *denotes significant difference Figure 5.7. Median proportions of macroinvertebrate functional feeding groups and habit types in high severity burn, low severity burn, and unburned sites in 2013. Scr: Scraper Bu: Burrower Shr: Shredder Cb: Climber C-G: Collector-Gatherer Cn: Clinger C-F: Collector-Filterer Sk: Skater Pred: Predator Sp: Sprawler Sw: Swimmer 63

PAGE 71

! Table 5.7. Key macroinvertebrate metrics of samples collected from high severity burn, low severity burn, and unburned sites in 2014. Category Metric Median p-value High Severity Low Severity Unburned Richness and Composition Taxa Richness   11 1 7.5 15 0.00054 *** % EPT 22.0 25.2 36.7 0.085 % Dominant Taxon 40.4 33.7 48.5 0.093 Tolerance Family Biotic Index   5.4 4. 3   4.8 0.015 % Intolerant Taxa   1.0   13.4 15.8 0.018 % Tolerant Taxa 0.7 5.6 6.2 0.044 Functional Feeding Groups % Scraper   0 .0   0 .0 0.8 0.00012 *** % Shredder   4.6 30.2   15.3 0.0033 ** % Collector-Filterer 17.0 15.5 11.3 1.0 Habit Type % Clinger 19.4 9.7 6.4 0.64   Identical symbols identify groups that do not differ significantly from each other *p<0.05 **p<0.01 ***p<0.001 high low unburned 5 10 15 20 25 Taxa Richness Burn Severity Number of Taxa high low unburned 0 20 40 60 80 Percent EPT Burn Severity Percent high low unburned 20 30 40 50 60 70 Percent Dominant Taxon Burn Severity Percent high low unburned 3 4 5 6 Family Biotic Index Burn Severity Index high low unburned 0 10 20 30 40 Percent Intolerant Taxa Burn Severity Percent high low unburned 0 5 10 15 20 Percent Tolerant Taxa Burn Severity Percent   Identical symbols identify groups that do not differ significantly from each other Figure 5.8. Key macroinvertebrate metrics of composition and tolerance in sites of high burn severity, sites of low burn severity, and unburned sites in 2014.           64

PAGE 72

Each of the three key measures of community tolerance exhibited differences as a function of burn severity (Table 5.7; Figure 5.8). Values of the Family Biotic Index were significantly lower in low severity burn sites (median = 4.3) than in high severity burn sites (median = 5.4) (p = 0.015). The percentage of intolerant taxa was significantly lower in high severity burn sites (median = 1.0%) than in unburned sites (median = 15.8%) (p = 0.018). The Kruskal-Wallis test detected a significant difference among burn categories in the value of percentage tolerant taxa (p = 0.044), but the test failed to identify the levels of burn that differed. The median values for percentage tolerant taxa were 0.7%, 5.6%, and 6.2% for sites of high severity burn, low severity burn, and no burn, respectively. Of the key metrics of functional feeding groups, percentage scraper and percentage shredder exhibited significant differences as a function of burn severity (Table 5.7; Figure 5.9). The values of percentage scraper were higher in unburned sites (median = 0.8%) than in low (median = 0.0%) and high severity burn sites (median = 0.0%) (p = 0.00012). The values of percentage shredder were significantly higher in low severity burn sites (median = 30.2%) than in high severity burn sites (median = 4.6%) (p = 0.0033). For macroinvertebrate habit types, values of the percentage of clingers in a community did not differ significantly among levels of burn severity. Sites of varying burn severities differed in one additional metric representing functional feeding groups for 2014: the percentage of predators (Figure 5.9). Values of percentage predator were significantly higher in high severity burn (median = 10.4%) and low severity burn (median = 8.5%) sites compared to unburned sites (median = 1.5%) (p < 0.001). The only significant difference in macroinvertebrate habit types among levels of burn 65

PAGE 73

severity appeared in the metric of percentage climber, which was higher in channels of low severity burn (median = 12.0%) compared to channels burned with high severity (median = 1.9%) (p < 0.05). Summarizing Key Results In 2012, one year following the Waldo Canyon Fire, channels burned with high severity differed from unburned channels with lower taxa richness, percentage EPT, percentage intolerant taxa, percentage tolerant taxa, percentage shredder, and percentage clinger. These channels also exhibited higher percentage dominant taxon and the Family Biotic Index compared to unburned channels. High severity burn sites differed from low severity burn sites in many of the same metrics, except taxa richness, the Family Biotic Index, and percentage clinger. Low severity burn sites differed from unburned sites Functional Feeding Groups Median Percentage 0 20 40 60 80 Scr Shr C-G C-F Pred Habit Types 0 10 20 30 40 50 Bu Cb Cn Sk Sp Sw High Severity Low Severity Unburned * *denotes significant difference Figure 5.9. Median proportions of macroinvertebrate functional feeding groups and habit types in high severity burn, low severity burn, and unburned sites in 2014. Scr: Scraper Bu: Burrower Shr: Shredder Cb: Climber C-G: Collector-Gatherer Cn: Clinger C-F: Collector-Filterer Sk: Skater Pred: Predator Sp: Sprawler Sw: Swimmer 66

PAGE 74

only with a lower percentage of collector-filterers. In 2013, sites of high burn severity differed from unburned sites in many of the same metrics as in 2012: percentage EPT, percentage intolerant taxa, and percentage clinger. High severity burn sites differed from low severity burn sites with lower taxa richness, percentage EPT, percentage intolerant taxa, and percentage shredder, and higher Family Biotic Index. Low severity burn sites differed from unburned sites only with a higher percentage of shredders. In 2014, high severity burn sites differed from unburned sites in fewer metrics: taxa richness, percentage intolerant taxa, and percentage scraper were lower in sites of high burn severity. Study reaches burned with high severity differed from reaches of low severity burn with a higher Family Biotic Index and lower percentage shredder. Low severity burn sites differed from unburned sites with a lower percentage of scrapers. Regarding Research Question 2 How does the response of macroinvertebrate communities to wildfire vary with the severity of burn? the results supported the hypothesis that the severity of burn influences the magnitude of ecological response. Overall, the differences identified in this analysis suggested lower community diversity and higher tolerance to pollution in high severity burn sites when compared with low severity burn and unburned sites. Sites of high burn severity also exhibited altered proportions of functional feeding groups and habit types compared to reference sites. Low severity burn sites differed from unburned sites only in metrics of functional feeding groups. These results support the hypothesis posed for this research question. 67

PAGE 75

Research Question (3): Stream Habitat Type The third research question introduced habitat type steps and pools as an independent variable contributing to variability in the response of macroinvertebrates following wildfire. First, did the median values of macroinvertebrate metrics differ significantly between steps and pools within each burn category? Second, were the values of macroinvertebrate metrics for each habitat type significantly different across burn categories? The first subsection of this section outlines the results of comparisons for samples collected from pools with those from steps within each burn category. The second subsection considers differences among samples from each burn category within each habitat type (e.g., pools in burned sites compared with pools in unburned sites). Each subsection is further divided to address differences at each time step (year). These subsections present the results of analysis of steps and pools within all burn categories. These categories are identified as follows: burned sites (referring to the inclusion of high and low severity burn sites), high burn severity sites, low burn severity sites, and unburned sites. The analysis utilized median values because the nonparametric tests were based on the ranks of the data values for each category of burn. Steps versus Pools This subsection addresses the following question: Did the values of macroinvertebrate metrics differ significantly between steps and pools within each burn category? To answer this question, pairwise comparison of the values of macroinvertebrate metrics by habitat type identified differences between the communities of steps and pools. This analysis addressed differences between habitat types within each category of burn: burned sites, sites burned with high severity, sites burned with low severity, and unburned sites 68

PAGE 76

Immediately after Burn: 2012 For burned sites immediately after the fire, statistical analysis revealed significant differences between steps and pools. These differences were reflected in the following metrics: percentage EPT, percentage dominant taxon, percentage intolerant taxa, and percentage shredder (Table 5.8). The median value of percentage EPT was higher in steps (19.5%) than in pools (3.6%) (p = 0.016), as was the median value of percentage intolerant taxa (12.8% compared to 1.9%) (p = 0.0078). The values of percentage dominant taxon were lower in steps (median = 52.9%) than in pools (median = 81.4%) (p = 0.0078). Percentage shredder was higher in steps (median = 12.7%) than in pools (median = 4.6%) (p = 0.039). Additional significant differences the values of metrics for steps and pools in burned sites included percentage Chironomidae (p < 0.05), percentage collector-gatherer (p < 0.05), percentage burrower (p < 0.05), and percentage sprawler (p < 0.05). The median value of percentage Chironomidae was higher in pools (81.4%) than in steps (43.9%). Percentages of collector-gatherers were higher in pools (median = 87.7%) compared to steps (median = Table 5.8. Key macroinvertebrate metrics of samples collected from steps and pools of burned sites in 2012, with results of the Wilcoxon Signed Rank test Category Metric Median p-value Pool Step Richness and Composition Taxa Richness 10 10 0.35 % EPT 3.6 19.5 0.016 % Dominant Taxon 81.4 52.9 0.0078 ** Tolerance Family Biotic Index 5.8 5.6 0.11 % Intolerant Taxa 1.9 12.8 0.0078 ** % Tolerant Taxa 1.9 2.7 0.55 Functional Feeding Groups % Scraper 0.0 0.0 0.37 % Shredder 4.6 12.7 0.039 % Collector-Filterer 0.0 0.5 0.11 Habit Type % Clinger 0.9 4.4 0.20 *p<0.05 **p<0.01 ***p<0.001 69

PAGE 77

59.5%), as were the percentages of burrowers (median = 84.1% in pools versus 53.2% in steps). The median value of percentage sprawler was higher in steps (8.7%) than in pools (1.6%). Analysis of the metrics of steps and pools within sites of high and low burn severity did not detect any significant differences in the composition of habit types in macroinvertebrate communities for 2012. Tables 5.9 and 5.10 display the results of these analyses for the ten key macroinvertebrate metrics. For unburned reaches in 2012, immediately after the fire, values for metrics representing overall richness and composition, community tolerance, and habit type exhibited no significant differences between steps and pools (Table 5.11). The only significant difference identified in metrics of functional feeding groups was a greater median value of percentage collector-filterer in steps than in pools (10.7% and 2.3%, respectively) Table 5.9. Key macroinvertebrate metrics of samples collected from steps and pools of high severity burn sites in 2012, with results of the Wilcoxon Signed Rank test Category Metric Median p-value Pool Step Richness and Composition Taxa Richness 6.5 10 0.25 % EPT 0.7 5.6 0.25 % Dominant Taxon 92.4 80.5 0.13 Tolerance Family Biotic Index 5.8 5.7 0.25 % Intolerant Taxa 0.6 2.8 0.13 % Tolerant Taxa 0.2 1.3 1.0 Functional Feeding Groups % Scraper 0.0 0.0 n/a % Shredder 3.0 3.7 0.63 % Collector-Filterer 0.3 3.2 0.25 Habit Type % Clinger 0.5 3.5 0.25 *p<0.05 **p<0.01 ***p<0.001 percentage scraper in all sites was zero 70

PAGE 78

One Year Post-Fire: 2013 Analysis of samples collected from steps and pools of burned sites in 2013, one year post-fire, yielded no significant difference in any macroinvertebrate metric (Table 5.12). Similarly, no differences were detected in the characteristics of macroinvertebrate samples from steps and pools in either high or low severity burn sites (Tables 5.13 and 5.14). Table 5.10. Key macroinvertebrate metrics of samples collected from steps and pools of low severity burn sites in 2012, with results of the Wilcoxon Signed Rank test Category Metric Median p-value Pool Step Richness and Composition Taxa Richness 11 12 1.0 % EPT 10.3 31.6 0.13 % Dominant Taxon 57.0 43.9 0.13 Tolerance Family Biotic Index 5.9 4.7 0.25 % Intolerant Taxa 7.7 31.6 0.13 % Tolerant Taxa 24.2 15.1 0.88 Functional Feeding Groups % Scraper 0.0 1.6 0.37 % Shredder 9.4 38.6 0.13 % Collector-Filterer 0.0 0.0 1.0 Habit Type % Clinger 3.7 5.3 0.63 *p<0.05 **p<0.01 ***p<0.001 Table 5.11. Key metrics of for steps and pools of unburned sites in 2012, with results of the Wilcoxon Signed Rank test Category Metric Median p-value Pool Step Richness and Composition Taxa Richness 14.5 13.5 1.0 % EPT 20.8 19.5 0.84 % Dominant Taxon 48.3 47.2 0.69 Tolerance Family Biotic Index 5.0 5.1 0.56 % Intolerant Taxa 10.3 7.7 0.44 % Tolerant Taxa 6.2 5.0 0.69 Functional Feeding Groups % Scraper 0.7 0.3 0.11 % Shredder 8.7 11.1 0.84 % Collector-Filterer 2.3 10.7 0.031 Habit Type % Clinger 9.0 21.2 0.063 *p<0.05 **p<0.01 ***p<0.001 71

PAGE 79

For unburned study reaches, statistical analysis did not detect significant differences in key metrics between steps and pools in 2013, one year after the fire (Table 5.15). The one exception was the median percentage collector-filterer, a key metric representing functional feeding groups. Steps contained higher percentage collector-filterer (14.8%) compared to pools (1.2%) (p = 0.031). Medians of the metrics of richness and composition, community Table 5.12. Key metrics for steps and pools of burned sites in 2013, with results of the Wilcoxon Signed Rank test Category Metric Median p-value Pool Step Richness and Composition Taxa Richness 9.5 12 0.28 % EPT 2.6 14.5 0.15 % Dominant Taxon 66.4 44.8 0.32 Tolerance Family Biotic Index 5.9 5.5 0.28 % Intolerant Taxa 1.4 4.9 0.21 % Tolerant Taxa 2.2 4.1 0.81 Functional Feeding Groups % Scraper 0.0 0.0 0.36 % Shredder 3.3 7.4 0.27 % Collector-Filterer 0.1 3.3 0.11 Habit Type % Clinger 0.8 4.6 0.080 *p<0.05 **p<0.01 ***p<0.001 Table 5.13. Key metrics for steps and pools of high severity burn sites in 2013, with results of the Wilcoxon Signed Rank test Category Metric Median p-value Pool Step Richness and Composition Taxa Richness 7.5 9 0.28 % EPT 0.0 1.4 0.18 % Dominant Taxon 35.7 42.9 1.0 Tolerance Family Biotic Index 6.0 5.8 0.56 % Intolerant Taxa 0.0 0.0 0.37 % Tolerant Taxa 0.5 4.1 1.0 Functional Feeding Groups % Scraper 0.0 0.0 0.37 % Shredder 0.0 1.3 0.80 % Collector-Filterer 0.0 9.5 0.18 Habit Type % Clinger 0.1 9.5 0.20 *p<0.05 **p<0.01 ***p<0.001 72

PAGE 80

tolerance, and macroinvertebrate habit types did not differ between steps and pools. Unburned sites did, however, show differences in one additional metric for 2013. The percentage of predators was higher in pools than in steps (median = 2.7% in pools compared to 1.2% in steps; p < 0.05). Table 5.14. Key metrics for steps and pools of low severity burn sites in 2013, with results of the Wilcoxon Signed Rank test Category Metric Median p-value Pool Step Richness and Composition Taxa Richness 11 14.5 0.88 % EPT 6.8 51.1 0.38 % Dominant Taxon 80.1 50.1 0.13 Tolerance Family Biotic Index 4.9 3.3 0.25 % Intolerant Taxa 4.4 48.2 0.38 % Tolerant Taxa 3.8 4.7 1.0 Functional Feeding Groups % Scraper 0.0 0.0 1.0 % Shredder 7.2 44.1 0.38 % Collector-Filterer 2.2 41.0 0.88 Habit Type % Clinger 1.1 4.6 0.13 *p<0.05 **p<0.01 ***p<0.001 Table 5.15. Key metrics for steps and pools of unburned sites in 2013, with results of the Wilcoxon Signed Rank test Category Metric Median p-value Pool Step Richness and Composition Taxa Richness 11.5 11.5 0.69 % EPT 26.5 6.4 0.84 % Dominant Taxon 60.7 75.5 0.84 Tolerance Family Biotic Index 5.3 5.8 0.14 % Intolerant Taxa 7.2 2.1 0.44 % Tolerant Taxa 5.4 1.1 0.31 Functional Feeding Groups % Scraper 0.2 0.04 0.10 % Shredder 4.2 2.5 0.84 % Collector-Filterer 1.2 14.8 0.031 Habit Type % Clinger 8.0 30.0 0.44 *p<0.05 **p<0.01 ***p<0.001 73

PAGE 81

Two Years Post-Fire: 2014 Steps and pools of burned sites only differed statistically in two metrics in 2014, two years after fire (Table 5.16). Of the key functional feeding groups, the percentage of collector-filterers was significantly higher in steps (median = 29.0%) than in pools (median = 4.1%) (p = 0.014). The percentage of clingers was also significantly higher in steps (median = 31.4%) compared to pools (median = 3.7%) (p = 0.0020). Analysis of additional macroinvertebrate metrics (Table 4.3) indicated that two of these metrics differed statistically between steps and pools of burned channels in 2014. Of the metrics of overall community composition, percentage Chironomidae was significantly higher in pools of burned reaches than in steps (median = 29.4% in pools versus 9.3% in steps; p < 0.05). In addition to percentage clinger, macroinvertebrate habit types differed between steps and pools in the percentage of burrowers. The median value of percentage burrower was greater in pools of burned channels (48.7%) than in steps (18.4%) (p < 0.005) For samples from high severity burn sites in 2014, statistical analysis detected Table 5.16. Key metrics for steps and pools of burned sites in 201 4, with results of the Wilcoxon Signed Rank test. Category Metric Median p-value Pool Step Richness and Composition Taxa Richness 14 13 0.84 % EPT 36.1 38.7 0.15 % Dominant Taxon 37.6 40.9 0.23 Tolerance Family Biotic Index 5.0 4.9 1.0 % Intolerant Taxa 5.0 10.3 0.44 % Tolerant Taxa 5.3 0.8 0.32 Functional Feeding Groups % Scraper 0.0 0.0 0.18 % Shredder 27.2 9.3 0.084 % Collector-Filterer 4.1 29.0 0.014 Habit Type % Clinger 3.7 31.4 0.0020 ** *p<0.05 **p<0.01 ***p<0.001 74

PAGE 82

significant differences only in the values of percentage collector-filterer, percentage clinger, and percentage sprawler which represent functional feeding groups and habit types (Table 5.17). The values of percentage collector-filterer and percentage clinger were significantly higher in steps than in pools (median = 48.7% in steps and 1.1% in pools for percentage collector-filterer, p = 0031; median = 53.8% in steps and 2.8% in pools for percentage clinger, p = 0.031). The value of percentage sprawler was significantly higher in pools (median = 37.8%) than in steps (median= 8.8%) (p < 0.05). For channels burned with low severity, statistical analysis detected no significant differences between steps and pools in any metric for 2014 (Table 5.18). For unburned sites for 2014, statistical analysis identified differences in only two key metrics (Table 5.19). Of the metrics of community tolerance, the median value of percentage tolerant taxa was significantly greater in pools than in steps (8.2% versus 4.3%, respectively) (p = 0.031). Macroinvertebrate habit types differed in median percentage clinger. Clingers were more prevalent in steps than in pools of unburned sites (median = 21.7% in steps Table 5.17. Key metrics for steps and pools of high severity burn sites in 201 4, with results of the Wilcoxon Signed Rank test. Category Metric Median p-value Pool Step Richness and Composition Taxa Richness 10.5 11 1.0 % EPT 40.7 15.9 0.11 % Dominant Taxon 39.4 51.5 0.31 Tolerance Family Biotic Index 5.1 5.6 0.059 % Intolerant Taxa 0.9 1.3 0.58 % Tolerant Taxa 2.1 0.5 0.56 Functional Feeding Groups % Scraper 0.0 0.0 1.0 % Shredder 4.6 4.0 0.094 % Collector-Filterer 1.1 48.7 0.031 Habit Type % Clinger 2.8 53.8 0.031 *p<0.05 **p<0.01 ***p<0.001 75

PAGE 83

compared to 4.3% in pools) (p = 0.031). Percentages of key functional feeding groups did not differ between habitat types. Of the additional metrics of functional feeding groups, however, percentage collector-gatherer was greater in pools (median = 71.4%) compared to steps (median = 53.4%) (p < 0.05) Table 5.18. Key metrics for steps and pools of low severity burn sites in 2014, with results of the Wilcoxon Signed Rank test Category Metric Median p-value Pool Step Richness and Composition Taxa Richness 17 17.5 0.88 % EPT 32.8 62.7 0.13 % Dominant Taxon 29.4 38.0 0.66 Tolerance Family Biotic Index 4.7 3.6 0.13 % Intolerant Taxa 8.4 25.2 0.13 % Tolerant Taxa 12.9 2.6 0.13 Functional Feeding Groups % Scraper 0.0 0.1 0.37 % Shredder 31.1 25.5 0.88 % Collector-Filterer 12.6 15.5 0.88 Habit Type % Clinger 5.6 22.6 0.13 *p<0.05 **p<0.01 ***p<0.001 Table 5.19. Key metrics for steps and pools of unburned sites in 2014, with results of the Wilcoxon Signed Rank test Category Metric Median p-value Pool Step Richness and Composition Taxa Richness 15 16 0.83 % EPT 34.7 37.0 0.69 % Dominant Taxon 58.3 39.9 0.44 Tolerance Family Biotic Index 5.1 4.6 1.0 % Intolerant Taxa 13.7 18.7 0.44 % Tolerant Taxa 8.2 4.3 0.031 Functional Feeding Groups % Scraper 0.7 0.8 0.69 % Shredder 13.5 18.3 0.44 % Collector-Filterer 8.9 15.7 0.22 Habit Type % Clinger 4.3 21.7 0.031 *p<0.05 **p<0.01 ***p<0.001 76

PAGE 84

Habitat Type across Burn Categories This subsection addresses the question: Were the characteristics of macroinvertebrate communities for each habitat type (step and pool) significantly different across burn categories? To answer this question, statistical analysis compared metrics (Table 4.3) within each habitat type (step and pool) across categories of burn severity. Subsections address the analysis at each time step (year). These subsections are further divided to present comparisons between burned and unburned sites (presence of burn) and between high severity, low severity, and unburned sites (burn severity). Immediately after Burn: 2012 Presence of burn. For 2012, immediately after the Waldo Canyon Fire, statistical analysis identified significant differences between samples from pools of burned and unburned sites in the three key metrics of overall richness and composition: taxa richness, percentage EPT, and percentage dominant taxon (Table 5.20). The median value of taxa richness was higher in pools of unburned sites than in pools of burned sites (14.5 versus 10, respectively) (p = 0.019). The percentage of EPT organisms was also higher in pools of unburned sites than in pools of burned sites (median = 20.8% compared to 3.6%) (p = 0.043). The median value of percentage dominant taxon was higher in pools of burned sites (81.4%) than in pools of unburned sites (48.3%) (p = 0.043). Two key metrics of community tolerance differed statistically between pools of burned and unburned sites for 2012 (Table 5.20). The median value of the Family Biotic Index was significantly lower in pools of unburned channels than in pools of burned channels (5.0 compared to 5.8, respectively) (p = 0.030). The percentage of intolerant taxa was 77

PAGE 85

significantly higher in pools of unburned reaches than in pools of burned reaches (median = 10.3% compared to 1.9%, respectively) (p = 0.043). Of the metrics representing functional feeding groups and habit types, median values of percentage collector-filterer and percentage clinger were greater in pools of unburned sites compared with pools of burned sites (2.3% versus 0.0% for collector-filterers, p = 0.022; 9.0% versus 0.9% for clingers, p = 0.020). Additionally, pools of burned sites exhibited higher median percentage Chironomidae (81.4%) compared to pools of unburned sites (39.2%) (p < 0.05). For 2012, immediately after fire, steps of burned and unburned sites differed statistically in three key metrics: taxa richness, percentage tolerant taxa, and percentage clinger (Table 5.21). The median value of taxa richness was significantly higher in steps of unburned reaches (13.5) compared to steps of burned reaches (10) (p = 0.0065). Percentage tolerant taxa was also higher in steps of unburned reaches than in steps of burned reaches Table 5.20. Key macroinvertebrate metrics in pools of burned and unburned sites in 2012, with results of the Wilcoxon Rank Sum test Category Metric Median p-value Burned Unburned Richness and Composition Taxa Richness 10 14.5 0.019 % EPT 3.6 20.8 0.043 % Dominant Taxon 81.4 48.3 0.043 Tolerance Family Biotic Index 5.8 5.0 0.030 % Intolerant Taxa 1.9 10.3 0.043 % Tolerant Taxa 1.9 6.2 0.33 Functional Feeding Groups % Scraper 0.0 0.7 0.060 % Shredder 4.6 8.7 0.18 % Collector-Filterer 0.0 2.3 0.022 Habit Type % Clinger 0.9 9.0 0.020 *p<0.05 **p<0.01 ***p<0.001 78

PAGE 86

(median = 5.0% versus 2.7%, respectively; p = 0.0012), as was percentage clinger (median = 21.2% versus 4.4%, respectively; p = 0.043). Burn severity. For 2012, analysis of pools in high severity burn, low severity burn, and unburned sites identified the following differences in the three key metrics of overall richness and composition: taxa richness, percentage EPT, and percentage dominant taxon (Table 5.22; Figure 5.10). The median value of taxa richness was lower in pools of high severity burn sites (6.5) than in pools of unburned sites (14.5) (p = 0.023). The percentage of EPT organisms was also significantly lower in pools of sites with high severity burn compared to pools of unburned sites (median = 0.7% compared to 20.8%, respectively) (p = 0.015). The median value of percentage dominant taxon was higher in pools of high severity burn sites (92.3%) than in pools of unburned sites (48.3%) (p = 0.015). One key metric of community tolerance differed among pools of varying burn severities for 2012 (Table 5.22; Figure 5.10). The median value of percentage intolerant taxa Table 5.21. Key macroinvertebrate metrics in steps of burned and unburned sites in 2012, with results of the Wilcoxon Rank Sum test Category Metric Median p-value Burned Unburned Richness and Composition Taxa Richness 10 13.5 0.0065 ** % EPT 19.5 19.5 1.0 % Dominant Taxon 52.9 47.2 0.35 Tolerance Family Biotic Index 5.6 5.1 0.14 % Intolerant Taxa 12.8 7.7 1.0 % Tolerant Taxa 2.7 5.0 0.012 Functional Feeding Groups % Scraper 0.0 0.3 0.50 % Shredder 12.7 11.1 0.95 % Collector-Filterer 0.5 10.7 0.14 Habit Type % Clinger 4.4 21.2 0.043 *p<0.05 **p<0.01 ***p<0.001 79

PAGE 87

was significantly lower in pools of channels burned with high severity (0.6%) than in pools of unburned channels (10.3%) (p = 0.015). Table 5.22. Key macroinvertebrate metrics in pools of high severity burn, low severity burn, and unburned sites in 2012, with results of the Kruskal-Wallis test Category Metric Median p-value High Severity Low Severity Unburned Richness and Composition Taxa Richness   6 .5   11 14. 5 0.023 % EPT   0.7   10.3 20.8 0.015 % Dominant Taxon   92.3   57 .0 48.3 0.015 Tolerance Family Biotic Index 5.8 5.9 5.0 0.090 % Intolerant Taxa   0.6   7.7 10.3 0.015 % Tolerant Taxa 0.2 24.2 6.2 0.073 Functional Feeding Groups % Scraper 0.0 0.0 0.7 0.12 % Shredder 3.0 9.4 8.7 0.076 % Collector-Filterer   0.3   0 .0 2.3 0.028 Habit Type % Clinger   0.5   3.7 9 .0 0.021   Identical symbols identify groups that do not differ significantly from each other *p<0.05 **p<0.01 ***p<0.001 high low unburned 6 8 10 14 Taxa Richness Burn Severity Number of Taxa high low unburned 0 20 40 60 Percent EPT Burn Severity Percent high low unburned 20 40 60 80 Percent Dominant Taxon Burn Severity Percent high low unburned 3 4 5 6 Family Biotic Index Burn Severity Index high low unburned 0 10 20 30 40 50 Percent Intolerant Taxa Burn Severity Percent high low unburned 0 10 20 30 40 Percent Tolerant Taxa Burn Severity Percent                   Identical symbols identify groups that do not differ significantly from each other Figure 5.10. Key metrics of composition and tolerance of macroinvertebrate communities in pools of sites of varying burn severities in 2012, with results of the Kruskal-Wallis test. 80

PAGE 88

Of the key functional feeding groups, percentage collector-filterer was lower in pools of low severity burn sites (median = 0.0%) than in pools of unburned sites (median = 2.3%) (p = 0028). Median percentage clinger was lower in pools of high severity burn sites (0.5%) compared to those of unburned sites (9.0%) (p = 0.021). Additional metrics of community composition that exhibited significant differences among pools of varying burn severities included the percentage of Chironomidae organisms (Table 4.3). Values of percentage Chironomidae were higher in pools of high severity burn sites than in pools of unburned sites (median = 92.3% versus 39.2%, respectively; p < 0.05). Chironomidae are a family of macroinvertebrates that are relatively tolerant of pollution and adapted to disturbance. For additional functional feeding groups, the values of percentage predator were significantly higher in pools of low severity burn sites than in pools of high severity burn sites (median = 12.3% versus 0.9%, respectively; p < 0.005). In addition to percentage clinger, the percentage of burrowers also differed between high severity burn sites and unburned sites. Values of percentage burrower were significantly higher in pools of high severity burn sites (median = 96.3% for sites of high burn severity versus 55.0% for unburned sites; p < 0.05). Statistical analysis for 2012 suggested similar significant differences among steps of varying burn severities. These differences were reflected in the three key metrics of community composition: taxa richness, percentage EPT, percentage dominant taxon (Table 5.23; Figure 5.11). Values of taxa richness were significantly higher in steps of unburned sites (median = 13.5) than in those of high severity burn sites (median = 10) (p = 0.013). Values of the percentage of EPT organisms were higher in steps of sites burned with low 81

PAGE 89

severity (median = 31.6%) than in steps of sites burned with high severity (median = 5.6%) (p = 0.026). Although the percentage of the dominant taxon was found to differ among levels of burn severity (p = 0.036), the multiple comparisons test failed to identify which levels differed. The median value for percentage dominant taxon in steps was 80.5% for sites burned with high severity, 43.9% for sites burned with low severity, and 47.2% for unburned sites. Of the key metrics of community tolerance, percentages of intolerant taxa differed statistically among steps of varying burn severities in 2012, immediately following the fire (Table 5.23; Figure 5.11). Values of the percentage of intolerant taxa were significantly lower in steps of high severity burn sites (median = 2.8%) compared to steps of low severity burn sites (median = 31.6%) (p = 0.0058). Values of the Family Biotic Index and the percentages of tolerant taxa showed no statistical differences among steps of varying burn severity in 2012. Table 5.23. Key macroinvertebrate metrics in steps of high severity burn, low severity burn, and unburned sites in 2012, with results of the Kruskal-Wallis test Category Metric Median p-value High Severity Low Severity Unburned Richness and Composition Taxa Richness   10   12 13.5 0.013 % EPT   5.6 31.6   19.5 0.026 % Dominant Taxon 80.5 43.9 47.2 0.036 Tolerance Family Biotic Index 5.7 4.7 5.1 0.072 % Intolerant Taxa   2.8 31.6   7.7 0.0058 ** % Tolerant Taxa 1.3 15.1 5.0 0.10 Functional Feeding Groups % Scraper 0.0 1.6 0.3 0.25 % Shredder   3.7 38.6   11.1 0.0075 ** % Collector-Filterer   3.2   0 .0 10.7 0.037 Habit Type % Clinger 3.2 5.3 21.2 0.11   Identical symbols identify groups that do not differ significantly from each other *p<0.05 **p<0.01 ***p<0.001 82

PAGE 90

! Of the key metrics representing functional feeding groups, statistical analysis identified significant differences among steps of varying burn severities in the percentages of shredders and collector-filterers (Table 5.23). The median value of percentage shredder was significantly higher in steps of low severity burn sites (38.6%) than in steps of high severity burn sites (3.7%) (p = 0.0075). The median value of percentage collector-filterer was significantly higher in steps of unburned sites (10.7%) than in those of low severity burn sites (0.0%) (p = 0.037). Macroinvertebrate habit types differed only in the percentages of sprawlers. Values of percentage sprawler were significantly greater in steps of low severity burn sites (median = 31.4%) than in those of either high severity burn or unburned sites (median = 2.2% and 2.6%, respectively) (p < 0.05). high low unburned 6 8 10 14 18 Taxa Richness Burn Severity Number of Taxa high low unburned 0 20 40 60 Percent EPT Burn Severity Percent high low unburned 30 50 70 90 Percent Dominant Taxon Burn Severity Percent high low unburned 4.0 4.5 5.0 5.5 6.0 Family Biotic Index Burn Severity Index high low unburned 10 20 30 40 Percent Intolerant Taxa Burn Severity Percent high low unburned 0 10 30 50 Percent Tolerant Taxa Burn Severity Percent               Identical symbols identify groups that do not differ significantly from each other Figure 5.11. Key metrics of composition and tolerance of macroinvertebrate communities in steps of varying burn severities in 2012, with results of the Kruskal-Wallis test. 83

PAGE 91

One Year Post-Fire: 2013 Presence of burn. For 2013, the key metrics of richness and composition, community tolerance, and functional feeding groups showed no statistically significant differences between burned and unburned sites in either pools or steps (Tables 5.24 and 5.25). In pools statistical analysis identified significant differences between burned and unburned sites in two metrics representing macroinvertebrate habit types: percentage clinger and percentage swimmer. Pools of unburned sites exhibited significantly higher values of the percentage of clingers than pools of burned sites (median = 8.0% versus 0.9%, respectively) (p = 0.011). Values of the percentage of swimmers were significantly higher in pools of unburned sites than in pools of burned sites (median = 18.6% compared to 9.6%, respectively) (p < 0.05). For steps statistical analysis identified no significant differences between burned and unburned sites in any of the ten key macroinvertebrate metrics (Table 5.25). Of all the metrics (Table 4.3), analysis only suggested a significant difference between communities in Table 5.24. Key macroinvertebrate metrics in pools of burned and unburned sites in 2013, with results of the Wilcoxon Rank Sum test Category Metric Median p-value Burned Unburned Richness and Composition Taxa Richness 9.5 11.5 0.21 % EPT 2.6 26.5 0.056 % Dominant Taxon 66.4 60.7 0.71 Tolerance Family Biotic Index 5.9 5.3 0.17 % Intolerant Taxa 1.4 7.2 0.14 % Tolerant Taxa 2.2 5.4 0.79 Functional Feeding Groups % Scraper 0.0 0.2 0.26 % Shredder 3.3 4.2 0.87 % Collector-Filterer 0.1 1.2 0.40 Habit Type % Clinger 0.9 8.0 0.011 *p<0.05 **p<0.01 ***p<0.001 84

PAGE 92

the steps of burned and unburned sites in one measure of macroinvertebrate habit types: the percentage of sprawlers. The values of percentage sprawler were significantly higher in steps of burned channels (median = 16.4%) than in steps of unburned channels (median = 1.6%) (p < 0.005). Burn severity. For 2013, one year after the fire, key measures of richness and composition in pools of varying burn severities differed only in percentage EPT (Table 5.26; Figure 5.12). The median value of percentage EPT was significantly greater in pools of unburned sites (26.5%) than in pools of high severity burn sites (0.0%). Of the measures of community tolerance, statistical analysis identified significant differences in the values of the Family Biotic Index (p = 0.044) and percentage intolerant taxa (p = 0.024) of pools across burn severities. The multiple comparisons test failed to identify between which categories these differences occurred. For the Family Biotic Index, the median values in pools were 6.0, 4.9, and 5.3 for sites of high severity burn, low severity burn, and no burn, respectively. For Table 5.25. Key macroinvertebrate metrics in steps of burned and unburned sites in 2013, with results of the Wilcoxon Rank Sum test Category Metric Median p-value Burned Unburned Richness and Composition Taxa Richness 12 11.5 0.87 % EPT 14.5 6.4 0.70 % Dominant Taxon 44.8 75.5 0.22 Tolerance Family Biotic Index 5.5 5.8 0.79 % Intolerant Taxa 4.9 2.1 0.87 % Tolerant Taxa 4.1 1.1 0.17 Functional Feeding Groups % Scraper 0.0 0.04 0.12 % Shredder 7.4 2.5 0.48 % Collector-Filterer 3.3 14.8 0.11 Habit Type % Clinger 4.6 30.0 0.057 *p<0.05 **p<0.01 ***p<0.001 85

PAGE 93

percentage intolerant taxa, median values in pools were 0.0%, 4.4%, and 7.2% for sites of high severity burn, low severity burn, and no burn, respectively. Table 5.26. Key macroinvertebrate metrics in pools of high severity burn, low severity burn, and unburned sites in 2013, with results of the Kruskal-Wallis test Category Metric Median p-value High Severity Low Severity Unburned Richness and Composition Taxa Richness 7.5 11 11.5 0.10 % EPT   0 .0   6.8 26.5 0.012 % Dominant Taxon 35.7 80.1 60.7 0.26 Tolerance Family Biotic Index 6.0 4.9 5.3 0.044 % Intolerant Taxa 0.0 4.4 7.2 0.024 % Tolerant Taxa 0.5 3.8 5.4 0.59 Functional Feeding Groups % Scraper 0.0 0.0 0.2 0.37 % Shredder 0.0 7.2 4.2 0.12 % Collector-Filterer 0.0 2.2 1.2 0.15 Habit Type % Clinger   0.1   1.1 8 .0 0.019   Identical symbols identify groups that do not differ significantly from each other *p<0.05 **p<0.01 ***p<0.001 high low unburned 5 10 15 20 Taxa Richness Burn Severity Number of Taxa high low unburned 0 20 40 60 80 Percent EPT Burn Severity Percent high low unburned 30 50 70 90 Percent Dominant Taxon Burn Severity Percent high low unburned 2 3 4 5 6 7 Family Biotic Index Burn Severity Index high low unburned 0 20 40 60 Percent Intolerant Taxa Burn Severity Percent high low unburned 0 5 10 15 20 Percent Tolerant Taxa Burn Severity Percent       Identical symbols identify groups that do not differ significantly from each other Figure 5.12. Key metrics of composition and tolerance of macroinvertebrate communities in pools in sites of varying burn severities in 2013, with results of the Kruskal-Wallis test. 86

PAGE 94

Metrics representing functional feeding groups in pools did not exhibit significant differences among levels of burn severity in 2013 (Table 5.26). Macroinvertebrate habit types differed in pools of high burn severity and unburned sites in percentage clinger (p = 0.019) and percentage swimmer (p < 0.05). The median percentages of these habit types were greater in pools of unburned reference sites than in pools of high severity burn sites (8.0% versus 0.1% for percentage clinger; 18.6% versus 0.0% for percentage swimmer). Analysis of macroinvertebrate communities in steps for 2013 identified statistically significant differences in one metric of overall richness and composition (Table 5.27; Figure 5.13). Values of the percentage of EPT organisms were significantly higher in steps of low severity burn sites (median = 51.1%) compared to steps of high severity burn sites (median = 1.4%) (p = 0.026). Table 5.27. Key macroinvertebrate metrics in steps of high severity burn, low severity burn, and unburned sites in 2013, with results of the Kruskal-Wallis test Category Metric Median p-value High Severity Low Severity Unburned Richness and Composition Taxa Richness 9 14.5 11.5 0.21 % EPT   1.4 51.1   6.4 0.026 % Dominant Taxon 42.9 50.1 75.5 0.31 Tolerance Family Biotic Index   5.8 3.3   5.8 0.026 % Intolerant Taxa   0 .0 48.2   2.1 0.0098 ** % Tolerant Taxa 4.1 4.7 1.1 0.36 Functional Feeding Groups % Scraper 0.0 0.0 0.04 0.17 % Shredder   1.3 44.1   2.5 0.017 % Collector-Filterer 9.5 3.3 14.8 0.25 Habit Type % Clinger 9.5 4.6 30.0 0.14   Identical symbols identify groups that do not differ significantly from each other *p<0.05 **p<0.01 ***p<0.001 87

PAGE 95

! Of the measures of community tolerance, statistical analysis identified differences in the Family Biotic Index and the percentage of intolerant taxa in steps as a function of burn severity for 2013 (Table 5.27; Figure 5.13). Values of the Family Biotic Index were significantly lower in steps of low severity burn sites (median = 3.3) than in those of high severity burn sites (median = 5.8) (p = 0.026). Values of the percentage of intolerant taxa were significantly greater in steps of low severity burn (median = 48.2%) than in steps of high severity burn (median = 0.0%) (p = 0.0098). Functional feeding groups differed among the steps of different levels of burn severity only in the percentage of shredders (Table 5.27). The median values of percentage shredder were significantly higher in steps of low burn severity (44.1%) than in steps of high high low unburned 5 10 15 Taxa Richness Burn Severity Number of Taxa high low unburned 0 10 30 50 Percent EPT Burn Severity Percent high low unburned 20 40 60 80 Percent Dominant Taxon Burn Severity Percent high low unburned 2 3 4 5 6 7 8 Family Biotic Index Burn Severity Index high low unburned 0 10 30 50 Percent Intolerant Taxa Burn Severity Percent high low unburned 0 10 20 30 40 50 Percent Tolerant Taxa Burn Severity Percent               Identical symbols identify groups that do not differ significantly from each other Figure 5.13. Key metrics of composition and tolerance in macroinvertebrate communities in steps of varying burn severities in 2013, with results of the Kruskal-Wallis test. 88

PAGE 96

burn severity (1.3%) (p = 0.017). Habit types differed only in the percentage of sprawlers. Median percentage sprawler was higher in low severity burn sites (47.8%) than in unburned sites (1.6%) (p < 0.01) Two Years Post-Fire: 2014 Presence of burn. For 2014, two years after the Waldo Canyon Fire, statistical analysis of pools in burned and unburned sites did not identify significant differences in any of the three key metrics of overall richness and composition (Table 5.28). Similarly, the three measures of community tolerance showed no differences between the pools of burned and unburned sites. Of the key functional feeding groups, values of the percentage of scrapers were higher in pools of unburned sites for 2014 (median = 0.7%) than in pools of burned sites (median = 0.0%) (p = 0.0012; Table 5.28). Other key measures of functional feeding groups (percentage shredder and percentage collector-filterer) did not differ between pools of burned and unburned study reaches. Additional metrics exhibiting differences in pools included only Table 5.28. Key metrics in pools of burned and unburned sites in 2014, with results of the Wilcoxon Rank Sum test Category Metric Median p-value Burned Unburned Richness and Composition Taxa Richness 14 15 0.27 % EPT 36.1 34.7 0.64 % Dominant Taxon 37.6 58.3 0.093 Tolerance Family Biotic Index 5.0 5.1 0.87 % Intolerant Taxa 5.0 13.7 0.12 % Tolerant Taxa 5.3 8.2 0.25 Functional Feeding Groups % Scraper 0.0 0.7 0.0012 ** % Shredder 27.2 13.5 0.79 % Collector-Filterer 4.1 8.9 0.17 Habit Type % Clinger 3.7 4.3 1.0 *p<0.05 **p<0.01 ***p<0.001 89

PAGE 97

the percentage of predators. Median percentage predator was significantly lower in pools of unburned sites (1.7%) compared to pools of burned sites (9.4%) (p < 0.001). Proportions of habit types did not differ significantly between pools of burned and unburned sites. In steps for 2014, metrics of overall richness and composition did not differ significantly between burned and unburned sites (Table 5.29). Metrics of community tolerance also showed no significant differences between burned and unburned sites. Of the key functional feeding groups, median percentage scraper was higher in steps of unburned sites (0.8%) than in steps of burned sites (0.0%) (p = 0.0039). Additionally, the median percentage of predators was higher in steps of burned channels (7.7%) than in those of unburned channels (1.2%) (p < 0.05) Analysis of macroinvertebrate habit types identified no statistically significant differences between steps of burned and unburned sites. Burn severity. In pools statistical analysis of metrics for 2014 by burn severity indicated that values of taxa richness were significantly higher in channels burned with low severity (median = 17) than in those burned with high severity (median = 10.5) (p = 0.013; Table 5.29. Key metrics in steps of burned and unburned sites in 2014, with results of the Wilcoxon Rank Sum test Category Metric Median p-value Burned Unburned Richness and Composition Taxa Richness 13 16 0.41 % EPT 38.7 37.0 1.0 % Dominant Taxon 40.9 39.9 0.79 Tolerance Family Biotic Index 4.9 4.6 0.96 % Intolerant Taxa 10.3 18.7 0.17 % Tolerant Taxa 0.8 4.3 0.22 Functional Feeding Groups % Scraper 0.0 0.8 0.0039 ** % Shredder 9.3 18.3 0.26 % Collector-Filterer 29.0 25.7 0.37 Habit Type % Clinger 31.4 21.7 0.22 *p<0.05 **p<0.01 ***p<0.001 90

PAGE 98

Table 5.30; Figure 5.14). Other key metrics of community composition did not differ significantly among levels of burn severity. Similarly, measures of community tolerance showed no statistically significant differences in pools of varying burn severities. Key functional feeding groups in pools for 2014 differed in the percentage of scrapers and the percentage of collector-filterers (Table 5.30). Values of percentage scraper were higher in pools of unburned sites (median = 0.7%) than in those of high severity burn sites (median = 0.0%) (p = 0.0044). Although a difference among burn categories was identified in percentage collector-filterer (p = 0.033), the difference could not be assigned to specific levels of burn severity. Median values of percentage collector-filterer were 1.1%, 12.6%, and 8.9% for sites of high severity burn, low severity burn, and no burn, respectively. Additionally, the median percentage of predators was higher in pools of sites of high severity burn (10.4%) than in unburned sites (1.7%) (p < 0.01). Table 5.30. Key macroinvertebrate metrics in pools of high severity burn, low severity burn, and unburned sites in 2014, with results of the Kruskal-Wallis test Category Metric Median p-value High Severity Low Severity Unburned Richness and Composition Taxa Richness   1 0.5 17   1 5 0.013 ** % EPT 40.7 32.8 34.7 0.83 % Dominant Taxon 39.4 29.4 58.3 0.11 Tolerance Family Biotic Index 5.1 4.7 5.1 0.61 % Intolerant Taxa 0.9 8.4 13.7 0.24 % Tolerant Taxa 2.1 12.9 8.2 0.073 Functional Feeding Groups % Scraper   0 .0   0 .0 0.7 0.0044 ** % Shredder 4.6 31.1 13.5 0.32 % Collector-Filterer 1.1 12.6 8.9 0.033 Habit Type % Clinger 2.8 5.6 4.3 0.48   Identical symbols identify groups that do not differ significantly from each other *p<0.05 **p<0.01 ***p<0.001 91

PAGE 99

! For macroinvertebrate habit types of pools for 2014, percentage clinger was not shown to differ among levels of burn severity (Table 5.30). Habit types did, however, exhibit differences in percentages of climbers and sprawlers. Values of percentage climber were higher in pools of low burn severity sites (median = 13.7%) than in pools of either unburned (median = 4.1%) or high severity burn sites (median = 4.1%) (p < 0.05). Median percentage sprawler was higher in pools of high burn severity sites (37.7%) than in those of low severity burn sites (4.1%) (p < 0.05). In steps for 2014, statistical analysis across burn severities identified significant differences in two metrics of overall richness and composition: taxa richness and percentage EPT (Table 5.31; Figure 5.15). Although the Kruskal-Wallis test detected a difference in taxa high low unburned 8 10 14 18 Taxa Richness Burn Severity Number of Taxa high low unburned 0 20 40 60 80 Percent EPT Burn Severity Percent high low unburned 20 30 40 50 60 70 Percent Dominant Taxon Burn Severity Percent high low unburned 3.5 4.5 5.5 Family Biotic Index Burn Severity Index high low unburned 0 10 20 30 40 Percent Intolerant Taxa Burn Severity Percent high low unburned 0 5 10 15 20 Percent Tolerant Taxa Burn Severity Percent       Identical symbols identify groups that do not differ significantly from each other Figure 5.14. Key metrics of composition and tolerance of macroinvertebrate communities in pools of varying burn severities in 2014, with results of the Kruskal-Wallis test. 92

PAGE 100

richness (p = 0.44), it failed to identify which levels of burn severity differed. The median values of taxa richness were 11 for sites of high burn severity, 17.5 for sites of low burn severity, and 16 for unburned sites. Median percentage EPT was higher in steps of low severity burn sites (62.7%) than in those of high severity burn sites (15.9%) (p = 0.012). Of the metrics representing community tolerance, two exhibited differences among steps in sites of varying burn severity: the Family Biotic Index and the percentage of intolerant taxa (Table 5.31). The values of the Family Biotic Index were lower in steps of low severity burn sites (median = 3.6) than in steps of high severity burn sites (median = 5.6) (p = 0.0041). Values of the percentage of intolerant taxa were higher in steps of low severity burn (median = 25.2%) and unburned sites (median = 18.7%) than in steps of high severity burn sites (median = 1.3%) (p = 0.0041). Table 5.31. Key macroinvertebrate metrics in steps of high severity burn, low severity burn, and unburned sites in 2014, with results of the Kruskal-Wallis test Category Metric Median p-value High Severity Low Severity Unburned Richness and Composition Taxa Richness 11 17.5 16 0.044 % EPT   15.9 62.7   37 .0 0.012 % Dominant Taxon 51.5 38.0 39.9 0.36 Tolerance Family Biotic Index   5.6 3.6   4.6 0.0041 ** % Intolerant Taxa   1.3 25.2 18.7 0.0059 ** % Tolerant Taxa 0.5 2.6 4.3 0.31 Functional Feeding Groups % Scraper   0 .0   0.1 0.8 0.012 % Shredder   4 .0 25.5 18.3 0.0049 ** % Collector-Filterer 48.7 15.5 25.7 0.036 Habit Type % Clinger 53.8 22.6 21.7 0.032   Identical symbols identify groups that do not differ significantly from each other *p<0.05 **p<0.01 ***p<0.001 93

PAGE 101

! Statistical analysis suggested that in steps each of the key metrics of functional feeding groups and habit types differed significantly among levels of burn severity (Table 5.31). Median percentage shredder was higher in steps of low burn severity (25.5%) and unburned sites (18.3%) than in steps of high severity burn sites (4.0%) (p = 0.0049). The median percentage of scrapers was significantly higher in steps of unburned sites (0.8%) than in steps of high severity burn sites (0.0%) (p = 0.012). Values of percentage collector-filterer differed among sites (p = 0.036), but the differences between each level of burn severity could not be distinguished. The median percentage of collector-filterers in steps was 48.7% in sites of high burn severity, 15.5% in sites of low burn severity, and 18.7% in unburned sites. Of the macroinvertebrate habit types, percentage clinger was detected to differ among high low unburned 5 10 15 20 Taxa Richness Burn Severity Number of Taxa high low unburned 0 20 40 60 80 Percent EPT Burn Severity Percent high low unburned 20 40 60 Percent Dominant Taxon Burn Severity Percent high low unburned 3 4 5 6 Family Biotic Index Burn Severity Index high low unburned 0 10 20 30 40 Percent Intolerant Taxa Burn Severity Percent high low unburned 0 5 10 15 Percent Tolerant Taxa Burn Severity Percent             Identical symbols identify groups that do not differ significantly from each other Figure 5.15. Key metrics of composition and tolerance of macroinvertebrate communities in steps of varying burn severities in 2014, with results of the Kruskal-Wallis test. 94

PAGE 102

sites (p = 0.032), but this difference also could not be identified. The median values of the percentage of clingers in steps were 53.8%, 22.6%, and 21.7% in steps of high severity burn, low severity burn, and no burn, respectively. Summarizing Key Results In 2012, steps differed from pools within burned reaches in four key metrics: percentage EPT, percentage dominant taxon, percentage intolerant taxa, and percentage shredder. These differences were largely not carried forward into later years. In 2012, pools across burned and unburned sites indicated significant differences in the following metrics: taxa richness, percentage EPT, percentage dominant taxon, Family Biotic Index, percentage intolerant taxa, percentage collector-filterer, and percentage clinger. These differences were largely not carried forward into later years. In contrast, in 2012, differences between steps of burned and unburned reaches included only taxa richness, percentage tolerant taxa, and percentage clinger. Pools across levels of burn severity generally showed differences between channels of high burn severity and unburned channels, such as lower percentage EPT and percentage intolerant taxa in high severity burn sites. Steps across levels of burn severity generally displayed differences between high and low severity burn sites. Percentage EPT, percentage intolerant taxa, and percentage shredder, for example, were consistently lower in channels burned with high severity. Few differences emerged between the metrics of steps and pools within each category of burn. These results addressed the third research question: How did the response of 95

PAGE 103

macroinvertebrate communities to wildfire vary as a function of habitat type (step or pool)? Although the direct comparison of steps and pools yielded few significant differences, the pools of burned reaches overall exhibited greater differences compared to pools of unburned reaches than the steps of burned reaches compared to steps of unburned reaches. These results supported the hypothesis that the response of macroinvertebrate communities would vary with habitat type. Research Question (4): Change over Time The fourth research question concerned changes in the response of macroinvertebrate communities over time following wildfire. Did the values of macroinvertebrate metrics of the study reaches change significantly from year to year, indicating a trend of recovery? Each subsection of this section focuses on a level of burn severity: high burn severity, low burn severity, and unburned. In other words, statistical analysis addressed changes in communities from 2012 to 2014 within each study reach independently in order to identify potential trajectories in the recovery of stream ecosystems impacted by wildfire. High Burn Severity Williams Canyon: Willis Reach For Willis Reach (Figure 3.3), statistical analysis identified significant differences over time in two of the three key metrics representing community richness and composition. These metrics were percentage EPT (p = 0.038) and percentage dominant taxon (p = 0.018) (Table 5.32; Figure 5.16). The Friedman multiple comparisons test failed to identify the years in which changes in the values of percentage EPT occurred. The median percentages of EPT organisms were 0.6% in 2012, 0.0% in 2013, and 28.6% in 2014. Values of the percentage of 96

PAGE 104

the dominant taxon significantly decreased from 2012 (median = 90.9%) to 2013 (median = 30.5%). Table 5.32. Key macroinvertebrate metrics of samples collected from Willis Reach (high burn severity) in 2012, 2013, and 2014, with results of the Friedman test Category Metric Median p-value 2012 2013 2014 Richness and Composition Taxa Richness 7 9 10.5 0.17 % EPT 0.6 0.0 28.6 0.038 % Dominant Taxon   90.9 30.5   55.1 0.018 Tolerance Family Biotic Index   5.8   6.2 5.4 0.039 % Intolerant Taxa 1.4 0.0 1.0 0.085 % Tolerant Taxa   1.3   9.9 0.02 0.024 Functional Feeding Groups % Scraper 0.0 0.0 0.0 n/a % Shredder 3.5 0.0 1.9 0.17 % Collector-Filterer 0.6 0.0 32.6 0.13 Habit Type % Clinger 0.6 7.1 32.6 0.42   Identical symbols identify groups that do not differ significantly from each other *p<0.05 **p<0.01 ***p<0.001 2012 2013 2014 5 6 7 8 9 10 Taxa Richness Year Number of Taxa 2012 2013 2014 0 10 20 30 40 Percent EPT Year Percent 2012 2013 2014 30 50 70 90 Percent Dominant Taxon Year Percent 2012 2013 2014 5.5 6.0 6.5 7.0 Family Biotic Index Year Index 2012 2013 2014 0 1 2 3 4 5 Percent Intolerant Taxa Year Percent 2012 2013 2014 0 5 10 15 20 Percent Tolerant Taxa Year Percent               Identical symbols identify groups that do not differ significantly from each other Figure 5.16. Key metrics of macroinvertebrate community composition and tolerance at Willis Reach in 2012, 2013, and 2014, with results of the Friedman test. 97

PAGE 105

Of the measures of community tolerance, the Family Biotic Index (p = 0.039) and the percentage of tolerant taxa (p = 0.024) decreased significantly from 2013 to 2014 (Table 5.32; Figure 5.16). The median value of the Family Biotic Index was 6.2 in 2013 and 5.4 in 2014. Percentage tolerant taxa decreased from 9.9% in 2013 to 0.02% in 2014. Key functional feeding groups did not exhibit significant changes in any years (Figure 5.17). Scrapers were absent from Willis Reach in all samples from all years. Median percentage shredder showed no clear trend over time. Although change over time in median percentage collector-filterer was not statistically significant, the median was notably higher in 2014 (32.6%) than in 2012 and 2013 (0.6% and 0.0%, respectively) (Table 5.32; Figure 5.17). Of the macroinvertebrate habit types in the samples from Willis Reach, clingers showed no significant change over time in proportion, despite a steady increase in median percentage (Figure 6.17). Statistical analysis of additional metrics (Table 4.3) at Willis Reach identified significant changes in several measures of functional feeding groups and habit types. For functional feeding groups, analysis indicated significant changes in the proportions of collector-gatherers (p < 0.05) and predators (p < 0.05), but these differences could not be 2012 2013 2014 0 1 2 3 4 5 Percent Scraper Year Percent 2012 2013 2014 0 2 4 6 8 Percent Shredder Year Percent 2012 2013 2014 0 20 40 60 Percent Collector! Filterer Year Percent 2012 2013 2014 0 20 40 60 Percent Clinger Year Percent Figure 5.17. Key functional feeding groups and macroinvertebrate habit types at Willis Reach in 2012, 2013, and 2014. 98

PAGE 106

isolated to specific years. The median values of percentage collector-gatherer were 94.9% in 2012, 44.7% in 2013, and 59.2% in 2014 (Appendix B). For the percentage of predators, median values were 1.%, 49.2%, and 5.3% in 2012, 2013, and 2014, respectively. Additional metrics of habit type that displayed significant changes from 2012 to 2013 included increases in median percentage climber from 0.3% to 7.3% (p < 0.05), and percentage sprawler from 0.6% to 49.2% (p < 0.05; Appendix B). Median percentage swimmer increased from 2013 (0.0%) to 2014 (26.7%) (p < 0.05). Change over time was detected in the proportion of burrowers (p < 0.05), but this change could not be isolated by year. Median percentage burrower decreased from 93.9% in 2012 to 32.8% in 2013 and 30.6% in 2014. Williams Canyon: Tributary Reach Macroinvertebrate samples from Tributary Reach in Williams Canyon (Figure 3.3) were collected in 2013 and 2014 only, because the channel was dry during the initial period of sampling in 2012. Of the metrics of overall richness and composition, statistical analysis identified a significant change in taxa richness (Table 5.33; Figure 5.18). The median taxa richness increased significantly from 3.5 in 2013 to 7.5 in 2014 (p = 0.046). Median percentage EPT and percentage dominant taxon did not change significantly. Measures of community tolerance did not exhibit significant changes. Individuals of intolerant taxa were absent in samples from Tributary Reach in both years. Key functional feeding groups and habit types did not differ significantly between years at Tributary Reach. Scrapers were absent in all samples from 2014. Shredders, collector-filterers, and clingers were absent in 2013 (Table 5.33; Figure 5.19). 99

PAGE 107

! Table 5.33. Key macroinvertebrate metrics of samples collected from Tributary Reach (high burn severity) in 2012, 2013, and 2014, with results of the Friedman test Category Metric Median p-value 2012 2013 2014 Richness and Composition Taxa Richness 3.5 7.5 0.046 % EPT 0.0 0.0 0.32 % Dominant Taxon 53.3 36.4 0.32 Tolerance Family Biotic Index 6.0 6.0 1.0 % Intolerant Taxa 0.0 0.0 n/a % Tolerant Taxa 0.0 8.6 0.32 Functional Feeding Groups % Scraper 0.0 0.0 0.32 % Shredder 0.0 3.5 0.083 % Collector-Filterer 0.0 22.0 0.083 Habit Type % Clinger 0.0 22.0 0.083 *p<0.05 **p<0.01 ***p<0.001 absent in all samples 2013 2014 4 6 8 10 Taxa Richness Year Number of Taxa 2013 2014 0 2 4 6 8 10 Percent EPT Year Percent 2013 2014 30 40 50 60 Percent Dominant Taxon Year Percent 2013 2014 5.5 6.0 6.5 7.0 7.5 8.0 Family Biotic Index Year Index 2013 2014 0 2 4 6 8 10 Percent Intolerant Taxa Year Percent 2013 2014 0 10 20 30 40 50 Percent Tolerant Taxa Year Percent     Identical symbols identify groups that do not differ significantly from each other Figure 5.18. Key metrics of macroinvertebrate community composition and tolerance at Tributary Reach in 2013 and 2014, with results of the Friedman test. 100

PAGE 108

Additional metrics (Table 4.3) exhibiting significant changes at Tributary Reach emerged in one functional feeding groups and one habit type. The median value of percentage collector-gatherer increased from 18.1% in 2013 to 45.6% in 2014 (p < 0.05) (Appendix B). The median value of percentage sprawler increased from 13.9% in 2013 to 35.5% in 2014 (p < 0.05). Camp Creek: Aussie Reach For Aussie Reach (a channel reach burned with high severity; Figure 3.3), statistical analysis identified significant changes in only one key metric of overall richness and composition and tolerance percentage EPT (Table 5.34; Figure 5.20). Values of the percentage of EPT organisms were significantly higher in 2014 (median = 47.8%) than in 2012 (median = 6.1%) (p = 0.039). For measures of community tolerance, statistical analysis indicated a significant change in the Family Biotic Index and the percentage of intolerant taxa (Table 5.34; Figure 5.20). The Friedman test indicated a change in the Family Biotic Index (p = 0.0498), but the multiple comparisons test failed to identify the years of change. The median value of the Family Biotic Index, however, remained stable at 5.8 in 2012 and 2013, and decreased to 4.2 2013 2014 0 5 10 15 Percent Scraper Year Percent 2013 2014 0 1 2 3 4 5 6 Percent Shredder Year Percent 2013 2014 0 5 15 25 35 Percent Collector! Filterer Year Percent 2013 2014 0 10 20 30 40 Percent Clinger Year Percent Figure 5.19. Key functional feeding groups and macroinvertebrate habit types at Tributary Reach in 2013 and 2014. 101

PAGE 109

in 2014 (Table 5.34). The median value of percentage intolerant taxa increased significantly between 2012 (0.2%) and 2014 (26.2%) (p = 0.039). 2012 2013 2014 6 8 10 14 Taxa Richness Year Number of Taxa 2012 2013 2014 0 10 20 30 40 50 Percent EPT Year Percent 2012 2013 2014 40 60 80 Percent Dominant Taxon Year Percent 2012 2013 2014 4.0 4.5 5.0 5.5 6.0 Family Biotic Index Year Index 2012 2013 2014 0 10 20 30 40 Percent Intolerant Taxa Year Percent 2012 2013 2014 0 1 2 3 4 5 Percent Tolerant Taxa Year Percent   Identical symbols identify groups that do not differ significantly from each other Figure 5.20. Key metrics of macroinvertebrate community composition and tolerance at Aussie Reach in 2012, 2013, and 2014, with results of the Friedman test.         Table 5.34. Key macroinvertebrate metrics of samples collected from Aussie Reach (high burn severity) in 2012, 2013, and 2014, with results of the Friedman test Category Metric Median p-value 2012 2013 2014 Richness and Composition Taxa Richness 10 12 14 0.057 % EPT   6.1   7.3 47.8 0.039 % Dominant Taxon 82.9 66.1 41.0 0.11 Tolerance Family Biotic Index 5.8 5.8 4.2 0.0498 % Intolerant Taxa   1.8   2.6 26.2 0.039 % Tolerant Taxa 0.2 1.5 2.3 0.17 Functional Feeding Groups % Scraper 0.0 0.0 0.0 0.61 % Shredder 3.3 5.8 27.6 0.11 % Collector-Filterer 2.9 12.8 12.7 1.0 Habit Type % Clinger 3.4 13.4 16.4 0.47   Identical symbols identify groups that do not differ significantly from each other *p<0.05 **p<0.01 ***p<0.001 102

PAGE 110

Metrics of functional feeding groups reflected no significant change over time at Aussie Reach (Table 5.34). Scrapers were absent from all samples in 2012, and the median percentage of scrapers remained at 0.0% in all years. Median percentage shredder was notably, but not significantly, higher in 2014 (27.6%) than in 2012 (3.3%) and 2013 (5.8%) (Figure 5.21). No significant change emerged in percentage collector-filterer. The key macroinvertebrate habit type (percentage clinger) also exhibited no change over time. Of the additional habit types, percentage burrower exhibited significant change (p < 0.05). This change could not be isolated to specific years. Median values of percentage burrower were 85.6% in 2012, 69.6% in 2013, and 24.2% in 2014. Low Burn Severity Camp Creek: Eagle Reach Statistical analysis identified significant changes in one key metric of overall richness and composition at Eagle Reach the percentage of the dominant taxon (Table 5.35; Figure 5.22). Values of percentage dominant taxon decreased significantly from 2013 (median = 57.2%) to 2014 (median = 26.8%) (p = 0.039). An overall increase in the median value of taxa richness from 12 in 2012 to 18.5 in 2014 was not found to be statistically significant. Figure 5.21. Key functional feeding groups and macroinvertebrate habit types in communities at Aussie Reach in 2012, 2013, and 2014. 2012 2013 2014 0.0 0.2 0.4 0.6 Percent Scraper Year Percent 2012 2013 2014 10 20 30 40 Percent Shredder Year Percent 2012 2013 2014 0 20 40 60 Percent Collector! Filterer Year Percent 2012 2013 2014 0 20 40 60 Percent Clinger Year Percent 103

PAGE 111

One metric of community tolerance exhibited a significant change over time at Eagle Reach. The median value of the Family Biotic Index decreased significantly from 2012 (5.6) to 2013 (2.2) (p = 0.018). The only key metric of functional feeding groups to exhibit statistically significant change over time at Eagle Reach was the proportion of collector-filterers (Table 5.35; Figure 5.23). Values of the percentage of collector-filterers were significantly higher in 2014 (median = 18.7%) than in 2012 (median = 0.0%) (p = 0.018). Of the additional metrics of functional feeding groups (Table 4.3), the values of the percentage of predators increased significantly from 2013 (median = 2.9%) to 2014 (median = 12.8%) (p < 0.05; Appendix B). Statistical analysis did not identify significant changes in any metrics of macroinvertebrate habit types. Table 5.35. Key macroinvertebrate metrics of samples collected from Eagle Reach (low burn severity) in 2012, 2013, and 2014, with results of the Friedman test Category Metric Median p-value 2012 2013 2014 Richness and Composition Taxa Richness 12 13.5 19 0.25 % EPT 22.9 59.2 25.2 0.37 % Dominant Taxon   47.5   57.2 26.8 0.039 Tolerance Family Biotic Index   5.6 2.2   4.1 0.018 % Intolerant Taxa 15.1 58.6 13.8 0.17 % Tolerant Taxa 21.0 10.2 14.2 1.0 Functional Feeding Groups % Scraper 2.1 0.0 0.0 0.23 % Shredder 18.2 52.2 30.2 0.37 % Collector-Filterer   0 .0   6.2 18.7 0.018 Habit Type % Clinger 5.4 2.8 15.3 0.17   Identical symbols identify groups that do not differ significantly from each other *p<0.05 **p<0.01 ***p<0.001 104

PAGE 112

! Camp Creek: Meadow Reach For Meadow Reach (a channel reach burned with low severity; Figure 3.3), the Friedman test identified significant changes over time in two key metrics of overall richness 2012 2013 2014 10 12 14 16 18 20 Taxa Richness Year Number of Taxa 2012 2013 2014 10 30 50 70 Percent EPT Year Percent 2012 2013 2014 20 30 40 50 60 70 Percent Dominant Taxon Year Percent 2012 2013 2014 2 3 4 5 6 Family Biotic Index Year Index 2012 2013 2014 10 30 50 70 Percent Intolerant Taxa Year Percent 2012 2013 2014 0 10 30 50 Percent Tolerant Taxa Year Percent           Identical symbols identify groups that do not differ significantly from each other Figure 5.22. Key metrics of macroinvertebrate community composition and tolerance at Eagle Reach in 2012, 2013, and 2014, with results of the Friedman test. 2012 2013 2014 0.0 1.0 2.0 3.0 Percent Scraper Year Percent 2012 2013 2014 10 30 50 70 Percent Shredder Year Percent 2012 2013 2014 0 5 10 15 20 Percent Collector! Filterer Year Percent 2012 2013 2014 0 5 10 15 20 Percent Clinger Year Percent       Identical symbols identify groups that do not differ significantly from each other Figure 5.23. Key functional feeding groups and macroinvertebrate habit types in communities at Eagle Reach in 2012, 2013, and 2014. 105

PAGE 113

and composition: taxa richness and percentage EPT (Table 5.36; Figure 5.24). The values of taxa richness were significantly higher in 2014 (median = 16) than in 2012 (median = 11). Values of the percentage of EPT organisms increased significantly from 2013 (median = 12.2%) to 2014 (median = 52.9%) (p = 0.039). For the measures of community tolerance, two metrics exhibited significant change over time at Meadow Reach: the Family Biotic Index and the percentage of tolerant taxa (Table 5.36; Figure 5.24). Values of the Family Biotic Index decreased from 2013 (median = 5.6) to 2014 (median = 4.4) (p = 0.039). Although the multiple comparisons test did not identify the specific years of change in percentage tolerant taxa, the median was notably higher in 2012 (16.6%) than in subsequent years (2.5% in 2013 and 2.3% in 2014) (p = 0.0498). Table 5.36. Key macroinvertebrate metrics of samples collected from Meadow Reach (low burn severity) in 2012, 2013, and 2014, with results of the Friedman test Category Metric Median p-value 2012 2013 2014 Richness and Composition Taxa Richness   11   12.5 16 0.039 % EPT   20.5   12.2 52.9 0.039 % Dominant Taxon 46.4 79.1 39.0 0.11 Tolerance Family Biotic Index   5.3   5.6 4.4 0.039 % Intolerant Taxa 20.5 8.3 13.1 0.17 % Tolerant Taxa 16.6 2.5 2.3 0.0498 Functional Feeding Groups % Scraper 0.0 0.0 0.0 0.37 % Shredder   23.4   9.9 27.7 0.039 % Collector-Filterer   0 .0   1.4 7.6 0.022 Habit Type % Clinger 3.6 3.0 9.0 0.17   Identical symbols identify groups that do not differ significantly from each other *p<0.05 **p<0.01 ***p<0.001 106

PAGE 114

! Functional feeding groups in communities of Meadow Reach exhibited significant differences among years in percentage shredder and percentage collector-filterer (Table 5.36; Figure 5.25). The median value of percentage shredder increased from 2013 (9.9%) to 2014 (27.7%) (p = 0.039). The median value of the percentage of collector-filterers was significantly higher in 2014 (7.6%) than in 2012 (0.0%) (p = 0.022). For the key macroinvertebrate habit type, percentage clinger did not change significantly. Additional changes in metrics of functional feeding groups (Table 4.3) at Meadow Reach included a significant change in the percentage of collector-gatherers (p < 0.05). The only habit type to change significantly over time was the percentage of climbers (p < 0.05). The multiple comparisons test failed to identify the years of change for these two metrics. 2012 2013 2014 6 8 10 14 18 Taxa Richness Year Number of Taxa 2012 2013 2014 0 20 40 60 80 Percent EPT Year Percent 2012 2013 2014 30 50 70 90 Percent Dominant Taxon Year Percent 2012 2013 2014 3.5 4.5 5.5 6.5 Family Biotic Index Year Index 2012 2013 2014 10 20 30 40 Percent Intolerant Taxa Year Percent 2012 2013 2014 0 10 20 30 40 Percent Tolerant Taxa Year Percent           Identical symbols identify groups that do not differ significantly from each other Figure 5.24. Key metrics of macroinvertebrate community composition and tolerance at Meadow Reach in 2012, 2013, and 2014, with results of the Friedman test. 107

PAGE 115

Median values of percentage collector-gatherer were 48.2% in 2012, 85.0% in 2013, and 53.2% in 2014 (Appendix B). The median values of percentage climber were 1.2% in 2012, 1.4% in 2013, and 7.1% in 2014. Reference Reaches Bear Creek: Hunter Reach At the unburned Hunter Reach within the Bear Creek watershed (Figure 3.3), statistical analysis identified significant change over time in two key metrics of overall richness and composition: percentage EPT and percentage dominant taxon (Table 5.37; Figure 5.26). The median value of the percentage of EPT organisms increased significantly from 2013 (4.5%) to 2014 (24.2%) (p = 0.039). Although a significant change was detected in the percentage of the dominant taxon (p = 0.0498), analysis failed to identify the years in which the change occurred. The median value of percentage dominant taxon was, however, notably higher in 2013 (91.4%) than in 2012 (57.8%) and 2014 (58.1%). Statistical analysis indicated change over time in each of the three metrics of community tolerance at Hunter Reach (Table 5.37; Figure 5.26). The median value of the 2012 2013 2014 0.0 0.1 0.2 0.3 0.4 Percent Scraper Year Percent 2012 2013 2014 10 20 30 40 Percent Shredder Year Percent 2012 2013 2014 0 5 10 15 20 Percent Collector! Filterer Year Percent 2012 2013 2014 0 5 10 15 20 Percent Clinger Year Percent           Identical symbols identify groups that do not differ significantly from each other Figure 5.25. Key functional feeding groups and macroinvertebrate habit types in communities of Meadow Reach in 2012, 2013, and 2014. 108

PAGE 116

Family Biotic Index decreased significantly from 2013 (5.9) to 2014 (4.9) (p = 0.039). Percentage intolerant taxa significantly increased from 2013 (median = 1.6%) to 2014 (median = 16.3%) (p = 0.018). Analysis of percentage tolerant taxa indicated significant change in over time (p = 0.0498), but did not identify the years in which this metric changed. However, median percentage tolerant taxa was notably low in 2013 (0.2%), compared to 2012 (3.2%) and 2014 (4.7%). Median taxa richness did not vary significantly with time. Of the key functional feeding groups, percentage shredder and percentage collectorfilterer exhibited significant changes over time (Table 5.37; Figure 5.27). Values of percentage shredder were higher in 2014 (median = 17.1%) than in 2013 (median = 2.6%) (p = 0.039). The median value of percentage collector-filterer was higher in 2014 (8.4%) than in 2012 (0.6%) (p = 0.039). The key metric of macroinvertebrate habit types, percentage clinger, did not change significantly over time (Table 5.37). Table 5.37. Key macroinvertebrate metrics of samples collected from Hunter Reach (unburned) in 2012, 2013, and 2014, with results of the Friedman test Category Metric Median p-value 2012 2013 2014 Richness and Composition Taxa Richness 12.5 10.5 13.5 0.22 % EPT   9 .0   4.5 24.2 0.039 % Dominant Taxon 57.8 91.4 58.1 0.0498 Tolerance Family Biotic Index   5.1   5.9 4.9 0.039 % Intolerant Taxa   8.6   1.6 16.3 0.018 % Tolerant Taxa 3.2 0.2 4.7 0.0498 Functional Feeding Groups % Scraper 0.3 0.1 1.5 0.42 % Shredder   12.8   2.6 17.1 0.039 % Collector-Filterer   0.6   2.2 8.4 0.039 Habit Type % Clinger 4.4 2.7 7.3 0.78   Identical symbols identify groups that do not differ significantly from each other *p<0.05 **p<0.01 ***p<0.001 109

PAGE 117

! Additional metrics (Table 4.3) reflecting significant changes from year to year at Hunter Reach included percentage predator, percentage climber, and percentage sprawler (p < 0.05; see Appendix B). Of these metrics, statistical analysis could only identify the years of 2012 2013 2014 9 10 12 14 Taxa Richness Year Number of Taxa 2012 2013 2014 5 10 20 30 Percent EPT Year Percent 2012 2013 2014 30 50 70 90 Percent Dominant Taxon Year Percent 2012 2013 2014 4.5 5.0 5.5 Family Biotic Index Year Index 2012 2013 2014 5 10 15 20 Percent Intolerant Taxa Year Percent 2012 2013 2014 0 5 10 15 Percent Tolerant Taxa Year Percent               Identical symbols identify groups that do not differ significantly from each other Figure 5.26. Key metrics of macroinvertebrate community composition and tolerance at Hunter Reach in 2012, 2013, and 2014, with results of the Friedman test.   Identical symbols identify groups that do not differ significantly from each other Figure 5.27. Key functional feeding groups and macroinvertebrate habit types in communities at Hunter Reach in 2012, 2013, and 2014. 2012 2013 2014 0 1 2 3 4 5 Percent Scraper Year Percent 2012 2013 2014 5 10 15 20 25 Percent Shredder Year Percent 2012 2013 2014 0 5 15 25 35 Percent Collector! Filterer Year Percent 2012 2013 2014 0 5 15 25 35 Percent Clinger Year Percent         110

PAGE 118

change for percentage sprawler. Median percentage sprawler was higher in 2014 (12.3%) than in 2013 (1.9%). The median value of the percentage of predators was 4.8% in 2012, 0.9% in 2013, and 1.0% in 2014. Medians value of the percentage of climbers were 5.0%, 0.8%, and 4.0% in 2012, 2013, and 2014, respectively. Bear Creek: Gage Reach Of the key metrics of overall macroinvertebrate richness and composition at Gage Reach (another unburned reference site; Figure 3.3 ), statistical analysis identified a significant change only in taxa richness (p = 0.0498) (Table 5.38; Figure 5.28). This change could not be isolated to specific years, but median richness was lowest in 2013 (11) and highest in 2014 (16.5). Metrics of community tolerance did not exhibit significant changes from year to year (Table 5.36). Analysis of key functional feeding groups metrics at Gage Reach indicated significant changes in proportions of scrapers (p = 0.0498) and collector-filterers (p = 0.050) (Table Table 5.38. Key macroinvertebrate metrics of samples collected from Gage Reach (unburned) in 2012, 2013, and 2014, with results of the Friedman test Category Metric Median p-value 2012 2013 2014 Richness and Composition Taxa Richness 14.5 11 16.5 0.0498 % EPT 70.6 54.3 64.8 0.78 % Dominant Taxon 32.1 38.1 36.3 0.78 Tolerance Family Biotic Index 3.5 4.1 4.2 0.37 % Intolerant Taxa 36.8 20.7 18.7 0.37 % Tolerant Taxa 10.8 6.9 5.1 0.78 Functional Feeding Groups % Scraper 4.1 2.8 1.1 0.0498 % Shredder 22.3 4.2 18.0 0.37 % Collector-Filterer 17.3 4.2 24.2 0.050 Habit Type % Clinger 30.0 30.5 25.8 0.47   Identical symbols identify groups that do not differ significantly from each other *p<0.05 **p<0.01 ***p<0.001 111

PAGE 119

5.38; Figure 5.29). The Friedman multiple comparisons test failed to determine in which years these changes occurred. Median percentage scraper decreased at each time step (4.1% in 2012, 2.8% in 2013, and 1.1% in 2014). Median percentage collector-filterer was notably low in 2013 (4.2%) compared with other years (17.3% in 2012 and 24.2% in 2014). Values of the percentage of clingers showed no significant change from year to year (Table 5.38). 2012 2013 2014 6 8 12 16 Taxa Richness Year Number of Taxa 2012 2013 2014 30 40 50 60 70 Percent EPT Year Percent 2012 2013 2014 20 30 40 50 Percent Dominant Taxon Year Percent 2012 2013 2014 3.0 3.5 4.0 4.5 5.0 Family Biotic Index Year Index 2012 2013 2014 10 20 30 40 50 Percent Intolerant Taxa Year Percent 2012 2013 2014 5 10 15 Percent Tolerant Taxa Year Percent Figure 5.28. Key macroinvertebrate metrics of community composition and tolerance at Gage Reach in 2012, 2013, and 2014. 2012 2013 2014 0 2 4 6 8 Percent Scraper Year Percent 2012 2013 2014 0 10 20 30 40 Percent Shredder Year Percent 2012 2013 2014 0 10 20 30 40 50 Percent Collector! Filterer Year Percent 2012 2013 2014 10 20 30 40 50 60 Percent Clinger Year Percent Figure 5.29. Key functional feeding groups and macroinvertebrate habit types of communities at Gage Reach in 2012, 2013, and 2014. 112

PAGE 120

Analysis of additional metrics (Table 4.3) at Gage Reach indicated significant changes in metrics of functional feeding groups and habit types. The percentage of predators decreased significantly between the years 2013 (median = 20.8%) and 2014 (median = 1.7/5) (p < 0.05; Appendix B). Percentage climber also changed with time (p < 0.05), but the years of change were not specified for this metric. The median values of percentage climber were 2.2%, 9.6%, and 2.2% in 2012, 2013, and 2014, respectively. West Monument Creek: Academy Reach At the unburned Academy Reach (Figure 3.3), two of the three key metrics of overall richness and composition exhibited significant changes over time: percentage EPT and percentage dominant taxon (Table 5.39; Figure 5.30). Although a change was detected in the percentage of EPT organisms (p = 0.0498), but the Friedman multiple comparisons test failed to identify the years in which it changed. Median values of percentage EPT were 19.2%, 12.6%, and 31.1% in 2012, 2013, and 2014, respectively. Values of the percentage of the dominant taxon was lower in 2014 (median = 48.7%) than in 2013 (median = 71.1%) (p = 0.039). Of the measures of community tolerance, only the percentage of tolerant taxa exhibited change over time at Academy Reach (p = 0.0498) (Table 5.39; Figure 5.30). Although the Friedman test detected a significant difference among years, the multiple comparisons test failed to specify the timing of the change. The median values of percentage tolerant taxa were 6.1% in 2012, 3.1% in 2013, and 12.2% in 2014. 113

PAGE 121

! Table 5.39. Key macroinvertebrate metrics of samples collected from Academy Reach (unburned) in 2012, 2013, and 2014, with results of the Friedman test Category Metric Median p-value 2012 2013 2014 Richness and Composition Taxa Richness 16 17 16.5 1.0 % EPT 19.2 12.6 31.1 0.0498 % Dominant Taxon   48.3   71.1 48.7 0.039 Tolerance Family Biotic Index 5.2 5.7 5.4 0.78 % Intolerant Taxa 6.2 3.5 8.0 0.37 % Tolerant Taxa 6.1 3.1 12.2 0.0498 Functional Feeding Groups % Scraper 0.0 0.1 0.2 0.58 % Shredder 6.6 4.1 8.7 0.37 % Collector-Filterer 7.8 38.8 11.3 0.47 Habit Type % Clinger   13.9   42.7 2.5 0.039   Identical symbols identify groups that do not differ significantly from each other *p<0.05 **p<0.01 ***p<0.001 2012 2013 2014 8 12 16 20 Taxa Richness Year Number of Taxa 2012 2013 2014 10 20 30 40 50 Percent EPT Year Percent 2012 2013 2014 40 50 60 70 80 Percent Dominant Taxon Year Percent 2012 2013 2014 4.5 5.0 5.5 Family Biotic Index Year Index 2012 2013 2014 5 10 20 30 Percent Intolerant Taxa Year Percent 2012 2013 2014 5 10 15 20 Percent Tolerant Taxa Year Percent       Identical symbols identify groups that do not differ significantly from each other Figure 5.30. Key metrics of macroinvertebrate community composition and tolerance at Academy Reach in 2012, 2013, and 2014, with results of the Friedman test. 114

PAGE 122

Proportions of the three key functional feeding groups did not differ among years at Academy Reach (Table 5.39; Figure 5.31). The key metric of habit types, percentage clinger, significantly decreased from 2013 (median = 42.7%) to 2014 (median = 2.5%) (p = 0.039). One additional metric (Table 4.3), percentage predator, exhibited significant change over time. The median proportion of predators decreased significantly from 2012 (12.5%) to 2013 (2.0%) (p < 0.05) (Appendix B). Summarizing Key Results Figure 5.35 illustrates temporal trends of the key macroinvertebrate metrics of overall richness and composition and community tolerance for all reaches. These plots supplement the statistical analysis by elucidating trends that the Friedman test may have failed to detect because of the small sample sizes. From these plots, several important points can be identified : In burned sites, taxa richness and percentage EPT increased over time. The Family Biotic Index generally decreased (i.e. ecological condition improved) with time. All study channels exhibited high variability in the percentage of the dominant taxon, 2012 2013 2014 0.0 0.2 0.4 0.6 0.8 Percent Scraper Year Percent 2012 2013 2014 0 5 10 15 20 25 30 Percent Shredder Year Percent 2012 2013 2014 0 20 40 60 80 Percent Collector! Filterer Year Percent 2012 2013 2014 0 20 40 60 80 Percent Clinger Year Percent       Identical symbols identify groups that do not differ significantly from each other Figure 5.31. Key functional feeding groups and macroinvertebrate habit types at Academy Reach in 2012, 2013, and 2014. 115

PAGE 123

but the temporal trends of this metric were distinct between the reaches of high burn severity and those of low burn severity and no burn. Percentage intolerant taxa generally increased or remained stable for all categories of burn. Trends in percentage tolerant taxa were variable. Macroinvertebrate communities in unburned reference sites were relatively stable in metrics of taxa richness and percentage EPT (two measures of overall richness and composition), as well as in the three metrics representing community tolerance. These results addressed the fourth research question: How does the response of macroinvertebrate communities to wildfire vary over time? In general, sites of high burn severity exhibited more variability from year to year, whereas patterns of change in low severity burn sites often mirrored those seen in reference sites. Aussie Reach can be 5 10 15 20 Taxa Richness Year Number of Taxa 2012 2013 2014 0 20 40 60 80 Percent EPT Year Percent 2012 2013 2014 20 40 60 80 Percent Dominant Taxon Year Percent 2012 2013 2014 2 3 4 5 6 7 8 Family Biotic Index Year Index 2012 2013 2014 0 20 40 60 Percent Intolerant Taxa Year Percent 2012 2013 2014 0 10 20 30 40 50 Percent Tolerant Taxa Year Percent 2012 2013 2014 Unburned Low burn severity High burn severity Figure 5.35. Change over time in key metrics of macroinvertebrate community composition and tolerance. 116

PAGE 124

distinguished among the sites of high burn severity burn because its metrics more closely resembled those for sites of low burn severity. Overall, the results support the hypothesis that burned study channels would show indications of ecological recovery by 2014, two years after the Waldo Canyon Fire. Discussion The results presented in this chapter characterize the ecological response of step-pool streams to the Waldo Canyon Fire of 2012. This section discusses these results and interprets them in the context of the existing literature. Presence and Severity of Burn The first research question addressed the response of macroinvertebrate communities to wildfire disturbance. Although analysis of macroinvertebrate metrics as a function of burn presence/absence revealed some significant differences, these were generally overshadowed by differences revealed with the classification of burned sites into sites of high and low burn severity (the second research question). This analysis suggested that the magnitude of ecological response was generally much greater in channels burned with high severity than in channels burned with low severity. This was expected as a result of the amplified effects of hydrologic events due to greater loss of vegetation and ground cover, and stronger soil hydrophobicity with increasing burn severity (Huffman et al. 2001; Rhoades et al. 2011). These results were consistent with the findings of Malison and Baxter (2010b) at a mid-term timeframe following a fire in central Idaho. This study found that five years after fire, insect communities of streams in areas of low burn severity exhibited greater differences to those in areas of high burn severity than did communities in unburned areas. The similarities between 117

PAGE 125

channels burned with low severity and unburned channels in the present study supported the idea proposed by Minshall (2003) that macroinvertebrate community response occurs above a threshold of burn extent and severity. This was in contrast to the findings of Bche et al. 2005 and Arkle et al. 2010 that the degree of macroinvertebrate response is proportional to the degree of burn severity. The nature of the ecological response appeared to differ between channels burned with high severity and those burned with low severity. When comparing sites of high burn severity to reference sites, differences occurred in metrics of overall richness and composition, community tolerance, functional feeding groups, and habit types. In contrast, sites of low burn severity tended to diverge from reference sites only in the composition of functional feeding groups (collector-filterers, shredders, and scrapers) and habit types (sprawlers). This suggested that habitat characteristics such as available food sources were altered in low severity burn sites with relatively little degradation of overall ecological condition. These findings were consistent with those of Minshall et al. (1997). This study reported changes in proportions of functional feeding groups, but not in metrics of richness, density, or biomass, following a fire in Yellowstone National Park. Similarly, Bche et al. (2005) found that a low severity prescribed fire in the Sierra Nevada range induced changes in overall macroinvertebrate community composition, but not in metrics of abundance, richness, and diversity. Stream Habitat Type The analysis of habitat types (steps and pools; the third research question) suggested that the ecological impact of fire was more pronounced in pools than in steps of burned study 118

PAGE 126

reaches, consistent with the hypothesized result. The ecological condition of pools in channels burned with high severity was degraded compared to that of pools in reference channels, whereas the condition of steps in channels of high burn severity was not degraded compared to steps in reference channels. The steps of channels burned with high severity did, however, differ from steps of channels burned with low severity. These results suggested that steps of low burn severity sites, mores than pools, maintained ecological quality similar to or exceeding that of reference steps. The difference in the ecological condition of steps and pools in 2012 could be attributed to the slow velocity of water in pools that would promote deposition and accumulation of fire-related materials such as smoke and ash (Wood and Armitage 1997). These materials have been linked with degraded habitat quality in the months following wildfire, particularly in small streams (Earl and Blinn 2003; Minshall 2003; Rhoades et al. 2011). In subsequent years, the differing response of steps and pools was likely a result of the stability of step clasts and relative instability of pool substrate. Geomorphically stable habitats have been associated with improved ecological condition (Sullivan et al. 2004) Change over Time Temporal trends of change, which were addressed by the fourth research question, also differed between sites of high and low burn severity. Contrary to the hypothesized result, high severity burn sites exhibited many signs of degraded ecological condition relative to reference sites in 2012. Differences between low severity burn and unburned sites, on the other hand, were minimal in 2012. The response of macroinvertebrate communities in 2012 resulted from the direct effects of wildfire because sampling occurred before secondary 119

PAGE 127

hydrological disturbances. Previous studies addressing the direct effects of fire in the absence of secondary disturbances have reported variable macroinvertebrate responses. The immediate response of communities in low severity burn sites was consistent with the findings of Rinne (1996) in Arizona. This study found that, in general, macroinvertebrate communities were not immediately affected by the direct effects of wildfire. The reduction of ecological quality seen in high severity burn sites in the absence of secondary disturbances, however, was consistent with the results of Minshall et al. (1998) in Yellowstone National Park and Oliver et al. (2012) in the Lake Tahoe basin. The small size of the studies streams in the present study likely made them susceptible to the direct effects of severe fire, such as intense heating and ash inputs (Minshall 2003). Earl and Blinn (2003) found that following moderate intensity fires in New Mexico, ash inputs into streams had variable impact on macroinvertebrate communities, which they attributed to characteristics of the ash itself. The most striking temporal trend in macroinvertebrate metrics was a tendency toward further degradation of ecological condition at high severity burn sites in 2013 (one year postfire), whereas sites of low burn severity tended toward maintenance or improvement. The degradation of high severity burn sites reflected the influence of secondary hydrological disturbances that occurred as a result of major precipitation events (Table 4.2). The responses of channels burned with high severity, however, were not consistent. Similarly, Minshall et al. (1998) reported widely variable responses to the indirect effects of wildfire in streams burned with similarly high severity and extent. The authors attributed this variability to differences in environmental characteristics (e.g., precipitation, topography, and geology). For example, small, steep watersheds tended to exhibit greater responses to post-fire 120

PAGE 128

hydrological disturbances. One year post-fire (in 2013), Willis Reach (a reach of high burn severity in Williams Canyon) exhibited large reductions in the percentage of intolerant taxa, with increases in the Family Biotic Index and percentage tolerant taxa. These changes were consistent with the findings of previous studies following fires of high burn severity (Rinne 1996; Minshall et al. 1998; Vieira et al. 2004). These studies documented decreases in macroinvertebrate density, taxa richness, and sensitive taxa in response to post-fire hydrological events. Percentage dominant taxon shifted from significantly higher to notably lower at high severity burn sites compared to sites of low burn severity and no burn between 2012 and 2013. This change in dominance was consistent with the theory of dynamic equilibrium in which a disturbance enables the displacement of dominant species by species that are capable of recolonizing perturbed habitats (Huston 1979). Despite the decrease in percentage dominant taxon, taxa richness remained stable, though comparatively low, at Willis Reach from 2012 to 2013. In contrast, communities at Aussie Reach (a reach of high burn severity in Camp Creek) exhibited a temporally similar response to communities in reaches burned with low severity. Aussie Reach exhibited signs of poorer habitat quality relative to Eagle Reach and Meadow Reach in 2012, including lower richness, percentage EPT, and percentage intolerant taxa, and higher percentage dominant taxon and Family Biotic Index. This suggested that Aussie Reach was more strongly impacted by the direct effects of fire. All three reaches, however, generally improved in ecological condition in 2013, indicating minimal impact from the indirect effects of fire following hydrologic disturbance. Biological metrics of high severity burn sites exhibited overall improvement in 121

PAGE 129

ecological condition by 2014 (two years post-fire). Similar improvement, however, was also seen in low severity burn and unburned sites at this time step. The relative ecological condition of channels burned with high severity therefore remained somewhat impaired two years after the fire, though to a lesser degree than in previous years. Percentage EPT, for example, was lower in high severity burn sites than in unburned sites 2012 and 2013, but attained reference levels in 2014. Taxa richness and percentage intolerant taxa, however, remained lower in high severity burn sites than in reference sites. This was consistent with many studies that reported no recovery of metrics such as taxa richness and percentage sensitive taxa at timeframes ranging from two to ten years (e.g., Richards and Minshall 1992; Vieira et al. 2004; Oliver et a. 2012). Some studies, however, have seen these ecological variables return to reference conditions within two years (e.g., Bche et al. 2005; Hall and Lombardozzi 2008). These differences in the timeline of recovery may be attributable to differences in post-fire hydrological and geomorphological changes, watershed characteristics (such as size, terrain, geology, or vegetation), availability of refugia, and accessibility to recolonizing organisms (Sedell et al. 1990; Richards and Minshall 1992; Vieira et al. 2004; Sueyoshi et al. 2014). The next chapter investigates the interactions between physical and ecological processes in the wake of a wildfire. It first presents the results of an exploratory statistical analysis of the relationships between ecological and geomorphological response, followed by an interpretation of the ecological response in the context of the changing geomorphology of the study channels following the fire. 122

PAGE 130

CHAPTER VI EXPLORING THE ECOLOGICAL RESPONSE WITHIN A CHANGING STEPPOOL MORPHOLOGY Integrating Bio-physical Characteristics Statistical analysis of the ecological data generally supported the hypotheses to the research questions discussed in the previous chapters. Two salient findings were that (1) channels burned with high severity exhibited significantly greater ecological degradation than channels burned with low severity, and (2) pool habitats showed the varying effects of fire with severity of burn more than step habitats. The ecological response following wildfire, however, results from complex interactions among post-fire hydrological, sedimentological, and geomorphological processes (Figure 1.2). This chapter explores some of these interactions in an effort to identify correlations among ecological responses and physical changes in the step-pool system. Additional data for the analysis discussed in this chapter come primarily from the larger sponsored project under which the ecological analyses for this thesis research were carried out (NSF EAR1254989 to Principal Investigator Dr. Anne Chin). These data include longitudinal profiles of the study reaches, sediment size distributions (median diameter or d 50 ), and measures of water quality (dissolved oxygen, temperature, and pH). These measurements were taken at selected time steps immediately after the Waldo Canyon Fire in 2012 and following storm events through 2014, in conjunction with the sampling of benthic macroinvertebrates reported in the previous chapters. 123

PAGE 131

To explore the biological data together with the physical characteristics of burned channels, ordination provides a powerful tool for integrative analysis. The statistical procedures that generated the results presented in Chapter V only allow testing of a single variable at a time (e.g., the severity of burn or habitat type). Ordination, on the other hand, is a multivariate statistical technique commonly used by ecologists that enables exploration of many parameters simultaneously (Zuur et al. 2007). For analysis of ecological data, such as the composition of biological communities, results of ordination include a visualization of the ecological distances among sample units. In other words, ordination generates a graphical distribution representing ecological similarities in the composition of biological communities, whereby communities with similar compositions are plotted near each other. Ordination may be constrained by environmental variables (such as burn severity and elevation) and other physical characteristics (such as sediment size), so that the distribution of samples correlates with these variables. These correlations indicate the influence of environmental and physical characteristics on the assemblages of macroinvertebrates. Previous studies have employed ordination methods to investigate interactions among the ecological and environmental characteristics of stream channels. Many environmental variables have correlated with the composition of aquatic communities, such as discharge, water depth, channel geometry, substrate particle size, water quality, and longitudinal position along the channel (Zuellig et al. 2010; Brown et al. 2012; Chung et al 2012; Troia and Gido 2012). The literature demonstrates, however, that the influence of such variables on community structure is highly variable. Studies using ordination to investigate the effects of wildfire specifically have found that the presence and severity of burn interact with 124

PAGE 132

hydrological processes to dictate flow conditions (e.g., peak streamflow or water velocity), which subsequently alters habitat characteristics such as sediment load and the availability of organic debris (e.g., Mellon et al. 2008; Arkle et al. 2010; Oliver et al. 2012). Such changes to the hydrological and physical environment of aquatic habitats ultimately govern ecological condition and the composition of macroinvertebrate communities in streams affected by wildfire. For the present study, ordination assessed similarities in taxonomic composition among all macroinvertebrate samples with consideration of several characteristics of the study streams and watersheds. Table 6.1 lists the specific variables considered in the analysis (see Table 3.1 for data). The ecological data included the abundance of organisms for each taxon present in a sample. Abundances were normalized by subsampling the total abundance of organisms in each sample to a count of 300 (see Chapter IV). Each taxon represented a different ecological variable. PC-ORD 6.08 was the software used to run nonmetric multidimensional scaling (NMS), which is a type of ordination analysis, on the ecological data (taxa abundances of macroinvertebrate communities) along gradients of these variables. Table 6.1. Variables considered in nonmetric multidimensional scaling analysis. Data for some of these variables are provided in Table 3.1. Ecological Physical Water Quality Environmental Taxa abundances Dissolved oxygen Study watershed Median particle size (d 50 ) Water temperature Burn severity Habitat (step or pool) pH Time since fire Drainage area % of area burned with high severity Elevation Steepness parameter ( ) 125

PAGE 133

The steepness parameter ( ) is the primary index used in this study to represent the physical character or degree of step-pool development (e.g., Lenzi 2001). In this analysis, H is the height from the top of a step identified in the longitudinal profile data to the deepest point of the adjacent pool. L is the length from one step crest to the next step crest. S is the slope of the 50-meter study reach. Higher values of indicate a more developed steppool morphology characterized by large (high) steps and deep pools spaced at relatively short intervals along the channel (Figure 6.1a). A ratio of one reflects a perfect staircase-like structure, in which step length over step height is equal to the overall channel slope (i.e., step treads are horizontal with no reverse slope; Figure 6.1b). Values of < 1 result when steps are spaced far apart with corresponding small heights, and without the scouring that produces reverse-sloped treads (Figure 6.1c). Values of for natural step-pool streams commonly fall between one and two because scouring in pools tends to exaggerate step heights relative to length. This range reflects a step-pool morphology with maximum flow resistance (Abrahams et al. 1995). (a) > 1 (c) < 1 (b) = 1 L S H Figure 6.1. Diagrams of step-pool sequences showing bed configurations where the steepness parameter is (a) greater than one; (b) equal to one; and (c) less than one. 126

PAGE 134

Previous studies have investigated the hydrological conditions associated with the alteration of step-pool morphology using the steepness parameter. Lenzi (2001) found that large floods with recurrence intervals of 30-50 years have the potential to destabilize steppool sequences, thereby decreasing to below one (i.e., step steepness < channel slope). Lenzi (2001) further found that smaller events will, over time, reorganize the channel and restore a steepness parameter of between one and two. Chin and Phillips (2007) additionally demonstrated how can change with development of step-pool channels from plane beds. They suggested that the formation of step-pool sequences generates a net increase in entropy that contributes to the flow resistance and energy dissipation provided by these morphological features. Although researchers have proposed other methods to quantify steppool sequences (e.g., Zimmerman et al. 2008), the ratio provides a simple metric that is possible to calculate from available data for entry into the ordination analysis in this study. Results of Ordination Analysis The best solution generated from ordination (NMS) analysis was a two-dimensional distribution of the macroinvertebrate samples distinguished according to the severity of burn of the sampling site (Figure 6.2; minimum stress = 16.0; final instability = 0.000010). The ordination plot (Figure 6.2) shows the distribution of all samples collected at all time steps along two axes. The horizontal axis (Axis 1) incorporates the percentage of the upstream drainage area burned with high severity (labeled "HiSev"), showing an overall gradient from large to small percentages (left to right). This axis therefore represents a measure of disturbance from high to low (left to right). The vertical axis incorporates the size of the drainage area of the upstream watershed (labeled "Area") from small to large areas (bottom 127

PAGE 135

to top; discussed in a later section). Each point on the plot represents a single sample (see Table 6.2 for definitions of the sample codes). For example, TR13HP on the left side of the plot refers to the sample collected from Tributary Reach (TR) in 2013 (13) from the high (i.e., farthest upstream) pool of the study reach (HP). Other ordination plots distinguished samples according to the study watershed, sampling year, and habitat type, but they were less informative. The correlation of the areal percentage of high burn severity with the horizontal axis (Axis 1 on Figure 6.2; shown by arrow symbol) indicates a significant influence of burn severity on the ecological character of the affected streams. Two visual observations pertaining to the horizontal distribution of samples on the ordination plot (Figure 6.2) are revealing. First, most of the points in the distribution cluster together on the right side of the plot (Figure 6.3a). This cluster includes nearly all of the points representing samples from unburned and lowly burned sites, as well as some of the points representing channels of high burn severity. The relatively short ecological distances between points within this cluster indicate similarities in the taxonomic composition of these macroinvertebrate samples. This plot therefore confirms the results presented in the previous chapter that the ecological character of macroinvertebrate samples from lowly burned sites are similar to those of unburned reference channels. In other words, channels burned with low severity are minimally disturbed. 128

PAGE 136

! Figure 6.2. Nonmetric multidimensional scaling (NMS) plot showing ecological distances among samples at all time steps. Samples are distinguished by the severity of burn of the sites from which they were collected. Table 6.2 provides definitions of the codes for each sample, representing the samp ling location. Burn Severity Unburned Low Severity High Severity 129

PAGE 137

Second, most of the points falling outside of this cluster, indicating ecological dissimilarity, represent samples from channels burned with high severity. These samples exhibit a much wider horizontal range on the ordination plot (Figure 6.3b). Different from the tight clustering of low severity burn and unburned sites along Axis 1 (Figure 6.3a), this broad distribution of points for severely burned sites suggests that the ecological responses of severely burned channels varied greatly. For example, samples from Aussie Reach (a reach of high burn severity in Camp Creek) fall mostly within the cluster of points representing low burn severity and unburned sites on the NMS plot (AU12MS, AU12LS, AU12MP, AU12LP, AU13MS, AU13LS, AU13MP, AU13LP, AU14LS, AU14MP, and AU14LP on Figure 6.3a; shown with diamond symbols). This result suggests that samples from Aussie Reach in Camp Creek, even though severely burned, are similar to those from lowly burned sites. On the other hand, samples from Willis Reach and Tributary Reach (two reaches of high burn severity in Williams Canyon) span a broad horizontal range (TR13HP, TR13HS, TR13LP, TR13LS, TR14LP, TR14LS, TR14HP, TR14HS, WL13HP, WL13LP, WL13HS, and WL13LS on Figure 6.3b; also shown with diamond symbols). The distances of the points Table 6.2. Explanation of codes for sites names used in NMS analysis (Figure 6.5). Code Reach Code Year Code Habitat* WL Willis 12 2012 LP Low Pool TR Tributary 13 2013 MP Middle Pool AU Aussie 14 2014 HP High Pool EA Eagle LS Low Step ME Meadow MS Middle Step HU Hunter HS High Step GA Gage AC Academy *Habitat codes indicate the specific cross section of macroinvertebrate sampling (see Chapter IV for sampling methods). For example, Low Pool and Low Step refer to the furthest downstream cross section of a given channel reach. 130

PAGE 138

representing Willis Reach and Tributary Reach away from the cluster of points for sites of low burn severity and unburned reference sites suggest their ecological dissimilarity. Figure 6.3. NMS plots showing ecological distances among samples at all time steps. (a) The circled area encompasses a cluster of points representing most of the samples from sites of low severity burn and no burn. This cluster also includes some of the points representing high burn severity. (b) The circled areas encompass the points representing high burn severity that diverged from the cluster of points representing low burn severity and no burn. Burn Severity Unburned Low Severity High Severity (a) Burn Severity Unburned Low Severity High Severity (b) 131

PAGE 139

Closer examination of the points representing samples from Willis Reach and Tributary Reach (that diverge from the cluster for low burn severity and unburned reaches) shows two distinct groups (Figure 6.3b; left side). One group at the far left side of the plot includes the four samples collected from Tributary Reach in Williams Canyon in 2013 (for example, TR13HP) and one sample collected from Willis Reach (Williams Canyon) in 2013 (WL13HP). These points are the farthest from those representing unburned reference sites, therefore suggesting the greatest ecological disturbance. A second group of points representing high burn severity fall only slightly to the left of the cluster of points that represent unburned reference samples (Figure 6.3b). These points represent three of the samples collected from Willis Reach in Williams Canyon in 2013 (one year after fire; for example, WL13LP) and the four samples collected from Tributary Reach in Williams Canyon in 2014 (two years after fire; for example, TR14LP). Taken together, the distribution of points representing samples from channels burned with high severity (Willis Reach, Tributary Reach, and Aussie Reach) suggests variations in ecological degradation and recovery for severely burned channels. First, all samples from Aussie Reach (Camp Creek) reflect little or no difference from reference conditions at any time step, given the proximity of points representing Aussie Reach to those of unburned sites. Second, although the points for Willis Reach and Tributary Reach in Williams Canyon both diverge from the data representing unburned reference reaches in 2013, the ecological distances are greatest for samples from Tributary Reach, suggesting more severe ecological degradation at this reach. Third, both Willis Reach and Tributary Reach exhibit signs of recovery in 2014. Points representing Willis Reach for 2014 (WL14LS, WL14HS, WL14LP, 132

PAGE 140

and WL14HP) fall within the cluster of points representing the unburned reaches (for example, GA14HP), suggesting recovery to reference condition. The clustering of points from Tributary Reach for 2014 (for example, TR14HS) with those from Willis Reach for 2013 (for example, WL13LS) suggests that the recovery of Tributary Reach lagged behind that of Willis Reach by about one year. Besides the severity of burn, the only other environmental variable that correlates significantly with the distribution of the ecological data is the drainage area of the upstream watershed (Figure 6.2; shown by arrow symbol). Because drainage area correlates with Axis 2, the vertical distribution of study sites on the NMS plot can be explained in part by a gradient of small to large drainage areas. This result suggests that the taxonomic composition of macroinvertebrate communities was influenced by the size of the upstream watershed. Samples from sites of each level of burn severity are similarly distributed along Axis 2, suggesting that burn severity does not help to explain vertical variability on the plot. A Changing Step-Pool Morphology Results of the ordination analysis discussed above point to varying ecological responses when channels are severely burned. That is, even though the severely burned sites as a whole generally show greater impact than sites of unburned reference channels (as well as sites burned by low severity), variation in this degradation is nevertheless evident by the spread in the points representing severely burned sites in the ordination plots (Figures 6.3a and 6.3b). The remainder of this chapter examines the physical character of the step-pool channels in severely burned sites in relation to the ecological response. In other words, as ecological conditions degraded with high burn severity, how did these conditions correlate 133

PAGE 141

with the physical structure of step-pool systems? The following discussion focuses on the three study reaches burned by high severity: Aussie Reach in Camp Creek, and Willis Reach and Tributary Reach in Williams Canyon. Results of ordination analysis suggest that Aussie Reach (Camp Creek) did not experience ecological degradation following the Waldo Canyon Fire. Instead, the ecological characteristics of benthic macroinvertebrate communities are similar to those of reference reaches (shown by the cluster of points in Figure 6.3a). Examination of surveyed longitudinal profiles shows that Aussie Reach also exhibited minimal change in its physical character through two post-fire years (Figure 6.4). The steepness ratio maintained an ideal range from 1.52 to 1.74. For Willis Reach in Williams Canyon, results of ordination analysis suggest ecological degradation from 2012 to 2013 (one year post-fire), followed by improvement (recovery) in ecological condition in 2014. Field observations and surveyed longitudinal profiles reveal dramatic aggradation of the channel bed following storms during the summer of 2013 (Figure 6.5; Table 4.2). Sediment filled in the pools and buried the step-pool Elevation (m) 2498 2500 2501 2503 2504 Distance (m) 0 10 20 30 40 2012 2013 Jul 2013 Nov 2014 Oct Aussie Reach H/L/S 1.74 1.52 1.53 1.59 Figure 6.4. Longitudinal profiles surveyed for Aussie Reach in Camp Creek illustrating minimal geomorphological change in the years following the Waldo Canyon Fire. 134

PAGE 142

morphology. By 2014, however, the step-pool morphology began to re-form. Correspondingly, values of the steepness parameter dropped from 1.72 to 0.98 (indicating low step-pool development) at the time of macroinvertebrate sampling 2013, when the reach experienced substantial aggradation of the channel bed. The steepness ratio increased to 1.77 when step-pools re-formed in 2014. For Tributary Reach in Williams Canyon, ordination analysis indicated that the ecological character was considerably degraded in 2013, and exhibited some signs of recovery toward reference conditions in 2014 (Figure 6.3b). Longitudinal profiles of Tributary Reach show substantial degradation of the channel bed from 2012 to 2013 and continued degradation into 2014 (Figure 6.6). The first post-fire storms of summer 2013 (Table 4.2) apparently produced substantial changes in the channel bed. Although the ratio remained at 0.93 for 2012 and 2013, the morphology changed from several small, regularly-spaced step-pool systems to only one or two large steps along the length of the reach. The steepness parameter fluctuated throughout the storms of 2013 and 2014, Elevation (m) 2264 2265 2266 2267 2268 Distance (m) 0 15 30 45 60 2012 Nov 2013 Jul 2013 Oct 2014 Aug Willis Reach H/L/S 1.72 0.98 1.37 1.77 Figure 6.5. Longitudinal profiles surveyed for Willis Reach in Williams Canyon illustrating the aggradation of the channel bed and the destruction of step-pool sequences from 2012 to 2013, followed by re-formation of a step-pool morphology by 2014. 135

PAGE 143

frequently dropping below a ratio of one. At the time of macroinvertebrate sampling in 2014, the value was 0.66. Qualitative assessments of the physical structure of the step-pool system suggest a correlation with ecological responses. When the step-pool morphology remained intact, such as in Aussie Reach, ecological quality also remained strong (Figure 6.3a). When step-pool sequences were obliterated or buried, such as in Willis Reach in 2013 or Tributary Reach in 2013 and 2014, ecological degradation is evident (Figure 6.3b). Correspondingly, ecological recovery was apparent in Willis Reach in 2014 when step-pool sequences re-formed. The reformation of a step-pool morphology presumably re-established the stability and diversity of habitat types that provide a high-quality environment for aquatic communities (Sullivan et al. 2004; Milner and Gilvear 2012). Exploring this point further, the ordination plot in Figure 6.7 shows the ecological character of step-pool channels, as distinguished arbitrarily by a steepness ratio of one (i.e., a perfect staircase-like structure; Figure 6.1b). The figure shows that ecological degradation correlates with poorly developed step-pool structure ; channel sites with steepness ratios less Elevation (m) 2248 2250 2252 2254 2256 Distance (m) 0 15 30 45 60 2012 2013 Jul 2013 Aug 2013 Sept 2014 Jul 2014 Aug Tributary Reach H/L/S 0.93 0.93 1.39 0.93 1.03 0.66 Figure 6.6. Longitudinal profiles surveyed for Tributary Reach in Williams Canyon illustrating the degradation of the channel bed and alteration of step-pool sequences from 2012 to 2013. This degraded channel morphology persisted into 2014. 136

PAGE 144

than one plot on the left side of the diagram. On the other hand, good ecological conditions occur when the step-pool morphology is well developed (steepness ratio greater than one; channel sites plot on the right side of Figure 6.7). These conditions occur at unburned reference channels, sites with low burn severity, and even a few sites severely burned. These data therefore suggest the significance of the step-pool morphology in maintaining the ecological integrity of the river system (Chin et al. 2009b), even in a post-fire environment. The next and final chapter reviews the key findings of this study in the context of previous literature. It outlines the limitations of the study and discusses the significance of the results. Steepness Parameter < 1 > 1 Figure 6.7. NMS plot distinguishing samples by the steepness parameters of the study sites from which they were collected. Sites with steepness parameters of one or greater, indicating maximum flow resistance, cluster together at the right side of the plot. Sites with steepness parameters of less than one diverge to the left of this cluster. 137

PAGE 145

CHAPTER VII SUMMARY AND CONCLUSIONS This chapter summarizes the major findings of this study. It also addresses the limitations of the study, identifies areas of future research, and discusses the significance and implications of the findings for management. Summary of Major Findings The objective of this study was to document the ecological response of step-pool mountain streams after the 2012 Waldo Canyon Fire of Colorado. The research questions were: (1) How so benthic macroinvertebrate communities in step-pool mountain streams respond to wildfire disturbance? (2) How does the response of macroinvertebrate communities to wildfire vary with the severity of burn? (3) How does the response of macroinvertebrate communities to wildfire vary as a function of habitat type (step or pool)? (4) How does the response of macroinvertebrate communities to wildfire vary over time? The analysis therefore addressed ecological response as a function of burn presence or absence, burn severity, habitat type, and time since fire. The ecological responses identified were also considered within a context of geomorphological change in the step-pool system. The major findings are as follows. First, burned streams showed degraded ecological conditions overall following wildfire compared to unburned reference streams. Second, in general, sites of high burn severity were clearly impacted by the effects of fire immediately after the disturbance. Sites of low burn severity, on the other hand, did not indicate poor ecological condition relative to reference sites immediately following the wildfire. They were, instead, more similar to the ecological conditions of unburned channels. One year post! 138

PAGE 146

fire, sites of high burn severity, particularly those in Williams Canyon, exhibited further degradation in ecological quality, although some severely burned sites in the Camp Creek watershed did not reflect this trend in 2013 (one year after the burn). These sites did, however, indicate some changes in the composition of functional feeding groups and habit types within macroinvertebrate communities. Third, pool habitats in burned channels exhibited poorer ecological conditions relative to reference sites than step habitats. This trend was especially evident in areas of high burn severity. Fourth, signs of improvement in the ecological conditions of burned channels were evident within two years post-fire (by 2014). In some cases, the degradation and improvement correlated with physical changes in the step-pool structure. In Willis Reach, for example, ecological condition degraded after fire along with alteration of the step-pool morphology, and then greatly improved from 2013 to 2014 as step-pool sequences were re-established. Tributary Reach, in which the degraded channel morphology of 2013 did not recover in 2014, did not reflect this ecological recovery. Drawing from these findings, the effects of wildfire on channel condition are spatially and temporally variable. This study found that the magnitude and nature of impact depended on characteristics of the fire itself, such as the extent and severity of burn. The response of stream channels in this study also related to the physical structure of the step-pool system. The ability of stream channels to retain a step-pool morphology through post-fire disturbances is apparently particularly important for the preservation of ecological integrity in aquatic communities. 139

PAGE 147

Limitations of Study Study Design Each stage of research in ecological study inevitably involves errors and/or limitations. Selection of the locations of study reaches, for example, was inherently limited by the spatial distribution of burn severity within the affected area, availability of stream channels within each burn severity, and accessibility for field data collection. These limitations led to variation among reaches in factors such as elevation, drainage area, aspect, riparian vegetation, and others, that were not included as explanatory variables in statistical analysis. To the extent possible, variation in these confounding factors was kept to a minimum by careful selection of study and reference reaches. Nevertheless, the drainage area of Willis Reach was somewhat larger than that of the other study reaches, followed by Aussie Reach (see Table 3.1). Eagle Reach, Meadow Reach, and Tributary Reach were more similar in drainage area. Of the reference reaches, Academy Reach was smaller than all study reaches and Gage Reach was substantially larger. Differences in drainage area indicate variations in the longitudinal position of reaches along the length of a stream. According to the River Continuum Concept, longitudinal positions are associated with specific characteristics of macroinvertebrate communities such as composition of functional feeding groups (Vannote et al. 1980). Elevations of the study reaches also varied (Table 3.1). The two reaches of low burn severity had the highest elevations, followed by the three reaches of high burn severity. Elevations of the three reference reaches were lower than all burned study reaches. Channel slopes were fairly consistent, although the slope of Meadow Reach was comparatively low, 140

PAGE 148

whereas that of Academy Reach was fairly high. Temporal variation in the sampling of macroinvertebrate communities likely also influenced the calculation of metrics (Linke et al. 1999; Bche et al. 2006). When multiple sampling efforts occurred for a given reach in a single year, samples were selected for analysis to minimize seasonal variation. Nonetheless, sample collection spanned from late summer (August) to late fall (December), undoubtedly contributing to differences detected among reaches as a result of expected seasonal trends in macroinvertebrate communities. Field and Laboratory Protocols In the field, some variation may have been introduced into the procedures for benthic sampling and longitudinal profile surveys. For benthic sampling, the vigor with which the substrate was disturbed likely varied from sample to sample. For surveying, the identification of steps and placement of the rod through step-pool structures could have varied from survey to survey. Such variations could have influenced the calculation of geomorphological dimensions, causing some uncertainty in the consistency of values of the steepness ratio ( ). In the laboratory, macroinvertebrates were primarily identified to family, providing a coarser picture of community composition than identification to genus or species. Assignment of tolerance values, functional feeding groups, and habit types for macroinvertebrate families is less accurate than assignment to lower taxonomic levels because genera and species within a family often vary greatly. Barbour et al. (1999), however, noted that identifications to family are more precise among different samples and taxonomists. Where possible, tolerance values, functional feeding groups, and habit types 141

PAGE 149

were assigned to macroinvertebrate taxa based on information specific to the region (Barbour et al. 1999). Some individuals were damaged to a degree that identification to the specified taxonomic level was difficult or impossible, resulting in uncertainties and unknowns. Unknown individuals and those without heads were eliminated from samples for the calculation of macroinvertebrate metrics. Uncertainties in which the order could be identified, but not the family, were treated as "not distinct" from other families within that order unless there were strongly divergent characteristics that could confirm a distinct taxon. Individuals classified as "not distinct" were included in total abundances, but omitted from the calculation of the macroinvertebrate metrics used in statistical analysis (CABW 2003). Statistical Analysis The small sample sizes and nonparametric tests used in the analysis resulted in low statistical power This situation likely led to type II errors, in which statistical tests failed to reject a false null hypothesis. This error was particularly likely in the analysis of macroinvertebrate metrics by habitat type and change over time because replicates in these analyses were so few. In many cases, statistical tests failed to identify significant differences for metrics despite apparently clear trends exhibited by quartile spreads. These trends may have reflected true differences among groups that were undetected by the analysis. Therefore, similar studies utilizing more replicates from a large number of study reaches could provide valuable supplementation to the results of this study. 142

PAGE 150

Significance and Implications for Management The findings of this study contribute to the scientific literature regarding step-pool streams, particularly augmenting the limited knowledge regarding the role of step-pool sequences in the response of stream channels to disturbance. Specifically, the research improves understanding of the ecological responses of step-pool streams to wildfire within a context of geomorphological change. The results of this study demonstrated that pool habitats were more susceptible to ecological degradation after wildfire compared to step habitats. This finding suggests that steps might represent refugia of ecological diversity and integrity following a disturbance to the stream environment. The interactions between the ecology and geomorphology of the study reaches after fire further indicate that the retention of step-pool morphology along a stream channel plays an important role in governing the magnitude of the ecological response and the timing of its recovery. In other words, channels that retained the original step-pool morphology exhibited little ecological impact and generally recovered to reference conditions within two years after the fire. In channels where post-fire floods demolished or buried the step-pool morphology, macroinvertebrate metrics indicated dramatic ecological impact and a delayed trajectory of recovery. Step-pool features could therefore be important components of successful management plans for the ecological restoration of stream channels affected by wildfire. The results of this study also provide insight into the variations in stream responses to wildfire as a function of the severity of burn. Channels burned with low severity exhibited little ecological and geomorphological impact following the Waldo Canyon Fire. On the 143

PAGE 151

other hand, the impacts on channels burned with high severity were great, with a delayed trajectory of recovery. This result has positive implications for the use of prescribed burns in ecological management. These findings suggest that low-severity prescribed fires can be used to control fuel loads as preventative measures for severe wildfires with minimal negative effects on stream ecosystems. Such measures may help to prevent or attenuate catastrophic post-fire secondary effects that may have more enduring impacts to the aquatic environment than the fire alone. Step-pool ecosystems are unique habitats for aquatic biota. Pools act as reservoirs of water, sediment, and organic materials to be transported downstream to larger lentic and lotic bodies, thus influencing ecosystem dynamics throughout the fluvial network. Disturbances occurring in step-pool headwater streams are therefore important events that can dictate water quality and habitat condition for widespread aquatic systems. This study provides insight into the effects of wildfire as a disturbance on the function of step-pool ecosystems. Integrative analysis of geomorphological and ecological changes contributes to the sparse body of literature documenting how step-pool systems can shape the response of streams to such disturbance and absorb harmful effects that might otherwise be transmitted downstream. In this regard, the findings from this study can assist the ecological management of step-pool systems in anticipation of, and in response to, inevitable disturbances. The outcomes can provide potential benefits to human and natural communities downstream. Future Work Future research on the impacts of wildfire on mountain stream channels should follow several fruitful lines of inquiry. First, a more thorough understanding of the environmental 144

PAGE 152

variables that influence the nature and magnitude of stream responses is needed for effective management of the potentially devastating impacts of wildfire. A better comprehension of the hydrological, geomorphological, sedimentological, and ecological characteristics of affected watersheds, and their complex interactions and feedbacks in response to fire, will assist in the accurate prediction of wildfire effects. Research along these lines is particularly needed for watersheds burned with moderate to high severity because this is where management efforts may be most effective. Improved knowledge on this front can help to minimize negative outcomes. Second, more research is needed on the midto long-term effects of wildfire on channel environments. Few studies have continuously documented the response of stream channels over multiple years to track the trajectory of recovery to or beyond reference conditions. Literature regarding stream responses extending beyond ten years post-fire is particularly sparse. Long-term studies that capture pre-fire conditions to serve as a baseline for comparison to the post-fire environment (such as Rugenski and Minshall 2014) are also needed to minimize the influence of spatial variability on the observed results. These studies will enable more accurate characterization of the response of stream channels to wildfire disturbances. Finally, as the ecological footprint of the human population grows, the urgency of investigating the impacts of anthropogenic activities also increases. Further research is necessary to fully understand the ways in which anthropogenic disturbances, including wildfire, interact with natural processes and contribute to environmental change. Understanding these interactions is a critical step toward mitigating and adapting to 145

PAGE 153

inevitable threats to humans and the natural world that will continue to emerge during the "Anthropocene". 146

PAGE 154

REFERENCES Abrahams AD, Li G, Atkinson JF. 1995. Step-pool streams: Adjustment to maximum flow resistance. Water Resources Research. 31(10): 2593-2602. Adams MA. 2013. Mega-fires, tipping points and ecosystem services: managing forests and woodlands in an uncertain future. Forest Ecology and Management. 294: 250-261. Allan JD, Castillo MM. 2007. Stream Ecology: Structure and Function of Running Waters. Springer, Dordrecht, The Netherlands. Anderson MD. 2002. Pinus edulis In: Fire Effects Information System, [Online]. U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station, Fire Sciences Laboratory (Producer). Retrieved from http:// www.fs.fed.us/database/feis/ plants/tree/pinedu/all.html Accessed [9 November 2015 ]. Anderson MD. 2003. Pinus contorta var. latifolia In: Fire Effects Information System, [Online]. U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station, Fire Sciences Laboratory (Producer). Retrieved from http://www.fs.fed.us/ database/feis/plants/tree/pinconl/all.html Accessed [9 November 2015]. Arkle RS, Pilliod DS, Strickler K. 2010. Fire, flow and dynamic equilibrium in stream macroinvertebrate communities. Freshwater Biology. 55: 299-314. Associated Press. 2013 June 21. Insurance cost of Waldo Canyon fire: $453 million. The Denver Post. Retrieved from http://www.denverpost.com/wildfires/ci_23513984/ insurance-cost-waldo-canyon-fire-453-million Accessed [9 November 2015]. Barbour MT, Gerritsen J, Snyder BD, Stribling JB. 1999. Rapid Bioassessment Protocols for Use in Streams and Wadeable Rivers: Periphyton, Benthic Macroinvertebrates, and Fish (2nd edition): Washington, D.C., U.A. Environmental Protection Agency, Office of Water, EPA 841-B-99-002, 339 p. Bche LA, McElravy EP, Resh VH. 2006. Long-term seasonal variation in the biological traits of benthic-macroinvertebrates in two Mediterranean climate streams in California, U.S.A. Freshwater Biology 51: 56-75. Bche LA, Stephens SL, Resh VH. 2005. Effects of a prescribed fire on a Sierra Nevada (California, USA) stream and its riparian zone. Forest Ecology and Management 218: 37-59. Benda L, Miller D, Bigelow P, Andras K. 2003. Effects of post-wildfire erosion on channel environments, Boise River, Idaho. Forest Ecology and Management. 178(1-2): 105-119. 147

PAGE 155

Biswas SR, Mallik AU. 2011. Species diversity and functional diversity relationship varies with disturbance intensity. Ecosphere. 2(4): Article 52. Boulton A. 2003. Parallels and contrasts in the effects of drought on stream macroinvertebrate assemblages. Freshwater Biology. 48(7): 1173-1185. Cairns, Jr. J, Pratt JR. 1993. A history of biological monitoring using benthic macroinvertebrates. In Freshwater Biomonitoring and Benthic Macroinvertebrates edited by Rosenberg DM and Resh VH, p. 10-27. New York: Chapman & Hall. California Aquatic Bioassessment Workshop. 2003. List of Californian Macroinvertebrate Taxa and Standard Taxonomic Effort. CA Department of Fish and Game. Chin A. 1989. Step pools in stream channels. Progress in Physical Geography. 13(3): 391-407. Chin A, An L, Florsheim JL, Laurencio LR, Marston RA, Solverson AP, Simon GL, Stinson E, Wohl E. 2015. Investigating feedbacks in human-landscape systems: Lessons following a wildfire in Colorado, USA. Geomorphology. doi: 10.1016/j.geomorph. 2015.07.030 Chin A, Anderson S, Collison A, Ellis-Sugai BJ, Haltiner JP, Hogervorst JB, Kondolf GM, O'Hirok LS, Purcell AH, Riley AL, Wohl E. 2009a. Linking Theory and Practice for Restoration of Step-Pool Streams. Environmental Management. 43: 645-661. Chin A, Phillips JD. 2007. The self-organization of step-pools in mountain streams. Geomorphology. 83: 346-358. Chin A, Purcell AH, Quan JWY. 2009b Assessing geomorphological and ecological responses in restored step-pool systems. GSA Special Papers, v. 451, p. 199-214. Chin A, Wohl E. 2005. Toward a theory for step pools in stream channels. Progress in Physical Geography. 29(3): 275-296. Chung N, Bae M-J, Li F, Kwon Y-S, Kwon T-S, Kim J-S, Park Y-S. 2012. Habitat characteristics and trophic structure of benthic macroinvertebrates in a forested headwater stream. Journal of Asia-Pacific Entomology. 15: 495-505. Church M, Zimmermann A. 2007. Form and stability of step-pool channels: research progress. Water Resources Research. 43(3): W03415, doi:10.1029/2006WR005037 Clausen B, Biggs B. 1997. Relationships between benthic biota and hydrological indices in New Zealand streams. Freshwater Biology. 38(2): 327-342. Coalition for the Upper South Platte. (n.d.) Waldo Canyon Fire. Retrieved from http:// 148

PAGE 156

waldofire.org/ Cole MB, Russell KR, Mabee TJ. 2003. Relation of headwater macroinvertebrate communities to in-stream and adjacent stand characteristics in managed secondgrowth forests of the Oregon Coast Range mountains. Canadian Journal of Forest Research. 33(8): 1433-1443. Colorado Climate Center. 2010. Colorado Temperature Monitoring. Colorado State University, Fort Collins, CO. Retrieved from http://ccc.atmos.colostate.edu/ CO_temp_status.php Accessed [9 November 2015]. Colorado Natural Heritage Program. 2011. "Terrestrial Ecological System Patches." Colorado State University, Fort Collins, CO. Retrieved from http:// www.cnhp.colostate.edu/download/gis/TESP_Map_2011.pdf Accessed [9 November 2015]. Connell JH. 1978. Diversity in tropical rain forests and coral reefs. Science. 199: 1302-1310. Corder GW, Foreman DI. 2009. Nonparametric Statistics for Non-Statisticians: A step-bystep approach. John Wiley & Sons, Inc., Hoboken, NJ. Danehy RJ, Ringler NH, Ruby RJ. 1999. Hydraulic and geomorphic influence on macroinvertebrate distribution in the headwaters of a small watershed. Journal of Freshwater Ecology. 14(1): 79-91. Doerr SH, Shakesby RA, Blake WH, Chafer CJ, Humphreys GS, Wallbrink PJ. 2006. Effects of differing wildfire severities on soil wettability and implications for hydrologic response. Journal of Hydrology. 319(1-4): 295-311. Doerr SH, Shakesby RA, MacDonald LH. 2009. Soil water repellency: a key factor in postfire erosion. In A Cerda and PA Robichaud (Eds.), Fire Effects on Soils and Restoration Strategies. Enfield, New Hampshire: Science Publishers. Doesken NJ, Pielke RA, Bliss OAP. 2003. Climate of Colorado. Climatography of the United States No. 60. Colorado Climate Center, Atmospheric Science Department, Colorado State University, Fort Collins, CO. Dupuis L, Friele P. 2006. The distribution of the Rocky Mountain tailed frog ( Ascaphus montanus ) in relation to the fluvial system: implications for management and conservation. Ecological Research. 21(4): 489-502. Earl SR, Blinn DW. 2003. Effects of wildfire ash on water chemistry and biota in SouthWestern U.S.A. streams. Freshwater Biology. 48: 1015-1030. Fountain Creek Watershed Flood Control and Greenway District. 2015. Retrieved from 149

PAGE 157

http://www.fountain-crk.org/watershed-district/general-information.html Accessed [29 August 2015]. Fountain Creek Watershed Study Watershed Management Plan. 2009. U.S. Army Corps of Engineers. Section 2. Fuller BM, Sklar LS, Compson ZG, Adams KJ, Marks JC, Wilcox AC. 2011. Ecogeomorphic feedbacks in regrowth of travertine step-pool morphology after dam decommissioning, Fossil Creek, Arizona. Geomorphology. 126(3-4): 314-332. Gresswell RE. 1999. Fire and aquatic ecosystems in forested biomes of North America. Transactions of the American Fisheries Society. 128(2): 193-221. Hall SJ, Lombardozzi D. 2008. Short-term effects of wildfire on montane stream ecosystems in the southern Rocky Mountains: one and two years post-burn. Western North American Naturalist. 68(4): 453-462. Harvey GL, Clifford NJ, Gurnell AM. 2008. Towards and ecologically meaningful classification of the flow biotope for river inventory, rehabilitation, design, and appraisal purposes. Journal of Environmental Management. 88(13): 638-650. Hilsenhoff WL. 1988. Seasonal correction factors for the biotic index. The Great Lakes Entomologist. 21(1): 9-13. Holsinger L, Keane RE, Isaak DJ, Eby L, Young MK. 2014. Relative effects of climate change and wildfires on stream temperatures: a simulation modeling approach in a Rocky Mountain watershed. Climate Change. 124:191-206. Howard, Janet L. 2003. Pinus ponderosa var. scopulorum In: Fire Effects Information System, [Online]. U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station, Fire Sciences Laboratory (Producer). Retrieved from http:// www.fs.fed.us/database/feis/plants/tree/pinpons/all.html Accessed [9 November 2015 ]. Huffman EL, MacDonald LH, Stednick JD. 2001. Strength and persistence of fire-induced hydrophobicity under ponderosa and lodgepole pine, Colorado Front Range. Hydrologic Processes. 15: 2877-2892. Huston M. 1979. A general hypothesis of species diversity. The American Naturalist. 113(1): 81-101. Inciweb: the Incident Information System. 2012 October 1. Waldo Canyon Fire. Retrieved from http://inciweb.nwcg.gov/incident/2929/ Accessed [9 November 2015]. Isaak DJ, Wollrab S, Horan D, Chandler G. 2012. Climate change effects on stream and river 150

PAGE 158

temperatures across the northwest U.S. from 1980-2009 and implications for salmonid fishes. Climate Change. 113: 499-524. Kaller MD, Hartman KJ. 2004. Evidence of a threshold level of fine sediment accumulation for altering benthic macroinvertebrate communities. Hydrobiologia. 518(1-3): 95-104. Legleiter CJ, Lawrence RL, Fonstad MA, Marcus WA, Aspinall R. 2003. Fluvial response a decade after wildfire in the northern Yellowstone ecosystem: a spatially explicit analysis. Geomorphology. 54(3-4): 119-136. Lenzi MA. 2001. Step-pool evolution in the Rio Cordon, northeastern Italy. Earth Surface Processes and Landforms. 26: 991-1008. Linke S, Bailey RC, Schwindt J. 1999. Temporal variability of stream bioassessments using benthic macroinvertebrates. Freshwater Biology. 42: 575-584. Lippitt CL, Stow DA, O'Leary JF, Franklin J. 2013. Influence of short-interval fire occurrence on post-fire recovery of fire-prone shrublands in California, USA. International Journal of Wildland Fire. 22(2): 184-193. Logan P, Brooker MP. 1983. The macroinvertebrate faunas of riffles and pools. Water Research. 17(3): 263-270. MacDonald LH, Huffman EL. 2004. Post-fire soil water repellency: persistence and soil moisture thresholds. Soil Science Society of America Journal. 68(5): 1729-1734. Malison RL, Baxter CV. 2010a. The fire pulse: wildfire stimulates flux of aquatic prey to terrestrial habitats driving increases in riparian consumers. Canadian Journal of Fisheries and Aquatic Sciences. 67(3): 570-579. Malison RL, Baxter CV. 2010b. Effects of wildfire of varying severity on benthic stream insect assemblages and emergence. Journal of the North American Benthological Society 29(4): 1324-1338. Mellon CD, Wipfli MS, Li JL. 2008. Effects of forest fire on headwater stream macroinvertebrate communities in eastern Washington, U.S.A. Freshwater Biology. 53: 2331-2343. Merritt RW, Cummins KW, Berg MG, eds. 2008. An Introduction to the Aquatic Insects of North America. 4th ed. Kendall Hunt Publishing. Mihuc TB, Minshall GW. 1995. Trophic generalists vs. trophic specialists: implication for food web dynamics in post-fire streams. Ecology. 76(8): 2361-2372. 151

PAGE 159

Mihuc TB, Minshall GW. 2005. The trophic basis of reference and post-fire food webs 10 years after wildfire in Yellowstone National Park. Aquatic Sciences 67(4): 541-548. Milner VS, Gilvear DJ. 2012. Characterization of hydraulic habitat and retention across different channel types; introducing a new field-based technique. Hydrobiologia. 694(1): 219-233. Minshall GW. 2003. Responses of stream benthic macroinvertebrates to fire. Forest Ecology and Management. 178(1-2): 155-161. Minshall GW, Minshall JN. 1976. Microdistribution of benthic invertebrates in a Rocky Mountain (U.S.A) stream. Hydrobiologia 55(3): 231-249. Minshall GW, Robinson CT, Lawrence DE. 1997. Postfire responses of lotic ecosystems in Yellowstone National Park, U.S.A. Canadian Journal of Fisheries and Aquatic Sciences. 54(11): 2509-2525. Minshall GW, Robinson CT, Royer TV. 1998. Stream ecosystem responses to the 1988 wildfires. Yellowstone Science. 6(3): 15-22. Minshall GW, Royer TV, Robinson CT. 2001. Response of the Cache Creek macroinvertebrates during the first 10 years following disturbance by the 1988 Yellowstone wildfires. Canadian Journal of Fisheries and Aquatic Sciences. 58(6): 1077-1088. Moody JA, Shakesby RA, Robichaud PR, Cannon SH, Martin DA. 2013. Current research issues related to post-wildfire runoff and erosion processes. Earth-Science Reviews. 122: 10-37. Moore RD, Richardson JS. 2003. Progress towards understanding the structure, function, and ecological significance of small stream channels and their riparian zones. Canadian Journal of Forest Research. 33(8): 1349-1351. Moss B. 2010. Ecology of Fresh Waters: A View of the Twenty-First Century. 4th ed. WileyBlackwell. West Sussex, UK. 480 p. Nakano S, Miyasaka H, Kuhara N. 1999. Terrestrial-aquatic linkages: riparian arthropod inputs alter trophic cascades in a stream food web. Ecology. 80(7): 2435-2441. National Water Information System: Web Interface. USGS surface-water daily data for the nation. U.S. Geological Survey. Retrieved from http://nwis.waterdata.usgs.gov/nwis/ dv? 152

PAGE 160

Natural Resources Conservation Service. 2013. United States Average Annual Precipitation. National Atlas of the United States, Reston, VA. Retrieved from http:// www.nationalatlas.gov/ Natural Resources Conservation Service, United States Department of Agriculture. Web Soil Survey. Retrieved from http://websoilsurvey.nrcs.usda.gov/ Accessed [29 August 2015]. O'Bryan KE, Prober SM, Lunt ID, Eldridge DJ. 2009. Frequent fire promotes diversity and cover of biological soil crusts in a derived temperate grassland. Oecologia. 159: 827-838. Oliver AA, Bogan MT, Herbst DB, Dahlgren RA. 2012. Short-term changes in stream macroinvertebrate communities following a severe fire in the Lake Tahoe basin, CA. Hydrobiologia 694(1): 117-130. Parker R. 2012 September 12. Waldo Canyon Fire human caused, but no one knows if it was intentional. The Denver Post. Retrieved from http://www.denverpost.com/wildfires/ ci_21527139/waldo-canyon-fire-human-caused-but-no-one Pearsons TN, Li HW, Lamberti GA. 1992. Influence of habitat complexity on resistance to flooding and resilience of stream fish assemblages. Transaction of the American Fisheries Society. 121(4): 427-436. Reice SR. 1980. The role of substratum in benthic macroinvertebrate microdistribution and litter decomposition in a woodland stream. Ecology 61(3): 580-590. Resh VH, Brown AV, Covich AP, Gurtz ME, Li HW, Minshall GW, Reice SR, Sheldon AL, Wallace JB, Wissman RC. 1988. The role of disturbance in stream ecology. Journal of the North American Benthological Society. 7(4): 433-455. Rhoades CC, Entwistle D, Butler D. 2011. The influence of wildfire extent and severity on stream water chemistry, sediment and temperature following the Hayman Fire, Colorado. International Journal of Wildland Fire. 20: 430-442. Richards C, Minshall GW. 1992. Spatial and temporal trends in stream macroinvertebrate communities: the influence of catchment disturbance. Hydrobiology 241: 173-184. Rinne JN. 1996. Short-term effects of wildfire on fishes and aquatic macroinvertebrates in the Southwestern United States. North American Journal of Fisheries Management. 16(3): 653-658. Roghair CN, Dolloff CA, Underwood MK. 2002. Response of a brook trout population and instream habitat to a catastrophic flood and debris flow. Transaction of the American Fisheries Society. 131(4): 718-730. 153

PAGE 161

Rosgen D, Rosgen B, Collins S. 2013. Waldo Canyon Fire Watershed Assessment: The WARSSS Results. Submitted to: Coalition for the Upper South Platte (CUSP). 5 April 2013. Rugenski AT, Minshall GW. 2014. Climate-moderated responses to wildfire by macroinvertebrates and basal food resources in montane wilderness streams. Ecosphere. 5(3): Article 25. Ryan SE, Dwire KA, Dixon MK. 2011. Impacts of wildfire on runoff and sediment loads at Little Granite Creek, western Wyoming. Geomorphology. 129(1-2): 113-130. Schoennagel T, Vleben TT, Kulakowski D, Holz A. 2007. Multidecadal climate variability and climate interactions affect subalpine fire occurrence, Western Colorado (USA). Ecology. 88(11): 2891-2902. Sedell JR, Reeves GH, Hauer FR, Hawkins CP. 1990. Role of refugia in recovery from disturbances: modern fragmented and disconnected river systems. Environmental Management. 14(5): 711-724. Simmoneaux V, Cheggour A, Deschamps C, Mouillot F, Cerdan O, Le Bissonnais Y. 2015. Land use and climate change effects on soil erosion in a seminars mountainous watershed (High Atlas, Morocco). Journal of Arid Environments 122: 64-75. Simonin KA. 2000. Quercus gambelii In: Fire Effects Information System. U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station, Fire Sciences Laboratory (Producer). Retrieved from http://www.fs.fed.us/database/feis/plants/tree/ quegam/all.html Accessed [9 November 2015]. Spencer CN, Gabel KO, Hauer FR. 2003. Wildfire effects of stream food webs and nutrient dynamics in Glacier National Park, USA. Forest Ecology and Management 178: 141-153. Steiner M. 2013 December 21. Languishing Colorado wildfire investigations feed rumor mill. The Gazette. Retrie ved from http://gazette.com/article/1511395 Stoeser DB, Green GN, Morath LC, Heran WD, Wilson AB, Moore DW, Van Gosen BS. 2005. Preliminary Integrated Geologic Map Databases for the United States Central States: Montana, Wyoming, Colorado, New Mexico, Kansas, Oklahoma, Texas, Missouri, Arkansas, and Louisiana, The State of Colorado: U.S. Geological Survey Open-File Report 2005-1351, U.S. Geological Survey, Denver, CO. Retrieved from http://pubs.usgs.gov/of/2005/1351/ Accessed [29 August 2015]. Stogner, Sr. RW. 2000. Trends in precipitation and streamflow in the Fountain Creek Watershed, southeastern Colorado, 1977-99. USGS Fact Sheet 136-00, 4 p. 154

PAGE 162

Sueyoshi M, Nakano D, Nakamura F. 2014. The relative contributions of refugium types to the persistence of benthic invertebrates in a seasonal snowmelt flood. Freshwater Biology 59: 257-271. Sullivan SMP, Watzin MC, Hession WC. 2004. Understanding stream geomorphic state in relation to ecological integrity: evidence using habitat assessments and macroinvertebrates. Environmental Management. 34(5): 669-683. TrŽmolires M. 2004. Plant response strategies to stress and disturbance: the case of aquatic plants. Journal of Biosciences. 4(29): 461-470. Troia MJ, Gido KB. 2012. Predicting community-environment relationships of stream fishes across multiple drainage basins: Insights into model generality and the effect of spatial extent. Journal of Environmental Management 128: 313-323. USDA Forest Service and U.S. Geological Survey. 2002. Forest Cover Types: National Atlas of the United States, Reston, VA. Retrieved from http://www.nationalatlas.gov/ U.S. Geological Survey. 2013. Cascade Quadrangle topographical map, Colorado El Paso Co. 7.5 minute series. The National Map. Vannote RL, Minshall GW, Cimmins KW, Sedell JR, Cushing CE. 1980. The River Continuum Concept. Canadian Journal of Fisheries and Aquatic Sciences. 37(1): 130-137. Vieira NKM, Clements WH, Guevara LS, Jacobs BF. 2004. Resistance and resilience of stream insect communities to repeated disturbances after a wildfire. Freshwater Biology. 49(10): 1243-1259. Wagenbrenner JW, MacDonald LH, Rough D. 2006. Effectiveness of three post-fire rehabilitation treatments in the Colorado Front Range. Hydrological Processes. 20: 2989-3006. Wallace JB, Webster JR. 1996. The role of macroinvertebrates in stream ecosystem function. Annual Review of Entomology. 41: 115-139. Wang Z, Lee JHW, Xu M. 2013. Eco-hydraulics and eco-sedimentation studies in China. Journal of Hydraulic Research. 51(1): 19-32. Wang Z, Melching CS, Duan X, Yu G. 2009. Ecological and hydraulic studies of step-pool systems. Journal of Hydraulic Engineering. 135(9): 705-717. Waters KA, Curran JC. 2012. Investigating step-pool sequence stability. Water Resources Research. 48(7): W07505, doi: 10.1029/2011WR011436. 155

PAGE 163

Westerling AL, Hidalgo HG, Cayan DR, Swetnam TW. 2006. Warming and earlier spring increase western U.S. forest wildfire activity. Science 313(5789): 940-943. Whitehead PG, Ilby RLW, Battarbee RW, Kernan M, Wade AJ. 2009. A review of the potential impacts of climate change on surface water quality. Hydrological Sciences Journal. 54(1): 101-123. Winterbourn MJ. 1978. The macroinvertebrate fauna of a New Zealand forest stream. New Zealand Journal of Zoology. 5: 157-169. Wood PJ, Armitage PD. 1997. Biological effects of fine sediment in the lotic environment. Environmental Management. 21(2): 203-217. Young D, Rust B. 2012. Waldo Canyon Fire Burned Area Emergency Response Soil Resource Assessment. Waldo Canyon BAER Team, USDA Forest Service, Region 5, Redding, CA. Zimmerman A, Church M, Hassan MA. 2008. Identification of steps and pools from stream longitudinal profile data. Geomorphology. 102: 395-406. Zuellig RE, Bruce JF, Stogner RW. 2010. Temporal change in biological community structure in the Fountain Creek Basin, Colorado, 2001-2008. U.S. Geologic Survey Scientific Investigations Report 2010-5103, 19 p. Zuur A, Ieno EN, Smith GM. 2007. Analysing Ecological Data. Springer Science & Business Media. New York, NY. 672 p. 156

PAGE 164

APPENDIX A ABBREVIATIONS Abbreviation Meaning Summary Statistics Med. Median SD Standard deviation IQR Interquartile range Macroinvertebrate Metrics Abun Macroinvertebrate abundance Rich Taxa richness EPT Ephemeroptera, Plecoptera, Trichoptera Dom Dominant taxon Chir Chironomidae FBI Family biotic index Intol Intolerant taxa Tol Tolerant taxa Scr Scraper Shr Shredder C-G Collector-gatherer C-F Collector-filterer Pred Predator Bu Burrower Cb Climber Cn Clinger Sk Skater Sp Sprawler Sw Swimmer Environmental Variables DO Dissolved Oxygen (mg/L) Temp Water Temperature (C) d 50 (blk) Median sediment size from bulk sample (mm) d 50 (peb) Median sediment size from pebble count (mm) 157

PAGE 165

APPENDIX B SUMMARY STATISTICS OF STUDY REACHES Willis Reach: Williams Canyon (high severity burn) 2012 2013 2014 Metric Mean Med. SD IQR Mean Med. SD IQR Mean Med. SD IQR Abun 690.6 595.2 559.9 739.8 24.75 27.50 12.66 8.25 1540.0 1391.5 1253.1 1886.0 Rich 7.25 7.00 2.06 1.25 8.25 9.00 1.50 0.75 10.00 10.50 1.41 1.50 EPT Abun 3.75 3.50 3.86 5.75 0.25 0.00 0.50 0.25 302.0 320.0 122.1 186.5 % EPT 0.78 0.57 0.84 0.50 0.68 0.00 1.35 0.68 28.31 28.65 14.48 22.83 % Dom 89.56 90.94 4.17 3.72 30.66 30.50 4.30 5.34 56.40 55.14 14.98 17.91 % Chir 89.56 90.94 4.17 3.72 20.74 25.03 16.59 22.14 20.74 28.32 19.43 29.00 FBI 5.81 5.83 0.10 0.075 6.41 6.26 0.56 0.46 5.38 5.37 0.28 0.45 % Intol 2.10 1.41 2.51 0.24 0.00 0.00 0.00 0.00 1.11 1.02 0.27 0.24 % Tol 1.35 1.32 1.56 2.67 10.78 9.94 7.04 6.60 0.03 0.015 0.042 0.045 % Scr 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 % Shr 3.44 3.49 1.47 1.66 2.03 0.00 4.06 2.03 2.32 1.89 1.02 0.69 % C-G 94.62 94.93 1.09 0.90 39.31 44.70 17.50 15.75 56.86 59.15 32.79 52.42 % C-F 0.59 0.62 0.51 0.71 4.73 0.00 9.46 4.73 35.62 32.56 38.52 61.99 % Pred 1.27 1.27 0.45 0.76 45.84 49.21 24.52 24.01 5.62 5.34 5.26 8.26 % Bu 93.91 94.63 5.07 7.12 32.78 39.88 23.14 22.66 30.58 30.79 18.96 28.21 % Cb 0.47 0.29 0.63 0.76 7.89 7.28 4.79 3.10 1.53 0.90 1.75 1.49 % Cn 0.59 0.62 0.51 0.71 8.30 7.15 9.77 15.45 35.17 32.56 38.54 62.00 % Sk 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.12 0.00 0.24 0.12 % Sp 0.73 0.62 0.73 0.85 42.95 49.21 18.93 19.76 5.92 5.20 4.69 6.84 % Sw 4.31 3.21 4.54 5.93 0.00 0.00 0.00 0.00 26.62 26.70 14.03 21.70 158

PAGE 166

Tributary Reach: Williams Canyon (high severity burn) 2013 2014 Metric Mean Med. SD IQR Mean Med. SD IQR Abun 6.25 6.00 2.06 1.25 42.00 44.50 17.57 18.50 Rich 3.50 3.50 0.58 1.00 7.75 7.50 2.50 2.25 EPT Abun 0.00 0.00 0.00 0.00 0.50 0.00 1.00 0.50 % EPT 0.00 0.00 0.00 0.00 2.63 0.00 5.27 2.63 % Dom 49.54 53.25 19.49 26.00 36.27 36.37 1.81 1.79 % Chir 6.25 0.00 12.50 6.25 24.71 25.82 10.75 6.81 FBI 6.38 6.00 1.11 0.63 6.06 5.98 0.40 0.56 % Intol 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 % Tol 12.50 0.00 25.00 12.50 8.79 8.60 7.50 7.38 % Scr 4.17 0.00 8.34 4.17 0.00 0.00 0.00 0.00 % Shr 0.00 0.00 0.00 0.00 3.23 3.58 2.87 4.19 % C-G 17.36 18.06 14.76 18.75 44.45 45.58 7.73 5.52 % C-F 0.00 0.00 0.00 0.00 19.98 22.00 17.78 26.98 % Pred 23.61 22.22 22.39 29.17 32.35 34.06 12.43 18.43 % Bu 21.53 18.06 21.56 22.92 36.26 38.42 5.57 5.32 % Cb 10.42 8.34 12.50 18.75 1.96 1.28 2.51 3.24 % Cn 0.00 0.00 0.00 0.00 22.54 22.00 21.25 29.54 % Sk 0.00 0.00 0.00 0.00 1.89 1.28 2.40 3.17 % Sp 13.20 13.89 10.49 10.42 36.04 35.50 15.25 22.28 % Sw 0.00 0.00 0.00 0.00 1.32 0.00 2.63 1.32 159

PAGE 167

Aussie Reach: Camp Creek (high severity burn) 2012 2013 2014 Metric Mean Med. SD IQR Mean Med. SD IQR Mean Med. SD IQR Abun 453.2 516.3 142.9 106.5 359.8 379.5 154.1 146.3 936.5 884.0 713.7 728.0 Rich 9.25 10.0 2.22 1.75 12.5 12.0 2.65 3.00 14.25 14.0 2.06 1.25 EPT Abun 31.95 28.3 31.40 47.75 33.0 25.0 31.59 35.50 345.5 409.5 191.5 200.0 % EPT 6.35 6.07 5.42 7.66 8.84 7.31 7.26 9.17 43.55 47.78 12.28 11.08 % Dom 78.14 82.92 18.09 18.39 66.27 66.15 27.05 45.80 45.90 40.97 15.61 9.77 % Chir 73.52 82.92 26.87 23.00 66.27 66.15 27.05 45.80 8.17 8.75 7.13 11.32 FBI 5.77 5.76 0.18 0.22 5.73 5.76 0.12 0.10 4.37 4.24 0.72 0.88 % Intol 1.87 1.81 1.18 1.52 3.39 2.62 2.60 1.56 25.36 26.20 17.65 27.97 % Tol 0.96 0.21 1.66 1.17 2.09 1.51 2.09 2.43 2.23 2.34 1.74 2.86 % Scr 0.00 0.00 0.00 0.00 0.16 0.00 0.32 0.16 0.18 0.00 0.35 0.18 % Shr 3.42 3.27 1.75 2.67 5.33 5.85 2.02 2.42 26.94 27.57 20.58 33.29 % C-G 79.60 89.53 25.45 17.74 73.17 74.51 21.59 35.12 41.60 47.23 13.68 8.15 % C-F 14.73 2.88 25.61 16.89 16.88 12.82 20.20 28.65 23.51 12.71 32.01 34.01 % Pred 2.24 1.49 2.26 2.52 3.20 2.56 3.27 2.90 6.89 6.32 4.65 7.32 % Bu 75.82 85.56 26.54 23.04 68.99 69.57 28.49 48.05 24.79 24.23 24.14 37.80 % Cb 1.06 0.25 1.79 1.01 3.08 2.31 1.81 1.00 5.21 5.24 3.80 3.91 % Cn 14.99 3.38 25.46 17.02 17.25 13.42 20.48 29.62 26.77 16.42 32.68 36.23 % Sk 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 % Sp 2.60 2.38 1.37 1.27 3.11 2.46 2.17 1.49 23.30 23.17 17.22 25.73 % Sw 5.54 5.29 3.40 4.64 6.55 5.94 6.23 8.48 19.74 15.42 18.30 17.78 160

PAGE 168

Eagle Reach: Camp Creek (low severity burn) 2012 2013 2014 Metric Mean Med. SD IQR Mean Med. SD IQR Mean Med. SD IQR Abun 145.50 126.50 64.43 77.00 374.25 296.00 310.91 276.75 174.25 170.50 37.32 49.75 Rich 11.50 12.00 1.00 0.50 14.25 13.50 4.65 6.25 18.00 19.00 2.71 1.50 EPT Abun 31.00 30.00 16.57 25.00 213.75 96.00 280.28 208.75 64.50 74.50 32.85 24.00 % EPT 25.83 22.92 18.74 16.45 51.36 59.22 29.26 17.56 23.20 42.04 25.13 28.98 % Dom 50.99 47.48 12.82 15.07 58.25 57.19 11.29 11.10 28.55 26.75 9.17 12.15 % Chir 41.53 40.49 21.10 13.12 31.78 25.62 19.61 15.74 18.59 16.75 9.64 9.90 FBI 5.31 5.58 1.24 1.23 2.49 2.17 0.99 0.82 3.95 4.11 1.08 1.11 % Intol 20.30 15.10 18.11 17.52 49.89 58.58 29.73 17.98 16.36 13.79 12.87 11.34 % Tol 23.90 21.03 28.00 44.93 11.42 10.15 7.60 9.73 13.97 14.22 8.54 11.84 % Scr 1.85 2.08 1.60 2.39 0.03 0.00 0.06 0.03 0.17 0.00 0.33 0.17 % Shr 21.39 18.17 13.68 10.77 47.76 52.22 27.46 22.64 27.22 30.20 8.06 5.05 % C-G 68.91 72.22 14.96 10.71 36.15 29.93 22.03 14.13 39.85 37.06 11.07 9.26 % C-F 0.16 0.00 0.33 0.16 7.95 6.24 6.74 7.39 18.07 18.75 4.14 2.80 % Pred 7.55 8.24 2.07 1.30 2.91 2.88 1.24 1.75 13.52 12.83 5.85 8.01 % Bu 46.46 45.85 21.64 16.31 41.96 32.74 26.74 17.82 47.39 42.60 30.60 45.44 % Cb 5.04 3.83 4.33 3.53 3.23 2.93 1.51 1.13 11.68 12.13 3.21 2.60 % Cn 5.41 5.40 4.50 6.01 2.68 2.81 1.65 2.46 14.24 15.27 9.88 14.90 % Sk 0.00 0.00 0.00 0.00 0.37 0.22 0.50 0.59 0.00 0.00 0.00 0.00 % Sp 18.26 12.68 12.16 6.65 49.54 58.05 29.11 17.48 8.00 3.71 10.66 6.54 % Sw 24.84 22.07 27.35 43.28 2.22 2.08 2.44 3.92 17.62 14.28 17.33 18.35 161

PAGE 169

Meadow Reach: Camp Creek (low severity burn) 2012 2013 2014 Metric Mean Med. SD IQR Mean Med. SD IQR Mean Med. SD IQR Abun 139.75 145 94.11 147.25 411.5 415 153.44 175.00 352.3 365.1 184.26 227.60 Rich 10.25 11.0 3.10 3.25 12.50 12.5 2.89 2.50 16.0 16.0 1.83 2.50 EPT Abun 27.25 11.0 33.17 16.75 61.5 57.0 55.46 89.50 211.5 227 147.28 216.00 % EPT 20.41 20.47 14.03 17.53 17.61 12.23 18.84 22.18 54.23 52.88 18.96 19.37 % Dom 50.32 46.39 17.04 11.22 73.27 79.07 22.62 25.10 38.60 39.03 9.42 10.38 % Chir 49.56 44.86 17.39 13.53 73.27 79.07 22.62 25.10 22.75 18.21 20.03 24.34 FBI 5.33 5.25 0.99 0.75 5.34 5.56 0.77 0.78 4.28 4.35 0.65 0.54 % Intol 19.97 20.47 14.71 17.97 14.20 8.35 17.26 18.76 18.46 13.13 15.79 12.42 % Tol 20.54 16.62 19.71 25.89 2.35 2.55 0.92 1.06 2.92 2.27 3.02 3.98 % Scr 0.00 0.00 0.00 0.00 0.11 0.00 0.22 0.11 0.06 0.00 0.11 0.06 % Shr 23.56 23.43 19.67 29.84 15.67 9.88 17.37 18.30 30.29 27.75 10.48 11.69 % C-G 51.22 48.18 16.79 10.16 78.81 85.04 19.98 22.30 54.23 53.24 10.25 6.44 % C-F 0.00 0.00 0.00 0.00 1.79 1.39 1.96 2.33 10.32 7.58 7.66 6.27 % Pred 25.23 22.33 17.70 20.48 3.02 3.02 0.76 0.79 4.79 3.61 2.86 1.90 % Bu 56.45 52.49 15.05 17.26 77.15 82.91 20.80 25.77 35.37 33.12 24.76 35.47 % Cb 1.54 1.15 1.65 1.37 1.43 1.44 0.49 0.65 7.99 7.05 7.53 11.62 % Cn 3.41 3.61 2.70 2.77 3.64 3.03 3.11 4.46 11.45 9.04 8.74 7.55 % Sk 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 % Sp 17.40 17.84 16.41 25.41 13.81 8.00 17.22 18.65 15.64 9.05 17.13 15.78 % Sw 21.21 16.62 20.51 26.33 3.88 3.44 1.86 2.39 29.37 29.82 11.95 15.06 162

PAGE 170

Hunter Reach: Bear Creek (unburned) 2012 2013 2014 Metric Mean Med. SD IQR Mean Med. SD IQR Mean Med. SD IQR Abun 173.25 138 96.80 84.25 774.25 682.5 294.37 311.25 1817.3 1759 1259.7 1731.3 Rich 12.25 12.5 1.71 1.75 10.50 10.50 1.29 1.50 13.00 13.50 2.16 2.0 EPT Abun 14.5 12.5 8.96 7.50 31.75 32 16.66 18.75 397.25 402.5 240.05 224.75 % EPT 8.84 9.00 4.03 3.06 4.00 4.54 1.31 1.05 23.79 24.15 8.22 10.09 % Dom 51.85 57.81 20.11 23.17 90.38 91.37 3.41 2.80 57.56 58.14 13.97 10.22 % Chir 47.15 48.41 22.83 37.27 90.38 91.37 3.41 2.80 57.56 58.14 13.97 10.22 FBI 5.17 5.10 0.25 0.20 5.86 5.86 0.03 0.03 4.93 4.94 0.50 0.56 % Intol 8.64 8.60 4.01 2.53 1.62 1.62 0.60 0.71 16.93 16.25 4.17 3.47 % Tol 6.37 3.21 8.29 7.34 0.24 0.23 0.24 0.35 4.57 4.70 4.16 6.89 % Scr 0.37 0.32 0.44 0.69 0.16 0.13 0.16 0.16 2.06 1.53 2.16 2.31 % Shr 14.23 12.82 11.44 15.22 2.69 2.56 0.53 0.35 18.66 17.14 2.40 5.47 % C-G 75.54 81.22 19.07 18.32 92.37 93.52 3.13 2.35 63.67 65.42 12.29 8.78 % C-F 0.54 0.59 0.45 0.64 3.66 2.26 3.32 2.40 14.54 8.41 13.77 7.70 % Pred 9.00 4.77 8.90 5.10 1.01 0.90 0.33 0.37 0.98 1.04 0.23 0.19 % Bu 77.52 85.04 20.30 13.13 91.57 92.61 3.92 3.42 65.75 66.10 16.58 17.26 % Cb 5.63 5.00 2.24 2.34 0.96 0.81 0.53 0.42 3.95 3.97 2.48 3.37 % Cn 6.33 4.43 5.76 5.50 4.10 2.67 3.29 2.15 12.66 7.26 15.59 10.28 % Sk 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 % Sp 4.40 2.07 5.49 4.15 1.68 1.94 0.67 0.50 12.94 12.31 5.67 4.71 % Sw 5.96 3.36 7.29 6.62 1.58 1.73 0.85 1.13 4.72 5.19 2.82 2.85 163

PAGE 171

Bear Creek: Gage Reach (unburned) 2012 2013 2014 Metric Mean Med. SD IQR Mean Med. SD IQR Mean Med. SD IQR Abun 111.75 104.5 28.14 32.25 53.75 52.5 26.09 40.75 1451.5 1435.5 1252.3 1850.0 Rich 14.25 14.5 1.71 1.75 11.50 11.0 5.20 6.0 16.50 16.50 2.38 3.50 EPT Abun 63.5 63.5 20.01 13.00 30.0 27.5 16.23 24.00 752.75 707.0 582.32 654.25 % EPT 60.47 70.57 22.97 14.11 54.80 54.36 4.37 3.86 62.12 64.80 18.01 26.62 % Dom 33.47 32.12 14.25 14.64 34.19 38.09 9.72 5.51 39.78 36.27 12.45 12.26 % Chir 3.73 4.14 1.00 0.70 6.34 5.65 6.36 8.28 10.06 10.23 2.59 3.48 FBI 3.80 3.53 1.00 1.27 4.13 4.13 0.42 0.38 4.14 4.23 0.58 0.62 % Intol 32.43 36.79 20.45 27.96 20.92 20.69 11.72 16.62 26.08 18.66 14.22 11.54 % Tol 10.99 10.78 7.23 9.94 7.80 6.90 2.12 1.13 5.36 5.12 1.21 1.11 % Scr 4.49 4.14 3.96 5.76 2.88 2.82 3.32 5.69 1.74 1.05 1.68 1.23 % Shr 20.37 22.30 12.76 17.25 6.40 4.16 7.84 8.44 22.43 17.98 12.69 11.47 % C-G 44.43 41.75 8.01 6.67 54.83 61.92 15.17 8.05 49.07 44.61 12.19 10.44 % C-F 22.44 17.35 20.92 19.03 7.36 4.23 9.92 11.59 23.92 24.25 17.48 25.42 % Pred 8.27 6.43 9.06 10.67 21.32 20.79 7.52 4.89 1.63 1.74 0.51 0.56 % Bu 15.58 16.09 6.66 8.76 7.62 7.50 6.42 5.86 16.28 15.78 3.91 5.28 % Cb 2.22 2.20 2.57 4.42 9.99 9.58 1.77 2.12 2.71 2.17 2.21 2.41 % Cn 35.28 30.01 13.78 8.07 37.12 30.52 17.14 15.26 25.45 25.81 17.05 23.19 % Sk 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.02 0.01 % Sp 17.06 17.86 12.95 17.72 4.61 2.43 6.15 4.93 20.16 15.28 11.84 9.75 % Sw 29.85 27.09 12.85 13.72 33.45 32.77 10.52 17.12 35.39 30.02 15.09 9.50 164

PAGE 172

West Monument Creek: Academy Reach (unburned) 2012 2013 2014 Metric Mean Med. SD IQR Mean Med. SD IQR Mean Med. SD IQR Abun 307.25 317.0 183.16 148.75 633.75 623.5 359.62 461.25 1832.0 1498.5 1268.3 1186.5 Rich 15.50 16.0 2.65 3.0 15.50 17.0 6.45 8.0 16.50 16.50 1.29 1.50 EPT Abun 59.20 34.50 60.25 37.70 80.25 60.00 53.86 51.25 483.00 427.50 183.76 202.50 % EPT 19.01 19.17 9.44 14.13 16.46 12.57 13.55 15.48 32.72 31.06 15.54 23.75 % Dom 51.32 48.26 10.20 6.84 68.08 71.08 11.39 9.03 49.88 48.71 16.08 24.75 % Chir 51.32 48.26 10.20 6.84 37.16 33.54 28.11 41.85 49.44 48.71 16.65 25.19 FBI 5.31 5.24 0.42 0.53 5.56 5.70 0.37 0.36 5.18 5.38 0.61 0.67 % Intol 7.22 6.25 3.32 2.54 4.27 3.48 3.62 4.86 12.51 7.97 11.10 6.70 % Tol 9.42 6.11 8.68 5.61 3.24 3.07 2.50 3.94 11.06 12.21 0.12 7.07 % Scr 0.15 0.00 0.29 0.15 0.09 0.07 0.11 0.15 0.31 0.20 0.37 0.34 % Shr 7.26 6.62 2.22 1.45 4.83 4.13 4.15 4.94 12.93 8.67 11.12 6.63 % C-G 72.18 73.39 8.09 7.02 51.83 51.71 36.10 60.66 72.87 74.37 9.73 8.09 % C-F 7.88 7.82 5.56 4.99 39.31 38.84 42.44 71.08 11.28 11.29 4.54 5.81 % Pred 12.54 11.89 4.59 3.92 1.95 1.89 0.88 1.37 2.56 2.53 0.97 0.86 % Bu 58.13 54.96 9.79 10.21 40.77 39.28 28.98 44.54 62.52 64.01 17.01 26.92 % Cb 4.88 4.89 2.58 2.22 2.61 1.92 2.77 3.61 3.04 2.67 0.98 0.84 % Cn 14.56 13.90 9.51 8.57 42.74 42.67 40.22 66.94 3.35 2.52 2.07 1.49 % Sk 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 % Sp 3.05 3.02 2.57 2.55 1.04 1.06 0.53 0.40 9.75 5.92 9.76 5.50 % Sw 19.39 20.98 8.59 10.27 12.74 9.54 10.51 10.93 21.35 17.54 11.61 10.81 165