Citation
Impact of anthropogenic activity and climate events during the late holocene in a subalpine lake ecosystem in central Colorado

Material Information

Title:
Impact of anthropogenic activity and climate events during the late holocene in a subalpine lake ecosystem in central Colorado
Creator:
Walker, Bethany Ann
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English

Thesis/Dissertation Information

Degree:
Master's ( Master of science)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
Department of Geography and Environmental Sciences, CU Denver
Degree Disciplines:
Environmental sciences
Committee Chair:
Briles, Christy E.
Committee Members:
Roane, Timberley
Spaulding, Sarah

Notes

Abstract:
This thesis examines how natural (e.g., climate) and human activity (e.g., mining) impacted a subalpine ecosystem in central Colorado. A lake sediment paleoecological record was developed based on sediment geochemistry, diatoms, pollen, and charcoal for the last 5000 years. Utilizing multiple proxies allowed for reconstruction of both the aquatic and terrestrial environment responses, and provided baseline information on ecosystem function prior, during and after the mining activity. The record also documents variability of the subalpine ecosystem during known climate changes of the late Holocene. This record shows similar responses to climate events during the last 5000 years as other paleo- records from high-elevation lakes in the region. For example, the transition from summer-wet to winter-wet precipitation patterns at ~2400 cal yr BP resulted in a more closed forest structure than before that was dominated by Pinus. The charcoal record indicates increased fire activity during the Medieval Climate Anomaly (MCA), that later decreased at the onset of the Little Ice Age (LIA). Human impacts have also affected the terrestrial and aquatic communities at the lake. The introduction of a nearby mine was preceded by timber harvest that altered the forest composition and fire regime. The change in vegetation also resulted in changes in allochthonous inputs into the lake, altering species abundance within the diatom community and increasing metal inputs into the pond. Since termination of mining activity, the terrestrial and aquatic systems have returned to a state similar to that seen at the start of the MCA. The results indicate that species abundances of both the terrestrial and aquatic communities varied throughout the late Holocene; however, there were no major community shifts, suggesting resistance to both climate and human activity.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
Copyright Bethany Ann Walker. Permission granted to University of Colorado Denver to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.

Downloads

This item has the following downloads:


Full Text
IMPACTS OF ANTHROPOGENIC ACTIVITY AND CLIMATE EVENTS DURING THE
LATE HOLOCENE IN A SUBAPLINE LAKE ECOSYSTEM IN CENTRAL COLORADO
by
BETHANY ANN WALKER
B.S., Allegheny College, 2013
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 Program
2019


©2019
BETHANY ANN WALKER
ALL RIGHTS RESERVED


This thesis for the Master of Science degree by Bethany Ann Walker Has been approved for the Environmental Sciences Program By
Christy E. Briles, Chair
Timberley Roane Sarah Spaulding


Walker, Bethany Ann (M.S., Environmental Sciences)
Impacts of Anthropogenic Activity and Climate Events during the late Holocene on a Subalpine Lake in Central Colorado
Thesis directed by Assistant Professor Christy E. Briles
ABSTRACT
This thesis examines how natural (e.g., climate) and human activity (e.g., mining) impacted a subalpine ecosystem in central Colorado. A lake sediment paleoecological record was developed based on sediment geochemistry, diatoms, pollen, and charcoal for the last 5000 years. Utilizing multiple proxies allowed for reconstruction of both the aquatic and terrestrial environment responses, and provided baseline information on ecosystem function prior, during and after the mining activity. The record also documents variability of the subalpine ecosystem during known climate changes of the late Holocene. This record shows similar responses to climate events during the last 5000 years as other paleo- records from high-elevation lakes in the region. For example, the transition from summer-wet to winter-wet precipitation patterns at -2400 cal yr BP resulted in a more closed forest structure than before that was dominated by Pinus. The charcoal record indicates increased fire activity during the Medieval Climate Anomaly (MCA), that later decreased at the onset of the Little Ice Age (LIA). Human impacts have also affected the terrestrial and aquatic communities at the lake. The introduction of a nearby mine was preceded by timber harvest that altered the forest composition and fire regime. The change in vegetation also resulted in changes in allochthonous inputs into the lake, altering species abundance within the diatom community and increasing metal inputs into the pond. Since termination of mining activity, the terrestrial and aquatic systems have returned to a state similar to that seen at the start of the MCA. The results indicate that species abundances of both the


terrestrial and aquatic communities varied throughout the late Holocene; however, there were no major community shifts, suggesting resistance to both climate and human activity.
The form and content of this abstract are approved. I recommend its publication.
Approved: Christy E. Briles
v


DEDICATION
I would like to dedicate this thesis to my husband. His consistent support and encouragement were essential to my success. I would also like to thank my parents and siblings for always encouraging me, no matter what adventure I’m embarking on. And lastly, thank you to the friends I have made here at the University of Colorado Denver for providing encouragement and laughs when they were needed the most.
vi


ACKNOWLEDGEMENTS
I would like to thank my advisor, Christy Briles, for her sharing her time, patience, and passion for paleoecology. Her guidance was incredibly helpful throughout this process, and I am inspired by her dedication to the subject. I would also like to thank my committee, Timberley Roane and Sarah Spaulding, as well as John Steelman for providing supporting data for this study. Lastly, thank you to all the faculty and administration at the University of Colorado Denver who supported me throughout my scholastic career.
Additionally, I would like to thank the Cushman Foundation for Foraminiferal Research and the American Association of Geographers Paleoenvironmental Change specialty group for supporting this research.
vn


TABLE OF CONTENTS
CHAPTER
I. INTRODUCTION..........................................................1
II. BACKGROUND INFORMATION...............................................5
Physical Environment...................................................5
Climate................................................................6
Climate Change Events................................................9
Proxy Data from Lake Sediments........................................10
Diatoms.............................................................11
Sediment Geochemistry...............................................14
Pollen..............................................................15
Charcoal............................................................16
Environmental Disturbances..........................................18
Sediment Core Chronology..............................................20
III. SITE DESCRIPTION...................................................23
Lily Pond.............................................................23
Forest Hill Mine......................................................24
IV. METHODS AM) DA I A ANALYSIS.........................................26
Field Methods.........................................................26
Laboratory Methods....................................................26
Core description and lithology......................................26
Chronology..........................................................27
Diatoms.............................................................27
Geochemistry........................................................28
Charcoal............................................................29
Pollen..............................................................30
Data Analysis.........................................................30
Chronology..........................................................30
Diatoms.............................................................31
Geochemistry........................................................31
Pollen..............................................................32
Charcoal............................................................33
viii


V. RESULTS
35
Core Lithology & Chronology.............................................35
Sediment Geochemistry...................................................40
Magnetic susceptibility & loss-on-ignition............................40
XRI...................................................................40
Diatoms.................................................................41
Pollen..................................................................45
Charcoal................................................................48
VI. DISCUSSION............................................................52
Climate of the late Holocene (5000 cal yr BP to present)................52
Ecosystem Response to Climate during the late Holocene..................55
Switch from summer- to winter-dominant precipitation regime...........55
Medieval Climate Anomaly & Little Ice Age.............................59
Human Impacts: Logging, Mining, Grazing and Recreation..................63
Ecosystem Response to Human Impacts.....................................64
Logging & Start of Forest Hill Mine (-70-30 cal yr BP; 1880-1920 CE)..64
Land Management Implications............................................72
Limitations.............................................................74
Future Research.........................................................75
VII. CONCLUSION...........................................................78
REFERENCES.......................................................................82
APPENDIX
A: Forest Hill Mining Complex.............................................91
B: Lead 210 Dating Results................................................93
C: Supplemental Age-Depth Model Materials.................................95
D: Diatom Species Used for Analysis.......................................95
E: Locations of paleoclimate and paleoecology studies referenced..........97
IX


LIST OF TABLES
TABLE
1. Overview of precipitation variation throughout Colorado............................7
2. Series of 210Pb and 14C dates used to generate the age depth model..................38
3. Zones within the sediment record with respective dates and ages.....................40
4. Summary of diatom abundances sorted by zone........................................45
5. Summary of pollen record organized by zones........................................48
6. Summary of charcoal record organized by zones......................................51
x


LIST OF FIGURES
FIGURE
1. Map of climate data locations.....................................................8
2. Overlap of proxies between biogeography and limnology research...................11
3. Different chemical and physical characteristics that can impact diatom survival..12
4. Site Map.........................................................................25
5. Lithology of LP07, LP15, and LP17 cores..........................................36
6. Age-depth model for Lily Pond Record.............................................37
7. Results from constrained cluster analysis using diatom data......................39
8. Results from diatom and geochemistry analyses....................................44
9. Results from pollen analysis.....................................................47
10. Results from charcoal analysis..................................................50
11. Images of mining remains around Lily Pond.......................................64
12. Environmental record from Lily Pond compared with human population..............66
xi


CHAPTER I
INTRODUCTION
Colorado has over 23,000 abandoned mine lands (AMLs) throughout the state (Colorado Geological Survey, 2013). Mining activity is very disruptive to its surroundings, by exposing heavy metals and other minerals and increasing the amount of loose sediments that can impact surrounding environments (Colorado Water Quality Control Division, 2016). Prior to 1975, Colorado mining companies were not required to remediate the mine site after the mine closed, resulting in a high number of toxic and dangerous AMLs (Colorado Division of Reclamation Mining & Safety, 2014). According to the Colorado Division of Reclamation Mining & Safety (2014), approximately 1,539 acres of AMLs have been reclaimed statewide since 1980. Even with remediation, AMLs continue to impact the surrounding environments due to remaining mine spoil and bedrock exposure (Brugam & Lusk, 1986). The long-term impacts of mining on the surrounding ecosystems are not well understood and is a focus of this thesis.
Mining activity is a source of anthropogenic disturbance that directly impacts the surrounding terrestrial and aquatic ecosystems. To understand how mining activity impacts an ecosystem, baseline conditions must be established. However, it is difficult to define these conditions because human activities have impacted most environments for decades and even centuries. Therefore, long-term historical records are crucial for setting reference conditions that examine changes in species communities, ecosystem productivity, and chemical conditions. Lake sediment records provide long-term historical records that contain information about both aquatic and terrestrial environments. Since sediment accumulation is always occurring, these records provide a continuous and measurable record of time that can be used to determine variability of the system on century- to millennia- time scales.
1


Internal lake processes and the surrounding terrestrial ecosystem can be reconstructed because of sediment internal and external sediment deposition through autochthonous organic accumulation and allochthonous runoff and atmospheric transport (Battarbee, 1999). Proxy indicators, such as diatoms, pollen, charcoal, and the chemical make-up of the sediments can be used to reconstruct past ecosystem conditions. Proxies are chosen based on the information they provide about the physical, biological, and chemical conditions of the surrounding environments, and how well they can be modeled to characterize modern ecosystems (McLaughlan et al.,
2014). Utilizing multiple proxies provides an enhanced picture of ecosystem change than one proxy alone, as each measures a different aspect of the system and therefore provides a clearer picture of change or response. Further, multiproxy record provide information about the ecosystem recovery rate and whether an ecosystem returns to its pre-disturbance condition (McLaughlan et al., 2014).
In Colorado, paleoecological studies have been conducted in the subalpine and alpine zones (e.g. Anderson, Brunelle, & Thomspon, 2015; Del Priore, 2015; Fall, 1997b, 1997a; Higuera, Briles, & Whitlock, 2014; Toney & Anderson, 2006; Wolfe, Baron, & Cornett, 2001). These studies have focused on the vegetation, fire, and lake level records, and have been used to reconstruct climate trends through the Holocene. For example, examining charcoal and pollen records has allowed for reconstruction of fire frequency and severity and the direct impact on forests during major changes in past climate (Briles et al, 2012; Fall, 1997a; Higuera et al, 2014). Similarly, oxygen-isotope climate reconstructions have identified changes in precipitation and moisture availability, and were compared with pollen, charcoal, and sediment geochemistry records to examine climate impacts on fire regimes, forest composition and structure, as well as allochthonous inputs into lake systems (Anderson, 2011; Anderson, 2012; Anderson et al, 2015).
2


Anthropogenic activities have also been shown to impact surrounding aquatic systems by altering the surrounding landscape. Mining, grazing, and logging have all impacted subalpine areas in Colorado, especially since Euro-American settlement. A study in Finland examined the diatom record in two lakes that were impacted by a Cu-Au mine that had been closed (Kihlman & Kauppila, 2010). Both records in the study document minor ecological effects during the active metal mining period. The majority of ecological effects occurred after mine closure. Results determined that changes in metal inputs into the system impacted diatom species abundances in both lake systems; however, changes in community abundances differed between sites.
Grazing and logging activities have been impacting both terrestrial and aquatic habitats in the western United States since Euro-American settlement. Approximately 70% of the western United States is impacted by livestock grazing (Fleischner, 1994). This activity has impacts on the surrounding landscape by altering species composition and structure of surrounding forests (i.e. density and biomass) and disrupting ecosystem functions (i.e. nutrient cycling and soil erosion). The terrestrial disruption also impacts nearby water features. Changes in soil erosion increases sedimentation rates into waterbodies, causing increases in turbidity, total suspended solids, and nutrient input (Fleischner, 1994). Similarly, land clearing due to timber harvesting also increases sedimentation rates into nearby water features (Arismendi et al., 2017). While terrestrial and aquatic communities impacted by grazing and timber harvest can recover, the amount of time it takes to do so varies.
To my knowledge, few studies have examined impacts of past climate, mining, and other anthropogenic activities on both aquatic and terrestrial ecosystems. Therefore, the main objective of this research is to reconstruct the paleoecological history of a subalpine lake, Lily Pond in
3


Taylor Park of central Colorado, where anthropogenic activities, including mining between 1880 and 1920 CE, are apparent on the modem landscape. A late-glacial period to early Holocene paleoenvironmental study was previously reported on by Briles et al. (2012) from a sediment core taken and radiocarbon dated in 2007. This study expands on the study by extending the pollen and charcoal record and adding new proxies, including diatoms and sediment geochemistry, from newly acquired cores taken in 2015 and 2017 that span the last 5000 years. The new record explores how the aquatic and forest ecosystem responds to late Holocene climate and more recent anthropogenic disturbance events, specifically the introduction of a metal mine nearby. The study aims to answer the following questions:
(1) How did climate variations during the late Holocene impact the terrestrial and aquatic environments at Lily Pond?
(2) How did the introduction and termination of mining activity affect the aquatic and terrestrial environments at Lily Pond?
This thesis is composed of six chapters. Chapter two provides information on Colorado’s physical environment, climate, and disturbance regimes specific to this study. It also introduces the use of multiple proxies for reconstructing past environments and introduces the concept of ecosystem resilience. Chapter three introduces the study site and provides information about known anthropogenic activity in the area. Chapter four outlines the field and laboratory methods used the carry out the research. Chapter five presents the results from the diatom, pollen, charcoal and sediment geochemistry analyses, as well as an updated 5000-year age-depth model for the Lily Pond lake sediment core. Chapter six summarizes the findings of the study. Chapter seven concludes the thesis.
4


CHAPTER II
BACKGROUND INFORMATION Physical Environment
The Rocky Mountains extend from New Mexico into Alaska. Colorado encompasses a portion of the southern Rocky Mountains, including the Sawatch Range in the central part of the state. The eastern portion of Colorado is relatively flat, with a maximum elevation of 1,980 meters where the plains meet the Rocky Mountains. At this point, the elevation increases dramatically into the foothills (elev. 2,100-2,700 m) before transitioning into the mountain ranges (elev. > 3,000 m) (Doesken, et al, 2003). The Sawatch Range is located in central Colorado, extends -160 km along a northwest-southeast axis, and is part of the Continental Divide. There are 15 peaks within the Sawatch Range measuring over 4,250 meters.
The Rocky Mountains were originally formed during the Pennsylvanian era about 300 million years ago. The original mountain range that was created during this period eventually eroded away before being reformed into the present-day Rocky Mountains. The Southern Rocky Mountains were formed during the Laramide orogeny, approximately 70-40 million years ago (mya). The exact process that formed the Rocky Mountains is unknown, but the range is thought to be the result of an oceanic plate subducting beneath a continental plate at a very low angle (US Geological Survey, 2017). The Sawatch Range experienced at least two periods of Pleistocene glaciation that shaped the landscape. The most recent period of glaciation ended -18,000 years ago (Brugger, 2006).
5


Climate
Colorado experiences a cool, dry climate with low humidity due to its inter-continental location. However, the mountains play a large role in Colorado’s variable climate. Due to the large differences in elevation between the eastern part of the state and the mountain ranges, there are substantial contrasts between the climate of the Eastern Plains and that in the mountain areas (Doesken et al, 2003). Table 1 and Figure 1 highlight the climate differences between cities and towns throughout Colorado (precipitation and temperate averages for Kit Karson and Denver from US Climate Data [https://www.usclimatedata.com/climate/colorado/united-states/3175]; averages for Taylor Park from Western Regional Climate Center; [https://wrcc.dri.edu/cgi-bin/cliMAIN.pl?cotayl].
Precipitation also varies greatly throughout the state due to elevation, topography, and large-scale precipitation mechanisms, specifically the inter-annual El Nino Southern Oscillation (ENSO) in the tropical Pacific, and Pacific Decadal Oscillation (PDO) in the north Pacific (~20 year cycles) (Kitzberger et al, 2007). Higher amounts of winter precipitation (i.e. snowfall) occur in the southern Rock Mountains when El Nino conditions (warmer Pacific SSTs) or positive PDO exist. Decreased precipitation amounts occur when La Nina conditions (cooler Pacific SSTs) or negative PDO conditions exist (Anderson, 2012; NOAA, 2014).
Precipitation also varies throughout Colorado based on elevation and topography of an area. Areas at lower elevations experience less precipitation than the mountain regions (Table 1). The eastern plains experience strong seasonal precipitation cycles. Annual precipitation (70-80%) falls from April through September. Winter months bring dry air and strong winds, resulting in very arid conditions. Therefore, areas on the eastern side of the Continental Divide are very dry especially during the winter months (Doesken et al, 2003). Areas at higher
6


elevations west of the Continental Divide experience more consistent precipitation throughout the year with most of the precipitation falling during the winter months in the form of snow (Doesken et al, 2003). This is highlighted in Table 1, where Taylor Park reports an average snowfall precipitation of 276.4 cm in the mountains, compared to only 54.9 cm in Kit Carson on the Eastern Plains.
During winter months, moist air masses from the Pacific Ocean bring snow to areas west of the Continental Divide. These air masses have minimal impacts on precipitation east of the Continental Divide (Doesken et al, 2003). The western slope receives more evenly distributed precipitation throughout the year than areas east of the Continental Divide. During summer months, the mountain ranges generate thunderstorms when there are high levels of moisture in the air. This is especially true for southwestern Colorado, where the southern “monsoon” occurs from July through September (Doesken et al, 2003). During the winter months, areas west of the Continental Divide receive higher amounts of precipitation in the form of snow than areas to the east (Fall, 1997a).
Table 1. Overview of precipitation variation throughout Colorado.
Location (Elevation) Avg. High Temperature Warmest Month Avg. Low Temperature Coldest Month Avg. Precipitation Avg. Snowfall
Kit Carson (1,306 m) 19.6 °C July 33.2 °C 1.3 °C January -11.7 °C 36.25 cm 44.5 cm
Denver IntT Airport (1,655 m) 18 °C July 31.2 °C 2.6 °C January -8.3 °C 39.32 cm 151.4 cm
Taylor Park (2,817 m) 9.5 °C July 22 °C -8.8 °C January -23.6 °C 42.39 cm 276.4 cm
7


105'W
i WY N E
*
Denver International * * Airport
Denver
U T Colorado ■ •. V '&&■ V 4 KS
• Taylor Park • Kit Carson

N
A
50 100 200 J
AZ N M OK
Copyright ©2014 Esri
Figure 1. Locations referenced for precipitation variation in Colorado.
Temperature throughout Colorado is controlled by differences in topography and elevation throughout the state. During winter months, areas at higher elevations experience much cooler temperatures than areas at lower elevations (Table 1). January is the coldest month on average, with temperatures ranging from -23.6 °C in Taylor Park (mountain area) to -8.3 °C at the Denver International Airport (eastern plains area). Spatial temperature ranges in the summer months mimic that of winter, with areas at lower elevations experiencing warmer temperatures than those at higher elevations. July is generally the warmest month, with average temperatures ranging from 22 °C in Taylor Park, to 33.2 °C in Kit Carson (Table 1; Figure 1).
8


Climate Change Events
Climate change events, or significant changes in temperature and precipitation, can impact ecosystems by altering growing seasons, water availability, and nutrient cycling. There are three major climate events that are known to have impacted the central Colorado Rocky Mountains during the late-Holocene (-5000 cal yr BP to present). The intensity of the El Nino Southern Oscillation (ENSO) increased around this time and corresponds to changes seen in the Rocky Mountain climate as identified in the oxygen isotope ratio record of lakes in northwestern Colorado, near Steamboat Springs (Anderson, 2011; Anderson, 2012; Anderson et al, 2015). The first climate shift occurred during the late-Holocene around -2400 cal yr BP, when there was a switch from a summer-dominated precipitation regime to a winter-dominated precipitation regime (Anderson, 2011; Anderson, 2012).
Prior to -2400 cal yr BP, Colorado received most of its annual precipitation as rain during the summer months. High-resolution pollen records from subalpine lakes in Colorado indicate subalpine forests had higher abundance of Picea (Higuera et al., 2014; Toney & Anderson, 2006). The charcoal records from the same lakes suggest low fire frequency leading up to -2400 cal yr BP. Although fire frequency was low, the few fire events that did occur were likely stand-replacing and high severity due to warmer summers and higher forest density (Higuera et al, 2014; Toney & Anderson, 2006).
Recent climate change events, Medieval Climate Anomaly and the Little Ice Age, are thought to be more extreme (i.e. higher temperatures and increased drought periods) than those that occurred earlier in the Holocene (Anderson, 2012; Higuera et al, 2014). The Medieval Climate Anomaly (MCA; -1200- 850 cal yr BP) is recognized as a natural warming period that resulted in droughts and higher temperatures in North America. The MCA was once again a rain-
9


dominated summer precipitation regime, similar to that seen prior to the precipitation switch at -3000 cal yr BP (Anderson, 2011). The increase in temperatures resulted in increased fire activity in subalpine areas. A composite study of 12 subalpine records from northern Colorado show that an increase in temperature (~0.5°C) from the previous centuries caused an increase in burning at 83% of the sites (Calder et al., 2015). Burning decreased towards the end of the MCA even though temperatures remained high. The decrease in fire through the end of the MCA is likely due to a decrease in biofuel availability because of continued drought conditions (Calder et al, 2015).
The Little Ice Age (LIA; -650-100 cal yr BP) is defined by an increase in winter precipitation in Colorado, with the lowest 5180 values of the Holocene (Anderson, 2011; Anderson, 2012). Wintertime snowfall was high during this time, but summers remained nearly as dry as during the MCA (Anderson, 2012). The charcoal record at Bison Lake indicates a decrease in fire activity at the start of the LIA. Picea also decreases, indicating shorter growing seasons due to higher amounts of winter precipitation (Anderson et al., 2015). These changes in fire and vegetation identified in paleo records during the MCA and LIA show how climate fluctuations have impacted natural systems during the late Holocene.
Proxy Data from Lake Sediments
Lake sediments provide a natural archive of proxies that can be used to reconstruct past ecosystems. These sediments are extracted using coring devices that remove a column of sediment from the lake bed. The extracted sediments are referred to as sediment cores and can contain thousands of years of sediment that are arranged chronologically. Several different types of proxy data can be extracted from these sediment cores (Figure 2). Multiple proxy records provide allow for more complete observations about historical environmental conditions. For
10


example, this study uses diatoms to infer how the aquatic ecosystem has changed, sediment geochemistry to examine elemental inputs into the lake that are both natural and anthropogenic, pollen for reconstructing forest composition and structure, and charcoal to determine fire regimes. Each proxy helps reconstruct the physical, biological, and chemical characteristics of the surrounding ecosystem and provides (Moser, 2004). The proxies analyzed using lake sediments provide insight about the terrestrial environment of the surrounding watershed (the land area where precipitation drains into the waterbody), as well as the surrounding airshed (the land area that may be a source of small particles that are deposited in the watershed through aerial transport). Airsheds often include a larger area than watersheds because there are no sharp boundaries (Penniman, n.d.). The size of an airshed is influenced by the topography of an area, as well as the specific particles and emissions being examined For example, a study in the Uinta Mountains (Utah) identified changes in elemental composition of subalpine lake sediments that was a result of increased dust deposition from areas 100+ km away from the source (Reynolds et al., 2010). The following sections introduce each of the proxies used in this study and identify how they have been used to recreate past conditions of both terrestrial and aquatic ecosystems.
BIOGEOGRAPHY
Paleobiogeography
LIMNOLOGY
Paleolimnology
Genetics ''"\Spectometry
Diatoms
Tree rings / Chironomids \ Mineralogical' (including / Cladocera
# isotopes) j I Chrysophytes 1 Ostracods ' 1 Elemental 1
1 Packrat 1 \ middens ' Phytoliths l Charcoal J | geochemistry 1
Historical
records
Pollen Macrofossils Stomates
Magnetics Isotopes
Figure 2. Venn diagram showing the overlap of proxies between biogeography and limnology
research (from Moser, 2004).
11


Diatoms
Diatoms are one of today’s most used biological proxies in paleolimnological studies because they provide insight about the aquatic habitat in which they are found (Moser et al, 2004; Stevenson et al, 2010). Diatoms are effective ecological indicators because they respond directly to physical, chemical, and even biological changes in aquatic ecosystems (Dixit et al, 1992; Reid, 1995; Stevenson et al, 2010; Lowe, 2011). Diatoms also reproduce rapidly, and therefore respond more quickly to environmental change than other aquatic organisms. Figure 3 highlights the different factors that impact diatom survival.
Figure 3. Overview of how different chemical and physical characteristics of an aquatic system can impact diatom survival (from Battarbee, 2000).
Diatoms can be classified as benthic or planktonic. Benthic diatoms live on lake surfaces including aquatic vegetation, mud and sediments, and rocks. Planktonic species live suspended in the water column. Species from both classifications are often found in the same aquatic
Water balance
Radiation balance
inflow
12


environment. Increases in benthic species would indicate an increase in benthic habitat that could be a result of decreased lake level, where as an increase in planktonic species would represent an increased lake level. Therefore, to reconstruct lake levels using lake sediments, a benthic to planktonic diatom ratio is often used. Diatoms can also be classified as epipelic, epilithic, and epiphytic, meaning they prefer to grow attached to clays/sediments, rocks, or plants, respectively. These classifications can help provide information about habitat availability within the lake system at different times. For example, an increase in epiphytic diatom species could signify an increase in aquatic vegetation, which could be a result of increased nutrient availability in the system. Benthic species often dominate diatom communities in shallow high-elevation lakes because of increased organic matter available in sediments, making benthic habitats less likely to be nutrient limited than planktonic habitats (Saros et al, 2005; Spaulding et al., 2015).
Diatom species are also representative of chemical characteristics of their habitat. For example, diatom analysis can be used to assess past pH levels of lacustrine environments. Some diatom species prefer acidic conditions while other species cannot survive in highly acidic waters (Battarbee et al, 2010). Diatom communities can also be monitored to identify the natural variation of water pH in environments that have not been as severely affected by acidic inputs and processes (Battarbee et al, 2010). Diatom communities are also representative of nutrient levels and can provide insight about land use in the surrounding area. Changes in inputs of nutrients, specifically nitrogen and phosphorus, alter freshwater production and impact the diatom community. A study done in the Grand Teton National park (northwestern WY) shows that diatom assemblages in shallow lakes that were dominated by benthic species did respond to changes in radioactive nitrogen levels from anthropogenic sources (Spaulding et al., 2015).
13


Sediment Geochemistry
Sediment geochemistry analyzes the metal and elemental composition of lake sediments using several different methods such as X-ray florescence (XRF), magnetic susceptibility, and loss-on-ignition (LOI). XRF analysis measures heavy elements present in the sediment.
Magnetic susceptibility measures the amount of iron-bearing minerals present in the sediment by testing the magnetic pull of the sediment (Gedye et al, 2000). An increase in magnetic susceptibility and heavy metals could signify a change in inputs that may be a result AMD or erosion due logging or grazing of the surrounding watershed. LOI is a measure of organic carbon production, and in part reflects lake productivity from both autochthonous and allochthonous inputs. All these analyses provide information about nutrient cycling and erosion (Gillson, 2015).
At Bison Lake (CO), increases in higher detrital element abundances (Fe, Ti, K), along with higher magnetic susceptibility, were indicators of greater detrital sedimentation during the LIA. The increases in elemental abundances signified an increase in allochthonous inputs and corresponded with decreased organic content, which suggested reduced biological productivity (Anderson et al, 2015). Anderson et al. (2015) used geochemical data to show that nutrient availability changed during the LIA, which was likely a result of the increased snowfall and more spring runoff.
Sediment geochemistry analyses can also be impacted by atmospheric deposition of dust. A study in the San Juan Mountains (Colorado) found that dust deposition in alpine areas increases nutrients (N, P), cations (Ca, Mg) and some metals (Cr, Cu, Ni) available in watersheds. The ecosystem response to these increases varies based on watershed size and bedrock geology (Ballantyne et al., 2011). A study in the Uinta Mountains (Utah) identified dust particles that had traveled -110 km and originated from a mine (Reynolds et al., 2010). This
14


study also found that an increase in aerial deposition of fine sediments corresponded with the settlement of the surrounding area, leading to increased urbanization and mining activity. Lakes located closer to urban and mining areas showed increased deposition than lakes located further away (Reynolds et al., 2010).
Pollen
Pollen found in sediment cores is an indication of the vegetation in the surrounding area. Changes in pollen may signify disturbance events such as fire and anthropogenic deforestation. It can also be indicative of climate variations, and changes in moisture availability and growing seasons (Briles et al, 2012; Higuera et al, 2014). Pollen can enter a lake system through runoff and aerial transport. Pollen data collected from small to medium size lakes represents vegetation within a 30-50 m radius. In general, this area provides -70% of the pollen source area for major taxa found in the sediment core (Sugita, 1994). Larger lakes are influenced more by extralocal and regional vegetation (Sugita, 1993). Aerial transport of pollen is also impacted by the size and shape of pollen grains, as well as environmental conditions such as wind, precipitation, temperature, and humidity. Lighter pollen grains (i.e. Quercus) can be transported further than heavier pollen grains (i.e. Picea, Abies). Research has suggested that the source radius of light pollen types could be lOOx larger than heavy types (Sugita, 1993).
In central Colorado, the forest composition in subalpine environments (>3000 m) was different during the early- and mid- Holocene (-11700- -5500 cal yr BP) than during the late Holocene (-4000 cal yr BP to present). Pollen records indicate an open forest that was dominated by Pinus, Picea, and Abies during the early Holocene (Briles et al, 2012; Fall, 1997a; Jimenez-Moreno & Anderson, 2012). Summer insolation began to decrease after -9000 cal yr BP, leading to climate changes in the mid-to-late Holocene. Changes in insolation lead to
15


increased ENSO variability, and decreased summer precipitation (Anderson, 2012). The mid-Holocene (-5500 cal yr BP) saw a switch to a more closed forest structure, with increased Picea and Abies species and fewer Pinus species (Briles et al, 2012; Fall, 1997a; Jimenez-Moreno & Anderson, 2012). This lasted until the late Holocene (-2400 cal yr BP), when a switch in the precipitation regime from summer-dominated to winter-dominated (Anderson et al, 2015) caused a switch in forest structure throughout Colorado to an open-canopy forest dominated by Pinus (Del Priore, 2015; Fall, 1997a; Higuera et al, 2014; Toney & Anderson, 2006). At Mirror Lake (elev > 3000 m) and Keystone Ironbog (elev <3000 m), both in central Colorado, the change in vegetation occurred at -2300 cal yr BP and -2600 cal yr BP, respectively (Del Priore, 2015; Fall, 1997a). Both locations experienced the shift to a more open forest structure, although the exact timing varied. Timing of vegetation shifts seen in regional records are discussed in more detail in Chapter 6.
Charcoal
Charcoal analysis provides information about the fire regime of an area, including fire frequency, severity, and amount of biomass burned. Fire is a natural disturbance; however the frequency and severity of fire events can be indicative of climate characteristics during that time (Calder et al, 2015). Fire is recognized as a common natural disturbance that is dependent on both climate and forest conditions (i.e. structure, composition, biomass availability) (Whitlock et al, 2010). Charcoal found in sediment cores is an indication of local fire occurrence and severity within 0-6 km of the study site (Whitlock & Larsen, 2001). Charcoal is deposited into lakes through airborne fallout during fire events. There is also a consistent input of charcoal into lake systems through streamflow and runoff. Macroscopic charcoal (>125 microns) can be analyzed in lake sediments to reconstruct fire activity throughout history (Whitlock, 2004).
16


Fires, both natural or human started, can be widespread and consume large amounts of forest biomass when weather conditions are right and have the ability to change forest composition and structure. In subalpine forests in Colorado, fire frequency is low due to cool, wet conditions and greater snowpack that result in increased fuel availably. The buildup of biofuel increases fire severity potential during drought periods (Whitlock et al, 2010). Therefore, subalpine forests are subject to high-severity, stand-replacing fires every -100 to 300 + years (Fall, 1997a; Sherriff et al, 2001). These high severity fires have been identified using tree ring records from subalpine forests. These fires impact forest structure and composition, nutrient cycling, wildlife habitat, and degrade surrounding water quality (Sherriff et al, 2001).
Charcoal records from the Rocky Mountains show that fire variability during the Holocene was influenced by climate trends and fuel availability (Fall, 1997; Briles et al, 2012; Higuera et al, 2014; Anderson, 2015). Records from central Colorado show that fire activity was high in the early Holocene and declined after -10000 cal yr BP. Activity then increased from -9000 - -7000 cal yr BP (Higuera et al, 2014). Fire events become more frequent in the region again between -2000- -1000 cal yr BP (Toney & Anderson, 2006; Jimenez-Moreno et al, 2008; Higuera et al, 2014; Calder et al, 2015). The later shift was again due to a climate event, with the onset of the Medieval Climate Anomaly that was characterized by increased temperatures and drought conditions in the region (Anderson et al., 2015; Calder et al., 2015; P. E. Higuera et al., 2014; Toney & Anderson, 2006).
Aside from fire activity, records from Rocky Mountain National Park show that fire severity also fluctuated throughout the Holocene. There was a transition from high severity fires with a high amount of biomass burned, to fires with lower severity that burned less biofuels around -2400 cal yr BP (Higuera et al, 2014). The change in severity and biomass burned was a
17


result of decreased forest density that was related to a shift in precipitation regime from winter-dominated to summer-dominated that occurred around the same time (Anderson et al, 2015).
Environmental Disturbances
Environmental, or ecological, disturbances are defined as discrete events that can have significant long-term impacts on the environment (White & Pickett, 1985; McLaughlan et al., 2014). A distinguishing characteristic of disturbance events is that they are identifiable, meaning that a change or event can be identified as a cause of the changes seen within the environment. This is also dependent on the spatial and temporal scale of the system being examined (White & Pickett, 1985). Impacts of a disturbance event depend on their type, severity, and frequency. While some impacts might be seen immediately, such as the loss of trees due to fire, others may not become apparent until after the disturbance took place, as was seen in the diatom study in Finland that was introduced previously (Kihlman & Kauppila, 2010). To understand how a disturbance event impacts the surrounding environments, researchers often look at species abundances prior to, and following, the disturbance event using proxy data from sediment cores. There are several types of disturbance events that have impacted central Colorado including fire, wind, insect, and anthropogenic activities (e.g., logging, mining, grazing, recreation). Of these, fire and anthropogenic disturbances are discernable in the Lily Pond paleorecord.
Anthropogenic disturbances refer to human activities that impact ecosystems either directly or indirectly. It is difficult to note exactly when humans began impacting natural systems, and at what intensity. There is evidence of Folsom Paleoindian groups occupying areas within central Colorado during the Younger Dryas Chronozone (-12.9 - 11.7 cal yr BP) (Briles et al, 2012). This early human activity was limited to fire, structures, and foraging that occurred at local scales, and therefore had a much smaller impact on the environment than more modern
18


anthropogenic activities (Lewis & Maslin, 2015). Recently, there has been discussion about the extent of anthropogenic activities and how they have impacted Earth’s natural systems. Some scientists have suggested that humans have caused enough disturbance to natural systems to warrant the designation of a new geological epoch (Marlon et al, 2012; Lewis & Maslin, 2015).
Anthropogenic activities have bene influencing western forests since settlement in the early 1800s, indicated by fire-scar and charcoal records from the western United States. Grazing and timber harvest (i.e. logging) increased with the settlement of the western United States throughout the 1800s and led to an overall decline in biomass burning (Marlon et al., 2012). Grazing was likely the first cause of fire reduction in the western United States because it resulted in less dense forest structure (Heyerdahl, Brubaker, & Agee, 2001; Mayer & Stockli, 2005; Savage & Swetnam, 1990). There is a peak in fire activity from 1850-1870 CE, which is likely a result of human-caused burning for clearing forests, railroad construction, agriculture, and lumbering (Marlon et al, 2012).
After settlement, fire suppression techniques impacting fire frequency as well, although has not impacted high elevation alpine areas extensively as lower elevations (Heyerdahl et al., 2001; Sherriff et al, 2001). There is, however, a lot of spatial variability in fire activity in the subalpine forests throughout Colorado during the late Holocene (Marlon et al., 2012; Sherriff et al., 2001). Given the current temperature trends (increased temperatures and droughts), biomass burning should be higher than what is actually occurring. In areas where human land management and fire suppression is less extensive, fire activity has remained high (Marlon et al., 2008; Marlon et al., 2012) While climate is a major drive of fire activity, human fire suppression efforts have altered trends in biomass burning since the late 1800s.
19


Recent paleolimnological studies have identified indirect impacts of human development on alpine lakes, which are often isolated from the direct impacts of anthropogenic activity. These studies have focused specifically on the indirect effects anthropogenic disturbances such as agriculture, mining, and urbanization, are having on natural systems because of atmospheric deposition of metals and nutrients (Hundey et al., 2014; Moser et al., 2010). The increased metal and nutrient deposition is causing changes in water chemistry, including pH, as well and nutrient cycling, specifically nitrogen and phosphorus, within lake systems.
In the Uinta Mountains (Utah), alpine lakes downwind of anthropogenic activities, including mining, industry, and agriculture, have seen increases in metals and nutrients (i.e. phosphorus, nitrogen, iron, calcium) since the 1800s (Hundey et al., 2014; Moser et al., 2010). The metals are being introduced by atmospheric inputs and have increased more since the 1900s. The increase in metal content has caused changes in the diatom communities in these lakes (Moser et al., 2010). These lakes have also seen an increase in nitrogen and phosphorus loading, likely due to increased use of nitrogen and phosphorus fertilizer, and phosphorus mining in the area. The increase in N and P into these lakes, which were previously thought to be N limited, have resulted increases in nitrophilous diatom species and increased primary production (Hundey et al., 2014). Similar results have been seen in other alpine areas within the western United States (CO, MT, WY) (Saros et al, 2003; Saros et al, 2011; Wolfe et al., 2001).
Sediment Core Chronology
Radiometric dating using lead-210 (210Pb) and radiocarbon (14C) dating are used to establish ages for the length of a sediment cores. Radiocarbon dating was developed by a group of chemists at the University of Chicago in the late 1940s (Libby, 1972). Early techniques used an anticoincidence counter to measure the radioactivity of solid carbon samples and used a half-
20


life of 5720 ± 47 years, which was then updated to 5568 ± 30 years (Libby, 1972; Libby et al, 1949). The latter value is referred to as the “Libby half-life" and is still used to calculate radiocarbon ages (Currie, 2004; University of Georgia, 2017). In 1965, the physical 14C half-life was accepted to be 5370 ± 40 years, which continues to be used today in some fields of radiocarbon research (Johnson, 1965; University of Georgia, 2017). In the 1970s, researchers began to use accelerator mass spectrometer (AMS) systems to measure 14C. AMS systems are able to analyze smaller sample sizes, allowing for analysis of amino acids, seeds, pollen, and other small macrofossils (University of Georgia, 2017).
The amount of atmospheric 14C has varied throughout time due to changes in production rates caused by the carbon cycle, and changes in the cosmic-ray flux (Libby, 1972; Reimer et al., 2013). Therefore, the resulting 14C date (yr BP) needs to be compared to an established calibration curve so an approximate calendar date can be determined (cal yr BP) (Blaauw & Christen, 2011; Reimer et al., 2013). Ideally, calibration should be based on a dated record that utilizes carbon levels from the atmosphere at the time of sample formation (Reimer et al, 2013). The IntCall3 curve represents the mid-latitude Northern Hemisphere atmospheric carbon record and was developed using tree-ring measurements and macrofossil data that extend to 13,900 cal BP and the end of the range of 14C dating methods, respectively (Reimer et al, 2013). The IntCall3 curve was updated from the IntCal09 curve, however the majority of differences between the two occurred prior to -12000 cal yr BP. Still, it is important to note that 14C calibration curves are still being developed as researchers gain increased understanding of atmospheric 14C processes (Reimer et al, 2013).
Radiocarbon dating is less reliable for the past 500 years due to the large injection of carbon into the atmosphere around AD 1955 (Gale, 2009). Therefore, more recent sediments
21


(within the last 150 years) must be dated using 210Pb dating techniques. 210Pb is a radiometric isotope that is formed as radon-222 decays. 210Pb then falls to the Earth’s surface, where it accumulates over time. Dates of the sediments are established by assessing the amount of 210Pb present at a certain depths, usually the last 100-200 years, and the results are used to calculate dates using a mathematical model based on 210Pb accumulation and sedimentation rates (Science Museum of Minnesota, 2018).
An age-depth model is generated by fitting a curve to the provided 210Pb and 14C dates, and using the IntCall3 radiocarbon calibration curve (Reimer et al., 2013). Recently, paleoecological studies have relied heavily on a methodology using Bayesian statistics rather than linear regression models (Blaauw & Christen, 2011). This methodology assumes a monotonic rate of accumulation and utilizes a gamma autoregressive process to establish sedimentation rates as a function of depth. Outliers are identified using a ‘shift outlier’ method, and a Markov Chain Monte Carlo (MCMC) “t-walk” algorithm to produce posterior distributions (Blaauw & Christen, 2011). Bacon splits the core into vertical sections (default = 5) and then uses the MCMC iterations to estimate the accumulation rates (or sedimentation times; default = years/centimeter). The result is a more accurate age-depth model that accounts for changes in accumulation/sedimentation rates over time (Blaauw & Christen, 2011). Each depth within the core is assigned a date that can be used in analyzing other proxy records.
22


CHAPTER III
SITE DESCRIPTION Lily Pond
Lily Pond (Lat. 38°56'3.87"N, Long. 106°38'48.26"W), located in Taylor Park in the Gunnison National Forest, lies within the Sawatch Mountain Range in central Colorado at an elevation of 3200 m (Figure 4). The pond is at an ecotone between montane and subalpine forest. The forest, primarily composed of Pinus contorta, Picea engelmannii, and Abies lasiocarpa, was historically responsive to changes in climate of the late-glacial period and Holocene (Briles et al, 2012; Fall, 1997). The pond was formed -12,000 years ago as a kettle lake by Pleistocene ice-age glaciers. A recessional moraine to the north of lake dams the site. The in- and out- flowing streams are intermittent and likely only move water during the spring snowmelt, and during major summer precipitation events. Today, the pond is shallow (-1 meter of water at the deepest point) and surrounded by sedge (Cyperaceae spp). During the winter months, Lily Pond ices over due to cold temperatures in the Taylor Park region. During the summer months, the pond is covered in lily pads (Nuphar spp.). The pH of Lily Pond at the end of August (2017) was 6.9, indicating circumneutral conditions. The modern diatom community is dominated by benthic species including Fragiliariaceae, Pinnularia, Sellaphora, Encyonema, and Stauroneis. The diatom community is discussed further in Chapter V. There are burned trees on the southwest slope of Lily Pond. Age estimates are -120 years, but addition dendrochronological work is needed to confirm the age.
Present day climate in the Taylor Park region is characterized by cold winters with heavy snowfall, and warmer summers with frequent thunderstorms during July and August. The high
23


concentration of tall mountain peaks influences the climate patterns in the region and causes temperature and precipitation to change with elevation, resulting in local microclimates (Fall, 1997). Taylor Park receives an average of 42 cm of precipitation each year. Average annual snowfall is -276 cm, the majority of which falls during November through April. Average temperatures range from a low of -8.8 °C to a high temperature of 9.5 °C. The average temperature during the summer months is 10°C (average low of 1,4°C, average high of 19°C), and average temperature during the winter months is -7°C (average low of -16°C, average high of 3°C) (average precipitation and temperature data from Western Regional Climate Center; [https://wrcc.dri.edu/cgi-bin/cliMAIN.pl7cotayl]).
Forest Hill Mine
The Forest Hill Mine (Lat. 38°55'24.78"N, Long. 106°38'28.12"W) is located -100 meters upslope and -1100 meters southeast of Lily Pond and was a small operation that mined lead, silver, and zinc (Figure 4). The mine operated upstream of Lily Pond and was in use from 1880-1920, with a short break from 1907-1916 CE. The mine is in a different hydrologic unit code (HUC) watershed than Lily Pond. However, there are smaller exploratory mines in the watershed, collectively called the Forest Hill Mining Complex (FHMC) (Appendix A). There are also remnants of cabins and other living structures in the area, including in the area immediately surrounding Lily Pond, along with old mining infrastructure including a boiler and metal pipes.
A preliminary dendrochronology study (Steelman, unpublished) identified three separate forest stands within the FHMC. Two of them indicate that there was an extensive amount of logging in the FHMC that corresponded to the establishment of Forest Hill Mine proper. Two of the forest stands have establishment dates between 1840-1870 CE, which corresponds with the start of mining in the area (Steelman, Unpublished). This indicates that anthropogenic activity
24


was taking place at Lily Pond beginning around 1850 CE, although it is not clear if all activity was directly related to mining operations.
Grand Junction
Lily
Pond
Coring
Sites
â–¡en Ml
L*J .
Lily Pond Buena Vista
A) State of Colorado showing location of Lily Pond (blue point). Black box shows area shown in B) & C). B) Aerial image of Lily Pond and Forest Hill Mine (white area). Image collected from GoogleEarth. looking from the south to the north. C) Location of Lily Pond in relation to Forest Hill Mine (black X). Red triangle shows location where cores were extracted. Dashed purple line shows watershed boundary. Contours are drawn in brown at 20-meter intervals.
Pictures show: D) Forest Hill Mine as it looks currently (picture taken looking towards the east) E) Lily Pond during August 2017 coring trip (picture taken from south looking north)
Watershed Boundary
Figure 4. Site Map
25


CHAPTER IV
METHODS AND DATA ANALYSIS Field Methods
A 1-meter continuous sediment core was extracted from Lily Pond in August 2017. The core (LP17) was taken from the east side of Lily Pond using a short corer designed by Steve Klein that captures the sediment-water interface. The core was measured before extraction and transportation. The top 20-cm of unconsolidated sediment was extracted at 1-cm increments onsite and placed in Whirl-Pak bags. The remaining part of the core was placed in PVC piping and transported back to the Paleoecology, Palynology and Climate Change Laboratory at University of Colorado Denver for analysis. Water temperature and pH were taken when the core was taken. Water samples from the sediment-water interface, the water surface, and the water column were also collected along with a sedge and lily pad for diatom analysis.
This study also uses data collected from a 2-meter sediment core taken at Lily Pond in August 2015. This core (LP15) was taken from the south-east side of Lily Pond using a D-section corer and did not capture the mud-water interface. This core was also taken back to the University of Colorado Denver for analysis.
Laboratory Methods
Core description and lithology
Once back in the lab, the LP15 and LP17 cores were measured to confirm length. Core lithology was described to account for any changes in appearance and texture of the sediment. The cores were then subsampled at 0.5 cm increments, stored in Whirl-Pak bags, and placed in
26


the refrigerator. Any macrofossils found while sampling cores were collected to be used for radiocarbon dating.
Chronology
There is an existing age model for Lily Pond that spans -17,100 years that was formulated using a series of radiocarbon dates obtained from a core taken in 2007 (LP07), and the LP15 core (Table 2; Briles et al, 2012). Samples from the LP07 core were sent to Beta Analytic for 14C analysis. One macrofossil from the LP15 core was sent to DirectAMS for 14C analysis and was used to link the LP07 and LP15 cores. However, the record was lacking a high-resolution chronology for the past 1000 years (top 55 cm of the core). Radiocarbon dating is less reliable for the past 500 years due to the large injection of carbon into the atmosphere around 1955 CE due to nuclear bomb testing (Gale, 2009). Therefore, to constrain the recent mining activity, a 210Pb chronology was developed. Sediments from the upper 25 cm of the LP17 core were sent to St. Croix Watershed Research Station for 210Pb dating. One macrofossil from the LP17 cores was sent to the University of Georgia Center for Applied Isotope Studies in Athens, Georgia for AMS-radiocarbon dating analysis to help link and constrain the three core chronologies.
Diatoms
Diatom species were used to reconstruct aquatic ecological conditions, including lake level and water chemistry. They can also provide insights about terrestrial land use changes that result in increased sedimentation and allochthonous inputs. Diatoms were sampled from both the LP15 and LP17 cores. Diatoms were initially sampled at 6-cm increments throughout the LP15 core. After 210Pb dating was complete, higher resolution sampling was done using the LP17 core with specific focus in the top -30 cm of the core. Sediment was dried at 90°C for 24 hours. Next,
27


0.4 g of dry sediment was weighed out and put into a 50-ml centrifuge tube. The samples were taken to the INSTAAR Sediment Laboratory at University of Colorado Boulder for processing. Approximately 5 ml of nitric acid was added to each sample, and then placed in a microwave digestor for one 35-minute cycle, to control temperature and pressure. The samples were heated to a maximum temperature of 200°C, and then allowed to cool before removing from the digestor. The remaining nitric acid was decanted, and the samples were rinsed with distilled water 6 times.
Diatoms were mounted on slides using Naphrax and counted using light microscopy. Valves were identified and counted using a Nikon Labphot microscope at lOOOx magnification with phase contrast & dark/bright field. Approximately 350 valves were counted per slide. Diatoms were identified to the genus level using the Taxon Identification Guide maintained by Diatoms of North America (2018; https://diatoms.org/).
Geochemistry
Magnetic susceptibility, loss-on-ignition (LOI), and X-ray fluorescence (XRF) analyses were performed to reconstruct sediment geochemistry and identify input changes. These analyses help determine the effects of mining on inputs into the aquatic ecosystem. Magnetic susceptibility was used to determine increases in iron-bearing minerals due to erosion into the lake (Gedye et al, 2000). Magnetic susceptibility was examined at 1-cm intervals using a Bartington MS2E point sensor meter. Measurements were recorded in units of centimeter-gram-second (cgs).
LOI measures changes in organic content of the sediments, which, in part, indicates changes in lake productivity (Dean, 1974). However, external/terrestrial organic inputs are possible (e.g., run-off events, human, wildlife, and livestock excrement), but these are likely
28


minimal at Lily Pond given limited stream inputs and the extensive sedge mat to the south of the pond. Samples were dried at 90°C for twenty-four hours and measured for weight loss to determine moisture content in each sample. Samples were then dried at 550°C and 900°C for two hours each, and again weighed between each drying period to determine amount of organic material, and carbonates present, respectively.
XRF analysis is used to determine elemental composition of sediment samples. This study used an Olympus model portable-XRF gun and a bench stand (Kenna et al, 2010; Burtt, et al, 2013). Processing samples with the laboratory bench stand decreased the amount of laser movement when taking the readings, resulting in more accurate results. Samples were also dried prior to analysis to eliminate risk of water absorption (Kido et al, 2006). XRF was initially measured at 4-cm intervals throughout the core. Additional XRF readings were conducted at 1-cm intervals through the mining period to present. Samples were dried at 90°C, crushed using an agate mortar and pestle, placed on top of super fine filter paper in a plastic XRF sample cup with polypropylene fill, and covered with polypropylene thin-film.
Charcoal
Macroscopic charcoal was used to understand the local fire history at Lily Pond. It also provides insight into both natural and anthropogenic fire activity. Charcoal data was collected and analyzed from the LP15 core. Charcoal was sampled at 0.5-cm increments throughout the length of the core using the sieve method (Whitlock & Larsen, 2001). Samples were treated with sodium hexametaphosphate to dissolve clay material and bleach to get rid of humic matter, and then filtered through a 125 pm sieve. Macroscopic charcoal pieces were counted in a gridded petri-dish using a stereomicroscope.
29


Pollen
Pollen analysis provides information regarding changes in local forest composition and structure. Pollen data in was collected and analyzed from the LP15 core at 6-cm increments.
After initial pollen analysis, additional samples were prepared from the top portion of the core (top ~30 cm). Laboratory methods outlined by the LacCore Pollen Preparation Procedure (University of Minnesota; http://lrc.geo.umn.edu/laccore) were followed for all samples. Potassium hydroxide and acetolysis were used to remove organic material, and hydrofluoric acid was used to remove silicates (Faegri & Iversen, 1989). Slides were prepared and counted using a light microscope. Lycopodium tracers were added to allow for pollen concentration calculation (grains cm"3). Grains were identified by Dr. Briles to the lowest taxonomic level possible using the PPCC lab reference library. Pinus grains were separated into Diploxylon (contorta-type) and Haploxylon (flexillis-type). All other grains were identified and counted at the genera level.
Data Analysis
Chronology
Age-depth models of lake sediments using Bayesian iterative models account for the probability distribution of each radiometric date, consider realistic sediment accumulation of lakes, weight known dates more heavily than less known dates, and provide error estimates for the entire chronology. Each date in the model influences other dates and the final chronology and allows for fewer dates to produce more accurate age-depth models (Blaauw & Christen, 2011). Radiocarbon and 210Pb dates were used to create an age-depth model using an R-based statistical package called BACON (Blaauw and Christen, 2011). BACON assumes monotonic accumulation, and then models sedimentation rates as a function of depth using a gamma autoregressive process. The accumulation rates were calculated at 0.5 cm/ year and all other
30


parameters within Bacon were left at the default values. 14C dates were calibrated using the IntCall3 calibration curve. 210Pb dates were not set to a calibration curve since they were already set to the calendar scale (Blaauw & Christen, 2011). Ages were calculated at 0.5 cm intervals from 0 cm to 139 cm.
Diatoms
Diatom counts were converted to percentages for each taxa at each sample depth using the following equation:
\£ all valves
Percentages were used for graphing and data analysis. Taxa were included in analysis if they
further analyzed using Paleontological Statistics Software Package (PAST 3.0; Hammer, Harper, & Ryan, 2001) software. Cluster analysis was used to identify different zones in the sediment core by comparing diatom genus percentages at the different sample depths. Diatom community composition in each sample was compared against the others using Principal Components Analysis (PCA). The trends and groupings between the samples were used to establish zone boundaries and discuss possible changes in ecological conditions throughout the sediment core. PCA was run an additional time for each zone to identify how the diatom genera changed between zones (Appendix D).
Geochemistry
valve count
contributed at least 1% of total relative frequency in more than one sample. Diatom results were
Results from magnetic susceptibility and LOI were plotting in C2 (Juggins, 2007) to visually identify trends in the data throughout the core. LOI data was used to calculate the percent of organic material present throughout the core. XRF data was collected using the
31


Olympus proprietary “Geochemistry Cycle” and were reported as percentages. Elemental ratios were calculated to determine any differences in sediment inputs into Lily Pond using the following equation:
element — Ti
element + Ti
Geochemistry at Lily Pond primarily focused on the levels of Iron (Fe) and Manganese (Mn). Both Fe and Mn levels were normalized against Titanium (Ti) because it is abundant, is not biologically important, and is resistant to weathering (Kylander et al, 2011). The Fe/Ti ratio provides information about allochthonous inputs into lake systems. Increases in the Fe/Ti ratio values indicate an increase in allochthonous inputs from the surrounding area. An increase or decrease may be indicative of changes in precipitation regimes or changes in land use in the surrounding area. The Mn/Ti ratio provides information about lake level changes. In oxygen-rich environments, Mn forms an insoluble oxide. Therefore, increases in the Mn/Ti may be indicative of increased oxygen levels due to a decrease in lake levels (Kylander et al, 2011).
Pollen
Pollen percentage data and accumulation rates were calculated to reconstruct forest structure and composition, respectively. Both were calculated at each sample depth using the following equations.
Pollen percentages:
32


Pollen accumulation rate:
Taxa concentration Sediment deposition rate
The arboreal to nonarboreal ratio was also calculated to help identify changes in the forest structure using the following equation:
£ arboreal pollen grains — £ nonarboreal pollen grains £ arboreal pollen grains + £ nonarboreal pollen grains
Higher amounts of arboreal pollen indicate a more closed canopy. A more open forest structure is indicated by increases in nonarboreal pollen from grasses, shrubs, and other steppe species. These changes can indicate possible logging events as well as large fire events.
Charcoal
Charcoal counts were analyzed using the statistical program CharAnalysis ([http://phiguera.gitbug.io/Chamalysis]) (P. Higuera, 2009). CharAnalysis uses the input information and interpolates it to the median sample resolution (yr sample"1). The program then uses this to distinguish between background charcoal (BCHAR) and charcoal accumulation rates (CHAR) using charcoal concentrations (particles cm'3) and the sedimentation rate (cm yr'1). CHAR is reported in units of particles per centimeter per year (particles cm"2 yr"1).
BCHAR represents a running average of charcoal accumulation rates through time. BCHAR reflects changes in the rate of total charcoal production (biomass) and changes in secondary charcoal deposition mainly from extra local sources (Higuera et al, 2009). For this study, BCHAR was estimated using a 800-year lowess smoother, robust to outliers. The 800-year smoothing window was selected based on the signal-to-noise index (Higuera, 2009). The signal-to-noise index (SNI) is a statistical measurement of the separation between BCHAR and peak
33


events. The larger the separation (generally above 3 for this record) the higher the confidence of a fire event occurred in the record at that time (Higuera et al, 2014). A local threshold value (99%) was used to identify CHAR peaks beyond BCHAR levels using a noise distribution model based upon a 1-mean Gaussian distribution. If a peak in CHAR exceeded the threshold, a local fire event was identified. Peak magnitude (particles cm"2 peak'1), or the amount of CHAR above BCHAR, has been used as an indicator of fire severity, although proximity of the fire events to the lake and wind direction would also influence peak magnitude. Pollen data around the time of an individual fire event should be examined to estimate fire severity (e.g., (Minckley et al, 2012; Del Priore, 2015; Morton et al., 2017). The fire return interval (FRI) was determined using a 1000-year running mean.
34


CHAPTER V
RESULTS
Core Lithology & Chronology
The top 20 cm from the LP17 core was unconsolidated sediments that were extracted in the field to preserve the sediment-water interface. The LP17 core consisted of unconsolidated, dark course detritus gyttja (CDG) for the first 20 cm and dark fine detritus gyttja (FDG) after 20 cm. In the LP15 core, the top ~6 cm of sediment was a dark CDG, and the remainder of the core was dark fine detritus gyttja (FDG). There was a sand layer at ~85 cm. The top portion of both cores had an abundance of roots and Cyperaceae (sedge) seeds, likely from the dense aquatic vegetation at Lily Pond. The LP15 and LP17 core lithologies are very similar, beginning with CDG and then transitioning into FDG. The LP07 core corresponds with the FDG in the other two cores. The LP15 and LP17 cores do not show the transition to silt that is seen in the LP07 core at ~75 cm. The transition may have occurred later in these cores due to variations in sediment accumulation within the basin.
35


LP07
LP15
LP17
0 1 50 -100 -150 -
1 200-_c
Q.
O 250-300 -350 -400 -
0
20
40
60
80
100
Coarse detritus gyttja (CDG) Fine detritus gyttja (FDG) Sand
Organic silt Organic clay Inorganic clay Sand & clay 14C Date (yr BP)
Figure 5. Lithology of the LP15 and LP17 cores.
Due to the extraction method (d-section corer) of the 2015 core, the sediment water interface was not preserved. The 2017 short core captured the lost upper sediments, which preserve information on anthropogenic period of the record and the modern proxy record of the terrestrial and aquatic communities. Three cores, 2007, 2015, and 2017, were linked based on an increase in Pinus percentages, a decrease in Artemisia percentages, an increase in CHAR between 90-100 cm depth, the sustained increase in magnetic susceptibility at ~28 cm in the 2015 and 2017 cores, a peak in CHAR at the top of the cores, and an overlapping 14C date at depth 97 cm in the 2007 core and 69 cm depth in the 2015 core (Table 2). An age-depth model was created for Lily Pond using five calibrated 14C dates and a series of 210Pb dates. Details about 210Pb and 14C dates are shown in Table 2. Additional information about the 210Pb dates is included in Appendix B. Four of the 14C dates were previously published, including a date that
36


overlapped with the 14C date on the 2015 core (Briles et al, 2012). The resulting BACON age model is shown in Figure 6.
6000 4000 2000 0
cal yr BP
Figure 6. a) Age-depth model for Lily Pond, b) Results from 210Pb dating before calibration.
37


Table 2. Series of 210Pb and 14C dates used to generate the age depth model.
210Pb Dates
Lab ID Material Age (yr BP) Error (yr) Depth1 (cm) Core
sediment 65.6 0.66 0 LP17
sediment 59.5 0.72 3 LP17
sediment 53.5 0.72 5 LP17
sediment 46.8 0.73 7 LP17
sediment 39.3 0.70 9 LP17
sediment 29.8 0.80 11 LP17
St. Croix sediment 16.8 0.79 13 LP17
Watershed sediment 8.5 0.85 14 LP17
Research sediment -1.3 0.97 15 LP17
Station sediment -13.3 1.25 16 LP17
sediment -31.3 2.30 17 LP17
sediment -51.6 3.70 18 LP17
sediment -68.8 6.10 19 LP17
sediment -87.8 11.55 20 LP17
sediment -122.5 10.51 21 LP17
sediment -148.5 25.36 22 LP17
14C Dates
Lab ID Material Age (yr BP) Error (yr) Depth1 (cm) Core
UGAMS34916 wood 940 25 41 LP17
D-AMS018629 wood 1627 29 69 LP15
Beta2448852 peat 1620 40 69 LP07
Beta244886 seed 2560 40 96 LP07
Beta242757 wood 5080 40 120 LP07
Beta244888 gyttja 6440 40 139 LP07
1 Depth reported in centimeters below the sediment-water interface after being linked to the LP15 and LP17 cores
2This date was used to link the LP07 and LP15 cores but was omitted from the age-depth model.
. Constrained cluster analysis was initially conducted on the diatom data. Similarity was measured using Euclidean distance to determine similarity in samples in chronological order. Results from the constrained cluster analysis are shown in Figure 7. The constrained cluster groupings and visual examination of diatom results were used to identify zones within the
38


sediment record. Cluster analysis identifies samples that are most similar to each other and places them into separate groups. The different groups identified do not carry more significance than others (Hammer et al., 2001). The zones identified using the diatom cluster analysis were then compared to results from constrained cluster analysis using pollen data as a comparison of significant changes between groups of samples. The grouping of the final zones were comparable between the two datasets. A total of four zones were established for discussion of the results. Zone number, depth, and age are listed in Table 3. The other data will also be discussed in the context of these zones.
Figure 7. Results from constrained cluster analysis using a) diatom data; b) pollen data.
39


Table 3. Zones within the sediment record with respective dates and ages.
Zone Depth (cm below sediment-water interface) Age (calyr BP)
1 115-92 5000-2400
2 92-71 2400-1600
3 71-30 1600-200
4 30-0 200-present
Sediment Geochemistry
Magnetic susceptibility & loss-on-ignition
Results from magnetic susceptibility and LOI analysis are shown in Figure 8. The magnetic susceptibility remains diamagnetic (< 0 cgs) through most of the bottom section of core (-5000 cal yr BP- -100 cal yr BP). Consistent paramagnetic trends (values > 0) are not recorded until -100 cal yr BP (1850 CE). Percent organic content in Zone 1 gradually increases from 41% to 64%. In Zone 2, organic content continues to increase to 71%. Organic content reaches its maximum value (84%) in Zone 3 at -337 cal yr BP. Organic content decreases towards the end of Zone 3 into Zone 4, from 84% to 60% beginning around -450 cal yr BP.
XRF
Results from XRF analysis are included in Figure 8. The Fe/Ti ratio decreases through Zone 1 from 0.89 to 0.84. The Fe/Ti ratio continues to decrease through Zone 2 down to 0.80. In Zone 3, the Fe/Ti ratio increases from 0.80 to 0.85, with a maximum value of 0.87 at 20 cm (82 cal yr BP). The Fe/Ti ratio continues to increase in Zone 4 to 0.91 at 0-2 cm (present day), which is the highest of the record. The Mn/Ti ratio increases slightly through Zone 1 from 0.81 to 0.84 and continues to increase through Zone 2 to 0.88. In Zone 3 the Mn/Ti ratio decreases from 0.88
40


to 0.81, with a minimum value of 0.78 at 36 cm (672 cal yr BP). The ratio continues to decrease through Zone 4 to 0.72 at 0-2 cm (present day), which is the lowest value of the record.
Diatoms
The diatom percentages are shown in Figure 8 and summarized in Table 4. A total of 16 taxa were identified and used in analysis (Appendix E), however this paper focuses on five taxa because of their high abundance throughout the record. The dominant diatom type in the Lily Pond record is from the Fragilariaceae family, including species of Fragilaria, Slaurosira, Stauroforma, and Staurosirella. Other dominant benthic diatom genera present in the record in descending abundance include: Pinnularia, Stauroneis, Sellaphora, and Encyonema. These species are the focus of the discussion of the Lily Pond record because they are the most common taxa throughout the record, and represent both epipelic (Pinnularia, Sellaphora, Stauroneis) and unattached (Encyonema) environments.
The diatom communities found in the sediment record were compared to the community found in modern-day samples from the sediment-water interface (0 cm depth), the water column, and surrounding aquatic vegetation. No planktonic taxa were found in the modern-day samples from the water column. The samples from the water column were dominated by the Fragilariaceae family and Sellaphora. Diatom samples from the lily pad and sedge were also composed of Fragilariaceae valves as well as Tabellaria, Encyonema, Cymhella, Sellaphora, and Navicula, however, not in large abundances. Tabellaria, Navicula, Encyonema, and Cymbella are all epiphytic species, meaning they are often found growing attached to vegetation (Patrick, 1977). Sellaphora is considered epipelic, meaning the species generally lives on sediments (Werner, 1977). However, it is not uncommon for diatoms to adapt to different environments and
41


survive in different habitats, especially in shallow lake environments where the open water environment is not dominated by planktonic species (Hellebust & Lewin, 1977; Patrick, 1977).
The community has been dominated by small benthic diatoms over the last 5000 years, which indicates a shallow lake level during the late Holocene at Lily Pond. Fragilariaceae are small, araphid species that are commonly found in aquatic systems located at high elevations (Lotter et al, 1999). They can survive in both planktonic and benthic habitats, and are referred to as tychoplanktonic (Patrick, 1977). Pinnularia, Sellaphora, and Stauroneis are all epipelic taxa, meaning they prefer to grow attached to sediments. Encyonema is a benthic taxon that does not attach to anything. All taxa identified at Lily Pond are considered circumneutral. Pinnularia has also been found in environments with low pH (< 3.5) (Patrick, 1977; DeNicola, 2000) as well as iron-rich environments (Wollmann et al, 2000). Only two planktonic genera were identified in the record, Aulacoseira and Cyclotella, and they were only found at abundances < 6%. Both are centric diatoms, require high levels of nutrients to survive, and are sensitive to changes in turbidity (Patrick, 1977). These taxa were graphed together to analyze for changes in the planktonic habitat through the record.
Zone 1 contains the lowest average abundance of Fragilariaceae (53%) and Encyonema (3%). Both groups gradually increase throughout the zone (to 79% and 22%, respectively). Pinnularia, Stauroneis, and Sellaphora have the highest average abundances (20%, 11%, and 6%, respectively) of the record in Zone 1 and decrease (to <4%) through the zone. The average percent of planktonic taxa is 0%, although the abundance increases to 1% towards the end of Zone 1.
Zone 2 is defined by increases in average abundance of Fragilariaceae (53% to 57%) and Encyonema, (3% to 4%). Sellaphora decreases (6% to 1%) throughout Zone 2. Pinnularia and
42


Stauroneis increase at the beginning of Zone 2 (to 35% and 28%, respectively), and stay fairly consistent through the end of Zone 2. However the averages are lower than Zone 1 (from 40% to 10%, and 28% to 3%, respectively). Planktonic taxa have the same average abundance as Zone 1 (1%).
The transition from Zone 2 to Zone 3 is defined by an increase in Fragilariaceae, Sellaphora, and planktonic species from 42% to 70%, 0 to 4%, and 0 to 3%, respectively. Pinnularia, Stauroneis, andEncyonema all decrease at this same time from 26% to 2%, 17% to 1%, and 6% to 3%, respectively. Fragiliariaceae, Sellaphora, and planktonic species decrease throughout the MCA (beginning around -1200 cal yr BP) to 59%, 2%, and 2% respectively. Pinnularia, Stauroneis, and Encyonema then increase through the LIA, hitting maximum abundances of 19%, 8%, and 10%, respectively.
Zone 4 is defined by an increase in Fragilariaceae species, Sellaphora, and planktonic taxa at the start of the Zone (to 80%, 10%, and 6%, respectively). The highest average abundances of Fragilariaceae (72%) and planktonic taxa (2%) occur in Zone 2 as well. Planktonic taxa follow similar trends as Zone 3 (lowest abundance 0%, highest abundance 6%). Pinnularia, Stauroneis, and Encyonema decrease at the beginning of Zone 4 (to <1%), but increase around 82 cal yr BP (-20 cm) to abundances >2%.
43


Sediment
Geochemistry
Diatoms
~T 1 T oo t-CNJ -100 -|
30- 300 -
40- 700 -
50- 1100 -
60- 70- 1500 -
80- 90- £• 1900-co >, 2300 -
100- To 2700 - 05 O) < 3100 -
3500 -
3900 -
110- 4300 -
4700 -
5100 -


cp‘





\K"




Figure 8. Sediment geochemistry and diatom record for Lily Pond. Sediment geochemistry graphs show magnetic susceptibility (blue line) with a gray line showing 0 cgs, percent organics (green line), and XRF ratios (black shading). Diatom graphs (shown in blue) are percents of Fragiliariaceae, Pinmdaria, Stauroneis, Encyonemci, and Sellaphora. The switch from summer-wet to winter-wet is shown by a blue dashed line; MCA is shown in red; LIA is shown in blue; mining
period is shown in gray.


Table 4. Summary of diatom abundances sorted by zone.
ZONE Fragilariaceae Pinnularia Stauroneis Encyonema Sellaphora Planktonic
4 Avg 72 6 2 4 4 2
High 97 19 8 17 18 6
Low 45 1 0 0 0 0
3 Avg 66 9 5 5 3 1
High 83 26 17 11 12 6
Low 42 3 0 1 0 0
2 Avg 57 21 15 4 1 0
High 79 35 28 6 3 1
Low 28 8 3 1 0 0
1 Avg 53 20 11 3 6 0
High 79 39 28 6 22 1
Low 22 4 1 0 0 0
Pollen
Lily Pond pollen percentages, accumulation rates, and ratios of arboreal to nonarboreal pollen are shown in Figure 9 and summarized in Table 5. All pollen analysis was conducted on the LP15 core. Pinus is the dominant pollen-type through the record, followed by Picea, Artemisia, Onerous, Amaranthaceae, and Abies. There a few notable long-term trends in the data. First, abundance of total Pinus increases (47.7% to 60.1%) from Zone 1 to Zone 4. Second, average relative abundances of Picea, Onerous, and Amaranthaceae decrease between Zone 1 and Zone 4 (17.9% to 8.8%, 5.2% to 0.5%, and 7.3% to 4.1%, respectively). Below more detailed changes of the dominant pollen types for each zone.
Zone 1 has the lowest Pinus percentages of the record (47.7%), while Picea percentages are the highest (17.9%). Abies percentages (1.8%), Onerous (3.4%), Artemisia (13.3%) and Amaranthaceae (7.3%) abundances are also the highest of the core. The AP/NAP ratio was lowest in Zone 1 (0.47). Pollen accumulation rates are also the lowest of the record for both Nuphar and total pollen in Zone 1 (48.5 grains cm"2 and 1270.9 grains cm'2, respectively).
45


The transition from Zone 1 to Zone 2 is defined by a decrease in average abundance of Abies, Picea, and Quercus (from 1.8% to 1.5%, 27.9% to 13.4%, 3.4% to 1.6%, respectively), and an increase in Pinus (from 47.7% to 59% average abundance). Abundance of Artemisia and Amaranthaceae also decrease in abundance (from 13.3% to 10.7%, and 7.3% to 5.3%, respectively) from Zone 1 into Zone 2. The AP/NAP ratio increases from 0.47 to 0.61 reflecting the increase in Pinus and decrease in nonarboreal species {Quercus, Artemisia, and Amaranthaceae). The average total total pollen accumulation rate increases from 1270.9 grains cm"2 to 3078.2 grains cm'2, while Nuphar also increases from 48.5 grains cm"2 to 142.2 grains cm"2.
Zone 3 has the highest average abundance of Pinus (64.5%), while Quercus and Amaranthaceae have the lowest abundance of the record (1.4% and 3.3%, respectively). Picea decreases (to 11%) Abies also decreases slightly (to 1.4%) from Zone 2 to Zone 3. The AP/NAP ratio reaches the highest value (0.64) of the record as well, representing the increase in Pinus abundance, and decrease in Quercus and Amaranthaceae. Pollen accumulation rates decrease between Zone 2 and Zone 3 for both Nuphar and total pollen (142.2 grains cm"2 to 135.1 grains cm"2and 3078.2 grains cm"2to 2628.9 grains cm"2, respectively).
The transition from Zone 3 to Zone 4 is defined by a decrease in Pinus abundance (64.5% to 60.1%). Picea also decreases (11.0% to 8.8%) while Abies remains the same as Zone 3 (1.4%). Quercus, Artemisia, and Amaranthaceae increase slightly (from 1.4% to 1.7%, 11.1% to 14.2%, and 3.3% to 4.1 %, respectively). The AP/NAP ratio dropped in Zone 4 to 0.56, representing the decrease in Pinus and Picea, and increase in Quercus, Artemisia, and Amaranthaceae. The pollen accumulation rates also increase dramatically during this period for both Nuphar (135.1 grains cm"2 to 700.1 grains cm"2) and total pollen (2628.9 grains cm"2 to 18488.8 grains cm"2). The AP/NAP ratio dropped in Zone 4 to 0.56, similar to Zone 1.
46


10-
20-
30-
40-
50-
60-
70-
80-
90-
100
110-
-100 300 700 1100 -1500 -
j£" 1900
CO
g 2300
8 2700
<1)
CD
< 3100 3500 3900 4300 4700 5100 -I


r i i i
i—1—r-1—i—i—r

k> °
Percent
b b bbbb o
o> â– ? & & â– > Ratio
-r --T Mine Zone 4

LIA
Zone 3
MCA

Winter wet Zone 2
Summer wet Zone 1

*o &o
7 x3 & 57
Accumulation Rate
(x1000 grains cnr2)
Figure 9. Pollen graphs for Lily Pond showing percentages for Firms, Picea, Abies, Onerous, Artemisia, Amaranthaceae. Arboreal species are shown in green. Herbaceous and shrub species are shown in brown and yellow. The AP/NAP ratio is shown in black. Accumulation rates for Niiphar and total Pollen are shown in gray. The switch from summer-wet to winter-wet precipitation is shown by dashed blue line; MCA is shown in red; LIA is shown in blue; mining period is shown in gray.


Table 5. Summary of Lily Pond pollen record organized by zones.
PERCENTAGES RATIO ACCUMULATION RATE (grains cm'2)
ZONE Pinus Picea Abies Quercus Artemisia Amaranthaceae AP/NAP Nuphar Total
Avg 60.1 8.8 1.4 1.7 14.2 4.1 0.56 700.1 18488.8
4 High 65.0 14.2 3.4 3.5 20.6 7.7 0.63 2019.3 58845.7
Low 52.9 5.6 0.0 0.5 9.8 1.1 0.39 0.0 2289.6
Avg 64.5 11.0 1.4 1.4 11.1 3.3 0.64 135.1 2628.9
3 High 70.9 14.0 2.3 2.7 13.1 5.0 0.75 239.6 3586.9
Low 57.8 8.0 0.7 0.0 8.4 2.1 0.57 48.8 1854.8
Avg 59.0 13.4 1.5 1.6 10.7 5.3 0.61 142.2 3078.2
2 High 65.9 17.8 1.9 2.7 12.1 7.0 0.70 190.6 4338.9
Low 53.0 10.3 0.6 0.7 8.9 2.2 0.51 72.4 2195.2
Avg 47.7 17.9 1.8 3.4 13.3 7.3 0.47 48.5 1270.9
1 High 62.2 25.6 4.2 5.2 14.4 10.7 0.59 109.4 1993.4
Low 34.5 9.8 0.5 1.2 10.7 3.7 0.39 0.0 595.0
Charcoal
Charcoal accumulation rates (CHAR), background charcoal accumulation rates (BCHAR), peak magnitude, and fire return interval (FRI) are shown in Figure 10, and summarized in Table 6. All charcoal analysis was conducted on the LP15 core. A total of 20 fire events were identified in the last 5000 years. The average signal-to-noise index at Lily Pond is 5.1; however, there is a short period between -2300 and 2000 cal yr BP where the SNI drops to levels below 3.0. Notable long-term trends in the charcoal data include an increase in average BCHAR from 0.1 particles cm"2 yr'1 at the start of the record to 0.3 particles cm"2 yr"1 in Zone 4. BCHAR and fire activity increase (0.6 particles cm"2 yr"1, 4 fire events, FRI average 138 yr fire'1) around -1200 cal yr BP during the height of the MCA and drop back to lower BCHAR levels and fire activity (0.3 1.3 particles cm"2 yr"1, 1 fire event, FRI average 276 yr fire'1) during the LIA. BCHAR increases dramatically -250 cal yr BP and the largest peak magnitudes (1.2
48


particles cm"2 yr'1, 50.6 particles cm'2 peak"1) occur thereafter. Below more specific trends in the charcoal record are discussed within individual zones.
In Zone 1, BCHAR and peak magnitude were the lowest of the record. BCHAR had a maximum value of 0.2 particles cm"2 yr"1 and a minimum of 0.0 particles cm"2 yr"1, with an average of 0.1 particles cm"2 yr"1. Maximum peak magnitude in Zone 1 was 8.1 particles cm"2 peak"1, with an average of 0.1 particles cm"2 peak"1. A total of 8 fire events were identified in Zone 1, with five occurring after -3000 cal yr BP. Around -2400 cal yr BP, the FRI decreases from 283 yr fire'1 to 125 yr fire'1 (average of 177 yr fire'1).
In Zone 2 there is an increase in CHAR accumulation from 0.1 to 0.3 particles cm"2 yr"1. BCHAR also increases to an average of 0.3 particles cm"2 yr"1, with a minimum of 0.2 particles cm"2 yr"1 and a maximum of 0.3 particles cm"2 yr"1. A total of three fire events were identified in this zone. The maximum peak magnitude in Zone 2 was 3.2 particles cm"2peak"1, with an average of 0.1 particles cm"2 peak"1. The average FRI increases between Zone 1 and Zone 2 from 177 yr fire'1 to 205 yr fire'1.
Zone 3 is defined an increase in peak magnitude and a decrease in the FRI. CHAR had a maximum value of 1.6 particles cm"2 yr"1 and minimum of 0.1 particles cm"2 yr"1, with an average of 0.4 particles cm"2 yr"1. Average BCHAR remained the same as in Zone 2 (0.3 particles cm"2 yr" J) with a maximum value of 0.4 particles cm"2 yr"1 and a minimum of 0.3 particles cm"2 yr"1.
Seven fire events were detected in Zone 3. There was a maximum peak magnitude of 15.4 particles cm"2 peak"1, however the average was only 0.7 particles cm"2 peak"1. Average FRI decreased in Zone 3 to 189 yr fire'1. However, Zone 3 includes both the MCA and LIA with increased fire activity (FRI 119 yr fire'1) and decreased fire activity (FRI 296 yr fire'1), respectively.
49


CHAR and peak magnitude increase in Zone 4. Averages were 1.2 particles cm"2 yr'1 and 15.9 particles cm"2peak"1, respectively. CHAR and peak magnitude also experienced the highest values of the record in Zone 3. The highest CHAR value was 4.3 particles cm'2 yr"1 and the highest peak magnitude was 230.4 particles cm"2 peak"1. BCHAR remains consistent (0.3 particles cm"2 yr"1), with a maximum value of 0.3 particles cm"2 yr"1 and a minimum value of 0.2 particles cm"2 yr"1. Two fire events were detected in Zone 4 around 87 cal yr BP and -49 cal yr BP. These two events have the highest peak magnitudes of the record. The average FRI increases during Zone 4 (from 189 yr fire'1 to 303 yr fire'1).
.
CHAR
(pieces cm^yr1)
i.
+
+
+
+
+
+
\
Mine
LIA
MCA
Winter
wet
Summer
wet
Zone 4
Zone 3
Zone 2
Zone 1
0 100 200
Peak Magnitude
(pieces cm 2peak-1)
500
FRI (yr fire1) 1000-yr mean 95% Cl
Figure 10. Charcoal graphs for Lily Pond showing charcoal accumulation, peak magnitude, and fire return intervals. Gray line in CHAR shows BCHAR level. Red crosses designate peak fire events. Horizontal gray bars show missing data. The switch from summer-wet to winter-wet precipitation is shown with a dashed blue line; the MCA is marked with red shading; the LIA is marked with blue shading; and the mining period is marked with gray shading.
50


Table 6. Summary of charcoal record organized by zones. 'SUM PEAK MAG is the peakMAG output from CharAnalysis. It is the sum of all samples exceeding the BCHAR threshold for a
given peak.
ZONE CHAR (particles cirUyr1) BCHAR (particles cirUyr'1) SUM PEAK MAG1 (particles cnr2peak-1)
4 Avg 1.2 0.3 15.9
High 4.3 0.3 230.4
Low 0.2 0.2 0.0
3 Avg 0.4 0.3 0.7
High 1.6 0.4 15.4
Low 0.1 0.3 0.0
2 Avg 0.3 0.3 0.2
High 0.6 0.3 3.2
Low 0.1 0.2 0.0
1 Avg 0.1 0.1 0.1
High 0.5 0.2 8.1
Low 0.0 0.0 0.0
51


CHAPTER VI
DISCUSSION
The following section outlines the late Holocene climate history of the Colorado Rocky Mountains and the surrounding region. The impacts of climate variations on the terrestrial and aquatic communities are then discussed. Finally, the introduction of Forest Hill Mine and its direct and indirect effects on the Lily Pond ecosystem are presented. Other paleoecological and paleoclimatic records exist for subalpine environments throughout the Rocky Mountains and provide an opportunity to compare the record at Lily Pondto separate local from regional responses. Records from central Colorado (Mirror Lake, Colorado; Keystone Ironbog, Colorado) (Del Priore, 2015; Fall, 1997a), northern Colorado (Bison Lake, CO; Hidden Lake, CO; Tiago Lake, CO) (Anderson et al., 2015; Jimenez-Moreno & Anderson, 2012; Shuman et al., 2009), southern Colorado (Little Molas Lake, CO; Toney & Anderson, 2006), New Mexico (Sangre de Cristo Mountains; Jimenez- Moreno et al, 2008), and southern Wyoming (Little Brooklyn Lake, WY; East Glacier Lake, WY) (Brunelle et al, 2013; Mensing et al, 2012) are used in the regional comparison(Appendix E).
Climate of the late Holocene (5000 cal yr BP to present)
The climate of North America during the late Holocene is characterized by increased variability in temperature and precipitation due to decreasing summer insolation and sea surface temperatures (SSTs), which altered the strength and position of the North American mid-latitude jet stream (Mann et al., 2008; Kitzberger et al., 2007; Anderson, 2012; Jimenez-Moreno & Anderson, 2012; Shuman et al., 2018; Shuman et al, 2014). The changes in insolation also resulted in intensified decadal and inter-annual variability (e.g., PDO, ENSO), cooler
52


temperatures than during the early Holocene, and increased effective moisture (Anderson, 2012; Briles et al, 2012).
Multiple proxies have been used to reconstruct temperature and precipitation regimes in North America through the Holocene. For example, lake levels from Emerald Lake (Colorado), Lake of the Woods (northwestern Wyoming), and Little Windy Hill Pond (southeastern Wyoming) were compared through the late Holocene using sediment stratigraphic changes across the lake (Shuman et al, 2014). Results show that periods of high-lake levels found at Emerald Lake correspond with periods of low-lake levels at Lake of the Woods. There were no statistically significant correlations between Little Windy Hill Pond, although the record did differ, especially during the early Holocene where Emerald Lake did not show near-modem lake levels that were found at the other two locations (Shuman et al, 2014). Thus, is it thought that a north-south moisture dipole exists in the Rocky Mountains that impacts moisture delivery to these areas in the past. Spatial variations in El-Nino Southern Oscillation (ENSO) are thought to be responsible for creating a north-south moisture dipole along the CO-WY border.
Further, a shift in precipitation regime around -2400 cal yr BP has been identified in central Colorado through several multiproxy analysis. Calcite -5180 analyses from Bison Lake (northwestern Colorado) indicate a shift to a snow-dominated precipitation regime in the region starting around -2400 cal yr BP (Anderson, 2012). Similarly, Hidden Lake near Steamboat Springs, Colorado, shows an increase in lake level beginning that peaked at -2020 cal yr BP , which is consistent in timing with the switch to a winter-dominated moisture regime. This climate shift is further supported by vegetation and fire records in the region, with a decrease in subalpine forest density around -2400 cal yr BP paired with a decrease in charcoal production (P. E. Higuera et al., 2014). The switch to a winter-dominated precipitation regime resulted in
53


more snow accumulation in the winter months which melted during the spring and summer. The decrease in summertime moisture resulted decreased effective moisture for plant growth in the summer and caused a structural shift in forests and a change in fire regime.
During the latter part of the late Holocene (past -1300 years), central Colorado experienced two climate events known as the Medieval Climate Anomaly (MCA; 1200 cal yr BP - 850 cal yr BP) and the Little Ice Age (LIA; 650 cal yr BP - 100 cal yr BP). The MCA is defined as a period of warming with drought-like conditions, where temperatures increased by ~0.5°C globally compared to the previous centuries. The increase in temperatures also resulted in higher frequency and severity of forest fires throughout the MCA (Calder et al, 2015). The LIA is characterized by cooler temperatures (-0.26 °C; Mann et al, 2009) than during the MCA and a return to wetter conditions. The climate differences between these two periods are thought to be the result of natural radiative forcing (i.e. solar variability and heightened volcanism) that caused notable changes in the temperatures throughout the Northern Hemisphere (Mann et al., 2009). A strong enriched-north/depleted-south precipitation-d180 dipole pattern existed in the North American Rocky Mountains during both the LIA and MCA, suggesting that precipitation behavior was very complex (Anderson, 2011; Anderson et al, 2016). The complexity of past climate changes and the limited information on the impacts on terrestrial and aquatic mountain ecosystems make high-resolution late-Holocene paleoecological records important. These records help inform ecosystem managers about how Rocky Mountain forests might respond to the rapid shift in climate we are seeing today and are predicted to experience into the next century.
In summary, the climate during the late Holocene was influenced by shifts in insolation and solar radiation that caused variation in precipitation and temperatures on various temporal
54


and spatial scales. Specifically, Colorado’s climate shifted to winter-dominated precipitation pattern -2400 cal yr BP with generally cooler temperatures than during the early Holocene, with a brief return to warmer drier conditions during the MCA and cooler and wetter conditions than today during the LIA. Paleoclimate and lake-level studies indicate variability across Colorado to Wyoming, suggesting that the southern Rockies will not respond similarly to future climate fluctuations.
Ecosystem Response to Climate during the late Holocene
Switch from summer-dominant to winter-dominant precipitation regime (-2400 cal yr BP)
Both the aquatic and forest ecosystem at Lily Pond responded to the climate regime shift at -2400 cal yr BP. The aquatic ecosystem is composed primarily of benthic genera during the last 5000 years suggesting that there were high amounts of light availability in the littoral zone, and likely shallow lake level during the late Holocene. There are consistently low levels of Encyonema, which are an unattached, benthic species. There are also low amounts of Nuphar pollen during this time, suggesting a lack of vegetative habitat for these species with lake water levels that did not support Nuphar growth (-1-2 m depth). The high abundances of Fragilariaceae, Pinnularia, Stauroneis, Sellaphora, and planktonic species, indicate stable habitat and nutrient availability from -5000 cal yr BP to -2400 cal yr BP. At -3100 cal yr BP, a few planktonic species are present, however the record is still dominated by benthic species. The data suggest a lake level that allowed for high levels of light penetration to the benthic zone, while also supporting some planktonic species. The Mn/Ti ratio is consistently high through -2400 cal yr BP, indicating no major variation in lake level. The Fe/Ti ratio is also high, indicating high levels of allochthonous inputs into Lily Pond. Magnetic susceptibility is paramagnetic, indicating low levels of metal-bearing minerals in lake sediments, although there is a brief period when
55


values are close to 0, indicating a brief increase in iron-bearing minerals into Lily Pond. Organic content is lower than today, likely indicating lower lake productivity.
The highest abundances of Picea and Abies occur prior to 2400 cal yr BP, suggesting open subalpine forest at Lily Pond. This section of the record also has the lowest AP/NAP along with low PAR, which suggests an open forest. The higher NAP species (e.g., those of Artemisia and Amaranthaceae) and Quercus are likely blowing upslope from lower elevation. Fewer trees would allow a greater abundance of these pollen types to reach Lily Pond. Fire frequency is the lowest of the record from -5000 cal yr BP to -3000 cal yr BP likely due to the wetter summer conditions than today. Unlike today, the forest around Lily Pond probably was similar in composition as Mirror Lake today (-147 m in elevation higher than Lily Pond; Del Priore,
2015). Fires were infrequent and those that did burn consumed very little biomass. Wet summer conditions likely supported more mesic conifers Picea and Abies and fewer fires at lower elevations than today.
The Lily Pond diatom records a transition from summer-dominant to winter-dominant precipitation regime around -2400 cal yr BP. Pinnularia, Stauroneis, and Encyonema increase after the shift in precipitation, while Fragilariaceae and Sellaphora decrease, suggesting a change in habitat at Lily Pond. The decrease in Fragilariaceae and Sellaphora around this time suggest decreased planktonic habitat suitability, which could be a result of longer periods of ice cover. The Mn/Ti ratio also increases after -2400 cal yr BP, indicated a possible decrease in lake level. The Fe/Ti ratio decreases after -2400 cal yr BP, indicating an overall reduction in allochthonous sedimentation into Lily Pond, perhaps due to a more closed forest (discussed below). The switch to winter-dominated precipitation regime may have also altered nutrient availability due to increased spring runoff, and lower amounts of summer precipitation. Magnetic susceptibility
56


remains paramagnetic (<0) indicating low levels of iron-bearing minerals present in the lake system. Organic matter increases gradually, indicating an increase in biological productivity. Pinus begins to increase around -3100 cal yr BP, while Picea, Abies, Artemisia, and Amaranthaceae decrease. The AP/NAP ratio increases, mainly due to the increase in Pinus, indicating the forest became more closed. The pollen data indicate a more closed forest that resembles the modern mixed-conifer forest currently at the transition between the montane and subalpine. Around -3000 cal yr BP, fire frequency and biomass burned increased, likely due to the increase in Pinus (more closed forest than before) and the decrease in moisture availability in the summer.
The vegetation shift at Lily Pond around -2400 cal yr BP at Lily Pond is also recorded in other locations in Taylor Park and in the Rocky Mountains. Mirror Lake (3347 m elev; Del Priore, 2015) in Taylor Park and Keystone Ironbog (2920 m elev; Fall 1997) near Crested Butte identified an abrupt change in vegetation from L’/'cea-dominated to /7/w.s-dominated forest composition at -2300 yr BP and -2600 cal yr BP, respectively. Records from southwestern Colorado at Little Molas Lake (3370 m elev; Toney & Anderson, 2005), and from the Sangre de Cristo Mountains in New Mexico (3100 m elev; Jimenez-Moreno etal, 2008) also record a similar abrupt change in vegetation at -2600 cal yr BP and -2800 cal yr BP, respectively. In northern Colorado, Picea began to decline steadily at Bison Lake (3255 m elev.) around -2500 cal yr BP (Anderson et al, 2015). A composite analysis of pollen and charcoal records from three lakes in Rocky Mountain National Park (>3000 m elev; Higuera et al, 2014) document a shift around -2400 cal yr BP to a more open forest structure indicated by increased Pinus abundances, and a decrease in fire activity and severity.
57


The Tiago Lake record (2700 m elev.), also in northern Colorado, records maximum Pinus at -5000 cal yr BP. The differences in time of Pinus establishment at Tiago Lake may be due to the lake’s lower elevation and difference in forest composition (e.g., montane forest). However, it may also indicate that moisture came mostly as snow in the winter rather than rain in the summer months through the entire late Holocene (Jimenez- Moreno et al, 2011).
Interestingly, vegetation records from southern Wyoming above 3000 m elevation also record high abundances of Pinus earlier than central Colorado to northern New Mexico records. For example, Little Brooklyn Lake (3153 m; Brunelle et al, 2013) and East Glacier Lake (3282 m; Mensing et al, 2012) record increases in Pinus that peak around -5000 cal yr BP and -5200 cal yr BP, respectively. The similarity between the Wyoming and Tiago Lake paleoecological records suggest the timing of the moisture transition may have occurred at least -2500 years earlier than at the sites in Rocky Mountain National Park and those in central and southern Colorado. The difference in timing of the shift in Pinus across Colorado and Wyoming, and the correspondence with lake level differences, is likely related to the difference in delivery of moisture related to decadal and interannual climate variations that are currently in operation today (Shuman et al. 2014).
The fire record at Mirror Lake, in Taylor Park just south of Lily Pond, suggests an increase in fire events leading up to -2400 cal yr BP, like Lily Pond, followed by a decrease in composite CHAR suggesting decreased fire activity (Del Priore, 2015). Elsewhere in the Colorado Rocky Mountains, records also identify decreased fire activity beginning between -2600 to -2300 cal yr BP, like trends in the Lily Pond record after -2400 cal yr BP. In Rocky Mountain National Park, composite CHAR decreased after -2400 cal yr BP, suggesting smaller fires with lower amounts of biomass burned until -1500 cal yr BP (Higuera et al, 2014). At Lake
58


Tiago, there was also an increase in fire activity from -3000 to -2500 cal yr BP that then decreases until -600 cal yr BP (Jimenez-Moreno et al, 2011). At Little Brooklyn Lake in Wyoming, fire activity is low from -4900 cal yr BP to -2000 cal yr BP likely due to a regional increase in moisture after -5000 cal yr BP (Brunelle et al, 2013). The variation in the fire record, like the vegetation record, from Colorado to Wyoming suggests that the shift in precipitation regime occurred at different times.
Medieval Climate Anomaly (MCA , -1200-850 cal yr BP) & Little Ice Age (LIA, 650-100 cal yr BP)
The change in summer- to winter- dominated precipitation at -2400 cal yr BP set the stage for the development of modern forests and disturbance regimes. However, century long climate events, such as the MCA and LIA, caused brief changes in the terrestrial and aquatic communities at Lily Pond. The diatom record indicates decreased Fragilariaceae species, and increased Pinnularia and Stauroneis abundances, suggesting a change in habitat availability possibly due to changes in lake level during the MCA. Sellaphora, planktonic species, and Enyconema remained consistent through the MCA. Nuphar rates were consistent through the MCA, suggesting low water levels that altered benthic habitat suitability. The Mn/Ti ratio increased throughout the MCA, indicating decreased lake level. The Fe/Ti ratio decreased during the MCA, indicating decreased in allochthonous inputs into Lily Pond. The change in diatom community composition may have been a result of loss of open-water habitat due to decreased lake level, and a change in nutrient availability due to decreased allochthonous inputs.
Around -1100 cal yr BP, Pinus decreased while lower elevation steppe species (Artemisia, Amaranthaceae) increased. The AP/NAP ratio also decreases, signifying a more open forest likely due to increased burning during the MCA. The FRI is the lowest of the record and
59


fire events are associated with larger charcoal peak magnitudes than seen in the earlier portion of the record. The increase in charcoal peak magnitude during the MCA may suggest more severe fires than were seen earlier in the Lily Pond record; however, higher resolution pollen sampling and chemical analysis is needed to examine variability in fire severity. Fire frequency and severity declines through the MCA, likely a result of a decrease in burnable biomass. Increased fire frequency led to a decrease in Pinus and an increase in steppe species blowing in from downslope. The increase in fire activity during the MCA is likely due to the warmer and drier conditions.
The transition between the MCA and LIA (-850 cal yr BP - -650 cal yr BP) is defined by a decrease in the Mn/Ti ratio, suggesting an increase in lake levels. The Fe/Ti ratio also increases suggesting an increase in allochthonous inputs. Fragilariaceae, Sellaphora, and planktonic species increase at this time, while Pinnularia, Stauroneis, and Encyonema decrease. Nuphar also decreases after the end of the MCA, suggesting increased lake level that limited its growth. The increase in certain diatom species is likely in response to an increase in lake level that resulted in a more suitable open-water habitat, while epiphytic and epipelic habitat suitability decreased. Pinus and Picea also increased at the end of MCA in response to increased moisture and the decrease in fire activity.
At the beginning of the LIA the percent organics increased, suggesting increased lake productivity. It could also indicate a decrease in allochthonous inorganic materials entering Lily Pond since the Fe/Ti ratio also decreases. The diatom record indicates an increase in Pinnularia, Stauroneis, and Enyconema at the start of the LIA, while Fragiliaraceae, Sellaphora, and planktonic species decrease. At -200 cal yr BP, there is an abrupt decrease in Pinnularia, Stauroneis, and Enyconema, and an increase in Fragilariaceae, Sellaphora, and planktonic
60


species. This shift indicates a potential change in epiphytic and epipelic habitat suitability that may have been related to a change in nutrient inputs into the system. Both Pinus and Picea increased throughout the LIA while steppe species decreased. Fire activity during the LIA was comparable to that prior to -2400 cal yr BP, with increased FRI and decreased charcoal peak magnitude. The AP/NAP ratio also increased, signifying a shift back to a more closed forest. There is one fire event at -560 cal yr BP that had a relatively high peak magnitude in comparison to the rest of the record. The high peak magnitude may have been a result of a buildup of forest biomass due to more effective moisture during the LIA.
The diatom and geochemistry record at Lily Pond indicate that there were changes in nutrient and aquatic habitat availability during the MCA, specifically decreased allocthonous inputs and a lower lake level, indicated by changes in sediment geochemistry. The changes in the aquatic habitat conditions are likely due to increased summer moisture, decreased ice cover, and warmer temperatures. Specifically, Fragiliariaceae species decreased through the MCA. Studies have shown that changes in ice cover and increased nutrient cycling due to warming temperatures have caused small, fragilariod species to decrease in alpine lakes in Canada (Ruhland et al, 2008), suggesting that the increased temperatures at Lily Pond during the MCA may have had a similar effect. Hidden Lake in northern Colorado (2710 m) also shows a decreased lake level during the MCA (Shuman et al, 2009), indicating that the water level at Lily Pond may have also decreased during this time. However, other lakes in the region (e.g. Little Molas Lake, Southern Colorado, 3330 m) did not record a similar lake level decrease during this time (Shuman et al, 2009). The Lily Pond diatom and geochemistry records indicate a shift back to conditions seen prior to the MCA at the start of the LIA, with a decreased Mn/Ti ratio, decreased Fragiliariaceae, and increased Pinnularia.
61


The fire record at Lily Pond during the MCA and LIA is similar to other records throughout the Colorado Rocky Mountains, with an increase in burning at the start of the MCA followed by a decline through the LIA. For example, a meta-analysis of fire histories from western United States lakes indicates increased burning at the beginning of the MCA that then decreased (Marlon et al, 2012). The records from Rocky Mountain National Park identified increases in fire beginning around -1500 cal yr BP (Higuera et al, 2014). Other records in northern Colorado show similar increases in fire activity. At Bison Lake there was increased burning at the start of MCA (-1200 cal yr BP), and an abrupt decline during the transition to the LIA at -650 cal yr BP (Anderson et al, 2015). At Lake Tiago, fire activity increased around -1400 cal yr BP, and then decreased around -600 cal yr BP (Jimenez-Moreno et al, 2001). The Lake Tiago record responded earlier than the Bison Lake record to the onset of the MCA, and then responded later to the onset of the LIA, suggesting spatial variations in fire response to periods of warming and cooling.
Overall trends show an increase in fire activity in the western United States after -2000 cal yr BP and into the start of the MCA (Marlon et al, 2012). In Wyoming, the fire record at Little Brooklyn Lake and Little Windy Hill show increases in charcoal beginning around -2000 cal yr BP (Brunelle et al., 2013; Minckley et al., 2012). The increase in burning in the western United States are likely a result of increased drought-like conditions and a high amount of biomass available to bum. The records from Wyoming indicate an increase in burning earlier (-500 years) than the Colorado records. The records from Colorado also show slight variations in timing. These temporal variations in timing are likely due to a combination of factors, including temperature, moisture availability, and forest composition, found at different elevations throughout the Wyoming and Colorado Rocky Mountains.
62


Human Impacts: Logging, Mining, Grazing and Recreation
In Colorado, the first substantial gold discovery was made at -100 cal yr BP (-1850 CE), which sparked the start of mining activity in the state that continues in some areas today. This also corresponds with an increase in population in the western United States that is tied to the Colorado Gold Rush (Encyclopedia Staff, 2016). Agriculture and ranching also became popular around this time as a way to sustain economic growth (BLM, 2008). The population increase in the western United States also resulted in changes in land use that impacted the surrounding ecosystems through changes in fire regime and forest structure (Marlon et al, 2006). The Forest Hill Mine (FHM) began operations near Lily Pond -70 cal yr BP. The mine was closed by -30 cal yr BP because it was not very successful and has had no remediation. There are remnants of old pipes, boilers, and other mining infrastructure scattered around the mine area (Figure 11). The mining footprint extends beyond the Forest Hill Mine area, with additional cabins, mine shafts, towers, and other infrastructure. The land around the main Forest Hill Mine area is still owned privately. Other mine remnants are owned by the United States Forest Service (USFS). A dendrochronology study in the watershed indicates that logging operations may have been impacting the area up to -100 years prior to the start of FHM operations (Appendix A).
63


b) Remains of building at FHM d) Main FHM f) Cabin at Lily Pond
Figure 11. Images of mining remains around Lily Pond. Photos: J. Steelman
Ecosystem Response to Human Impacts Logging & Start of Forest Hill Mine (-70-30 cal yr BP; 1880-1920 CE)
The diatom, geochemistry, pollen, and charcoal record from Lily Pond all suggest whole scale ecosystem responses since -180 cal yr BP (1770 CE), prior to the start of FHM operations (-70 cal yr BP; 1880 CE). The Lily Pond record is compared with paleoclimate and human population data for the Northern Hemisphere in Figure 12. Since these shifts predate the introduction of the mine, it suggests that other anthropogenic activity, likely logging, was taking place around Lily Pond prior to the mining. The Gunnison National Forest was not added to the Grand Mesa Uncompahgre and Gunnison (GMUG) National Forest until -45 cal yr BP (1905 CE), so there are limited records available regarding logging activity before the mining. Lily
64


Pond records a shift in the diatom, sediment geochemistry, and pollen records around -180 cal yr BP likely due to logging. A preliminary dendrochronology analysis identified the earliest widespread timber stand release date around 1777 CE, supporting the theory that anthropogenic activities began at Lily Pond prior to the FHM operations (Steelman, unpublished).
Timber harvest at Lily Pond prior to mining activity is indicated by increased magnetic susceptibility, increased Fe/Ti ratios, decreased Mn/Ti ratio, and decreased organic content, suggesting increased allochthonous inputs, increased lake level, and decreased lake productivity, respectively. Fragilariaceae, Sellaphora, and planktonic species increase to high abundances at -200 cal yr BP, while Pinnularia, Stauroneis and Encyonema decrease. An increase in lake level may have impacted light penetration to the benthic areas, resulting in decreased abundances of benthic diatoms. There is also a decrease in Nuphar abundance, indicating a decrease in aquatic vegetation. Planktonic species, Fragilariaceae, and Sellaphora begin to decrease around -150 cal yr BP, prior to the start of the mine, while Pinnularia and Encyonema slowly increase. The change at -150 cal yr BP corresponds with an increase in Nuphar and an increase in the Mn/Ti ratio, suggesting decreased lake levels and increased amounts of aquatic vegetation. This shift overlaps with the end of the LIA, so the changes seen at -150 cal yr BP may be a combination of human impacts and increased summer drought conditions (Anderson et al, 2015).
65


(cal yr BP)
300-
1100-
1300-
1500-
1700-
1900-
Allochthonous Inputs
Decrease *-------â–º Increase
Lake Level
Increase *------â–º Decrease
i â–  i ........................
“i—i—i—I—i—i
I—i-1—i—I—i--1
°o3 '°oy°Oj°o °0j 0.> Og o9 1C
Magnetic Susceptibility
(cgs)
Mn/Ti & Fe/Ti (ratio)
i « i ■ i ■ » ■ i ■ i i ■ » ■ i ■ i ■ i
l—i—i—i—i—i—i—i—i—i—i i—i—i—i—i—i—i—'—i
0 «?0 4o Sq Bq Jqq 0 *0 60 $0
Pinnularia & Picea & Pinus
Fragilariaceae (%)
(%)
Forest Structure
Open*-----------â–º Closed
» ■ i ■ i ■ i ■ i « i i ■ i ■ » ■ i ■ i ■ i
AP/NAP CHAR & Fire Events
(ratio) (pieces cm-2 yr1)
Winter Insolation
(%)
'■?-0 ~l-6 °S '°-4 °0
0 gO ’60 * Population of North Summer Insolation America (%)
(in millions)
Figure 12. Lily Pond environmental history, population growth in North America (HYDE 3.1; Goldewijk et al, 2010), and
insolation changes in North America at 40° N (Berger & Loutre, 1991).
On
On


The AP/NAP ratio also decreases around -200 cal yr BP, along with all pollen species abundances. This decrease signifies an overall reduction in vegetation around Lily Pond. This may have been the result of timber harvest activity in relation to the FMH establishment. There is also a fire event at -87 cal yr BP (1863 CE). The peak magnitude of this event is larger than the previous fire events identified in the record, suggesting a severe fire occurred at Lily Pond and may be the result of land clearance. In order to facilitate mine construction and operations, vegetation was likely harvested (arboreal species) or burned (nonarboreal species) to clear the surrounding landscape. There are remnants of log cabins and other wooden structures around the mine, likely constructed using the Picea and Pinus harvested in the area. There is a brief period of increased Picea, Abies, and Quercus leading up to the start of FHM operations. The AP/NAP ratio also increased briefly, suggesting a possible short-term reestablishment of forest between logging and the start of mining activity.
The introduction of FHM (-70 cal yr BP; -1880 CE) saw an increase in magnetic susceptibility, where values become consistently diagenetic for the first time in the last 5000 years, suggesting increased in iron-bearing allochthonous inputs into Lily Pond. Since mining activity disturbs sediments through digging, increased metal content is likely due to runoff of loose sediments around Lily Pond. All diatom abundances increase slightly through the mining period. All diatom taxa present in the Lily Pond record are considered circumneutral, suggesting no sudden changes in pH with the introduction of the mine. This is further supported by the taxa similarities between the modern-day diatom community, and those seen through the sediment core. Based on the diatom community structure during the mining period, FMH mining operations likely did not result in AMD that directly impacted Lily Pond. However, increased metals were detected through magnetic susceptibility and XRF analyses, specifically increased
67


Fe levels. The changes in diatom abundances are likely the result of disturbances in the terrestrial environment surrounding Lily Pond and nutrient availability due to increased sedimentation and allochthonous inputs.
There is a decrease in Picea, Abies, Quercus, and steppe species at the start mining, indicating increased land clearing. Pinus abundance stays consistent through this period, however it is lower than prior to -180 cal yr BP. The AP/NAP ratio also decreases indicating a more open forest. CHAR decreases after the fire event -87 cal yr BP (1863 CE) and remains low through the time of FHM operations. The decrease in CHAR is likely due to absence of biofuels and anthropogenic fire suppression efforts since people were living in the area. The additional land clearing likely contributed to iron-bearing minerals entering Lily Pond at the beginning of the mine. A reduction in surrounding vegetation, combined with sediment disturbance associated with mining activity, would result in increased amounts of sediment runoff during precipitation events, increasing allochthonous inputs into the aquatic system.
The anthropogenic changes at Lily Pond have been documented elsewhere. A study in the Uinta Mountains of Utah found that an increase in anthropogenic activity in the region impacted alpine lakes through atmospheric deposition, where even small amounts of metal and nutrient enrichment in lake systems were enough to affect the diatom communities (Moser et al., 2010). Similar results were seen in other alpine lakes in the Colorado Front Range, where increased nitrogen deposition in alpine areas was a result of increase agricultural practices in surrounding areas (Wolfe et al, 2001). Thus, Lily Pond may have also been impacted by other anthropogenic activities, including surrounding mining and agriculture, prior to the establishment of Forest Hill Mine. Other fire records from the western United States indicate an increase in fire activity during the mid to late 1800s AD (-150 cal yr BP), with highest levels between 1850-1870 AD
68


(-100-80 cal yr BP) (Marlon et al., 2012). The decrease in CHAR in the Lily Pond record is likely a result of anthropogenic land management techniques given the high levels of anthropogenic activity in the area (Marlon et al., 2012).
Termination of Forest Hill Mine (>~30 cal yr BP; 1920 CE) & Present-Day Activities
The FHM operations ended at -30 cal yr BP (1920 CE). There was a brief break in mining activity from -43 to -34 cal yr BP (-9 years; 1907-1916 CE), however the break was not detected by any of the proxy data at Lily Pond. Since closure of the FHM, other anthropogenic activities such as recreation, wildfire suppression, and grazing, have increased around Lily Pond. There is a system of 4x4 and ATV trails that run around Lily Pond, the Forest Hill Mine, and the surrounding area, that are very popular during the summer months. Additionally, the USFS allows horse and cattle grazing in the forest around Lily Pond (United States Forest Service, 2018).
Since mine closure, sediment geochemistry indicates a continued input of iron-bearing minerals into Lily Pond than was seen earlier in the record. The Mn/Ti ratio decreases, suggesting increased lake levels. The diatom record shows a decrease in Fragiliariaceae, Sellaphora, and planktonic species, while I'innularia, Stauroneis, and Encyonema increase. These changes indicate a possible period of increased nutrient availability immediately following the closure of Forest Hill Mine. This may be due to increased runoff, and therefore nutrients entering the system due to limited vegetation. It could also be indicative of a change in habitat availability due to an increase in lake level. During the last -40 years, Pinnularia, Sellaphora and planktonic species increased, while Fragilariaceae species decreased, suggesting a possible increase in nitrogen at Lily Pond. While some species of Fragiliariaceae are tolerant of increased
69


nitrogen levels, smaller species like the ones found at Lily Pond, have been found to decrease as nitrogen levels increase (Saros et al, 2003).
The pollen record shows a decrease in the AP/NAP ratio immediately after mine closure, indicating a more open forest. Quercus, Artemisia, and Amaranthaceae increased immediately after the termination of mine activity, also indicating an open forest allowing for the deposition of pollen taxa from downslope, while Picea increased slowly. The diatom and pollen record show that the species present in both the terrestrial and aquatic communities did not change with the introduction of the mine, and that these species continue to be present at Lily Pond after mine termination. While relative abundances did shift during the mining period, the shifts were not enough to cause either system to shift in ways that were comparable to those of the MCA or LIA. The presence of mine remnants, including mine tailings, do not seem to be continuing to alter the terrestrial and aquatic environments at Lily Pond.
There is a decrease in Pinus, which may be a result of a fire event identified at ~2 cal yr BP (-1948 CE). The fire event is associated with the largest charcoal peak in the Lily Pond record, indicating a substantial fire in comparison to those recorded during the past millennia. The fire recorded in the lake sediments is likely the one that left bum scares on trees on the southwest side of the watershed. Visual evidence of a recent fire was recorded in this area (Steelman, unpublished). Additional tree coring my reveal the exact timing of this event beyond that observed in the lake sediment record. The modem forest composition is similar to that recorded prior to the mine and logging with increased amounts of Pinus and increased AP/NAP ratios within the last 40 years, suggesting a more open forest that is dominated by lodgepole pine.
70


Since 1950, other diatom records from alpine lakes in the Colorado Front Range have identified shifts in diatom communities due to increases in anthropogenic activity. Specifically, a study on mountain lakes in the Colorado Front Range found that an increase in agriculture within the region resulted in increased nitrogen levels in the adjacent mountain areas. The increased nitrogen levels resulted in increased abundances of opportunistic diatom species (Wolfe et al., 2001). Given the sensitive nature of diatoms, even small increases in nutrient inputs into lake systems, especially through atmospheric deposition, is enough to alter diatom communities. However there are several other factors that are believed to impact community response to nitrogen. A study in northwestern Wyoming suggests that benthic species in shallow lakes are more likely to be impacted by light availability than nutrients, so the effects of increased nitrogen levels vary between systems (Spaulding et al, 2015). Environmental sensitivity of diatoms is species specific, which is why some species respond more strongly than others to changes in their environments. Since this study only identified the community to the genus level, species sensitivity cannot be determined. Fluctuations in genera abundances throughout the last 200 years suggest that the diatom species found in this environment less sensitive to nutrient changes and are representative of changes in habitat suitability.
Over the last century, there has been a decrease in biomass burned as well as fire frequency due to suppression efforts throughout the western United States (Marlon et al., 2012). The decrease in burning is counter to what is expected given the warmer and drier conditions than seen during the LIA. It is likely that anthropogenic factors such as grazing and land management, has limited forest biomass needed to carry and sustain fires. In areas where land management has not been as extensive, fire frequency has increased through the 20th century (Marlon et al., 2008).
71


Land Management Implications
Management of lands in the western United States is a complex issue that must take both climate and anthropogenic activities into consideration. The research presented in this thesis can provide insights about how terrestrial and aquatic environments in a subalpine forest in central Colorado respond to climate conditions as well as anthropogenic activity. The 21st century has experienced warmer surface, atmospheric, and ocean temperatures (~0.85°C; IPCC, 2014), especially over the last 30 years, with a substantial contribution coming from anthropogenic activity. Colorado annual average temperatures have increased by 1°C in the last 50 years (Lukas et al, 2014).
Models predict continued warming globally throughout the remainder of the 21st century, with increases of over 1°C expected (Alexander et al., 2018; IPCC, 2014). In Colorado, future warming is projected to increase by up to 3°C by 2050, with summer temperatures projected to increase more than winter temperatures (Lukas et al, 2014). Precipitation trends for Colorado indicate increased winter extreme precipitation events, but do not necessarily show the same change for summer precipitation (Lukas et al, 2014). Climate projections also predict increased frequency and severity of heat waves, droughts, and wildfires in Colorado by the mid-21st century due to the warming climate (Lukas et al, 2014). There is still uncertainty as to how forests will respond to the increase in temperature, making it difficult for forest managers to plan appropriately. However, the paleoecological and paleolimnological records provide insights.
The recent record at Lily Pond (past -200 years) indicates conditions similar to those seen during the MCA. The similarities suggest that climate drivers, specifically changes in timing of precipitation and drought severity, impact the terrestrial and aquatic communities at Lily Pond more than anthropogenic activities. There has been a decrease in precipitation and
72


increase in temperatures in the western United States (including Colorado) beginning in the late 20th century, resulting in increased drought conditions during the summer months (Knowles et al, 2006; Mote et al, 2005). Continued warming in mountain catchments, both in summer and winter, will result in overall decrease in annual runoff, which could impact temperature, lake level, and nutrient availability (Dierauer, Whitfield, & Allen, 2018). If warming continues to increase through the 21st century as predicted, it is likely to have greater impact on the terrestrial and aquatic communities than the MCA.
Given the current warming trends of the 21st century, forest composition and structure will likely continue to forests and fire regimes seen in the MCA with a more closed Pinus dominated forest with higher fire activity due to periodic droughts due to increased temperatures during summer months (Calder et al., 2015; Marlon et al., 2012). However, temperatures are expected to exceed those of the MCA. If the temperature increase result in a shift to more summer precipitation, conditions like those prior to those seen prior to 2400 cal yr BP may prevail. However, current future projections do not indicate increased summer precipitation. Therefore, tree line could shift to a higher elevation under warmer drier summer conditions. Additionally, anthropogenic fire suppression efforts have increased forest biomass in some forests at the local level, which will likely result in more severe bums when they do occur. Temperature and precipitation are also likely to spatially vary from Colorado to Wyoming and forest managers need to keep in mind that forest management practices needs be flexible to account for the climate gradient.
The terrestrial and aquatic systems at Lily Pond are under pressure from combined effects of climate change and anthropogenic changes that have altered the function of the system throughout the late Holocene. Beyond managing for climate variations in the region, Colorado
73


has thousands of AMLs scattered though the state. Abandoned mines can continue to impact the surrounding systems even after activity has stopped, specifically due to high amounts of acid mine drainage leaving the mine (Brugam & Lusk, 1986). Since runoff from Forest Hill Mine does not drain directly into Lily Pond, is it unlikely that remediation of that specific site would have an impact on the aquatic conditions. However, it could be impacting water quality in a different area downstream.
Further research needs to be done to assess that possibility and make recommendations specific to Forest Hill Mine remediation. The 2007 forest management plan states that AMLs in the GMUG National Forest areas will not be remediated unless they are significantly degrading water quality or are a safety hazard to the public (United States Forest Service, 2007). Based on the current management plan, combined with the results from this research, the Forest Hill Mine is likely not a priority for remediation with regard impacts to the Lily Pond watershed. While the diatom community at Lily Pond has maintained the same genera composition, abundances have fluctuated throughout the record, especially in response to terrestrial disturbances. Therefore, it is important for land managers to continue to monitor aquatic conditions, to determine how future logging, grazing and recreational pressures will impact the Lily Pond conditions and habitat.
Limitations
One limitation for this study is the unconsolidated nature of sediments in the top ~20 cm of the core, and the amount of microscope time required to produce a high-resolution multiproxy record. The limited number of samples for the mining period make it difficult to assess how the brief break in mining activity impacted the aquatic and terrestrial systems, so the amount of time for the system to recover could not be assessed. It would also be beneficial to add additional diatom and pollen samples in the lower portion of the core (between 70-115 cm) to
74


better characterize the summer-dominated precipitation regimes impact on terrestrial and aquatic ecosystems.
Another limitation of this study was the inability to use an electron microscope to identify diatoms down to species level. While genus was sufficient to gain insight about the communities throughout the Lily Pond record, it would have been helpful to be able to distinguish between the different species, specifically the small araphid Fragiliariaceae species, present in the sediment record. It is possible to identify species with light microscopy; however the majority of diatoms present in each sample were very small (<10 microns) making it difficult to identify down to species level given the equipment available.
One final limitation of this study is the lack of information about activity related to the Forest Hill Mine. There are very limited records regarding the history about mine establishment and construction, as well as possible timber harvest activity related to it. This makes it difficult to understand the extent of disturbance as it relates to the paleoecological record. Since Forest Hill Mine is not located within the same watershed as Lily Pond, the direct impacts from mining activity are likely a result of timber harvesting that may or may not have been related to the mine. For example, other natural disturbances, specifically wind, can blow down large spans of forest. The dendrochronological study has helped with an understanding of human vs natural disturbances in the Lily Pond watershed; however, further research is needed in the area.
Future Research
A goal of this thesis was to determine how aquatic systems responded to and recovered from anthropogenic disturbance events and climate change. The Lily Pond record has proved to be a valuable location to reconstruct paleoenvironmental conditions. All three cores record have shown to have similar stratigraphic tends and radiocarbon dates across cores correspond. The site
75


is well-dated through the last glacial period; however, there is a gap in dating between 4000-6000 years. An additional date would strengthen the chronology. Higher resolution pollen and diatom data is also needed between 2000-5000 cal yr BP to characterize the summer-dominated precipitation regime.
Additional proxies could be added to the Lily Pond record to better understand the local changes that were happening. Lake level analysis would be helpful in further evaluating the impacts of climate variations on Lily Pond. There is visual evidence of different lake shores around Lily Pond, especially along the south part of the catchment area. Lake level could be reconstructed by taking a series of cores through the area and examining core lithology and sediment geochemistry. To further assess the impacts of anthropogenic impacts around Lily Pond, fecal stanol and dung spore analyses would be beneficial. Fecal stanol analysis would help identify when anthropogenic activity started impacting Lily Pond, and dung spores would help determine when grazing activity may have begun in the area and its extent. Additional dendrochronology sampling could provide a more complete understanding of the age of the current forest and would provide insight on fire events identified in the record after the termination of the FHM. Lastly, a hydrological model would be useful to understand how runoff around the mine influences the Lily Pond.
To better understand how anthropogenic disturbances may have influenced the terrestrial and aquatic systems at Lily Pond, additional information regarding the start of mining and agriculture in the region is needed. The introduction of anthropogenic activities, such as mining, have been shown to impact subalpine and alpine systems indirectly through atmospheric deposition (Moser et al, 2010; Saros et al, 2003; Wolfe et al, 2001). Therefore, it would be beneficial to research nitrogen and phosphorus levels throughout the Lily Pond core. It would
76


also be helpful to map wind patterns around Lily Pond to better understand how atmospheric deposition may have impacted the area during the late Holocene.
Additional studies that examine multiple proxies throughout the Rocky Mountains would be beneficial in understanding variations along the north-south precipitation gradient. Given the sensitivity of forests and fire regimes historically, additional paleoecological records between 38-43° N latitude across the western United States could help ecosystem models better predict future climate change responses. Lily Pond currently lies at an ecotone between the montane and subalpine forest ecosystems and provides a historical perspective on the rate and trajectory ecosystems respond to abrupt climate change events. Additional sites in the Taylor Park and Southern Rocky Mountains on these gradients will help better estimate forest and aquatic response to present-day and future rapid climate changes.
77


CHAPTER VII
CONCLUSION
In conclusion, the Lily Pond paleo-record provides insights into how subalpine and montane forests in central Colorado are impacted by climate variations and anthropogenic activity (i.e. mining, logging, recreation) during the late Holocene. This study used multiple proxies including diatoms, sediment geochemistry, pollen, and charcoal to reconstruct both terrestrial and aquatic conditions during the last 5000 years, with specific focus on how a period of known mining activity nearby influenced the ecosystem. The research focused on the following two questions:
(1) How did the terrestrial and aquatic environments at Lily Pond respond to climate variations during the late Holocene?
The aquatic and terrestrial systems at Lily Pond were influenced by regional climate shifts that impacted forest and diatom community composition. Precipitation patterns shifted around -2400 cal yr BP from summer- to winter-dominated precipitation. Temperatures also decreased, resulting in increased ice cover, increased spring runoff, and shorter growing seasons based on the increase in Pinus in the pollen record. The forest initially was asubalpine forest dominated by Picea and Abies during a period when precipitation fell mainly in the summer. However, forest structure responded to the change to a winter-wet precipitation regime at -2400 cal yr BP, which was a result of decreasing summer insolation seen throughout the late Holocene. This switch resulted in an overall decrease in net moisture availability and resulted in a switch from a more open canopy structure to a more closed forest canopy dominated by Pinus. Fire activity at Lily Pond also intensified at -2400 cal yr BP with the switch in precipitation regime. The increase in
78


fire activity was due to decreased net moisture and increased drought-like conditions during the summer months.
Temperatures increased, and precipitation decreased around -1200 cal yr BP with the onset of the MCA that lasted until -850 cal yr BP. Conditions returned to high amounts of winter precipitation and cooler temperatures during the LIA from -650 to -100 cal yr BP. The forest around Lily Pond continued to be an open canopy and was dominated by Pinus through the MCA. It switched back to a more closed forest canopy when moisture availability increased again during the LIA. Fire frequency increased at the start of the MCA due to increased temperature and drought conditions, before decreasing in the LIA due to cooler temperatures.
The Lily Pond diatom community remained primarily benthic through the record. Planktonic species were present beginning around -3100 cal yr BP, however they remained at small abundances (<6%) throughout the record. Species abundances fluctuated in response to changes in lake level (Mn/Ti ratio) and allochthonous inputs (Fe/Ti ratio), indicating that habitat (influenced by lake levels) and nutrient availability were primary drivers in determining the diatom community composition. Organic matter increased through the core until -200 cal yr, indicating that anthropogenic disturbance may have had a larger impact on system productivity than climate change. Magnetic susceptibility remained paramagnetic (<0) until -100 cal yr BP and magnetic thereafter, suggesting iron-bearing inputs of allochthonous materials likely from mining activity (e.g., logging and mine tailings).
(2) How did the introduction and termination of mining activity affect the aquatic and terrestrial environments?
The Lily Pond record indicates that anthropogenic activity began prior to the actual start of mining and impacted both the terrestrial and aquatic communities. The dendrochronology
79


record at Lily Pond indicated that the surrounding vegetation was cleared, resulting in increased runoff into the aquatic environment that is reflected in the sediment geochemistry and diatom records. It is not clear if all timber harvest during this period was directly related to the establishment of FHM, but the remnant living structures and boiler indicate that some harvest may have been used to build mining infrastructure, as well as for burning. There was also an increase in CHAR immediately before the start of FHM operations, which potentially indicates the use of fire to clear the landscape. This is followed by a decrease in CHAR through the mining period, likely a result of anthropogenic fire suppression.
Since termination of the mine, the sediment record shows a return to both a diatom community and forest present prior to the introduction of the mine. Pinus dominated the landscape and Fragilariaceae species were abundant in the diatom community. The diatom community seems to have been influenced by changes in both habitat and nutrient availability throughout the record in response to changes in land cover and precipitation patterns. It does not appear that Lily Pond is being impacted by AMD from the exposed minerals and remaining mine tailings; however, this could be impacting water resources in a different area. Although mining operations have ended, anthropogenic activity, mainly recreation, is still present at Lily Pond.
The Lily Pond record presented in this thesis provides information about how both the terrestrial and aquatic communities respond to both climatic and anthropogenic events specific to the late Holocene. Reconstructing conditions at Lily Pond during the late Holocene using multiple proxies provides an additional record for comparison within the southern Rocky Mountains. The terrestrial and aquatic responses to climate variations and anthropogenic activities at Lily Pond are similar to other records in Colorado, indicating the widespread impacts; however, the timing of past change varies along a latitudinal gradient. Paleoecological data places more recent trends in the context of long-term variability and provides land
80


management strategies in subalpine forests in the western United States. This is especially important given the warming temperatures and anthropogenic activity that have occurred over the last several decades, and which are expected to continue in the future. The combined records of both terrestrial and aquatic change provide the necessary context for understanding how Rocky Mountain ecosystems should be managed to ensure continued ecosystem function and health.
81


REFERENCES
Alexander, M. A., Scott, J. D., Friedland, K. D., Mills, K. E., Nye, J. A., Pershing, A. J., &
Thomas, A. C. (2018). Projected sea surface temperatures over the 21st century: Changes in the mean, variability and extremes for large marine ecosystem regions of Northern Oceans. Elem Sci Anth, 6(1), 9. https://doi.org/10.1525/elementa.191
Anderson, L. (2011). Holocene record of precipitation seasonality from lake calcite 8180 in the central Rocky Mountains, United States. Geology, 39(3), 211-214. https://doi.org/10.! 130/G31575.1
Anderson, L. (2012). Rocky Mountain hydroclimate: Holocene variability and the role of
insolation, ENSO, and the North American Monsoon. Global and Planetary Change, 92-93, 198-208. https://doi.Org/10.1016/j.gloplacha.2012.05.012
Anderson, L., Brunelle, A., & Thomspon, R. S. (2015). A multi-proxy record of hydroclimate, vegetation, fire, and post-settlement impacts for a subalpine plateau, central Rocky Mountains, U.S.A. The Holocene, 25(6), 932-943.
Arismendi, I., Groom, J. D., Reiter, M., Johnson, S. L., Dent, L., Meleason, M., ... Skaugset, A. E. (2017). Suspended sediment and turbidity after road construedon/improvement and forest harvest in streams of the Trask River Watershed Study, Oregon. Water Resources Research, 53(8), 6763-6783. https://doi.org/10.1002/2016WR020198
Ballantyne, A. P., Brahney, J., Fernandez, D., Lawrence, C. L., Saros, J., & Neff, J. C. (2011). Biogeochemical response of alpine lakes to a recent increase in dust deposition in the Southwestern, US. Biogeosciences, 8(9), 2689-2706.
Battarbee, R. W., Charles, D. F., Bigler, C., Cumming, B. F., & Renberg, I. (2010). Diatoms as indicators of surface-water acidity. In J. P. Smol & E. F. Stoermer (Eds.), The Diatoms: Applicaitons for the Environmental and Earth Sciences, 2nd Edition (2nd ed., pp. 98-121). Cambridge University Press.
Battarbee, Richard W. (1999). The importance of palaeolimnology to lake restoration.
Hydrobiologia, 395-396(0), 149-159. https://doi.Org/10.1023/A:1017093418054
Blaauw, M., & Christen, J. A. (2011). Flexible paleoclimate age-depth models using an autoregressive gamma process. Bayesian Analysis, 6(3), 457-474. https://doi.org/10.1214/ba/1339616472
Berger, A., & Loutre, M. F. (1991). Insolation values for the climate of the last 10 million years. Quaternary Science Reviews, 10(A), 291-311. https://doi.org/10.1016/0277-3791(91)90033-Q
82


BLM. (2008). Ranching and Agriculture: The Economic Diversification of Southwestern
Colorado. In Frontier in Transition: A History of Southwestern Colorado. Retrieved from https://www.nps.gov/parkhistory/online_books/blm/co/10/chap9.htm
Briles, C. E., Whitlock, C., & Meltzer, D. J. (2012). Last glacial-interglacial environments in the southern Rocky Mountains, USA and implications for Younger Dryas-age human occupation. Quaternary Research, 77, 96-103.
Brugam, R. B., & Lusk, M. (1986). Diatom evidence for neutralizaiton in acid surface mine lakes. Diatoms and Lake Acidity, 115-129.
Brugger, K. A. (2006). Late Pleistocene climate inferred from the reconstruction of the Taylor
River glacier complex, southern Sawatch Range, Colorado. Geomorphology, 75(3), 318-329. https://doi.Org/10.1016/j.geomorph.2005.07.020
Brunelle, A., Minckley, T. A., Lips, E., & Burnett, P. (2013). A record of Lateglacial/Holocene environmental change from a high-elevation site in the Intermountain West, USA. Journal of Quaternary Science, 25(1), 103-112. https://doi.org/10.1002/jqs.2600
Burtt, D. G., Addison, J. A., & Foster, A. (2013). Reconstructing Holocene fluvial dischage in the marine environment near Eureka, California, using a portable XRF method.
Calder, W. J., Parker, D., Stopka, C. J., Jimenez-Moreno, G., & Shuman, B. N. (2015). Medieval warming initiated exceptionally large wildfire outbreaks in the Rocky Mountains. Proceedings of the National Academy of Sciences, 112(43), 13261-13266. https://doi.org/10.1073/pnas.1500796112
Colorado Division of Reclamation Mining & Safety. (2014). Colorado Division of Reclamation Mining & Safety. Retrieved March 10, 2017, from
http://mining.state.co.us/Programs/Abandoned/Pages/impwelcomepage.aspx
Colorado Geological Survey. (2013, June 20). Abandoned Mine Lands. Retrieved February 26, 2017, from http://coloradogeologicalsurvey.org/mineral-resources/abandoned-mine-lands/
Colorado Water Quality Control Division. (2016). Gold King Mine spill; Animas River Basin, Southwest Colorado, August 2015. Department of Public Health and Environment. Retrieved from https://www.colorado.gov/pacific/sites/default/files/Gold-King-Mine-Spill-Report-01 -22-16-Digital .pdf
Currie, L. A. (2004). The Remarkable Metrological History of Radiocarbon Dating [II], Journal of Research of the National Institute of Standards and Technology, 109(2), 185-217.
83


Dean, W. E. (1974). Determination of carbonate and organic matter in calcareous sediments and sedimentary rocks by loss on ignition; comparison with other methods. Journal of Sedimentary Research, 44(1), 242-248. https://doi.org/10.1306/74D729D2-2B21-llD7-8648000102C1865D
Del Priore, T. (2015). 4000 years of environmental change in central Colorado: A paleoecologicalperspective. University of Colorado Denver.
DeNicola, D. M. (n.d.). A review of diatoms found in highly acidic environments.
Hydrobiologia, 433(1-3), 111-122. https://doi.Org/10.1023/A:1004066620172
Dierauer, J. R., Whitfield, P. H., & Allen, D. M. (2018). Climate Controls on Runoff and Low Flows in Mountain Catchments of Western North America. Water Resources Research, 54(10), 7495-7510. https://doi.org/10.1029/2018WR023087
Doesken, N. J., Pielke, Sr., R. A., & Bliss, O. A. P. (2003). Climatography of the United States No. 60. Retrieved from http://climate.colostate.edu/climate_long.html
Faegri, K. (1989). Textbook of Pollen Analysis (4th ed.). Chichester: John Wiley & Sons.
Fall, P. L. (1997a). Fire History and Composition of the Subalpine Forest of Western Colorado During the Holocene. Journal of Biogeography, 24(3), 309-325.
Fall, P. L. (1997b). Timberline fluctuations and late Quaternary paleoclimates in the Southern Rocky Mountains, Colorado. GSA Bulletin, 709(10), 1306-1320.
Fleischner, T. L. (1994). Ecological Costs of Livestock Grazing in Western North America. Conservation Biology, 8(3), 629-644.
Gale, S. J. (2009). Dating the recent past. Quaternary Geochronology, 4(5), 374-377. https://doi.Org/10.1016/j.quageo.2009.05.011
Goldewijk, K. K., Beusen, A., & Janssen, P. (2010). Long-term dynamic modeling of global
population and built-up area in a spatially explicit way: HYDE 3.1. The Holocene, 20(4), 565-573. https://doi.org/10.1177/0959683609356587
Hammer, O., Harper, D. A., & Ryan, P. D. (2001). PAST: Paleontological statistics software package for education and data analysis. Palaeontologica Electronica, 4(1), 9pp.
Hellebust, J. A., & Lewin, J. (1977). Heterotropgic Nutrition. In D. Werner (Ed.), The Biology of Diatoms (pp. 169-197). Los Angeles, California: University of California Press.
84


Heyerdahl, E. K., Brubaker, L. B., & Agee, J. K. (2001). Spatial Controls of Historical Fire
Regimes: A Multiscale Example from the Interior West, USA. Ecology, 52(3), 660-678. https://doi.org/10.2307/2680187
Higuera, P. (2009). CharAnalysis 0.9: Diagnostic and analytical tools for sediment-charcoal analysis. ResearchGate. Retrieved from
https://www.researchgate.net/publication/254406379_CharAnalysis_09_Diagnostic_and
analyticaltoolsforsediment-charcoalanalysis
Higuera, P. E., Briles, C. E., & Whitlock, C. (2014). Fire-regime complacency and sensitivity to centennial-through millennial-scale climate change in Rocky Mountain subalpine forests, Colorado, USA. Journal of Ecology, 102(6), 1429-1441. https://doi.org/10.llll/1365-2745.12296
Higuera, P. E., Brubaker, L. B., Anderson, P. M., Hu, F. S., & Brown, T. A. (2009). Vegetation Mediated the Impacts of Postglacial Climate Change on Fire Regimes in the South-Central Brooks Range, Alaska. Ecological Monographs, 79(2), 201-219.
Hundey, E. J., Moser, K. A., Longstaffe, F. J., Michelutti, N., & Hladyniuk, R. (2014). Recent changes in production in oligotrophic Uinta Mountain lakes, Utah, identified using paleolimnology. Limnology and Oceanography, 59(6), 1987-2001. https://doi.Org/10.4319/lo.2014.59.6.1987
IPCC. (2014). Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, (p. 151). IPCC, Geneva, Switzerland.
Jimenez-Moreno, G., & Anderson, R. S. (2012). Pollen and macrofossil evidence of Late
Pleistocene and Holocene treeline fluctuations from an alpine lake in Colorado, USA. The Holocene, 23(1), 68-77.
Jimenez-Moreno, G., Anderson, R. S., & Toney, J. L. (2011). A high-resoultion record of
climate, vegetation, and fire in the mixed confier forest of northern Colorado, USA. GSA Bulletin, 123(1/2), 240-254.
Johnson, F. (1965). Half-Life of Radiocarbon. Science, 149(3690), 1326.
Juggins, S. (2007). C2: Software for ecological and palaeoecological data analysis and
visualisation. Retrieved from https://www.staff.ncl.ac.uk/stephen.juggins/software.htm
Kido, Y., Koshikawa, T., & Tada, R. (2006). Rapid and quantitative major element analysis method for wet fine-grained sediments using an XRF microscanner. Marine Geology, 229, 209-225.
85


Kihlman, S., & Kauppila, T. (2010). Tracking the Aquatic Impacts of a Historical Metal Mine Using Lacustrine Protists and Diatom Algae. Mine Water and the Environment, 29(2), 116-134. https://doi.org/10.1007/sl0230-009-0096-2
Kitzberger, T., Brown, P. M., Heyerdahl, E. K., Swetnam, T. W., & Veblen, T. T. (2007).
Contingent Pacific-Atlantic Ocean influence on multicentury wildfire synchrony over western North America. Proceedings of the National Academy of Sciences, 104(2), 543-548. https://doi.org/10.1073/pnas.0606078104
Knowles, N., Dettinger, M. D., & Cayan, D. R. (2006). Trends in Snowfall versus Rainfall in the Western United States. Journal of Climate; Boston, 79(18), 4545-4559.
Kylander, M. E., Ampel, L., Wohlfarth, B., & Veres, D. (2011). High-resolution X-ray
fluorescence core scanning analysis of Les Echets (France) sedimentary sequence: new insights from chemical proxies. Journal of Quaternary Science, 26(1), 109-117. https://doi.org/10.1002/jqs.1438
Libby, W. (1972). Dating by Radiocarbon. Accounts of Chemical Research, 5(9), 289-295.
Libby, W. F., Anderson, E. C., & Arnold, J. R. (1949). Age Determination by Radiocarbon Content: World-Wise Assay of Natural Radiocarbon. Science, 709(2827), 227-228.
Lukas, J., Barsugli, J., Doesken, N., Rangwala, I., & Wolter, K. (2014). Climate Change in Colorado: A Synthesis to Support Water Resources Management and Adaptation. Colorado Water Conservation Board.
Mann, M. E., Zhang, Z., Hughes, M. K., Bradley, R. S., Miller, S. K., Rutherford, S., & Ni, F. (2008). Proxy-based reconstructions of hemispheric and global surface temperature variations over the past two millennia. Proceedings of the National Academy of Sciences, 105(36), 13252-13257. https://doi.org/10.1073/pnas.0805721105
Mann, M. E., Zhang, Z., Rutherford, S., Bradley, R. S., Hughes, M. K., Shindell, D., ... Ni, F. (2009). Global Signatures and Dynamical Origins of the Little Ice Age and Medieval Climate Anomaly. Science, 326(5951), 1256-1260. https://doi.org/10.! 126/science. 1177303
Marlon, J. R., Bartlein, P. J., Carcaillet, C., Gavin, D. G., Harrison, S. P., Higuera, P. E., ...
Prentice, I. C. (2008). Climate and human influences on global biomass burning over the past two millennia. Nature Geoscience, 7(10), 697-702. https://doi.org/10.1038/ngeo313
86


Marlon, Jennifer R., Bartlein, P. J., Gavin, D. G., Long, C. J., Anderson, R. S., Briles, C. E., ... Walsh, M. K. (2012). Long-term perspective on wildfires in the western USA. Proceedings of the National Academy of Sciences, 109(9), E535-E543. https://doi.org/10.1073/pnas.1112839109
Mayer, A. C., & SUfickli, V. (2005). Long-Term Impact of Cattle Grazing on Subalpine Forest Development and Efficiency of Snow Avalanche Protection. Arctic, Antarctic, and Alpine Research, 37(4), 521-526.
McLaughlan, K. K., Higuera, P. E., Gavin, D. G., Perakis, S. S., Mack, M. C., Alexander, EL, ... Williams, J. J. (2014). Reconstructing Disturbances and Their Biogeochemical Consequences over Multiple Timescales. BioScience. https://doi. org/10.1093/biosci/bit017
Mensing, S., Korfmacher, J., Minckley, T., & Musselman, R. (2012). A 15,000 year record of vegetation and climate change from a treeline lake in the Rocky Mountains, Wyoming, USA. The Holocene; London, 22(1), 739-748.
http://dx.doi.org.aurarialibrary.idm.oclc.Org/10.l 177/0959683611430339
Meurant, G. (2012). The Ecology of Natural Disturbance and Patch Dynamics. Academic Press.
Minckley, T. A., Shriver, R. K., & Shuman, B. (2012). Resilience and regime change in a southern Rocky Mountain ecosystem during the past 17000 years. Ecological Monographs, 52(1), 49-68. https://doi.Org/10.1890/ll-0283.l
Moser, K. A. (2004). Paleolimnology and the Frontiers of Biogeography. Physical Geography, 25(6), 453-480. https://doi.Org/10.2747/0272-3646.25.6.453
Moser, K. A., Mordecai, J. S., Reynolds, R. L., Rosenbaum, J. G., & Ketterer, M. E. (2010). Diatom changes in two Uinta mountain lakes, Utah, USA: responses to anthropogenic and natural atmospheric inputs. Hydrobiologia, 648( 1), 91-108. http://0-dx.doi.org.skyline.ucdenver.edu/10.1007/sl0750-010-0145-7
Mote, P. W., Hamlet, A. F., Clark, M. P., & Lettenmaier, D. P. (2005). Declining mountain
snowpack in western north am erica*. Bulletin of the American Meteorological Society, 56(1), 39-50. https://doi.org/10.1175/BAMS-86-l-39
National Oceanic and Atmospheric Administration (NOAA). (2014). ENSO Impacts on United States Winter Precipitation and Temperatures. Retrieved November 29, 2018, from http://www.cpc.ncep.noaa.gov/
87


Patrick, R. (1977). Ecology of Freshwater Diatoms and Diatom Communities. In D. Werner (Ed.), The Biology of Diatoms (pp. 284-332). Los Angeles, California: University of California Press.
Penniman, C. (n.d.). Atmospheric Fertilization of Estuaries—Watersheds and Airsheds. Retrieved January 17, 2019, from
http://www.prenhall.com/esm/app/nebel/html/updates/update20.html
Reimer, P. J., Bard, E., Bayliss, A., Beck, J. W., Blackwell, P. G., Ramsey, C. B., ... van der
Plicht, J. (2013). IntCall3 and Marinel3 Radiocarbon Age Calibration Curves 0-50,000 Years cal BP. Radiocarbon, 55(04), 1869-1887. https://doi.org/10.2458/azujs_rc.55.16947
Reynolds, R. L., Mordecai, J. S., Rosenbaum, J. G., Ketterer, M. E., Walsh, M. K., & Moser, K. A. (2010). Compositional changes in sediments of subalpine lakes, Uinta Mountains (Utah): evidence for the effects of human activity on atmospheric dust inputs. Journal of Paleolimnology; Dordrecht, 44(1), 161-175.
http://dx.doi.org.aurarialibrary.idm.oclc.org/10.1007/sl0933-009-9394-8
Saros, J. E., Clow, D. W., Blett, T., & Wolfe, A. P. (2011). Critical Nitrogen Deposition Loads in High-elevation Lakes of the Western US Inferred from Paleolimnological Records.
Water, Air, & Soil Pollution, 216(1), 193-202. https://doi.org/10.1007/sll270-010-0526-6
Saros, J. E., Interlandi, S. J., Wolfe, A. P., & Engstrom, D. R. (2003). Recent Changes in the
Diatom Community Structure of Lakes in the Beartooth Mountain Range, U.S.A. Arctic, Antarctic, and Alpine Research, 55(1), 18-23.
Saros, J. E., Michel, T. J., Interlandi, S. J., & Wolfe, A. P. (2005). Resource requirements of
Asterionella formosa and Fragilaria crotonensis in oligotrophic alpine lakes: implications for recent phytoplankton community reorganizations. Canadian Journal of Fisheries and Aquatic Sciences; Ottawa, 62(1), 1681-1689.
Savage, M., & Swetnam, T. W. (1990). Early 19th-Century Fire Decline Following Sheep Pasturing in a Navajo Ponderosa Pine Forest. Ecology, 71(6), 2374-2378. https://doi.org/10.2307/1938649
Science Museum of Minnesota. (2018). Laboratories & Equipment. Retrieved from https://www.smm.org/scwrs/facilities/laboratories
Sherriff, R. L., Veblen, T., & Sibold, J. (2001). Fire history in high elevation subalpine forests in the Colorado Front Range. Ecoscience, 8(3), 369-380. https://doi.org/10.1080/11956860.2001.11682665
88


Shuman, B., Henderson, A. K., Colman, S. M., Stone, J. R., Fritz, S. C., Stevens, L. R., ... Whitlock, C. (2009). Holocene lake-level trends in the Rocky Mountains, U.S.A. Quaternary Science Reviews, 25(19), 1861-1879. https://doi.Org/10.1016/j.quascirev.2009.03.003
Shuman, B. N., Routson, C., McKay, N., Fritz, S., Kaufman, D., Kirby, M. E., ... St-Jacques, J-M. (2018). Placing the Common Era in a Holocene context: millennial to centennial patterns and trends in the hydroclimate of North America over the past 2000 years. Climate of the Past, 14, 665-686.
Shuman, Bryan N., Carter, G. E., Hougardy, D. D., Powers, K., & Shinker, J. J. (2014). A north-south moisture dipole at multi century scales in the Central and Southern Rocky Mountains during the late Holocene. Rocky Mountain Geology, 49(1), 33-49.
Spaulding, S. A., Otu, M. K., Wolfe, A. P., & Baron, J. S. (2015). Paleolimnological Records of Nitrogen Deposition in Shallow, High-Elevation Lakes of Grand Teton National Park, Wyoming, U.S.A. Arctic, Antarctic, and Alpine Research, 47(4), 703-717. http://0-dx.doi.org.skyline.ucdenver.edu/10.1657/AAAR0015-008
Steelman, J. (Unpublished). A Dendrochronology Study of Mining & Logging Impacts in a Small Mountainous Watershed.
Sugita, S. (1993). A Model of Pollen Source Area for an Entire Lake Surface. Quaternary Research, 39(2), 239-244. https://doi.org/10.1006/qres.1993.1027
Sugita, S. (1994). Pollen Representation of Vegetation in Quaternary Sediments: Theory and Method in Patchy Vegetation. Journal of Ecology, 52(4), 881-897. https://doi.org/10.2307/2261452
Toney, J. L., & Anderson, R. S. (2006). A postglacial palaeoecological record from the San Juan Mountains of Colorado USA: fire, climate and vegetation history. The Holocene, 16(4), 505-517. https://doi.org/10.1191/0959683606hl946rp
United States Forest Service. (2007). Grand Mesa, Uncompahgre, Gunnison National Forest Proposed Land Management Plan: Program Priorities and Proposed Possible Actions (Appendix C).
University of Georgia: Center for Applied Isotope Studies. (2017). Radiocarbon: Background and History. University of Georgia: Center for Applied Isotope Studies. Retrieved from https://cais.uga.edu/submit/Radiocarbon_intro.pdf
US Geological Survey. (2017). Geologic Provinces of the United States: Rocky Mountains. Retrieved from https://geomaps.wr.usgs.gov/parks/province/rockymtn.html
89


Full Text

PAGE 1

IMPACTS OF ANTHROPOGENIC ACTIVITY AND CLIMATE EVENTS DURING THE LATE HOLOCENE IN A SUBAPLINE LAKE ECOSYSTEM IN CENTRAL COLORADO by BETHANY ANN WALKER B.S., Allegheny College, 2013 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 Program 201 9

PAGE 2

ii © 201 9 BETHANY ANN WALKER ALL RIGHTS RESERVED

PAGE 3

iii This thesis for the Master of Science degree by Bethany Ann Walker Has been approved for the Environmental Sciences Program By Christy E. Briles, Chair Timberley Roane Sarah Spaulding Date: May 1 8 , 2019

PAGE 4

iv Walker, Bethany Ann (M.S., Environmental Sciences) Impacts of Anthropogenic A ctivity and C limate E vents during the late Holocene on a S ubalpine L ake in C entral Colorado Thesis directed by Assistant Professor Christy E. Briles ABSTRACT This thesis examines how natural ( e.g., climate) and human activity (e.g., mining) impacted a subalpine ecosystem in central Colorado. A lake sediment p aleoecological record was developed based on sediment geochemistry, diatoms, pollen, and charcoal for the last 5000 years. Utilizing multiple proxies allow ed for reconstruction of both the aquatic and terrestrial environment responses , and provide d baseline information on ecosystem function prior, du ring and after the min ing activity . The record also documents variability of the subalpine ecosystem during known climate change s of the late Holocene. This record show s similar responses to climate events during the last 5000 years as other paleo records from high elevation lakes in the region . For example, t he transition from summer wet to winter wet precipitation patterns at ~2400 cal yr BP resulted in a more closed forest structure than before that was dominated by Pinus. The c harcoal record indicates increase d fire activity during the Medieval Cli m ate Anomaly (MCA), that later decreased at the onset of the Little Ice Age ( LIA ) . Human impacts have also affected the terrestrial and aquatic communities at the lake . The introduction of a nearby mine was pr eceded by timber harvest that altered the forest composition and fire regime. The change in vegetation also resulted in changes in allochthonous inputs into the lake, altering species abundance within the diatom communit y and increasing metal inputs into t he pond . Since termination of mining activity, the terrestrial and aquatic systems have returned to a state similar to that seen at the start of the MCA. The results indicate that species abundances of both the

PAGE 5

v terrestrial and aquatic communities v aried th roughout the late Holocene; however, there were no major community shifts, suggesting resistance to both climate and human activity. The form and content of this abstract are approved. I recommend its publication. Approved: Christy E. Briles

PAGE 6

vi DEDICATION I would like to dedicate this thesis to my husband . His consistent support and encouragement were essential to my success. I would also like to thank my parents an d siblings for always encouraging me lastly, thank you to the friends I have made here at the University of Colorado Denver for providing encouragement and laughs when they were needed the most.

PAGE 7

vii ACKNOWLEDGEMENTS I would like to thank my advisor, Christy Briles, for her sharing her time, pa tience , and pa ssion for paleoecology. Her guidance was incredibly helpful throughout this process, and I am inspired by her dedication to the subject. I would also like to thank my committee, Timberley Roane and Sarah Spaulding , as well as John Steelman f or providing supporting data for this study. Lastly, thank you to all the faculty and administration at the University of Colorado Denver who supported me throughout my scholastic career. Additionally, I would like to thank the Cushman Foundation for Fora miniferal Research and the American Association of Geographers Paleoenvironmental Change specialty group for supporting this research.

PAGE 8

viii TABLE OF CONTENTS C HAPTER I. INTRODUCTION ................................ ................................ ................................ ........... 1 II. BACKGROUND INFORMATION ................................ ................................ ............... 5 Physical Environment ................................ ................................ ................................ ..... 5 Climate ................................ ................................ ................................ ............................ 6 Climate Change Events ................................ ................................ ............................... 9 Proxy Data from Lake Sediments ................................ ................................ ................. 10 Diatoms ................................ ................................ ................................ ..................... 11 Sediment Geochemistry ................................ ................................ ............................ 14 Pollen ................................ ................................ ................................ ........................ 15 Charcoal ................................ ................................ ................................ .................... 16 Environmental Disturbances ................................ ................................ ..................... 18 Sediment Core Chronology ................................ ................................ ........................... 20 III. SITE DESCRIPTION ................................ ................................ ................................ . 23 Lily Pond ................................ ................................ ................................ ....................... 23 Forest Hill Mine ................................ ................................ ................................ ............ 24 IV. METHODS AND DATA ANALYSIS ................................ ................................ ...... 26 Field Methods ................................ ................................ ................................ ............... 26 Laboratory Methods ................................ ................................ ................................ ...... 26 Core description and lithology ................................ ................................ .................. 26 Chronology ................................ ................................ ................................ ............... 27 Diatoms ................................ ................................ ................................ ..................... 27 Geochemistry ................................ ................................ ................................ ............ 28 Charcoal ................................ ................................ ................................ .................... 29 Pollen ................................ ................................ ................................ ........................ 30 Data Analysis ................................ ................................ ................................ ................ 30 Chronology ................................ ................................ ................................ ............... 30 Diatoms ................................ ................................ ................................ ..................... 31 Geochemistry ................................ ................................ ................................ ............ 31 Pollen ................................ ................................ ................................ ........................ 32 Charcoal ................................ ................................ ................................ .................... 33

PAGE 9

ix V. RESULTS ................................ ................................ ................................ .................... 35 Core Lithology & Chronology ................................ ................................ ...................... 35 Sediment Geochemistry ................................ ................................ ................................ 40 Magnetic susceptibility & loss on ignition ................................ ............................... 40 XRF ................................ ................................ ................................ ........................... 40 Diatoms ................................ ................................ ................................ ......................... 41 Pollen ................................ ................................ ................................ ............................ 45 Charcoal ................................ ................................ ................................ ........................ 48 V I . DISCUSSION ................................ ................................ ................................ ............. 52 Climate of the late Holocene (5000 cal yr BP to present) ................................ ............ 52 Ecosystem Response to Climate during the late Holocene ................................ ........... 55 Switch from summer to winter dominant precipitation regime .............................. 55 Medieval Climate Anomaly & Little Ice Age ................................ ........................... 59 Human Impacts: Logging, Mining, Grazing and Recreation ................................ ........ 63 Ecosyst em Response to Human Impacts ................................ ................................ ...... 64 Logging & Start of Forest Hill Mine (~70 30 cal yr BP; 1880 1920 CE) ................ 64 Land Management Implications ................................ ................................ ................... 72 Limitations ................................ ................................ ................................ .................... 74 Future Research ................................ ................................ ................................ ............ 75 VII. CONCLUSION ................................ ................................ ................................ ......... 78 REFERENCES ................................ ................................ ................................ ............................. 82 APPENDI X A: Forest Hill Mining Complex ................................ ................................ ........................ 91 B: Lead 210 Dating Results ................................ ................................ .............................. 93 C: Supplemental Age Depth Model Materials ................................ ................................ . 95 D: Diatom Species Used for Analysis ................................ ................................ .............. 95 E: Locations of paleoclimate and paleoecology studies referenced ................................ . 97

PAGE 10

x LIST OF TABLES TABLE 1. Overview of precipitation variation throughout Colorado 7 2. Series of 210 Pb and 14 C dates used to generate the age depth model . . .38 3. Zones within the sediment record with respective dates and ages 4 0 4. Summary of diatom abundances sorted by zone 5 5. Summary of pollen record organized by zones 48 6. Summary of charcoal record organized by zones 5 1

PAGE 11

xi LIST OF FIGURES FIGURE 1. . 8 2. O verlap of proxies between b iogeography and limnology research . 1 1 3. D ifferent chemical and physical characteristics that can impact diatom survival . 1 2 4. Site Map . 2 5 5. Lithology of LP07, LP15 , and LP17 cores .. .3 6 6. Age depth model for Lily Pond Record . 3 7 7. Results from constrained cluster analysis using diatom data .... . 39 8. Results from diatom and geochemistry analys e 4 9. . 4 7 10. . . . 5 0 11. . 6 4 12. 6

PAGE 12

1 CHAPTER I INTRODUCTION Colorado has over 23,000 abandoned mine lands (AMLs) throughout the state (Colorado Geological Survey, 2013) . Mining activity is very disruptive to its surroundings , by expos ing heavy metals and other minerals and increas ing the amount of loose sediments that can impact surrounding environments (Colorado Water Quality Control Division, 2016) . Prior to 1975, Colorado mining companies were not required to remediate the mine site after the mine closed, resulting in a high number of toxic and dangerous AMLs (Colorado Division of Reclamation Mining & Safety, 2014 ) . According to the Colorado Division of Reclamation Mining & Safety (2014), approximately 1,539 acres of AMLs have been reclaimed statewide since 1980. Even with remediation, AMLs continue to impact the surrounding environments due to remaining mine spoil and bedrock exposure (Brugam & Lusk, 1986) . T he long term impacts of mining on the surrounding ecosystems are not well understood and is a focus of this thesis . Mining activity is a source of anthropogenic disturbance that directly impacts the surrounding terrestrial and aquatic ecosystems. To understand how mining activity impacts an ecosystem, baseline conditions must be established. However, it is difficult to define these conditions because human activities have impacted most environments for decades and even centuries. Therefore, long term historical records are crucial for setting reference conditions that examine changes in species communities, ecosystem pro ductivity, and chemical conditions . Lake sediment records provide long term historical records that contain information about both aquatic and terrestrial environments. Since sediment accumulation is always occurring, these records provide a continuous and measurable record of time that can be used to determine variability of the system on century to millennia time scales.

PAGE 13

2 I nternal lake processes and the surrounding terrestrial ecosystem can be reconstructed because of sediment internal and external sediment deposition through autochthonous organic accumulation and allochthonous runoff and atmospheric transport (Battarbee, 1999) . P rox y indicators , such as diatoms, pollen, charcoal, and the chemical make up of the sediments can be used to reconstruct past ecosystem conditions . Proxies are chosen based on the information they provide about the physical, biolo gical, and chemical conditions of the surrounding environments , and how well they can be modeled to characterize modern ecosystems (McLaughlan et al., 2014) . U tilizing multi ple proxies provide s an enhanced picture of ecosystem change than one proxy alone , as each measures a different aspect of the system and therefore provide s a clearer picture of change or response . Further, multiproxy record provide informatio n about the ecosystem recovery rate and whether an ecosystem returns to its pre disturbance condition (McLaughlan et al., 2014) . In Colorado, paleoecological studies have been conducted in the subalpine and alpine zones ( e.g. Anderson, Brunelle, & Thomspon, 2015; Del Priore, 2015; Fall, 1997b, 1997a; Higuera, Briles, & Whitlock, 2014; Toney & Anderson, 2006; Wolfe, Baron, & Cornett, 2001) . These studies have foc us ed on the vegetation, fire, and lake level records, and have been used to reconstruct climate trends through the Holocene. For example, examining charcoal and pollen record s has allowed for reconstruction of fire frequency and severity and the direct imp act on forests during major changes in past climate (Briles et al , 2012; Fall, 1997a; Higuera et al , 2014) . Similarly , oxygen isotope climate reconstructions have identified changes in precipitation and moisture availability , and were compared with pollen, charcoal, and sediment geochemistry records to examine climate impacts on fire regimes, f o rest composition and structure, as well as allochthonous inputs into lake systems (Anderson, 2011 ; Anderson, 2012; Anderson et al, 2 015) .

PAGE 14

3 Anthropogenic activities have also been shown to impact surrounding aquatic systems by altering the surrounding landscape . Mining, grazing, and logging have all impacted subalpine areas in Colorado , especially since Euro American settlement . A study in Finland examined the diatom record in two lakes that were impacted by a Cu Au mine that had been c losed (Kihlman & Kauppila, 2010) . Both records in the study document minor ecological effects during the active metal mining pe riod . The majority of ecological effects occurred after mine closure . Results determined that changes in metal inputs into the system impacted diatom species abundances in both lake systems ; however , changes in community abundances differed between sites. Grazing and logging activit ies have been impacting both terrestrial and aquatic habitats in the western United States since Euro American settlement . Approximately 70% of the western United States is impacted by livestock grazing (Fleischner, 1994) . This activity has impacts on the surrounding landscape by altering species composition and structure of surrounding forests (i.e. density and biomass) and disrupting ecosystem functions (i.e. nutrie nt cycling and soil erosion ) . The terrestrial disruption also impacts nearby water features. Changes in soil erosion increases sedimentation rates into waterbodies, causing increases in turbidity , total suspended solids, and nutrient i nput (Fleischner, 1994) . Similarly, land clearing due to tim b er harvest ing also increases sedimentation rates in to nearby water features (Arismendi et al., 2 017) . While terrestrial and aquatic communities impacted by grazing and timber harvest can recover, the amount of time it takes to do so varies . To my knowledge, few studies have examined impacts of past climate , mining , and other anthropogenic activities on both aquatic and terrestrial ecosystem s. Therefore, t he main objective of this research is to reconstruct the paleoecological history of a subalpine lake , Lily Pond in

PAGE 15

4 Taylor Park of central Colorado , where anthropoge nic activities, including mining between 1880 and 1920 CE , are apparent on the modern landscape . A late glacial period to early Holocene paleoenvironmental study was previously reported on by Briles et al. (2012) from a sediment core taken and radiocarbon dated in 2007. This study expands on the study by extending the pollen and charcoal record and adding new proxies, including diatoms and sediment geochemistry , from newly acquired core s taken in 2015 and 2017 that span the last 5000 years . The new record explore s how the aquatic and forest ecosystem responds to late Holocene climate an d more recen t anthropogenic disturbance events , specifically the introduction of a metal mine nearby . Th e study aims to answer the following questions: (1) How did climate variations during the late Holocene impact the terrestrial and aquatic environments at Lily Pond ? (2) How did the introduction and termination of mining activity affect the aquatic and terrestrial environment s at Lily Pond ? This thesis is composed of six chapters. Chapter two provides information on physical environment, climate, and disturbance regimes specific to this study. It also introduces the use of multiple proxies for reconstructing past environments and introduces the concept of ecosystem resilience. Chapter three introduces the study site and provides information about known anthropogenic activity in the area. Chapter four outlines the field and laboratory methods use d the carry out the research. Chapter five presents the results from the diatom , pollen, charcoal and sediment geochemistry analyses, as well as an updated 5000 year age depth model for the Lily Pond lake sediment core. Chapter six summarizes the findings of the study. Chapter seven concludes the thesis.

PAGE 16

5 CHAPTER II BACKGROUND INFORMATION Physical Environment The Rocky Mountain s extend from New Mexico into Alaska. Colorado encompasses a portion of the southern Rocky Mountains, including the Sawatch Range in the central part of the state. The eastern portion of Colorado is relatively flat, with a maximum elevation of 1,980 meters where the plains meet the Rocky Mountains. At this point, the elevation increases dramatically into the foothills (elev. 2,100 2,70 0 m) before transitioning into the mountain ranges (elev. > 3,000 m) (Doesken, et al , 2003) . The Sawatch Range is located in central Colorado , extends ~160 km along a northwest southeas t axis , and is part of the Continental Divide . There are 15 peaks within the Sawatch Range measuring over 4,250 meters. T he Rocky Mountains were originally formed during the Pennsylvanian era about 300 million years ago. The original mountain range that was created during this period eventually eroded away before being reformed into the present day Rocky Mountains. The Southern Rocky Mountains were formed during the Laramid e orogen y , approximately 70 40 million years ago (mya). The exact process that formed the Rocky Mountains is unknown, but the range is thought to be the result of an oceanic plate subducting beneath a continen tal plate at a very low angle (US Geologic al Survey, 2017) . The Sawatch Range experienced at least two periods of Pleistocene glaciation that shaped the landscape. The most recent period of glaciation ended ~18,000 years ago (Brugger, 2006) .

PAGE 17

6 Climate Colorado experiences a cool, dry climate with low humidity due to its inter continental location. However, the mountains play a large role in variable climate. Due to th e large differences in elevation between the eastern part of the state and the mountain ranges, there are substantial contrasts between the climate of the Eastern Plains and that in the mountain areas ( Doesken et al, 2003 ). Table 1 and Figure 1 highlight the climate differences between cities and towns throughout Colorado ( precipitation and temperate averages for Kit Karson and Denver from US Climate Data [https://www.usclimatedata.com/climate/colorado/united states/3175] ; averages for Taylor Park from Wes tern Regional Climate Center; [https://wrcc.dri.edu/cgi bin/cliMAIN.pl?cotayl] . Precipitation also varies greatly throughout the state due to elevation , topography , and large scale precipitation mechanisms , specifically the inter annual El Nino Southern Oscillation (ENSO) in the tropical Pacific, and Pacific Decadal Oscillation (PDO) in the north Pacific (~20 year cycles) (Kitzberger et al , 2007) . Higher amounts of winter precipitation (i.e. snowfall) occur in the southern Rock Mountains when El Nino conditions (warmer Pacif i c SSTs) or positive PDO exist. Decreased precipitation amounts occur when La Nina conditions (cooler Pacific SST s) or negative PDO conditions exist (Anderson, 2012; NOAA, 2014) . Precipitation also varies throughout Colorado based on elevation and topography of an area. Areas at lower elevations experienc e less precipitation than the mountain regions (Table 1). The eastern plains experience strong seasonal precipitation cycles. Annual precipitation ( 70 80% ) falls from April through September. Winter months bring dry air and strong winds, resulting in very arid conditions. Therefore, areas on the eastern side of the Continental Divide are very dry especially during the winter months (Doesken et al, 2003). Areas at higher

PAGE 18

7 elevations west of the Continental Divide experience more consistent precipitation throu ghout the year with most of the precipitation falling during the winter months in the form of snow (Doesken et al, 2003). This is highlighted in Table 1, where Taylor Park reports an average snowfall precipitation of 276.4 cm in the mountains, compared to only 54.9 cm in Kit Carson on the Eastern Plains . Du ring winter months, moist air masses from the Pacific Ocean bring snow to areas west of the Continental Divide. These air masses have minimal impacts on precipitation east of the Continental Divide (Does ken et al, 2003). The western slope receives more evenly distributed precipitation throughout the year than areas east of the Continental Divide. During summer months, the mountain ranges generate thunderstorms when there are high levels of moisture in the from July through September (Doesken et al, 2003). During the winter months, areas west of the Continental Divide receive higher amounts of precipitation in the f orm of snow than areas to the east (Fall, 1997a) . Table 1 . Overview of precipitation variation throughout Colorado. Location (Elevation) Avg. High Temperature Warmest Month Avg. Low Temperature Coldest Month Avg. Precipitation Avg. Snowfall Kit Carson (1,306 m) July January 36.25 cm 44.5 cm Airport (1,655 m) July January 39.32 cm 151.4 cm Taylor Park (2,817 m) July January 42.39 cm 276.4 cm

PAGE 19

8 Figure 1 . Locations referenced for precipitation variation in Colorado. Temperature throughout Colorado is controlled by differences in topography and elevation throughout the state. During winter months, areas at higher elevations experience much cooler temperatures tha n areas at low er elevations (Table 1). January is the coldest month on average, with temperatures ranging from the Denver International Airport (eastern plains area). Spatial t emperature ranges in the summer months m imic that of winter, with areas at lower elevations experiencing warmer temperatures than those at higher elevations. July is generally the warmest month, with average temperatures e 1).

PAGE 20

9 Climate Change Events Climate change events , or significant changes in temperature and precipitation , can impact ecosystems by altering growing seasons, water availability, and nutrient cycling. There are three major climate events that are known to have impacted the central Colorado Rocky Mountains during the late Holocene (~5000 cal yr BP to present). The intensity of the El Nino Southern Oscillation (ENSO) increased around this time and corresponds to changes seen in the Rocky Mountain climate as identified in the oxygen isotope ratio record of lakes in northwestern Colorado, near Steamboat Springs (Anderson, 2011; Anderson, 2012; Anderson et al, 2015) . The first climate shift occurred during the late Holocene around ~ 2 4 00 cal yr BP, when there was a switch from a summer dominated precipitation regime to a winter dominated precipitation regime (Anderson, 2011; Anderson, 2012) . Prior to ~2 4 00 cal yr BP , Colorado received most of its annual precipitation as rain during the summer m onths . H igh resolution pollen records from subalpine lakes in Colorado indicate subalpine forests had higher abundance of Picea (Higuera et al., 2014; Toney & Anderson, 2006) . The charcoal record s from the same lakes suggest low fire frequency leading up to ~2 4 00 cal yr BP. Although fi re frequency was low, the few fire events that did occur were likely stand replacing and high severity due to warmer summers and higher forest density (Higuera et al, 2014 ; Toney & Anderson, 2006 ). Recen t climate change events , Medieval Climate Anomaly and the Little Ice Age, are thought to be more extreme (i.e. higher temperatures and increased drought periods) than those that occurred earlier in the Holocene (Anderson, 2012 ; Higuera et al, 2014 ). The Me dieval Climate Anomaly (MCA; ~1200 850 cal yr BP) is recognized as a natural warming period that resulted in droughts and higher temperatures in North America. The MCA was once again a rain -

PAGE 21

10 dominated summer precipitation regime , similar to that seen prior to the precipitation switch at ~3000 cal yr BP (Anderson, 2011) . The increase in temperatures resulted in increased fire activity in subalpine areas. A composite study of 12 subalpine records from northern Colorado show that an i ncrease in temperature (~0.5 C) from the previous centuries caused an increase in burning at 83% of the sites (Calder et al., 2015). Burning decreased towards the end of the MCA even though temperatures remained high. The decrease in fire through the end of the MCA is likely due to a decrease in biofuel availability because of continued drought conditions (Calder et al , 2015) . T he Little Ice Age (LIA; ~650 10 0 cal yr BP) is defined by an increase in winter precipitation in Colorado, with the lowest 18 O values of the Holocene ( Anderson, 2 011; Anderson, 2012) . Wintertime s nowfall was high during this time , but summers remained nearly as dry as during the MCA (Anderson, 2012 ) . The ch arcoal record at Bison Lake indicates a decrease in fire activity at the start of the LIA . Picea also decreases , indicating shorter growing season s due to higher amounts of winter precipitation ( Anderson et al., 2015) . T hese changes in fire and vegetation identified in paleo records during the MCA and LIA show how climate fluctuations have impacted natural systems during the late Holocene. Proxy Data from Lake Sediments Lake sediments provide a natural archive of proxies that can be used to reconstruct pas t ecosystems . These sediments are extracted using coring devices that remove a column of sediment from the lake bed. The extracted sediments are referred to as sediment cores and can contain thousands of years of sediment that are arranged chronologically. Several different types of proxy data can be extracted from these sediment cores (Figure 2 ). Multiple proxy records provide allow for more complete observations about historical environmental conditions. For

PAGE 22

11 example, this study uses diatoms to infer how t he aquatic ecosystem has changed, sediment geochemistry to examine elemental inputs into the lake that are both natural and anthropogenic, pollen for reconstructing forest composition and structure, and charcoal to determine fire regimes. Each proxy helps re construct the physical, biological, and chemical characteristic s of the surrounding ecosystem and provide s (Moser, 2004). The proxies analyzed using lake sediments provide insight about the terrestrial environment of the surrounding watershed (the land a rea where precipitation drains into the waterbody) , as well as the surrounding airshed ( the land area that may be a source of small particles that are deposited in the watershed through aerial transport ) . Airsheds often include a larger area t han watershed s because there are no sharp boundaries (Penniman, n .d.) . The size of an airshed is influenced by the topography of an area, as well as the specific particles and emissions being examined For example, a study in the Uinta Mountains (Utah) identified changes in elemental composition of subalpine lake sediments that was a result of increased dust deposition from areas 100 + km away from the source (Reynolds et al., 2010) . The following sections introduce each of the proxies used in this study and identify how they have been used to recreate past conditions of both terrestrial and aquatic ecosystems. Figure 2. Venn diagram showing the overlap of proxies between biogeography and limnology research (from Moser, 2004).

PAGE 23

12 Diatoms D because they provide insight about the aquatic habitat in which they are found (Moser et al, 2004; Stevenson et al, 2010). Diatoms are effective ecological indicators because they respond directly to physical, chemical, and even biological changes in aquatic ecosystems (Dixit et al, 1992; Reid, 1995; Stevenson et al, 2010; Lowe, 2011 ). Diatoms also reproduce rapidly, and therefore respond more quickly to environmental change than other aquatic organisms. Figure 3 highlights the different factors that impact diatom survival. Figure 3. Overview of how different chemical and physical characteristics of an aquatic system can impact diatom survival (from Battarbee, 2000). Diatoms can be classified as benthic or planktonic. Benthic diatoms live on lake surfaces including aquatic v egetation, mud and sediments, and rocks. Planktonic species live suspended in the water column. Species from both classifications are often found in the same aquatic

PAGE 24

13 environment. Increases in benthic species would indicate an increase in benthic habitat th at could be a result of decreased lake level, where as an increase in planktonic species would represent an increased lake level. Therefore, to reconstruct lake levels using lake sediments, a benthic to planktonic diatom ratio is often used. Diatoms can al so be classified as epipelic, epilithic, and epiphytic, meaning they prefer to grow attached to clays/sediments, rocks, or plants, respectively. These classifications can help provide information about habitat availability within the lake system at differe nt times. For example, an increase in epiphytic diatom species could signify an increase in aquatic vegetation, which could be a result of increased nutrient availability in the system. Benthic species often dominate diatom communities in shallow high elev ation lakes because of increased organic matter available in sediments, making benthic habitats less likely to be nutrient limited than planktonic habitats (Saros et al , 2005; Spaulding et al., 2015) . Diatom species are also representative of chemical characteristics of their habitat. For example, d iatom analysis can be used to assess past pH levels of lacustrine environments . Some diatom species prefer acidic conditions while other species cannot survive in highly acidic waters ( Battarbee et al , 2010) . Diatom communities can also be monitored to id entify the natural variation of water pH in environments that have not been as severely affected by acidic inputs and processes ( Battarbee et al , 2010) . Diatom communities are al so representative of nutrient levels and can provide insight about land use in the surrounding area. Changes in inputs of nutrients, specifically nitrogen and phosphorus, alter freshwater production and impact the diatom community. A study done in the Gran d Teton National park (northwestern WY) shows that diatom assemblages in shallow lakes that were dominated by benthic species did respond to changes in radioactive nitrogen levels from anthropogenic sources (Spaulding et al., 2015) .

PAGE 25

14 Sediment Geochemistry Sediment geochemistry analyzes the metal and elemental composition of lake sediments using several different methods such as X ray florescence (XRF), magnetic susceptibility, and loss on ignition (LOI). XRF analysis measures heavy elements present in the sediment. Magnetic susceptibility measures the amount of iron bearing minerals present in the sediment by test ing the magnetic pull of the sediment (Gedye et al, 2000). An increase in magnetic susceptibility and heavy metals could signify a change in inputs that may be a result AMD or erosion due logging or grazing of the surrounding watershed. LOI is a measure of organic carbon production, and in part reflects lake productivity from both autochthonous and allochthonous inputs. All these analyses provide information about nutrient cycling and erosion (Gillson, 2015). At Bison Lake (CO), increases in higher detrita l element abundances (Fe, Ti, K), along with higher magnetic susceptibility, were indicators of greater detrital sedimentation during the LIA. The increases in elemental abundances signified an increase in allochthonous inputs and corresponded with decreas ed organic content, which suggested reduced biological productivity (Anderson et al, 2015). Anderson et al. (2015) used geochemical data to show that nutrient availability changed during the LIA, which was likely a result of the increased snowfall and more spring runoff. Sediment geochemistry analyses can also be impacted by atmospheric deposition of dust. A study in the San Juan Mountains (Colorado) found that dust deposition in alpine areas increases nutrients (N, P), cations (Ca, Mg) and some metals (Cr, Cu, Ni) available in watersheds. The ecosystem response to these increase s varies based on watershed size and bedrock geology (Ballantyne et al., 2011) . A study in the U inta Mountains (Utah) identified dust particles that had traveled ~110 km and originated from a mine (Reynolds et al., 2010) . This

PAGE 26

15 study also found that an increase in aerial deposition of fine sediments corresponded with the settlement of the surrounding area, leading to increased urbanization and mining activity. Lakes located closer to urban and mining areas showed increased deposi tion than lakes located further away (Reynolds et al., 2010) . Pollen Pollen found in sediment cores is an indication of the vegetation in the surrounding area . Changes in pollen may signify disturbance events such as fire and anthropogenic deforestation. It can also be indicative of climate variations, and changes in moisture availability and growing seasons (Briles et al, 2012; Higuera et al, 2014). Pollen can enter a lake system through runoff and aerial transport. Pollen data collected from small to medium size lakes represents vegetation within a 30 50 m radius. In general, t his area p rovides ~ 70% of the pollen source area for major taxa found in the sediment core (Sugita, 1994) . Larger lakes are influenced more by extralocal and regional vegetation (Sugita, 1993) . Aerial trans port of pollen is also impacted by the size and shape of pollen grains , as well as environmental conditions such as wind, precipitation, temperature, and humidity. L ighter pollen grains (i.e. Quercus ) can be transported further than heavier pollen grains ( i.e. Picea , Abies ) . Research has suggested that the source radius of light pollen types could be 100x larger than heavy types (Sugita, 1993) . In central Colorado, the forest composition in subalpine environments (>3000 m) was different during the early and mid Holocene (~11700 ~5500 cal yr BP) than during the late Holocene (~4000 cal yr BP to present). Pollen records indicate an open forest that was dominated by Pinus, Picea , and Abies during the early Holocene (Briles et al , 2012; Fall, 1997a; Jimenez Moreno & Anderson, 2012) . Summer insolation began to decrease after ~9000 cal yr BP, leading to climate changes in the mid to late Holocene . Changes in insolation lead to

PAGE 27

16 increased ENSO variability, and decreased summer precipitation (Anderson, 2012) . The mid Holocene (~5500 cal yr BP) saw a switch to a more closed forest structure, with increased Picea and Abies species and fewer Pinus species (Briles et al , 2012; Fall, 1997a; Jimenez Moreno & Anderson, 201 2) . This lasted until the late Holocene (~2 4 00 cal yr BP), when a switch in the precipitation regime from summer dominated to winter dominated (Anderson et al, 2015) caused a switch in forest structure throughout Colorado to an open canopy forest domina ted by Pinus ( Del Priore, 2015; Fall, 1997a; Higuera et al , 2014; Toney & Anderson, 2006) . At Mirror La ke (elev > 3000 m) and Keystone Ironbog (elev <3000 m), both in central Colorado, the change in vegetation occurred at ~2300 cal yr BP and ~2600 cal yr BP, respectively (Del Priore, 2015; Fall, 1997a). Both locations experienced the shift to a more open fo rest structure, although the exact timing varied. Timing of vegetation shifts seen in regional records are discussed in more detail in Chapter 6. Charcoal Charcoal analysis provides information about the fire regime of an area, including fire frequency, severity, and amount of biomass burned . Fire is a natural disturbance; however the frequency and severity of fire events can be indicative of climate characteristics during that time (Calder et al, 2015). Fire is recognized as a common natural d isturbance that is dependent on both climate and forest conditions (i.e. structure, composition, biomass availability) (Whitlock et al , 2010) . Charcoal found in sediment cores is an indication of local fire occurrence and severity within 0 6 km of the study site (Whitlock & Larsen, 2001) . Charcoal is deposited into lakes through airborne fallout during fire events. There is also a consistent in put of charcoal into lake systems through streamflow and runoff. Macroscopic charcoal (>125 microns) can be analyzed in lake sediments to reconstruct fire activity throughout history (Whitlock, 2004).

PAGE 28

17 Fires, both natural or human started, can be widesprea d and consume large amounts of forest biomass when weather conditions are right and have the ability to change forest composition and structure. In subalpine forests in Colorado, fire frequency is low due to cool, wet conditions and greater snowpack that r esult in increased fuel availably. The buildup of biofuel increases fire severity potential during drought periods (Whitlock et al, 2010). Therefore, subalpine forests are subject to high severity, stand replacing fires every ~100 to 300 + years (Fall, 1997 a; Sherriff et al, 2001) . These high severity fires have been identified using tree ring records from subalpine forests. These fires impact forest structure and composition, nutrient cycling, wildlife habitat, and degrade surrounding water quality (Sherriff et al, 2001). Charcoal records from the Rocky Mountains show that fire variability during the Holocene was influenced by climate trends an d fuel availability (Fall, 1997; Briles et al, 2012; Higuera et al, 2014; Anderson, 2015). Records from central Colorado show that fire activity was high in the early Holocene and declined after ~10000 cal yr BP. Activity then increased from ~9000 ~7000 cal yr BP (Higuera et al , 2014) . Fire events become more frequent in the region again between ~2000 ~1000 cal yr BP (Toney & Anderson, 2006; Jimenez Moreno et al, 2008; Higuera et al, 2014; Calder et al, 2015). The later shift was again due to a climate event , with the onset of the Medieval Climate Anomaly that was characterized by increased temperatures and drought conditions in the region (Anderson et al., 2015; Ca lder et al., 2015; P. E. Higuera et al., 2014; Toney & Anderson, 2006) . Aside from fire activity, records from Rocky Mountain National Park show that fire severity also fluctuated throughout the Holocene. There was a transition from high severity fires with a high amount of biomass burned, to fires with lower severity that burned less biofuels around ~2400 cal yr BP (Higuera et al, 2014). The change in severity and biomass burned was a

PAGE 29

18 result of decreased forest density that was related to a shift in pre cipitation regime from winter dominated to summer dominated that occurred around the same time (Anderson et al, 2015). Environmental Disturbances Environmental, or ecological, disturbances are defined as discrete events that can have significant long term impacts on the environment ( White & Pickett, 1985; McLaughlan et al., 2014 ). A distinguishing characteristic of disturbance events is that they are identifiable, meaning that a change or event can be identified as a cause of the changes seen within the environment. This is also dependent on the spatial and temporal scale of the system being examined (White & Pickett, 1985). Impacts of a disturbance event depend on their type, severity , and frequency. While some impacts might be seen immediately, such as the loss of trees due to fire, others may not become apparent until after the disturbance took place , as was seen in the diatom study in Finland that was introduced previously (Kihlman & Kauppila, 2010) . To understand how a disturbance event impacts the surrounding environments, researchers often look at species abundances prior to, and following, the disturbance event using proxy data from sediment cores . There are several types of disturbance events that have impacted central Colorado including fire , wind, insect, a nd anthropogenic activities (e.g., logging, mi ning, grazing, recreation) . Of these , fire and anthropogenic disturbances are discernable in the Lily Pond paleorecord. Anthropogenic disturbances refer to human activities that impact ecosystems either directly or indirectly. It is difficult to note exa ctly when humans began impacting natural systems, and at what intensity. There is evidence of Folsom Paleoindian groups occupying areas within central Colorado during the Younger Dryas Chronozone (~12.9 11.7 cal yr BP) (Briles et al, 2012). This early hu man activity was limited to fire, structures, and foraging that occurred at local scales, and therefore had a much smaller impact on the environment than more modern

PAGE 30

19 anthropogenic activities (Lewis & Maslin, 2015). Recently, there has been discussion about the scientists have suggested that humans have caused enough disturbance to natural systems to warrant the designation of a new geological epoch (Marlon et al, 201 2; Lewis & Maslin, 2015). Anthropogenic activities have bene influencing western forests since settlement in the early 1800s, indicated by fire scar and charcoal records from the western United States. Grazing and timber harvest (i.e. logging) increased with the settlement of the w estern United States throughout the 1800s and led to an overall decline in biomass burning (Marlon et al., 2012) . Grazing was likely the first cause of fire reduction in the western United States because it resulted in less dense forest structure (Heyerdahl, Brubaker, & Agee, 2001; Mayer & St o ckli, 2005; Savage & Swetnam, 1990) . There is a peak in fire activity from 1850 1870 CE, which is likely a result of human caused burning for clear ing forests, railroad construction, agriculture, and lumbering (Marlon et al, 2012). After settlement, f ire suppression techniques impacting fire frequency as well, although has not impacted high elevation alpine areas extensively as lower elevations (Heyerdahl et al., 2001; Sherriff et al , 2001) . There is, howe ver, a lot of spatial variability in fire activity in the subalpine forests throughout Colorado during the late Holocene (Marlon et al., 2012; Sherriff et al., 2001) . Given t he current temperature trends (increased temperatures and droughts), biomass burning should be higher than what is actually occurring. In areas where human land management and fire suppression is less extensive, fire activity has remained high (Marlon et al., 2008; Marlon et al., 2012) While climate is a ma jor drive of fire activity, human fire suppression efforts have altered trends in biomass burning since the late 1800s.

PAGE 31

20 Recent paleolimnological studies have identified indirect impacts of human development on alpine lakes, which are often isolated from th e direct impacts of anthropogenic activity. These studies have focused specifically on the indirect effects anthropogenic disturbances such as agriculture, mining, and urbanization, are having on natural systems because of atmospheric deposition of metals and nutrients (Hundey et al., 2014; Moser et al. , 2010) . The increased met al and nutrient deposition is causing changes in water chemistry, including pH, as well and nutrient cycling, specifically nitrogen and phosphorus, within lake systems. In the Uinta Mountains (Utah), alpine lakes downwind of anthropogenic activities, incl uding mining, industry, and agriculture, have seen increases in metals and nutrients (i.e. phosphorus, nitrogen, iron, calcium) since the 1800s (Hundey et al., 2014; Moser et al., 2010 ) . The metals are being introduced by atmospheric inputs and have increased more since the 1900s. The increase in metal content ha s caused changes in the diatom communities in these lakes (Moser et al., 2010) . These lakes have also seen an increase in nitrogen and phosphorus loading, likely due to increased use of nitrogen and phosphorus fertilizer, and phosphorus mining in the area. The increase in N and P into these lakes, which were previously thought to be N limited, have resulted increases in nitrophilous diatom species a nd increased primary production (Hundey et al., 2014) . Similar results have been seen in other alpine areas within the western United States (CO, MT, WY) ( Saros et al, 2003; Saros et a l, 2011 ; Wolfe et al., 2001) . Sediment Core Chronology Radiometric dating using lead 210 ( 210 Pb) and radiocarbon ( 14 C) dating are used to establish ages for the length of a sediment cores. Radiocarbon dating was developed by a group of chemists at the University of Chicago in the late 1940s (Libby, 1972) . Early techniques used an anticoincidence counter to measure the radioactivity of solid carbon samples and used a half -

PAGE 32

21 life of 5720 ± 47 years, which was then updated to 5568 ± 30 years (Libby, 1972; Libby et al, 1949) . life" and is still used to calculate radiocarbon ages (Currie, 2004; University of Georgia, 2017) . In 1965, the physical 14 C half life was accepted to be 5370 ± 40 years, which continues to be used today in some fields of radiocarbon research (Johnson, 1965; University of Georgi a , 2017) . In the 1970s, researchers began to use accelerat or mass spectrometer (AMS) systems to measure 14 C. AMS systems are able to analyze smaller sample sizes, allowing for analysis of amino acids, seeds, pollen, and other small macrofossil s (University of Georgia, 2017). T he amount of atmospheric 14 C has varied throughout time due to changes in production rates caused by the carbon cycle, and changes in the cosmic ray flux (Libby, 1972; Reimer et al., 2013) . The refore, the resulting 14 C date (yr BP) needs to be compared to an established calibration curve so an approximate calendar date can be determine d (cal yr BP) (Blaauw & Christen, 2011; Reimer et al., 2013) . Ideally, calibration should be based on a dated record that utilizes carbon levels from the atmosphere at the ti me of sample formation (Reimer et al, 2013). The IntCal13 curve represents the mid latitude Northern Hemisphere atmospheric carbon record and wa s developed using tree ring measurements and macrofossil data that extend to 13,900 cal BP and the end of the range of 14 C dating methods, respectively (Reimer et al, 2013). The IntCal13 curve was updated from the IntCal09 curve, however the majority of di fferences between the two occurred prior to ~12000 cal yr BP. Still, it is important to note that 14 C calibration curve s are still being developed as researchers gain increased understanding of atmospheric 14 C processes (Reimer et al, 2013) . Radiocarbon d ating is less reliable for the past 500 years due to the large injection of carbon into the atmosphere around AD 1955 (Gale, 2009). Therefore, more recent sediments

PAGE 33

22 (within the last 150 years) must be dated using 210 Pb dating techniques. 210 Pb is a radiometric isotope that is formed as radon 222 decays. 210 Pb accumulates over time. Dates of the sediments are established by assessing the amount of 210 Pb present at a certain depths, usually the last 100 200 years, and the results are used to calculate dates using a mathematical model based on 210 Pb accumulation and sedimentation rates (Science Museum of Minnesota, 2018) . An age depth model is generated by fitting a curve to the provided 210 Pb and 14 C dates, and using the IntCal13 radiocarbon calibration cur ve (Reimer et al., 2013) . Recently, paleoecological studies have relied heavily on a methodology using Bayesian statistics rather than linear regression models (Blaauw & Christen, 2011) . This m ethodology assumes a monotonic rate of accumulation and utilizes a gamma autoregressive process to establish and a Markov Chain Monte Carlo (MCMC) algorithm to produce posterior distributions (Blaauw & Christen, 2011) . Bacon splits the core into vertical sections (default = 5) and then uses the MCMC iterations to estimate the accumulation rates (or sedimentation times ; default = years/centimeter ) . The resu lt is a more accurate age depth model that accounts for changes in accumulation/ sedimentation rates over time (Blaauw & Christen, 2011) . Each depth within the core is assigned a date that can be used in analyzing other proxy records.

PAGE 34

23 CHAPTER III SITE DESCRIPTION Lily Pond Lily Pond (Lat. 38°56'3.87"N , Long . 106°38'48.26"W ), located in Taylor Park in the Gunnison National Forest , lies within the Sawatch Mountain Range in central Colorado at an elevation of 3200 m (Figure 4) . The pond is at an ecotone be tween montane and subalpine forest. The forest , primarily composed of Pinus contorta , Picea engelmannii , and Abies las i o carpa , was historically responsive to changes in climate of the late glacial period and Holocene ( Briles et al, 2012 ; Fall , 1997 ) . The pond was formed ~1 2 , 000 years ago as a kettle lake by Pleistocene ice age glaciers . A recessional moraine to the north of lake dams the site . The in and out flowing streams are intermittent and likely only move water during the spring snowmelt , and during major summer precipitation events . Today, the pond is shallow ( ~1 meter of water at the deepest point ) and surrounded by sedge ( Cyperaceae spp ). During the winter months, Lily Pond ices ove r due to cold temperatures in the Taylor Park region. During the summer months, the pond is covered in lily pads ( Nupha r spp .). The pH of Lily Pond at the end of August (2017) was 6.9, indicating circumneutral conditions. The modern diatom community is dom inated by benthic species including Fragiliariaceae, Pinnularia, Sellaphora, Encyonema, and Stauronei s . The diatom community is discussed further in Chapter V. There are burned trees on the southwest slope of Lily Pond . Age estimates are ~120 years , but addition dendrochronological work is needed to confirm the age. Present day climate in the Taylor Park region is characterized by cold winters with heavy snowfall, and warmer summers with frequent thunderstorms during July and August. The high

PAGE 35

24 conce ntration of tall mountain peaks influences the climate patterns in the region and causes temperature and precipitation to change with elevation, resulting in local microclimates (Fall, 1997). Taylor Park receives an average of 42 cm of precipitation each y ear. Average annual snowfall is ~276 cm, the majority of which falls during November through April. Average temperatures range from a low of and average temperature during the winter months is [https://wrcc.dri.edu/cgi bin/cliMAIN.pl?cot ayl]). Forest Hill Mine The Forest Hill Mine (Lat. 38°55'24.78"N , Long . 106°38'28.12"W ) is located ~100 meters upslope and ~1100 meters southeast of Lily Pond and was a small operation that mined lead, silver, and zinc (Figure 4 ) . The mine operated upstr eam of Lily Pond and was in use from 1880 1920, with a short break from 1907 1916 CE . The mine is in a different hydrologic unit code ( HUC ) watershed than Lily Pond. However, there are smaller exploratory mines in the watershed , collectively called the For est H ill Mining Complex (FHMC) ( Appendix A ). There are also remnants of cabins and other living structures in the area, including in the area immediately surrounding Lily Pond, along with old mining infrastructure including a boiler and metal pipes. A preliminary dendrochronology study (Steelman, unpublished) identified three separate forest stands within the FHMC. Two of them indicate that there was an extensive amount of logging in the FHMC that corresponded to the establishment of Forest Hill Mine proper. Two of the forest stands have establishment dates between 1840 1870 CE, which corresponds with the start of mining in the area (Steel man, Unpublished) . This indicates that anthropogenic activity

PAGE 36

25 was taking place at Lily Pond beginning around 1850 CE, although it is not clear if all activity was directly related to mining operations. Figure 4. Site Map

PAGE 37

26 CHAPTER IV METHODS AND DATA ANALYSIS Field Methods A 1 meter continuous sediment core was extracted from Lily Pond in August 2017. The core (LP17) was taken from the east side of Lily Pond using a short corer designed by Steve Klein that captures the sediment water interface. Th e core was measured before extraction and transportation. The top 20 cm of unconsolidated sediment was extracted at 1 cm increments onsite and placed in Whirl Pak bags. The remaining part of the core was placed in PVC piping and transported back to the Pal eoecology , Palynology and Climate Change Laboratory at University of Colorado Denver for analysis. Water temperature and pH were taken when the core was taken. Water samples from the sediment water interface, the water surface, and the water column were al so collected along with a sedge and lily pad for diatom analysis. This study also uses data collected from a 2 meter sediment core taken at Lily Pond in August 2015. This core (LP15) was taken from the south east side of Lily Pond using a D section corer and did not capture the mud water interface . This core was also taken back to the University of Colorado Denver for analysis. Laboratory Methods Core description and lithology Once back in the lab, the LP15 and LP17 cores were measured to confirm length. Core lithology was described to account for any changes in appearance and texture of the sediment. The cores were then subsampled at 0.5 cm increments , stored in Whirl Pak bags, and placed in

PAGE 38

27 the refrigerator. Any macrofossils found while s ampling cores were collected to be used for radiocarbon dating. Chronology There is an existing age model for Lily Pond that spans ~17,100 years that was formulated using a series of radiocarbon dates obtained fro m a core taken in 2007 (LP07), and the LP1 5 core (Table 2; Briles et al, 2012). Samples from the LP07 core were sent to Beta Analytic for 14 C analysis . One macrofossil from the LP15 core was sent to DirectAMS for 14 C analysis and was used to link the LP07 and LP15 cores . However, the record was lacking a high resolution chronology for the past 1000 years (top 55 cm of the core). Radiocarbon dating is less reliable for the past 500 years due to the large injection of carbon into the atmosphere around 1955 CE due to nuclear bomb testing (Gale, 2009) . Therefore, to constrain the recent mining activity, a 210 Pb chr onology was developed . Sediments from the upper 25 cm of the LP17 core were sent to St. Croix Wat ershed Research Station for 210 Pb dating. One macrofossil from the LP17 cores was sent to the University of Georgia Center for Applied Isotope Studies in Athens, Georgia for AMS radiocarbon dating analysis to help link and constrain the three core chronolo gies . Diatoms Diatom species were used to reconstruct aquatic ecological conditions , including lake level and water chemistry. They can also provide insight s about terrestrial land use changes that result in increased sedimentation and allochthonous inputs. Diatoms were sampled from both the LP15 and LP17 cores. Diatoms were initially sampled at 6 cm increments throughout the LP15 core. After 210 Pb da ting was complete , higher resolution sampling was done using the LP17 core with specific focus in the top ~30 cm of the core . Sediment was dried at 90 C for 24 hours. Next,

PAGE 39

28 0.4 g of dry sediment was weighed out and put into a 50 ml centrifuge tube. The samples were taken to the INSTAAR Sediment Laboratory at University of Colorado Boulder for processing. Approximately 5 ml of nitric acid was added to each sample, and then placed in a microwave digestor for one 35 minute cycle , to control temperature and pressure. Th e samples were heated to a maximum temperature of 200 C , and then allowed to cool before removing from the digestor. The r emaining nitric acid was decanted, and the samples were rinsed with distilled water 6 times. D iatoms were mounted on slides using Naphrax and counted using light microscopy. Valves were identified and counted using a Nikon Labphot microscope at 1000x magnification with phase contrast & dark/bright field. Approximately 3 5 0 valves were counted per slide. Diatoms were identified to the genus level using the Taxon Identification Guide maintained by Diatoms of North America (2018; https://diatoms.org/ ). Geochemistry Magnetic susceptibility, loss on ignition (LOI), and X ray fluorescence (XRF) analyses were performe d to reconstruct sediment geochemistry and identify input changes. These analyses help determine the effects of mining on inputs into the aquatic ecosystem. Magnetic susceptibility was used to determine increases in iron bearing minerals due to erosion int o the lake ( Gedye et al, 2000 ) . Magnetic susceptibility was examined at 1 cm intervals using a Bartington MS2E point sensor meter. Measurements were recorded in units of centimeter gram second (cgs). LOI measures changes in organic content of the sediments , which, in part, in dicates changes in lake productivity (Dean, 1974) . However, external /terrestrial organic inputs are possible (e.g., run off events, human, wildlife, and livestock excrement ), but these are likely

PAGE 40

29 minimal at Lily Pond given limited stream inputs and the extensive sedge mat to the south of the pond . Samples four hours and measured for weight loss to hours each, and again weighed between each drying period to determine amount of organic material, and carbonates present, respectively. XRF analysis is used to determine elemental composition of sediment samples. This study used an Olympus model portable XRF gun and a bench stand ( Kenna et al, 2010; Burtt, et al, 2013) . Processing samples with the laboratory bench stand decrease d the amount of laser movement when taking the readings, resulting in more accurate results. Sam ples were also dried prior to analysis to eliminate risk of water absorption (Kid o et al , 2 006) . XRF was initially measured at 4 cm intervals throughout the core. Additional XRF readings were conducted at 1 cm intervals through the mining period to present. Samples were dried at 90 C , crushed using an agate mo r t a r and pestle , placed on top of super fine filter paper in a plastic XRF sample cup with polypropylene fill, and c overed with polypropylene thin film . Charcoal Macroscopic charcoal was used to understand the local fire history at Lily Pond. It also provides insight into both natural and anthropogenic fire activity. Charcoal data was collected and analyzed from the LP15 core . Charcoal was sampled at 0.5 cm increments throughout the length of the core using the sieve method ( Whitlock & Larsen, 2001) . Samples were treated with sodium hexametaphosphate to dissolve clay material and bleach to get rid of humic matter, and then filtered through a 125 µm sieve. Macroscopic charcoal pieces were counted in a gridded petri dish usin g a stereomicroscope.

PAGE 41

30 Pollen Pollen analysis provides information regarding changes in local forest composition and structure . Pollen data in was collected and analyzed from the LP15 core at 6 cm increments . After initial pollen analysis, additional sampl es were prepared from the top portion of the core (top ~30 cm). Laboratory methods outlined by the LacCore Pollen Preparation Procedure (University of Minnesota; http://lrc.geo.umn.edu/laccore ) were followed for all samples. Potassium hydroxide and acetolysis were used to remove organic material, and hydrofluoric acid was used to remove silicates (Faegri & Iversen , 1989) . Slides were prepared and counted using a light microscope. Lycopodium tracers were added to allow for pollen concentration calculation (grains cm 3 ). Grains were identified by Dr. Briles to the lowest taxonomic level possible using the PPCC lab reference library . Pinus grains were separated into Diploxylon ( contorta type ) and Haploxylon ( flexillis type ). All other grains were identified and counted at the genera level. D ata Analysis Chronology A ge depth models of lake sediments using Bayesian iterative models account for the probability distribution of each radiometric date , consider realistic sediment accumulation of lakes , weight known dates more heavily than less known dates, and provide error estimates for the entire chronology . E ach date in the model influences other dates and the final chronology and allows for fewer dates to produce more accurate age depth models (Blaauw & Christen, 20 11) . Radiocarbon and 210 Pb dates were used to create an age depth model u sing an R based statistical package called BACON (Blaauw and Christen, 2011) . BACON assumes monotonic accumulation, and then models sedimentation rates as a function of depth using a gamma autoregressive process. The accumulation rates were calculated at 0.5 cm/ year and all other

PAGE 42

31 parameters within Bacon were left at the default v alues. 14 C dates were calibrated using the IntCal13 calibration curve. 210 Pb dates were not set to a calibration curve since they were already set to the calendar scale (Blaauw & Christen, 2011). Ages were calculated at 0.5 cm intervals from 0 cm to 139 c m. Diatoms Diatom counts were converted to percentages for each taxa at each sample depth using the following equation: P ercentages were used for graphing and data analysis . Taxa were included in analysis if t hey contributed at least 1% of total relative frequency in more than one sample. Diatom results were further analyzed using Paleontological Statistics Software Package (PAST 3.0; Hammer, Harper, & Ryan, 2001) software. Cluster analysis was used to identify different zones in the sediment core by comparing diatom genus percentages at the different sample depths. Diatom community composition in each sample was compared against the others using Principal Compon ents Analysis (PCA). The trends and groupings between the samples were used to establish zone boundaries and discuss possible changes in ecological conditions throughout the sediment core. PCA was run an additional time for each zone to identify how the diatom genera changed between zones ( Appendix D ). Geochemistry Results from magnetic susceptibility and LOI were plotting in C2 (Juggins, 2007) to visually identify trends in the data throughout the core. LOI data was used to calculate the percent of organic material presen t throughout the core. XRF data was collected using the

PAGE 43

32 Olympus proprietary Geochemistry C ycle reported as percentages. Elemental ratios were calculated to determine any differences in sediment inputs into Lily Pond using the following equation: Geochemistry at Lily Pond primarily focused on the levels of Iron (Fe) and Manganese (Mn). Both Fe and Mn l evels were normalized against Titanium (Ti) bec ause it is abundant, is not biologically important, and is resistant to weathering (Kylander et al, 2011). Th e Fe/Ti ratio provides information about allochthonous inputs into lake systems. Increases in the Fe/Ti ratio values indicate an increase in allochthonous inputs from the surrounding area. An increase or decrease may be indicative of changes in precipitation regimes or changes in land use in the surrounding area. The Mn/Ti ratio provides information about lake level changes. In oxygen rich environments, Mn forms an insoluble oxide. Therefor e, increases in the Mn /Ti may be indicative of increased oxygen levels due to a decrease in lake level s (Kylander et al , 2011) . Pollen Pollen percentage data and a ccumulation rates were calculated to reconstruct forest structure and composition , respectively . Both were calculated at each sample depth using the following equations. Pollen percentages :

PAGE 44

33 Pollen accumulation rate: The arboreal to nonarboreal ratio was also calculate d to help identify changes in the forest structure using the following equation: Higher amounts of arboreal pollen indicat e a more closed canopy. A more open forest structure is indicated by increases in nonarboreal pollen from grasses, shrubs, and other ste p p e species. These changes can indicate possible logging events as well as large fire events. Charcoal Charcoal counts w ere analyzed using the statistical program CharAnalysis ([http://phiguera.gitbug.io/Charnalysis]) (P. Higuera, 2009) . CharAnalysis uses the input information and interpolates it to the median sample resolution (yr sample 1 ). The program then uses this to distinguish between background charcoal (BCHAR) and charcoal accumulation rates (CHAR) using charcoal concentrations (particles cm 3 ) and the sedimentation rate (cm yr 1 ). CHAR is reported in units of particles per centimeter per year (particles cm 2 yr 1 ). BCHAR represents a running average of charcoal accumulation rates through time . BCHAR reflects changes in the rate of total char coal production (biomass) and changes in secondary charcoal deposition mainly from extra local sources (Higuer a et al, 2009) . For this study, BCHAR was estimated using a 800 year lowess smoothe r, robust to outliers . The 800 year smoothing window was selected based on the signal to noise index (Higuera, 2009) . The signal to noise index (SNI) is a statistical measurement of the separation between BCHAR and peak

PAGE 45

34 events. The larger the separation (generally above 3 for th is record ) the higher the confidence of a fire event occurred in the record at that time (Higuera et al, 2014) . A local threshold value (99%) was used to identify CHAR peaks beyond BCHAR levels using a noise distribution model based upon a 1 mean Gaussian distribution. If a peak in CHAR exceeded the threshold, a local fire event was identified . P eak magnitude (particles cm 2 peak 1 ), or the amount of CHAR above BCHAR, has been used as an indicator of fire severity , although proximity of the fire events to the lake and wind direction would also influence peak magnitude . Pollen data around the time of an individual fire event should be examined to estimate fire severity ( e.g., (Minckley et al , 2012 ; Del Priore, 2015; Morton et al., 201 7 ). The fire return interval (FRI) was determined using a 1000 year running mean.

PAGE 46

35 CHAPTER V RESULTS Core Lithology & Chronology The top 20 cm from the LP17 core was unconsolidated sediments that were extracted in the field to preserve the sediment water interface . The LP17 core consisted of unconsolidated, dark course detritus gyttja (CDG) for the first 20 cm and dark fine detritus gyttja ( F DG) after 20 cm . In the LP15 core, the top ~ 6 cm of sediment was a dark C DG, and the remainder of the core was dark fine detritus gyttja (FDG). There was a sand layer at ~85 cm . The top portion of both cores had an abundance of roots and Cyperaceae (sedge) seeds , likely from the dense aquatic vegetation at Lily Pond . The LP15 and LP17 core lithologies are very similar, beginning with CDG and then transitioning into FDG. The LP07 core corresponds with the FDG in the other two cores. T he LP15 and LP17 cores do not show the transition to silt that is seen in the LP07 co re at ~75 cm. The transition may have occurred later in these cores due to variations in sediment accumulation within the basin .

PAGE 47

36 Figure 5. Lithology of the LP15 and LP17 cores. Due to the extraction method (d section corer) of the 2015 core, the sediment water interface was not preserved. T he 2017 short core captured the lost upper sediments , which preserve information on anthropogenic period of the record and the modern proxy re cord of the terrestrial and aquatic communit ies . T hree c ores , 2007, 2015, and 2017, were linked based on an increase in Pinus percentages , a decrease in Artemisia percentages, an increase in CHAR between 90 100 cm depth, the sustained increase in magnetic susceptibility at ~28 cm in the 2015 and 2017 cores, a peak in CHAR at the top of the cores , and an overlapping 14 C date at depth 97 cm in the 2007 core and 69 cm depth in the 2015 core (Table 2) . An age depth model was created for Lily Pond using five calibrated 14 C dates and a series of 210 Pb dates . Details about 210 Pb and 14 C dates are shown in Table 2 . Additional information about the 210 Pb dates is included in Appendix B. Four of the 14 C dates were previously published , including a date that

PAGE 48

37 overlapped with the 14 C date on the 2015 core (Briles et al, 2012). The resulting BACON age model is show n in Figure 6 . F igure 6 . a) A ge depth model for Lily Pond. b) Results from 210 Pb dating before calibration.

PAGE 49

38 Table 2 . Series of 210 Pb and 14 C dates used to generate the age depth model. 210 Pb Dates Lab ID Material Age (yr BP) Error (yr) Depth 1 (cm) Core St. Croix Watershed Research Station sediment 65.6 0.66 0 LP17 sediment 59.5 0.72 3 LP17 sediment 53.5 0.72 5 LP17 sediment 46.8 0.73 7 LP17 sediment 39.3 0.70 9 LP17 sediment 29.8 0.80 11 LP17 sediment 16.8 0.79 13 LP17 sediment 8.5 0.85 14 LP17 sediment 1.3 0.97 15 LP17 sediment 13.3 1.25 16 LP17 sediment 31.3 2.30 17 LP17 sediment 51.6 3.70 18 LP17 sediment 68.8 6.10 19 LP17 sediment 87.8 11.55 20 LP17 sediment 122.5 10.51 21 LP17 sediment 148.5 25.36 22 LP17 14 C Dates Lab ID Material Age (yr BP) Error (yr) Depth 1 (cm) Core UGAMS34916 wood 940 25 41 LP17 D AMS018629 wood 1627 29 69 LP15 Beta244885 2 peat 1620 40 69 LP07 Beta244886 seed 2560 40 96 LP07 Beta242757 wood 5080 40 120 LP07 Beta244888 gyttja 6440 40 139 LP07 . Constrained cluster analysis was initially conducted on the diatom data . Similarity was measured using Euclidean distance to determine similarity in samples in chronological order . Results from the constrained cluster analysis a re shown in Figure 7 . The co nstrained cluster grouping s and v isual examination of diatom results w ere used to identify zones within the 1 Depth reported in centimeters below the sediment water interface after being linked to the LP15 and LP17 cores 2 This date was used to link the LP07 and LP15 cores but was omitted from the age depth model.

PAGE 50

39 sediment record . Cluster analysis identifies samples that are mo st similar to each other and places them into separate group s . The different groups identified do not carry more significance than others (Hammer et al., 2001) . Th e zones identified using the diatom cluster analysis were then compared to results from constrained cluster analysis using pollen data as a comparison of significant changes between groups of samples. The grouping of the final zones were comparable between the two datasets . A total of four zones were established for discussion of the results . Zone number, depth, and age are listed in Tabl e 3 . The other data will also be discussed in the context of these zones. Figure 7. Results from constrained cluster analysis using a) diatom data ; b) pollen data.

PAGE 51

40 Table 3 . Zones within the sediment record with respective dates and ages. Zone Depth (cm below sediment water interface) Age (cal yr BP) 1 115 92 5000 2400 2 92 71 2400 1600 3 71 30 1600 200 4 30 0 200 present Sediment Geochemistry Magnetic susceptibility & loss on ignition Results from magnetic susceptibility and LOI analysis are shown in Figure 8 . The magnetic susceptibility remains diamagnetic (< 0 cgs) through most of the bottom section of core (~5000 cal yr BP ~100 cal yr BP). Consistent paramagnetic trends (values > 0) are not recorded until ~100 cal yr BP (1850 CE) . Percent organic content in Zone 1 gradually increases from 4 1 % to 6 4 %. In Zone 2, organic content continues to increase to 7 1 %. Organic content reaches its maximum value (8 4 % ) in Zone 3 at ~337 cal yr BP . Organic content decreases towards the end of Zone 3 into Zone 4 , from 84 % to 60 % beginning around ~450 cal yr BP. XRF Results from XRF analysis are included in Figure 8 . T h e Fe/Ti ratio decreases through Zone 1 from 0.89 to 0.84 . The Fe/Ti ratio continues to decrease through Zone 2 down to 0.80. In Zone 3 , the Fe/Ti ratio increase s from 0.80 to 0.8 5, with a maximum value of 0.87 at 20 cm ( 82 cal yr BP ). The Fe/Ti ratio continues to increase in Zone 4 to 0.91 at 0 2 cm ( present day) , which i s the highest of the record . T he Mn/Ti ratio increase s slightly through Zone 1 from 0.81 to 0.84 and continue s to increase through Zone 2 to 0.8 8 . In Zone 3 the Mn/Ti ratio decreases from 0.88

PAGE 52

41 to 0.81, with a minimum value of 0.78 at 36 cm ( 672 cal yr BP ). The ratio continues to decrease through Zone 4 to 0.72 at 0 2 cm (present day), which is the lowest value of the record. Diatom s The diatom percentages are shown in Figure 8 and summarized in Table 4 . A total of 1 6 taxa were identified and used in analysis (Appendix E) , however this paper focuses on five taxa because of their high abundance throughout the record. The dominant diatom type in the Lily Pond record is from the Fragilariaceae family, including species of Fragilaria , Staurosi ra , Stauroforma , and Staurosirella . Other dominant benthic diatom genera present in the record in descending abundance include: Pinnularia , Stauroneis , Sellaphora , and Encyonema . These species are the focus of the discussion of the Lily Pond record because they are the most common taxa throughout the record , and represent both epipelic ( Pinnularia, Sellaphora, Stauroneis ) and unattached ( Encyonema ) environments . The diatom communities found in the sediment record were compared to the community found in modern day samples from the sediment water interface (0 cm depth), the water column, and surrounding aquatic vegetation. No planktonic taxa were found in the modern day samples from the water column. The samples from the water column were dominated by the Fragilariaceae family and Sellaphora . Diatom samples from the lily pad and sedge were also composed of Fragilariaceae valves as well as Tabellaria, Encyonema, Cymbe lla , Sellaphora, and Navicula , however , not in large abundances. Tabellaria , Navicula , Encyonema , and Cymbella are all epiphytic species, meaning they are often found growing attached to vegetation (Patrick, 1977). Sellaphora is considered epipelic, meanin g the species generally lives on sediments (Werner, 1977) . However, it is not uncommon for diatoms to adapt to different environments and

PAGE 53

42 survive in different habitats, especially in shallow lake environments where the open water envir onment is not dominated by planktonic species (Hellebust & Lewin, 1977; Patrick, 1977) . The community ha s been dominated by small benthic diatoms over the last 5000 years , which indicates a shallow lake level during the late Holocene at Lily Pond. Fragilariaceae are small, araphid species that are commonly found in aquatic systems located at high elevations ( Lotter et al, 1999 ). They can survive in both planktonic and benthic habitats, and are referred to as tychoplanktonic (Patrick , 1977 ). Pinnularia, Sellaphora, and Stauroneis are all epipelic taxa, meaning they prefer to grow attached to sediments. Encyonema is a benthic taxon that does not attach to anything . All taxa identified at Lily Pond are considered circumneutral. Pinnularia has also been found in environments with low pH (< 3.5) (Patrick, 1977 ; DeNicola, 2000 ) as well as iron rich environments (Wollmann et al , 2000) . Only two planktonic genera we re identified in the record , Aulacoseira and Cyclotella , and they were only found at abundances 6%. Both are centric diatom s, require high levels of nutrients to survive , and are sensitive to changes in turbidity (Patrick , 1977 ). These taxa were graphed together to analyze for changes in the planktonic habitat through the record . Zone 1 contains the lowest average abundance of Fragilariaceae (53%) and Encyonema (3%). Both groups gradually increase throughout the zone (to 79% and 22%, respectively) . Pinnularia , Stauroneis , and Sellaphora have the highest average abundance s ( 20%, 11%, and 6% , respectively ) of the record in Zone 1 and decrease (to < 4 %) through the zone. The a verage percent of plankton ic taxa is 0%, although the abundance increases to 1% towards the end of Zone 1 . Zone 2 is defined by increases in average abundance of Fragilariaceae (53% to 57%) and Encyonema , (3% to 4%) . Sellaphora decreases ( 6% to 1%) throughout Zone 2. Pinnularia and

PAGE 54

43 Stauroneis increase at the beginning of Zone 2 ( to 35% and 28%, respectively), and stay fairly consistent through the end of Zone 2. However the averages are lower than Zone 1 (from 40% to 10%, and 28% to 3%, respectively) . Planktonic taxa have the same average abundance as Zone 1 (1%). The transition from Zone 2 to Zone 3 is defined by an increase in Fragilariaceae, Sellaphora, and planktonic species from 42% to 70%, 0 to 4%, and 0 to 3%, respectively. Pinnularia, Stauro neis, and Encyonema all decrease at this same time from 26% to 2%, 17% to 1%, and 6% to 3%, respectively. Fragiliariaceae, Sellaphora , and planktonic species decrease throughout the MCA ( beginning around ~1200 cal yr BP) to 59%, 2%, and 2% respectively . Pinnularia , Stauroneis, and Encyonema then increase through the LIA, hitting maximum abundances of 19%, 8%, and 10%, respectively. Zone 4 is defined by an increase in Fragilariaceae specie s, Sellaphora, and planktonic taxa at the start of the Zone (to 8 0 % , 10%, and 6%, respectively). The highest average abundances o f Fragilariaceae (72%) and planktonic taxa (2%) occur in Zone 2 as well . Planktonic taxa follow similar trends as Zone 3 (lowest abundance 0%, highest abund ance 6%). Pinnularia, Stauroneis, and Encyonema decrease at the beginning of Zone 4 (to 1% ) , but increase around 82 cal yr BP (~20 cm) to abundances 2%.

PAGE 55

44 Figure 8 . Sediment geochemistry and diatom record for Lily Pond . Sediment geochemistry graphs show magnetic susceptibility (blue line) with a gray line showing 0 cgs, percent organics (green line), and XRF ratios (black shading). Diatom graphs (shown in blue) are percents of Fragiliariaceae, Pinnularia, Stauroneis, Enc yonema, and Sellaphora . The switch from summer wet to winter wet is shown by a blue dashed line; MCA is shown in red ; LIA is shown in blue ; mining period is shown in gray.

PAGE 56

45 Table 4 . Summary of diatom abundances sorted by zone. ZONE Fragilariaceae Pinnularia Stauroneis Encyonema Sellaphora Planktonic 4 Avg 72 6 2 4 4 2 High 97 19 8 17 18 6 Low 45 1 0 0 0 0 3 Avg 66 9 5 5 3 1 High 83 26 17 11 12 6 Low 42 3 0 1 0 0 2 Avg 57 21 15 4 1 0 High 79 35 28 6 3 1 Low 28 8 3 1 0 0 1 Avg 53 20 11 3 6 0 High 79 39 28 6 22 1 Low 22 4 1 0 0 0 Pollen Lily Pond p ollen percentages, accumulation rates, and ratio s of arboreal to nonarboreal pollen are shown in Figure 9 and summarized in Table 5 . All pollen analysis was conducted on the LP15 core. Pinus is the dominant pollen type through the record, followed by Picea , Artemisia , Quercus , Amaranthaceae, and Abies . There a few notable long term trends in the data . First, abundance of total Pinus in crease s ( 47.7% to 60.1 % ) from Zone 1 to Zone 4 . Second, average r elative abundances of Picea , Quercus , and Amaranthaceae decrease between Zone 1 and Zone 4 ( 17.9% to 8.8%, 5.2% to 0.5% , and 7.3% to 4.1%, respectively). Below more detailed changes of the dominant pollen types for each zone . Zone 1 has the lowest Pinus percentages of the record ( 47.7 %) , while Picea percentages are the highest ( 17.9 %). Abies percentages (1. 8 %), Quercus ( 3.4 %), Artemisia (13.3%) and Amarant haceae ( 7.3 %) abundances are also the highest of the core. The AP/NAP ratio was lowest in Zone 1 (0. 47 ) . Pollen accumulation rates are also the lowest of the record for both Nuphar and total pollen in Zone 1 ( 48.5 grains cm 2 and 1270.9 grains cm 2 , respectively).

PAGE 57

46 The transition from Zone 1 to Zone 2 is defined by a decrease in average abundance of Abies, Picea, and Quercus (from 1.8% to 1.5%, 27.9% to 13.4%, 3.4% to 1.6%, respectively) , and an increase in Pinus (from 47.7% to 59% average abundance) . Abundance of Artemisia and Amaranthaceae also decrease i n abundance (from 13.3% to 10.7%, and 7.3% to 5.3%, respectively) from Zone 1 into Zone 2. The AP/NAP ratio increases from 0.47 to 0.61 reflecting the increase in Pinus and decrease in nonarboreal sp ecies ( Quercus, Artemisia , and Amaranthaceae). The average total total pollen accumulation rate increases from 1270.9 grains cm 2 to 3078.2 grains cm 2 , while Nuphar also increases from 48.5 grains cm 2 to 142.2 grains cm 2 . Zone 3 has the highest average abundance of Pinus ( 6 4.5 %) , while Quercus and Amaranthaceae have the lowest abundance of the record ( 1.4% and 3.3%, respectively). Picea decreases (to 11%) Abies also decreases slightly (to 1.4%) from Zone 2 to Zone 3. The AP/NAP ratio reaches the high est value (0.64 ) of the record as well, representing the increase in Pinus abundance, and decrease in Quercus and Amaranthaceae . Pollen accumulation rates decrease between Zone 2 and Zone 3 for both Nuphar and total pollen (142 .2 grains cm 2 to 135.1 grains cm 2 and 3078.2 grains cm 2 to 2628.9 grains cm 2 , respectively). The transition from Zone 3 to Zone 4 is defined by a decrease in Pinus abundance ( 64.5% to 60.1%). Picea also decreases ( 11.0% to 8.8%) while Abies remains the same as Zone 3 (1.4%) . Quercus , Artemisia , and Amaranthaceae increase slightly ( from 1.4% to 1.7%, 11.1% to 14.2%, and 3.3% to 4.1 %, respectively). The AP/NAP ratio dropped in Zone 4 to 0.56, representing the decrease in Pinus an d Picea , and increase in Quercus, Artemisia , and Amaranthaceae. The pollen accumulation rates also increase dramatically during this period for both Nuphar (135.1 grains cm 2 to 700.1 grains cm 2 ) and total pollen ( 2628.9 grains cm 2 to 18488.8 grains cm 2 ). The AP/NAP ratio dropped in Zone 4 to 0.56 , similar to Zone 1 .

PAGE 58

47 Figure 9 . Pollen graphs for Lily Pond showing percentages for Pinus, Picea, Abies, Quercus , Artemisia , Amaranthaceae. Arboreal species are shown in green. Herbaceous and shrub species are shown in brown and yellow. The AP/NAP ratio is shown in black. Accumulation rates for Nuphar and total Pollen are shown in gray . The switch from summer wet to winter wet precipitation is shown by dashed blue line; MCA is shown in red ; LIA is shown in blue ; mining period is shown in gray.

PAGE 59

48 Table 5 . Summary of Lily Pond pollen rec ord organized by zones. PERCENTAGES RATIO ACCUMULATION RATE (grains cm 2 ) ZONE Pinus Picea Abies Quercus Artemisia Amaranthaceae AP/NAP Nuphar Total 4 Avg 60.1 8.8 1.4 1.7 14.2 4.1 0.56 700.1 18488.8 High 65.0 14.2 3.4 3.5 20.6 7.7 0.63 2019.3 58845.7 Low 52.9 5.6 0.0 0.5 9.8 1.1 0.39 0.0 2289.6 3 Avg 64.5 11.0 1.4 1.4 11.1 3.3 0.64 135.1 2628.9 High 70.9 14.0 2.3 2.7 13.1 5.0 0.75 239.6 3586.9 Low 57.8 8.0 0.7 0.0 8.4 2.1 0.57 48.8 1854.8 2 Avg 59.0 13.4 1.5 1.6 10.7 5.3 0.61 142.2 3078.2 High 65.9 17.8 1.9 2.7 12.1 7.0 0.70 190.6 4338.9 Low 53.0 10.3 0.6 0.7 8.9 2.2 0.51 72.4 2195.2 1 Avg 47.7 17.9 1.8 3.4 13.3 7.3 0.47 48.5 1270.9 High 62.2 25.6 4.2 5.2 14.4 10.7 0.59 109.4 1993.4 Low 34.5 9.8 0.5 1.2 10.7 3.7 0.39 0.0 595.0 Charcoal Charcoal accumulation rates (CHAR), background charcoal accumulation rates (BCHAR), peak magnitude , and fire return interval (FRI) are shown in Figure 10 , and summarized in Table 6 . All charcoal analysis was conducted on the LP15 core. A total of 20 fire events were identified in the last 5000 years. The average signal to noise index at Lily Pond is 5.1 ; however, there is a short period between ~2300 and 2000 cal yr BP where the SNI drops to levels below 3 . 0. Notable long term trends in the charcoal data include a n increase in average BCHAR from 0 .1 particles cm 2 yr 1 at the start of the record to 0 .3 particles cm 2 yr 1 in Zone 4 . BCHAR and fire activity increase (0.6 particles cm 2 yr 1 , 4 fire events, F R I average 138 yr fire 1 ) around ~1200 cal yr BP during the height of the MCA and drop back to low er BCHAR levels and fire activity ( 0.3 1.3 particles cm 2 yr 1 , 1 fire event, F R I average 276 yr fire 1 ) during the LIA . BCHAR increase s dramatically ~250 cal yr BP and the largest peak magnitudes ( 1.2

PAGE 60

49 particles cm 2 yr 1 , 50.6 particles cm 2 peak 1 ) occur thereafter . Below more specific trends in the charcoal record are discussed within individual zones. In Zone 1, BCHAR and peak magnitude were the lowest of the record. BCHAR had a maximum value of 0. 2 particles cm 2 yr 1 and a minimum of 0.0 particles cm 2 yr 1 , with an average of 0.1 particl es cm 2 yr 1 . Maximum peak magnitude in Zone 1 was 8.1 particles cm 2 peak 1 , with an average of 0.1 particles cm 2 peak 1 . A total of 8 fire events were identified in Zone 1, with five occurring after ~ 3000 cal yr BP. Around ~2 4 00 cal yr BP, the FRI decrease s from 2 83 yr fire 1 to 1 25 yr fire 1 ( average of 177 yr fire 1 ). In Zone 2 there is a n increase in CHAR accumulation from 0.1 to 0.3 particles cm 2 yr 1 . BCHAR also increases to an average of 0. 3 particles cm 2 yr 1 , with a minimum of 0. 2 particles cm 2 yr 1 and a maximum of 0 . 3 particles cm 2 yr 1 . A total of three fire events were identified in this zone. The m aximum peak magnitude in Zone 2 was 3.2 particles cm 2 peak 1 , with an average of 0.1 particles cm 2 peak 1 . The average FRI increases b etween Zone 1 and Zone 2 from 177 yr fire 1 to 205 yr fire 1 . Zone 3 is defined an increase in peak magnitude and a decrease in the FRI. CHAR had a maximum value of 1.6 particles cm 2 yr 1 and minimum of 0.1 particles cm 2 yr 1 , with an a verage of 0.4 particles cm 2 yr 1 . Average BCHAR rem ained the same as in Zone 2 (0.3 particles cm 2 yr 1 ) with a maximum value of 0. 4 particles cm 2 yr 1 and a minimum of 0.3 particles cm 2 yr 1 . Seven fire events were detected in Zone 3. There was a maxim um peak magnitude of 15. 4 particles cm 2 peak 1 , however the average was only 0. 7 particles cm 2 peak 1 . Average FRI decreased in Zone 3 to 1 89 yr fire 1 . However, Zone 3 includes both the MCA and LIA with in creased fire activity ( FRI 1 19 yr fire 1 ) and decreased fire activity ( FRI 2 96 yr fire 1 ) , respectively .

PAGE 61

50 CHAR and peak magnitude increase in Zone 4 . Averages were 1.2 particles cm 2 yr 1 and 15.9 particles cm 2 peak 1 , r espectively . CHAR and peak magnitude also experienced the highest values of the record in Zone 3. The highest CHAR value was 4.3 particles cm 2 yr 1 and the highest peak magnitude was 230.4 particles cm 2 peak 1 . BCHAR re ma ins consisten t (0.3 particles cm 2 yr 1 ), with a maximum value of 0.3 particles cm 2 yr 1 and a minimum value of 0.2 particles cm 2 yr 1 . Two fire e vent s w ere detected in Zone 4 around 87 cal yr BP and 49 cal yr BP . These two events have the highest peak magnitude s of the record. The av erage FRI increases during Zone 4 (from 189 yr fire 1 to 303 yr fire 1 ) . Figure 10 . Charcoal graphs for Lily Pond showing charcoal accumulation, peak magnitude, and fire return intervals. Gray line in CHAR shows BCHAR level. Red crosses designate peak fire events. Horizontal gray bars show missing data. The switch from summer wet to winter wet precipitation is shown with a dashed blue line; the MCA is marked with red shading; the LIA is marked with blue shading; and the mining period is marked with gray shading.

PAGE 62

51 Table 6 . Summa ry of charcoal record organized by zones. 1 SUM PEAK MAG is the peakMAG output from CharAnalysis. It is the sum of all samples exceeding the BCHAR threshold for a given peak. ZONE CHAR (particles cm 2 yr 1 ) BCHAR (particles cm 2 yr 1 ) SUM PEAK MAG 1 (particles cm 2 peak 1 ) 4 Avg 1.2 0.3 15.9 High 4.3 0.3 230.4 Low 0.2 0.2 0.0 3 Avg 0.4 0.3 0.7 High 1.6 0.4 15.4 Low 0.1 0.3 0.0 2 Avg 0.3 0.3 0.2 High 0.6 0.3 3.2 Low 0.1 0.2 0.0 1 Avg 0.1 0.1 0.1 High 0.5 0.2 8.1 Low 0.0 0.0 0.0

PAGE 63

52 CHAPTER VI DISCUSSION The following section outlines the late Holocene climate history of the Colorado Rocky Mountains and the surrounding region. The impacts of climate variations on the terrestrial and aquatic communities are then discussed. Finally, the introduction of Forest Hill Mine and its direct and indirect effects on the Lily Pond ecosystem are presented . Other paleo ecological and paleoclimatic records exist for su balpine environments throughout the Rocky Mountains and provide an oppor tunity to compare the record at Lily Pond to separate local from regional responses . Records from central Colorado (Mirror Lake, Colorado ; Keystone Ironbog, Col orado ) (Del Priore, 2015; Fall, 1997a) , norther n Colorado (Bison Lake, CO; Hidden L ake, CO; Tiago Lake, CO) (Anderson et al., 2015; Jimenez Moreno & Anderson, 2012; Shuman et al., 2009) , southern Colorado (Little Molas Lake, CO ; Toney & Anderson, 2006) , New Mexico (Sangre de Cristo Mountains ; Jimenez Moreno et al, 2008) , and southern Wyoming ( Little Brooklyn Lake, WY; E ast G lacier L ake, WY ) (Brunelle et al , 2013; Mensing et al , 2012) are used in the regional comparison (Appendix E) . Climate of the late Holocene (5000 cal yr BP to present) The climate of North America during the late Holocene is characterized by increased variability in temperature and precipitation due to decreasing summer insolation and sea surface temperatures (SST s) , which alter ed the strength and position of the North Americ an mid latitude jet stream (Mann et al., 2008 ; Kitzberger et al., 2007 ; Anderson, 2012; Jimenez Moreno & Anderson, 2012; Shuman et al., 2018; Shuman et al , 2014 ) . The changes in inso lation also resulted in intensified decadal and inter annual v ariability ( e.g., PDO, ENSO ) , cooler

PAGE 64

53 temperatures than during the early Holocene , and increased effective moisture (Anderson, 2012; Briles et al, 2012). Multiple proxies have been used to recon struct temperature and precipitation regimes in North America through the Holocene. For example, l ake levels from Emerald Lake (Colorado), Lake of the Woods (northwestern Wyoming), and Little Windy Hill Pond (southeastern Wyoming) were compared through the late Holocene using sediment stratigraphic changes across the lake (Shuman et al, 2014) . Results show that periods of high lake levels found at Emerald Lake correspond with periods of low lake levels at Lake of the Woods. There were no statistically significant correlations between Little Windy Hill Pond, although the record did differ, especially during the early Holocene where Emerald Lake did not show near modern lake levels that were found at the other two locations (Shuman et al, 2014). Thus, is it thought that a north south moisture dipole exists in the Rocky Mountains that impacts moisture delivery to these areas in the past . Spatial variations in El Nino Southern Oscillation (ENSO) are thought to be responsible for creating a north south moisture dipole along the CO WY border . Further, a shift in precipitation regime around ~2 4 00 cal yr BP has been identified in central Colorado through several multiproxy analysis. Calcite 18 O analyses from B ison Lake (northwestern Colorado) indicate a shift to a snow dominated precipitation regime in the region starting around ~2 4 00 cal yr BP (Anderson, 2012). Similarly , Hidden Lake near Steamboat Springs, Colorado, shows an increase in lake level beginning t hat peaked at ~2020 cal yr BP , which is consistent in timing with the switch to a winter dominated moisture regime . This climate shift is further supported by vegetation and fire record s in the region, with a decrease in subalpine forest density around ~ 2400 cal yr BP paired with a decrease in charcoal production (P. E. Higuera et al., 2014) . The switch to a winter dominated precipitation regime resulted in

PAGE 65

54 more snow accumulation in the winter months which melted during the spring and summer . The decrease in summertime moisture resulted decreased effective moisture for plant growth in the summer and caused a structural shift in forests and a change in fire regime. During the latter part of the late Holocene (past ~1300 year s) , c entral Colorado experienced two climate events known as the Medieval Climate Anomaly (MCA; 1200 cal yr BP 850 cal yr BP) and the Little Ice Age (LIA; 650 cal yr BP 10 0 cal yr BP). The MCA is defined as a period of warming with drought like conditi ons, where temperatures increased by ~0.5 C globally compared to the previous centuries. The increase in temperatures also resulted in higher frequency and severity of forest fires throughout the MCA (Calder et al, 2015) . The LIA is characterized by cooler temperatures (~0.26 C; Mann et al, 2009) than during the MCA and a return to wetter conditions. The climate differences bet ween these two periods are thought to be the result of natural radi ative forcing (i.e. solar variability and heightened volcanism ) that caused notable changes in the temperatures through out the Northern Hemisphere (Mann et al., 2009) . A strong enriched north/depleted south precipitation 18 O dipole pattern existed in the North American Rocky Mountains during both the LI A and MCA, suggesting that precipitation behavior was very complex ( Anderson, 2011; Anderson et al, 2016). T he complexity of past climate changes and the limited information on the impacts on terrestrial and aquatic mountain ecosystem s make high resolution late Holocene paleoecological record s important . These records help inform ecosystem managers about how Rocky Mountain forests might respon d to the rapid shift in climate we are seeing today and are predicted to experience into the next cen tury. In summary , the climate during the late Holocene was influenced by shifts in insolation and solar radiation that caused variation in precipitation and temperatures on various temporal

PAGE 66

55 and spatial scales. Specifically, Colorado shift ed to w inter dominated precipitation pattern ~2 4 00 cal yr BP with generally cooler temperatures than during the early Holocene , with a brief return to warmer drier conditions during the MCA and cooler and wetter conditions than today during the LIA . Paleoclimate and lake level s tudies indicate variability across Colorado to Wyoming , suggesting that the southern Rockies will not respond similarly to future climate fluctuations . Ecosystem Response to Climate during the late Holocene Switch from summer dominant to winter dominant precipitation regime (~2400 cal yr BP) Both the aquatic and forest ecosystem at Lily Pond responded to the climate regime shift at ~ 2400 cal yr BP . The aquatic ecosystem is composed primarily of benthic genera during the last 50 00 years suggesting that the re were high amounts of light availability in the littoral zone, and likely shallow lake level during the late Holocene . There are consistently low levels of Encyonema , which are an unattached, benthic species. There are also lo w amounts of Nuphar pollen during this time, suggesting a lack of vegetative habitat for these species with lake water levels that did not support Nuphar growth ( ~ 1 2 m depth ) . The high abundances of Fragilariaceae, Pinnularia, Stauroneis, Sellaphora, and planktonic species , indicate stable habitat and nutrient availability from ~5000 cal yr BP to ~2400 cal yr BP. At ~3100 cal yr BP, a few planktonic species are present, however the record is still dominated by benthic species. The data sugg est a lake level that allowed for high levels of light penetration to the benthic zone , while also supporting some planktonic species . The Mn/Ti ratio is consistent ly high through ~2400 cal yr BP, i ndicating no major variation in lake level. The Fe/Ti rati o is also high, indicating high levels of allochthonous inputs into Lily Pond. Magnetic susceptibility is paramagn etic, indicating low levels of metal bearing minerals in lake sediments, although there is a brief period when

PAGE 67

56 values are close to 0 , indicati ng a brief increase in iron bearing minerals into Lily Pond. Organic conten t is low er than today , likely indicating low er lake productivity. The highest abundance s of Picea and Abies occur prior to 2400 cal yr BP , suggesting open subalpine forest at Lily Pond. This section of the record also has the lowest AP/NAP along with low PAR , which suggests an open forest . The higher NAP species (e.g., those of Artemisia and Amaranthaceae ) and Quercus are likely blowing upslope from lower elevation. Fewer trees would allow a greater abundance of these pollen types to reach Lily Pond . Fire frequency is the lowest of the record from ~5000 cal yr BP to ~3000 cal yr BP likely due to the wetter summer conditions than today . Unlike today, the forest around Lily Pond probably was similar in composition as Mirror Lake today (~ 147 m in elevation higher than Lily Pond ; Del Priore , 2 015 ). Fires were infrequent and those that did burn consumed very little biomass. W et summer conditi ons likely supported more mesic conifers Picea and Abies and fewer fires at lower elevations than today . The Lily Pond diatom record s a transition from summer dominant to winter dominant precipitation regime around ~2400 cal yr BP. Pinnularia, Stauroneis, a nd Encyonema increase after the shift in precipitation, while Fragilariaceae and Sellaphora decrease, suggesting a change in habitat at Lily Pond. The decrease in Fragilariaceae and Sellaphora around this time suggest decreased planktonic habitat suitability, which could be a result of longer periods of ice cover. The Mn/Ti ratio also increases after ~2400 cal yr BP, indicated a possible decrease in lake level. The Fe/Ti ratio decreases after ~2400 cal yr BP, indicating an overall reductio n in allochthonous sedimentation into Lily Pond, perhaps due to a more closed forest (discussed below) . The switch to winter dominated precipitation regime may have also altered nutrient availability due to increased spring runoff, and lower amounts of sum mer precipitation. Magnetic susceptibility

PAGE 68

57 remains paramagnetic (<0) indicating low levels of iron bearing minerals present in the lake system. Organic matter increases gradually, indicating an increase in biological productivity. Pinus begins to increase around ~3100 cal yr BP, while Picea, Abies, Artemisia, and Amaranthaceae decrease. The AP/NAP ratio increases, mainly due to the increase in Pinus, indicating the forest became more closed. The pollen data indicate a more closed forest that resembles the modern mixed conifer forest currently at the transition between the montane and subalpine. Around ~3000 cal yr BP, fire frequency and biomass burned increased, likely due to the increase in Pinus ( more closed forest than before ) and the decrease in moistur e availability in the summer. The vegetation shift at Lily Pond a round ~2400 cal yr BP at Lily Pond is also recorded in other locations in Taylor Park and in the Rocky Mountains . Mirror Lake (3347 m elev; Del Priore, 2015 ) in Taylor Park and Keystone Iron bog ( 2920 m elev ; Fall 1997 ) near Crested Butte identified an abrupt change in vegetation from Picea dominated to Pinus dominated forest composition at ~2300 yr BP and ~2600 cal yr BP , respectively. Records from southwestern Colorado at Little Molas Lake (3370 m elev ; Toney & Anderson, 2005 ), and from the Sangre de Cristo Mountains in New Mexico (3100 m elev ; Jimenez Moreno etal, 2008 ) also record a similar abrupt change in vegetation a t ~2600 cal yr BP and ~2800 cal yr BP, respectively. In northern Colorado, Picea began to decline steadily at Bison Lake (3255 m elev. ) around ~2500 cal yr BP (Anderson et al, 2015). A composite analysis of pollen and charcoal records from three lakes in R ocky Mountain National Park (>3000 m elev; Higuera et al, 2014) document a shift around ~2400 cal yr BP to a more open forest structure indicated by increased Pinus abundances , and a decrease in fire activity and severity .

PAGE 69

58 The Tiago Lake record (2700 m el ev.), also in northern Colorado, records maximum Pinus at ~5000 cal yr BP. The differences in time of Pinus establishment at Tiago Lake may be However, it may al so indicate that moisture came mostly as snow in the winter rather than rain in the summer months through the entire late Holocene (Jimenez Moreno et al, 2011). Interestingly, vegetation records from southern Wyoming above 3000 m elevation also record hig h abundances of Pinus earlier than central Colorado to northern New Mexico records. For example, Little Brooklyn Lake (3153 m; Brunelle et al, 2013) and East Glacier Lake (3282 m; Mensing et al, 2012) record increases in Pinus that peak around ~5000 cal yr BP and ~5200 cal yr BP, respectively. The similarity between the Wyoming and Tiago Lake paleoecological records suggest the timing of the moisture transition may have occurred at least ~2500 years earlier than at the sites in Rocky Mountain National Park and those in central and southern Colorado. The difference in timing of the shift in Pinus across Colorado and Wyoming, and the correspondence with lake level differences, is likely related to the difference in delivery of moisture related to decadal and i nterannual climate variations that are currently in operation today (Shuman et al. 2014). The fire record at Mirror Lake, in Taylor Park just south of Lily Pond, suggests an increase in fire events leading up to ~2400 cal yr BP, like Lily Pond, followed by a decrease in composite CHAR suggesting decreased fire activity (Del Priore, 2015). Elsewhere in the Colorado Rocky Mountain s , records also identify decreased fire activi ty beginning between ~2600 to ~2300 cal yr BP , like trends in the Lily Pond record after ~2400 cal yr BP. In Rocky Mountain National Park, composite CHAR decreased after ~2400 cal yr BP, suggesting smaller fires with lower amounts of biomass burned until ~ 1500 cal yr BP ( Higuera et al, 2014 ). At Lake

PAGE 70

59 Tiago, there was also an increase in fire activity from ~3000 to ~2500 cal yr BP that then decreases until ~600 cal yr BP (Jimenez Moreno et al, 2011) . At Little Brooklyn Lake in Wyoming, fire activity is low from ~4900 cal yr BP to ~2000 cal yr BP likely due to a regional increase in moisture after ~5000 cal yr BP (Brunelle et al, 2013). The variation in the fire record, like the vegetation record, from Colorado to Wyoming suggests that the shift in precipitation regime occurred at different times. Medieval Climate Anomaly (MCA , ~1200 850 cal yr BP) & Little Ice Age (LIA, 650 100 cal yr BP) The change in summer to winter dominated precipitation at ~2400 cal yr BP set the stage for the development of modern forests and disturbance regimes. However , century long climate events, such as the MCA and LIA , caused brief changes in the terrestrial and aquatic communities at Lily Pond. The diatom record indicates decrease d Fragilariaceae species , and increase d Pinnularia and Stauroneis abundances , suggesting a change in habitat availability possibly due to cha n g es in lake level during the MCA. Sellaphora , planktonic species, and Enyconema remain ed consistent through the MCA. Nuphar rates were consistent through the MCA, suggesting low water levels that altered benthic habitat suitability . The Mn/Ti ratio increa se d throughout the MCA, indicating decrease d lake level. The Fe/Ti ratio decrease d during the MCA, indicating decrease d in allochthonous inputs into Lily Pond. The change in diatom community composition may have been a result of loss of open water habitat due to decreased lake level , and a change in nutrient availability due to decreased alloc hthonous inputs . Around ~1100 cal yr BP, Pinus decrease d while lower elevation steppe species ( Artemisia , Amaranthaceae) increase d . The AP/NAP ratio also decreases, signifying a more open forest likely due to increased burning during the MCA . The FRI is the lowest of the record and

PAGE 71

60 fire events are associated with larger charcoal peak magnitudes than seen in the earlier portion of the record. The increase in charcoal peak magnitude during the MCA may suggest more severe fires than were seen earlier in the Lily Pond record ; however, higher resolution pollen sampling and chemical analysis is needed to examine variability in fire severity . Fire frequency and sev erity declines through the MCA , likely a result of a decrease in burnable biomass. Increased fire frequency led to a decrease in Pinus and an increase in step pe species blowing in from downslope . The increase in fire activity during the MCA is likely due t o the warmer and drier conditions. The transition between the MCA and LIA (~850 cal yr BP ~650 cal yr BP) is defined by a decrease in the Mn/Ti ratio, suggesting an increase in lake levels. The Fe/Ti ratio also increases suggesting an increase in allochthonous inputs. Fragilariaceae, Sellaphora , and planktonic species increase at this time , while Pinnularia, Stauroneis, and Encyonema decrease. Nuphar also decreases after the end of the MCA, suggesting increased lake level that limited its growth. The increase in certain diatom species is likely in response to an increase in lake level that resulted in a more suitable open water habitat , while epiphytic and epipelic habitat suitability decreased. Pinus and Picea also increase d at the end of MCA in response to increased moisture and the decrease in fire activity . At the beginning of the LIA the percent organics increased, suggest ing increase d lake productivity. It could also indicate a decrease in allochthon ou s inorganic materials entering Lily Pond since the Fe/Ti ratio also decr eases. The diatom record indicates an increase in Pinnularia, Stauroneis, and Enyconema at the start of the LIA, while Fragiliaraceae, Sellaphora, and planktonic species decrease. At ~200 cal yr BP, there is an abrupt decrease in Pinnularia, Stauroneis, an d Enyconema, and an increase in Fragilariaceae, Sellaphora , and planktonic

PAGE 72

61 species . This shift indicat es a potential change in epiphytic and epipelic habitat suitability that may have been related to a change in nutrient inputs into the system . Both Pinus and Picea increased throughout the LIA while steppe species decreased. Fire activity during the LIA was comparable to that prior to ~2400 cal yr BP, with increase d FRI and decrease d charcoal peak magnitude. The AP/NAP ratio also increased, signifying a shi ft back to a more closed forest. There is one fire event at ~560 cal yr BP that ha d a relatively high peak magnitude in comparison to the rest of the record. The high peak magnitude may have been a result of a buildup of forest biomass due to more effectiv e moisture during the LIA . The diatom and geochemistry record at Lily Pond indicate that there were changes in nutrient and aquatic habitat availability during the MCA , specifically decrease d allocthonous inputs and a lower lake level , indicated by changes in sediment geochemistry . The changes in the aquatic habitat conditions are likely due to increased summer moisture, decreased ice cover, and warmer temperatures. Specifically, Fragiliariaceae species decreased through the MCA. Studie s have shown that changes in ice cover and increased nutrient cycling due to warming temperatures have caused small, fragilariod species to decrease in alpine lakes in Canada (Ruhland et al, 2008), suggesting that the increased temperatures at Lily Pond du ring the MCA may have had a similar effect. Hidden Lake in northern Colorado (2710 m) also shows a decrease d lake level during the MCA (Shuman et al, 2009), indicating that the water level at Lily Pond may have also decreased during this time. However, oth er lakes in the region ( e.g. Little Molas Lake, Southern Colorado, 3330 m ) did not record a similar lake level decrease during this time (Shuman et al, 2009). The Lily Pond diatom and geochemistry record s indicate a shift back to conditions seen prior to t he MCA at the start of the LIA, with a decreased Mn/Ti ratio, decreased Fragiliariaceae, and increased Pinnularia .

PAGE 73

62 The fire record at Lily Pond during the MCA and LIA is similar to ot her records through out the Colorado Rocky Mountain s , with an increase in burning at the start of the MCA followed by a decline through the LIA . For example, a meta analysis of fire histories from western United States lakes indicates increased burning at the beginning of the MCA that then decreased (Marlon et al, 20 12). The record s from Rocky Mountain National Park identified increases in fire beginning around ~1500 cal yr BP (Higuera et al, 2014). Other records in northern Colorado show similar increases in fire activity . At Bison Lake there was increased burning at the start of MCA (~1200 cal yr BP) , and an abrupt decline during the transition to the LIA at ~650 cal yr BP ( Anderson et al, 2015) . At Lake Tiago, fire activity increased around ~1400 cal yr BP, and then decreased around ~600 cal yr BP (Jimenez Moreno et al, 2001). The Lake Tiago record responded earlier than the Bison Lake record to the onset of the MCA, and then responded later to the onset of the LIA , suggesting spatial variations in fire response to periods of warming and cooling. Overall trends sh ow an increase in fire activity in the western United States after ~2000 cal yr BP and into the start of the MCA (Marlon et al, 2012). In Wyoming, the fire record at Little Brooklyn Lake and Little Windy Hill show increase s in charcoal beginning around ~ 2000 cal yr BP (Brunelle et al., 2013; Minckley et al., 2012) . The increase in burning in the western United States are likely a r esult of increase d drought like conditions and a high amount of biomass available to burn . The records from Wyoming indicate an increase in burning earlier (~500 years) than the Colorado records. The records from Colorado also show slight variations in timing. T he se temporal variations in timing are likely due to a combination of factors, including temperature, moisture availability, and forest composition, found at different elevations throughout the Wyoming and Colorado Rocky Mountains.

PAGE 74

63 Human Impact s : Logging , Mining, Grazing and Recreation In Colorado, the first substantial gold discovery was made at ~100 cal yr BP (~1850 CE), which sparked the start of mining activity in the state that continues in some areas today. This also corresponds with an increase in population in the western United States that is tied to the Colorado Gold Rush (Encyclopedia Staff, 2016). Agriculture and ranching also became popular around this time as a way to sustain economic growth (BLM, 2008) . The population increase in the western United States also resulted in changes in land use that impacted the surrounding ecosystems through changes in fire regime and forest structure (Marlon et al, 2006). The Forest Hill Mi ne (FHM) began operations near Lily Pond ~70 cal yr BP. The mine was closed by ~30 cal yr BP because it was not very successful and has had no remediation. There are remnants of old pipes, boilers, and other mining infrastructure scattered around the mine area (Figure 1 1 ). The mining footprint extends beyond the Forest Hill Mine area, with additional cabins, mine shafts, towers, and other infrastructure. The land around the main Forest Hill Mine area is still owned privately. Other mine remnants are owned b y the United States Forest Service (USFS). A dendrochronology study in the watershed indicates that logging operations may have been impacting the area up to ~100 years prior to the start of FHM operations (Appendix A).

PAGE 75

64 Ecosystem Response to Human Impacts Logging & Start of Forest Hill Mine (~70 30 cal yr BP ; 1880 1920 CE ) The diatom, geochemistry, pollen, and charcoal record from Lily Pond all suggest whole scale ecosystem response s since ~ 1 8 0 cal yr BP ( 1770 CE) , prior to the start of FHM operations (~70 cal yr BP; 1880 CE) . The Lily Pond record is compared with paleoclimate and human population data for the Northern Hemisphere in Figure 12 . Since these shifts predate the introduction of the mine, it suggests that other anthropogenic activity, likely logging, was taking place around Lily Pond prior to the mining . The Gunnison National Forest was not added to the Grand Mesa Uncompahgre and Gunnison (GMUG) National Forest until ~45 cal yr BP (1905 CE), so there are limited records available regarding logging activity before the mining . Lily Figure 11. Images of mining remains around Lily Pond. Photos: J. Steelman

PAGE 76

65 Pond records a shift in the diatom, sediment geochemistry, and pollen records around ~ 18 0 cal yr BP likely due to loggi ng . A preliminary dendrochronology analysis identified the earlies t widespread timber stand release date around 1777 CE , supporting the theory that anthropogenic activities began at Lily Pond prior to the FHM operations (Steelman, u npublished) . Timber harvest at Lily Pond prior to mining activity is indicated by increased magnetic susc eptibility, increased Fe/Ti ratios, decreased Mn/Ti ratio, and decreased organic content, suggesting increased allochthonous inputs, increased lake level, and decreased lake productivity, respectively. Fragilariaceae, Sellaphora , and planktonic species inc rease to high abundances at ~200 cal yr BP, while Pinnularia, Stauroneis and Encyonema decrease. An increase in lake level may have impacted light penetration to the benthic areas, resulting in decreased abundances of benthic diatoms. There is also a decre ase in Nuphar abundance, indicating a decrease in aquatic vegetation. Planktonic species, Fragilariaceae, and Sellaphora begin to decrease around ~150 cal yr BP, prior to the start of the mine, while Pinnularia and Encyonema slowly increase. The change at ~150 cal yr BP corresponds with an increase in Nuphar and an increase in the Mn/Ti ratio, suggesting decreased lake levels and increased amounts of aquatic vegetation. This shift overlaps with the end of the LIA, so the changes seen at ~150 cal yr BP may b e a combination of human impacts and increased summer drought conditions (Anderson et al, 201 5 ).

PAGE 77

66 Figure 12. Lily Pond environmental history, population growth in North America (HYDE 3.1; Goldewijk et al, 2010 ) , and insolation changes in North America

PAGE 78

67 The AP/NAP ratio also decreases around ~ 200 cal yr BP, along with all pollen species abundances . This decrease signifies an overall reduction in vegetation around Lily Pond. This may have been the result of timber harvest activity in relation to the FMH establishment. There is also a fire event at ~87 cal yr BP (1863 CE). The peak magnitude of this event is larger than the previous fire events identified in the record, suggesting a severe fire occurred at Lily Pond and may be the result of land clearance . In order to facilitate mine construction and opera tions, vegetation was likely harvested (arboreal species) or burned (nonarboreal species) to clear the surrounding landscape. There are remnants of log cabins and other wooden structures around the mine, likely constructed using the Picea and Pinus harvest ed in the area. There is a brief period of increase d Picea, Abies, and Quercus leading up to the start of FHM operations. The AP/NAP ratio also increase d briefly , suggesting a possible short term reestablishment of forest between logging and the start of m ining activity . The i n troduction of F HM (~70 cal yr BP; ~1880 CE) saw a n increase in magnetic susceptibility, where values become consistently diagenetic for the first time in the last 5000 years, suggesting increase d in iron bearing allochthonous inputs into Lily Pond. Since mining activity disturbs sediments through digging, increased metal content is likely due to runoff of loose sediments around Lily Pond. All diatom abundances increase slightly through the mining period. All diatom taxa present in the Lily Pond record are considered circumneutral, suggesting no sudden chan ges in pH with the introduction of the mine. This is further supported by the taxa similarities between the modern day diatom community, and those seen throug h the sediment core. Based on the diatom community structure during the mining period, FMH mining operations likely did not result in AMD that directly impacted Lily Pond. However, increased metal s were detected through magnetic susceptibility and XRF analyses , specifically increased

PAGE 79

68 Fe levels. The changes in diatom abundances are likely the result of disturbances in the terrestrial environment surrounding Lily Pond and nutrient availability due to increased sedimentation and allochthonous inputs. There is a decrease in Picea, Abies, Quercus, and steppe species at the start mining, indicating increased land clearing. Pinus abundance stays consistent through this period, however it is lower than prior to ~ 18 0 cal yr BP. The AP/NAP ratio also decreases indicating a more open forest . CHAR decreases after the fire event ~87 cal yr BP (1863 CE) and remains low through the time of FHM operations . The d ecrease in CHAR is likely due to absence of biofuels and anthropogenic fire suppression efforts since people were living in the area. The additional land clearing likely contributed to iron bearing miner als entering Lily Pond at the beginning of the mine. A reduction in surrounding vegetation, combined with sediment disturbance associated with mining activity, would result in increased amounts of sediment runoff during precipitation events, increasing allochthonous inputs into the aq uatic system. The anthropogenic changes at Lily Pond have been documented elsewhere . A study in the Uinta Mountains of Utah found that an increase in anthropogenic activity in the region impacted alpine lakes through atmospheric deposition, where even sm all amounts of metal and nutrient enrichment in lake systems were enough to affect the diatom communities (Moser et al., 2010). Similar results were seen in other alpine lakes in the Colorado Front Range, where increased nitrogen deposition in alpine areas was a result of increase agricultural practices in surrounding areas (Wolfe et al, 2001). Thus, Lily Pond may have also been impacted by other anthropogenic activities, including surrounding mining and agriculture, prior to the establishment of Forest Hil l Mine. Other fire records from the western United States indicate an increase in fire activity during the mid to late 1800s AD (~150 cal yr BP), with highest levels between 1850 1870 AD

PAGE 80

69 (~100 80 cal yr BP) (Marlon et al., 2012). The decrease in CHAR in th e Lily Pond record is likely a result of anthropogenic land management techniques given the high levels of anthropogenic activity in the area (Marlon et al., 2012). Termination of Forest Hill Mine ( > ~30 cal yr BP ; 1920 CE ) & Present Day Activities The FHM operations ended at ~30 cal yr BP (1920 CE). There was a brief break in mining activity from ~43 to ~34 cal yr BP (~9 years; 1907 1916 CE), however the break was not detected by any of the proxy data at Lily Pond. Since closure of the FHM, other an thropogenic activities such as recreation, wildfire suppression, and grazing, have increased around Lily Pond. There is a system of 4x4 and ATV trails that run around Lily Pond, the Forest Hill Mine, and the surrounding area, that are very popular during t he summer months. Additionally, the USFS allows horse and cattle grazing in the forest around Lily Pond (United States Forest Service, 2018). Since mine closure, sediment geochemistry indicates a continued input of iron bearing minerals into Lily Pond than was seen earlier in the record. The Mn/Ti ratio decreases, suggesting increased lake level s . The diatom record shows a decrease in Fragiliariaceae, Sellaphora , and planktonic species, while Pinnularia, Stauroneis, and Encyonema increase. These changes ind icate a possible period of increased nutrient availability immediately following the closure of Forest Hill Mine. This may be due to increase d runoff, and therefore nutrients entering the system due to limited vegetation. It could also be indicative of a c hange in habitat availability due to an increase in lake level. During the last ~40 years, Pinnularia , Sellaphora and planktonic species increased , while Fragilariaceae species decreased, suggesting a possible increase in nitrogen at Lily Pond. While some species of Fragiliariaceae are tolerant of increased

PAGE 81

70 nitrogen levels, smaller species like the ones found at Lily Pond, have been found to decrease as nitrogen levels increase (Saros et al, 2003) . The pollen record shows a decrease in the AP/NAP ratio immediately after mine closure, indicating a more open forest . Quercus, Artemisia, and Amara nthaceae increased immediately after the termination of mine activity , also indicating an open forest allowing for the deposition of pollen taxa from downslope , while Picea increased slowly. The diatom and pollen record show that the species present in bot h the terrestrial and aquatic communities did not change with the introduction of the mine, and that these species continue to be present at Lily Pond after mine termination. While relative abundances did shift during the mining period, the shifts were not enough to cause either system to shift in ways that were comparable to those of the MCA or LIA . The presence of mine remnants, including mine tailings, do not seem to be continuing to alter the terrestrial and aquatic environments at Lily Pond. There is a decrease in Pinus , which may be a result of a fire event identified at ~2 cal yr BP (~1948 CE). The fire event is associated with the largest charcoal peak in the Lily Pond record, indicating a substantial fire in comparison to those recorded during the past millennia . The fire recorded in the lake sediments is likely the one that left burn scares on trees on the south west side of the watershed. Visual evidence of a recent fire was recorded in th is area (Steelman, unpublished). Additional tree coring my reveal the exact timing of this event beyond that observed in the lake sediment record . The modern forest composition is similar to that recorded prior to the mine and logging with increased amount s of Pinus and increased AP/NAP ratios within the last 4 0 years , suggesting a more open forest that is dominated by lodgepole pine .

PAGE 82

71 Since 1950, other diatom records from alpine lakes in the Colorado Front Range have identified shifts in diatom communitie s due to increases in anthropogenic activity. Specifically, a study on mountain lakes in the Colorado Front Range found that an increase in agriculture within the region resulted in increased n itrogen levels in the adjacent mountain areas . The increased nitrogen levels resulted i n increased abundances of opportunistic diatom species (Wolfe et al., 2001). Given the sensitive nature of diatoms, even small increases in nutrient inputs into lake systems, especially through atmospheric deposition, is enough to alter diatom communit ies . However there are several other factors that are believed to impact community response to nitrogen. A study in northwestern Wyoming suggests that b enthic species in shallow lakes are more likely to be impacted by light availability than nutrients , so the effects of increased nitrogen levels vary betw een systems (Spaulding et al, 2015) . Environmental sensitivity of diatoms is species specific , which is why some species respond more strongly than others to changes in their environments . Since this study only identified the community to the genus level, species sensitivity cannot be determined . Fluctuations in g enera abundance s throughout the last 200 years suggest that the diatom species found in this environment less sensitive to nutrient changes and are representative of changes in habitat suitability . Over the last century, there has been a decrease in biomass burned as well as fire frequency due to suppression efforts throughout the western United States (Marlon et al., 2012). The decrease in burning is counter to what is expected give n the warmer and drier conditions than seen during the LIA. It is likely that anthropogenic factors such as grazing and land management, has limited forest bio mass needed to carry and sustain fires . In areas where land management has not been as extensive, fire frequency has increased through the 20 th century (Marlon et al., 2008) .

PAGE 83

72 Land Management Implications Management of lands in the western U nited S tates is a complex issue that must take both climate and anthropogenic activities into consideration. The research presented in this thesis can provide insight s about how terrestrial and aquatic environments in a subalpine forest in central Colorado respond to c limate conditions as well as anthropogenic activity. The 21 st century has experienced warmer surface , atmospheric, and ocean temperatures ( ~0.85 C ; IPCC, 2014) , especially over the last 30 years , with a substantial contribution coming from anthropogenic activity . Colorado annual avera ge temperatures have increased by 1 C in the last 50 years (Lukas et al, 2014) . M odels predict continued warmi ng globally throughout the remainder of the 21 st century, with increases of over 1 C expected (Alexander et al., 2018; IPCC, 2014) . In Colorado, future warm ing is proj ected to increase by up to 3 C by 20 50 , with summer temperatures projected to increase more than winter temperatures (Lukas et al, 2014). Precipitation trends for Colorado indicate increased winter extreme precipitation events, but do not necessarily show the same change for summer precipitation (Lukas et al, 2014). Climate projections also predict increased frequency and severity of heat waves, droughts, and wildfires in Colorado by the mid 2 1 st century due to the warming climate (Lukas et al, 2014). There is still uncertainty as to how forests will respond to the increase in temperature, making it difficult for forest managers to plan appropriately. However, the paleoecological and paleolimno logical records provide insights. The recent record at Lily Pond (past ~200 years) indicates conditions similar to those seen during the MCA. The simi larities suggest that climate drivers, specifically changes in timing of precipitation and drought severit y, impact the terrestrial and aquatic communities at Lily Pond more than anthropogenic activities. There has been a decrease in precipitation and

PAGE 84

73 increase in temperatures in the western United States (including Colorado) beginning in the late 20 th century , resulting in increased drought conditions during the summer months (Knowles et a l , 2006; Mote et al , 2005) . Continued warming in mountain catchments, both in summer and winter, will result in overall decrease in annual runoff , which could impact temperature, lake level, and nutrient availability (Dierauer, Whitfield, & Allen, 2018) . If warming continues to increase through the 21 st century as predicted , it is likely to have greater impact on the terrestrial and aquatic communities than the MCA. G iven the current warming trends of the 21 st century, forest composition and structure will likely continue to forests and fire regimes seen in the MCA with a mo re closed Pinus dominated forest with higher fire activity due to periodic droughts due to increased temperatures during summer months (Calder et al., 2015; Marlon et al., 2012) . However, temperat ures are expected to exceed those of the MCA . If the temperature increase result in a shift to more summer precipitation, conditions like those prior to those seen prior to 2 4 00 cal yr BP may pr evail. However, current future projections do not indicate inc reased summer precipitation. Therefore, t ree line could shift to a higher elevation under warmer drier summer conditions . Additionally, anthropogenic fire suppression efforts ha ve increased forest biomass in some forests at the local level , which will like ly result in more severe burns when they do occur . Temperature and precipitation are also likely to spatially vary from Colorado to Wyoming and forest managers need to keep in mind that forest management practices needs be flexible to account for the climate gradient. The terrestrial and aquatic systems at Lily Pond are under pressure from combined effects of climate change and anthropogenic changes that have altered the function of the system throughout the late Holocene. Beyond managing for climate variations in the region, Colorado

PAGE 85

74 has thousands of AMLs scattered though the state. Abandoned mines can continue to impact the surrounding syst ems even after activity has stopped, specifically due to high amounts of acid mine drainage leaving the mine (Brugam & Lusk, 1986) . S ince runoff from Forest Hill Mine does not drain directly into Lily Pond, is it unlikely that remediation of that specific site would have an impact on the aquatic conditions. However, it could be impacting water quality in a different area downstream . Further research needs to be do ne to assess that possibility and make recommendations specific to Forest Hill Mine remediation. The 2007 forest management plan states that AMLs in the GMUG National Forest areas will not be remediated unless they are significantly d egrading water quality o r are a safety hazard to the public (Un ited States Forest Service, 2007) . Based on th e current management plan , combined with the results from this research, the Forest Hill Mine is likely not a priority for remediation with regard impacts to the Lily Pond watershed . While the diatom community at Lily Pond has maintained the same genera composition, abundances have fluctuated throughout the record, especially in response to terrestrial disturbances. Therefore, it is important for land manag ers to continue to monitor aquatic conditions, to determine how future logging, grazing and recreational pressures will impact the Lily Pond conditions and habitat. Limitations One limitation for this study is the unconsolidated nature of sediments in the top ~20 cm of the core , and the amount of microscope time required t o produce a high resolution multi proxy record . The limited number of samples for the mining period make it difficult to assess how the brief break in mining activity impacted the aquatic and terrestrial systems, so the amount of time for the system to recover could not be assessed. It would also be beneficial to add additional diatom and pollen samples in the lower portion of the core ( between 70 115 cm) to

PAGE 86

75 better characterize the summer dominated precipitation regime s impact on terrestrial and aquatic ecosystems . Another limitation of this study was the inability to use an electron microscope to identify diatoms down to species level. While genus was sufficient to gain insight about the c ommunities throughout the Lily Pond record, it would have been helpful to be able to distinguish between the different species , specifically the small araphid Fragiliariaceae species, present in the sediment record. It is possible to identify species with light microscopy; however the majority of diatoms present in each sample were very small (<10 microns) making it difficult to identify down to species level given the equipment available. One final limitation of this study is the lack of information about activity related to the Forest Hill Mine. There are very limited records regarding the history about mine establishment and construction, as well as possible timber harvest activity related to it . This makes it difficult to understand the extent of distur bance as it relates to the paleoecological record . Since Forest Hill Mine is not located within the same watershed as Lily Pond, the direct impacts from mining activity are likely a result of timber harvesting that may or may not have been related to the mine . For example, other natural disturbances, spe cifically wind, can blow down large spans of forest. The dendrochronological study has helped with an understanding of human vs natural disturbances in the Lily Pond watershed ; however, further research is needed in the area . Future Research A goal of this thesis was to determine how aquatic systems responded to and recovered from anthropogenic disturbance events and climate change . The Lily Pond record has proved to be a valuable location to reconstruct paleoenvironmental conditions . All three cores recor d have shown to have similar stratigraphic tends and radiocarbon dates across cores correspond. The site

PAGE 87

76 is well dated through the last glacial period ; however, there is a gap in dating between 4000 6000 years . An additional date would strengthen the chron ology. Higher resolution pollen and diatom data is also needed between 2 000 5000 cal yr BP to characterize the summer dominated precipitation regime. Additional proxies could be added to the Lily Pond record to better understand the local changes that we re happening. Lake level analysis would be helpful in further evaluating the impacts of climate variations on Lily Pond. There is visual evidence of different lake shores around Lily Pond, especially along the south part of the catchment area. Lake level c ould be reconstructed by taking a series of cores through the area and examining core lithology and sediment geochemistry. To further assess the impacts of anthropogenic impacts around Lily Pond, fecal stanol and dung spore analyses would be beneficial. Fecal stanol analysis wo uld help identify when anthropogenic activity started impacting Lily Pond, and dung spores would help determine when grazing activity may have begun in the area and its extent . A dditional dendrochronology sampling could provide a more complete understandin g of the age of the current forest and would provide insight on fire event s identified in the record after the termination of the FHM. Lastly, a hydrological model would be useful to understand how runo ff around the mine influences the Lily Pond. To bette r understand how anthropogenic disturbances may have influenced the terrestrial and aquatic systems at Lily Pond, additional information regarding the start of mining and agriculture in the region is needed. The introduction of anthropogenic activities, su ch as mining, have been shown to impact subalpine and alpine systems indirectly through atmospheric deposition (Moser et al, 2010; Saros et al , 2003; Wolfe et al , 2001) . Therefore, it would be beneficial to research nitrogen and phosphorus levels throughout the Lily Pond core. It would

PAGE 88

77 also be helpful to map wind patterns around Lily Pond to better understand how atmospheric deposition may have impacted the area during the late Holocene. Additional studies that examine multiple proxies thr oughout the Rocky Mountains would be beneficial in understanding variations along the north south precipitation gradient . Given the sensitivity of forests and fire regimes historically, additional paleoecological rec ords between 38 4 3 ° N latitude across the western United States c ould help ecosystem models better predict future climate change responses. Lily Pond currently lies at an ecotone between the montane and subalpine forest ecosystems and provides a historical perspective on t he rate and trajectory ecosystems respond to abrupt climate change events. Additional sites in the Taylor Park and Southern Rocky Mountains on these gradients will help better estimate forest and aquatic response to present day and future rapid climate cha nges.

PAGE 89

78 C HAPTER VII CONCLUSION In conclusion, the Lily Pond paleo record provides insight s into how subalpine and montane forests in c entral Colorado are impacted by climate variations and anthropogenic activity (i.e. mining, logging, recreation) during the late Holocene . This study used multiple proxies including diatoms, sediment geochemistry, pollen, and charcoal to reconstruct both terrestrial and aquatic conditions during the last 5000 years , with specific focus on how a period of known mining activity nearby influenced the ecosystem . The research focused on the following two questions: (1) How did the terrestrial and aquatic environments at Lily Pond respond to climate variations during the late Holocene? The aquatic and terrestrial sy stems at Lily Pond w ere influenced by regional climate shifts that impacted forest and diatom community composition . Precipitation patterns shifted around ~2400 cal yr BP from summer to winter dominated precipitation . Temperatures also decreased, resulting in increased ice cover , increased spring runoff, and shorter growing seasons based on the increase in Pinus in the pollen record. The forest initially was asubalpine forest dominated by Picea and Abies during a period when precipitati on fell mainly in the summer. However, forest structure responded to the change to a winter wet precipitation regime at ~2400 cal yr BP, which was a result of decreasing summer insolation seen throughout the late Holocene. This switch resulted in an overal l decrease in net moisture availability and resulted in a switch from a more open canopy structure to a more closed forest canopy dominated by Pinus . F ire activity at Lily Pond also intensified at ~2400 cal yr BP with the switch in precipitation regime . The increase in

PAGE 90

79 fire activity was due to decreased net moisture and increased drought like conditions during the summer months. Temperatures increased, and precipitation decreased around ~1200 cal yr BP with the on set of the MCA that lasted until ~8 5 0 cal yr BP. Conditions returned to high amounts of winter precipitation and cooler temperatures during the LIA from ~650 to ~100 cal yr BP . The f orest around Lily Pond continued to be an open canopy and was dominated by Pinus through the MCA . It switched back to a more closed forest canopy when moisture availability increased again during the LIA. Fire frequency increased at the start of the MCA due to increased temperature and drought conditions, before decreasing in the LIA due to cooler temperatures . The Lily Pond diatom community remained primarily benthic through the record. Planktonic species were present beginning around ~3100 cal yr BP, however they remained at small abundances (<6%) throughout the record. Spe cies abundances fluctuated in response to changes in lake level (Mn/Ti ratio) and allochthonous inputs (Fe/Ti ratio), indicating that habitat (influenced by lake levels) and nutrient availability were primary drivers in determining the diatom community com position. Organic matter increased through the core until ~200 cal yr, indicating that anthropogenic disturbance may have had a larger impact on system productivity than climate change. Magnetic susceptibility remained paramagnetic (<0) until ~100 cal yr B P and magnetic thereafter, suggesting iron bearing inputs of allochthonous materials likely from mining activity (e.g., logging and mine tailings). (2) How d i d the introduction and termination of mining activity affect the aquatic and terrestrial environm ents? The Lily Pond record indicates that anthropogenic activity began prior to the actual start of mining and impacted both the terrestrial and aquatic communitie s. The dendrochronology

PAGE 91

80 record at Lily Pond indicated that the surrounding vegetation was cle ared, resulting in increa sed runoff into the aquatic environment that is reflected in the sediment geochemistry and diatom records. It is not clear if all timber harvest during this period was directly related to the establishment of FHM, but the remnant living structures and boiler indicate that some harvest may have been used to build mining infrastructure, as well as for burning. There was also an increase in CHAR immediately before the start of FHM operations, which potentially indicates the us e of fire to clear the landscape. This is followed by a decrease in CHAR through the mining period , likely a result of anthropogenic fire suppression . Since termination of the mine, the sediment record shows a return to both a diatom community and forest present prior to the introduction of the mine . Pinus dominate d the landscape and Fragilariaceae species were abundant in the diatom community . The diatom community seems to have been influenced by changes in both habitat and nutrient availability throughou t the record in response to changes in land cover and precipitation patterns. It does not appear that Lily Pond is being impacted by AMD from the exposed minerals and remaining mine tailings ; however , this could be impacting water resources in a different area. Although mining operations have ended, anthropogenic activity , mainly recreation, is still present at Lily Pond. The Lily Pond record presented in this thesis provides information about how both the terrestrial and aquatic communities respond to both climatic and anthropogenic events specific to the late Holocene. Reconstructing conditions at Lily Pond during the late Holocene using multiple proxies provides an additional record for comparison within the southern Rocky Mountains. The terrestrial a nd aquatic responses to climate variations and anthropogenic activities at Lily Pond are similar to other records in Colorado, indicating the widespread impacts ; however , the timing of past change varies along a latitudinal gradient . Paleoecological data p laces more recent trends in the context of long term variability and provides land

PAGE 92

81 management strategies in subalpine forests in the western United States. This is especially important given the warming temperatures and anthropogenic activity that have occ urr ed over the last several decades , and which are expected to continue in the future . The combined records of both terrestrial and aquatic change provide the necessary context for understanding how Rocky Mountain ecosystems should be managed to ensure continued ecosystem function and health.

PAGE 93

82 REFERENCES Alexander, M. A., Scott, J. D., Friedland, K. D., Mills, K. E., Nye, J. A., Pershing, A. J., & Thomas, A. C. (201 8). Projected sea surface temperatures over the 21 st century: Changes in the mean, variability and extremes for large marine ecosystem regions of Northern Oceans. Elem Sci Anth , 6 (1), 9. https://doi.org/10.1525/elementa.191 Anderson, L. (2011). Holocene re central Rocky Mountains, United States. Geology , 39 (3), 211 214. https://doi.org/10.1130/G31575.1 Anderson, L. (2012). Rocky Mountain hydroclimate: Holocene variability and the role of insolat ion, ENSO, and the North American Monsoon. Global and Planetary Change , 92 93 , 198 208. https://doi.org/10.1016/j.gloplacha.2012.05.012 Anderson, L., Brunelle, A., & Thomspon, R. S. (2015). A multi proxy record of hydroclimate, vegetation, fire, and post s ettlement impacts for a subalpine plateau, central Rocky Mountains, U.S.A. The Holocene , 25 (6), 932 943. E. (2017). Suspended sediment and turbidity after road construction/improvement and forest harvest in streams of the Trask River Watershed Study, Oregon. Water Resources Research , 53 (8), 6763 6783. https://doi.org/10.1002/2016WR020198 Ballantyne, A. P., Brahney, J., Fernandez, D., Lawrence, C. L., Saros, J., & Neff, J. C. (2011). Biogeochemical response of alpine lakes to a recent increase in dust deposition in t he Southwestern, US. Biogeosciences , 8 (9), 2689 2706. Battarbee, R. W., Charles, D. F., Bigler, C., Cumming, B. F., & Renberg, I. (2010). Diatoms as indicators of surface water acidity. In J. P. Smol & E. F. Stoermer (Eds.), The Diatoms: Applicaitons for the Environmental and Earth Sciences, 2nd Edition (2nd ed., pp. 98 121). Cambridge University Press. Battarbee, Richard W. (1999). The importance of palaeolimnology to lake restoration. Hydrobiologia , 395 396 (0), 149 159. https://doi.org/1 0.1023/A:1017093418054 Blaauw, M., & Christen, J. A. (2011). Flexible paleoclimate age depth models using an autoregressive gamma process. Bayesian Analysis , 6 (3), 457 474. https://doi.org/10.1214/ba/1339616472 Berger, A., & Loutre, M. F. (1991). Insolatio n values for the climate of the last 10 million years. Quaternary Science Reviews , 10 (4), 297 317. https://doi.org/10.1016/0277 3791(91)90033 Q

PAGE 94

83 BLM. (2008). Ranching and Agriculture: The Economic Diversification of Southwestern Colorado. In Frontier in Tra nsition: A History of Southwestern Colorado . Retrieved from https://www.nps.gov/parkhistory/online_books/blm/co/10/chap9.htm Briles, C. E., Whitlock, C., & Meltzer, D. J. (2012). Last glacial interglacial environments in the southern Rocky Mountains, USA a nd implications for Younger Dryas age human occupation. Quaternary Research , 77 , 96 103. Brugam, R. B., & Lusk, M. (1986). Diatom evidence for neutralizaiton in acid surface mine lakes. Diatoms and Lake Acidity , 115 129. Brugger, K. A. (2006). Late Pleisto cene climate inferred from the reconstruction of the Taylor River glacier complex, southern Sawatch Range, Colorado. Geomorphology , 75 (3), 318 329. https://doi.org/10.1016/j.geomorph.2005.07.020 Brunelle, A., Minckley, T. A., Lips, E., & Burnett, P. (2013) . A record of Lateglacial/Holocene environmental change from a high elevation site in the Intermountain West, USA. Journal of Quaternary Science , 28 (1), 103 112. https://doi.org/10.1002/jqs.2600 Burtt, D. G., Addison, J. A., & Foster, A. (2013). Reconstruc ting Holocene fluvial dischage in the marine environment near Eureka, California, using a portable XRF method. Calder, W. J., Parker, D., Stopka, C. J., Jiménez Moreno, G., & Shuman, B. N. (2015). Medieval warming initiated exceptionally large wildfire out breaks in the Rocky Mountains. Proceedings of the National Academy of Sciences , 112 (43), 13261 13266. https://doi.org/10.1073/pnas.1500796112 Colorado Division of Reclamation Mining & Safety. (2014). Colorado Division of Reclamation Mining & Safety. Retrie ved March 10, 2017, from http://mining.state.co.us/Programs/Abandoned/Pages/impwelcomepage.aspx Colorado Geological Survey. (2013, June 20). Abandoned Mine Lands. Retrieved February 26, 2017, from http://coloradogeologicalsurvey.org/mineral resources/aband oned mine lands/ Colorado Water Quality Control Division. (2016). Gold King Mine spill; Animas River Basin, Southwest Colorado, August 2015 . Department of Public Health and Environment. Retrieved from https://www.colorado.gov/pacific/sites/default/files/Go ld King Mine Spill Report 01 22 16 Digital.pdf Currie, L. A. (2004). The Remarkable Metrological History of Radiocarbon Dating [II]. Journal of Research of the National Institute of Standards and Technology , 109 (2), 185 217.

PAGE 95

84 Dean, W. E. (1974). Determinati on of carbonate and organic matter in calcareous sediments and sedimentary rocks by loss on ignition; comparison with other methods. Journal of Sedimentary Research , 44 (1), 242 248. https://doi.org/10.1306/74D729D2 2B21 11D7 8648000102C1865D Del Priore, T. (2015). 4000 years of environmental change in central Colorado: A paleoecological perspective . University of Colorado Denver. DeNicola, D. M. (n.d.). A review of diatoms found in highly acidic environments. Hydrobiologia , 433 (1 3), 111 122. https://doi.or g/10.1023/A:1004066620172 Dierauer, J. R., Whitfield, P. H., & Allen, D. M. (2018). Climate Controls on Runoff and Low Flows in Mountain Catchments of Western North America. Water Resources Research , 54 (10), 7495 7510. https://doi.org/10.1029/2018WR023087 Doesken, N. J., Pielke, Sr., R. A., & Bliss, O. A. P. (2003). Climatography of the United States No. 60. Retrieved from http://climate.colostate.edu/climate_long.html Faegri, K. (1989). Textbook of Pollen Analysis (4th ed.). Chichester: John Wiley & Sons. Fall, P. L. (1997a). Fire History and Composition of the Subalpine Forest of Western Colorado During the Holocene. Journal of Biogeography , 24 (3), 309 325. Fall, P. L. (1997b). Timberline fluctuations and late Quaternary paleoclimates in the Southern Rocky Mountains, Colorado. GSA Bulletin , 109 (10), 1306 1320. Fleischner, T. L. (1994). Ecological Costs of Livestock Grazing in Western North America. Conservation Biology , 8 (3), 629 644. Gale, S. J. (2009). Dating the recent past. Quaternary Geochronology , 4 (5), 374 377. https://doi.org/10.1016/j.quageo.2009.05.011 Goldewijk, K. K., Beusen, A., & Janssen, P. (2010). Long term dynamic modeling of global population and built up area in a spa tially explicit way: HYDE 3.1. The Holocene , 20 (4), 565 573. https://doi.org/10.1177/0959683609356587 Hammer, O., Harper, D. A . , & Ryan, P. D. (2001). PAST: Paleontological statistics software package for education and data analysis. Palaeontologica Electronica , 4 (1), 9pp. Hellebust, J. A., & Lewin, J. (1977). Heterotropgic Nutrition. In D. Werner (Ed.), The Biology of Diatoms (pp. 169 197). Los Angeles, California: University of California Press.

PAGE 96

85 Heyerdahl, E. K., Brubaker, L. B., & A gee, J. K. (2001). Spatial Controls of Historical Fire Regimes: A Multiscale Example from the Interior West, USA. Ecology , 82 (3), 660 678. https://doi.org/10.2307/2680187 Higuera, P. (2009). CharAnalysis 0.9: Diagnostic and analytical tools for sediment ch arcoal analysis. ResearchGate . Retrieved from https://www.researchgate.net/publication/254406379_CharAnalysis_09_Diagnostic_and_ analytical_tools_for_sediment charcoal_analysis Higuera, P. E., Briles, C. E., & Whitlock, C. (2014). Fire regime complacency an d sensitivity to centennial through millennial scale climate change in Rocky Mountain subalpine forests, Colorado, USA. Journal of Ecology , 102 (6), 1429 1441. https://doi.org/10.1111/1365 2745.12296 Higuera, P. E., Brubaker, L. B., Anderson, P. M., Hu, F. S., & Brown, T. A. (2009). Vegetation Mediated the Impacts of Postglacial Climate Change on Fire Regimes in the South Central Brooks Range, Alaska. Ecological Monographs , 79 (2), 201 219. Hundey, E. J., Moser, K. A., Longstaffe, F. J., Michelutti, N., & Hla dyniuk, R. (2014). Recent changes in production in oligotrophic Uinta Mountain lakes, Utah, identified using paleolimnology. Limnology and Oceanography , 59 (6), 1987 2001. https://doi.org/10.4319/lo.2014.59.6.1987 IPCC. (2014). Climate Change 2014: Synthesi s Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. (p. 151). IPCC, Geneva, Switzerland. Jimenez Moreno, G., & Anderson, R. S. (2012). Pollen and macrofossil evidence of La te Pleistocene and Holocene treeline fluctuations from an alpine lake in Colorado, USA. The Holocene , 23 (1), 68 77. Jimenez Moreno, G., Anderson, R. S., & Toney, J. L. (2011). A high resoultion record of climate, vegetation, and fire in the mixed confier f orest of northern Colorado, USA. GSA Bulletin , 123 (1/2), 240 254. Johnson, F. (1965). Half Life of Radiocarbon. Science , 149 (3690), 1326. Juggins, S. (2007). C2: Software for ecological and palaeoecological data analysis and visualisation. Retrieved from h ttps://www.staff.ncl.ac.uk/stephen.juggins/software.htm Kido, Y., Koshikawa, T., & Tada, R. (2006). Rapid and quantitative major element analysis method for wet fine grained sediments using an XRF microscanner. Marine Geology , 229 , 209 225.

PAGE 97

86 Kihlman, S., & Kauppila, T. (2010). Tracking the Aquatic Impacts of a Historical Metal Mine Using Lacustrine Protists and Diatom Algae. Mine Water and the Environment , 29 (2), 116 134. https://doi.org/10.1007/s10230 009 0096 2 Kitzberger, T., Brown, P. M., Heyerdahl, E. K ., Swetnam, T. W., & Veblen, T. T. (2007). Contingent Pacific Atlantic Ocean influence on multicentury wildfire synchrony over western North America. Proceedings of the National Academy of Sciences , 104 (2), 543 548. https://doi.org/10.1073/pnas.0606078104 Knowles, N., Dettinger, M. D., & Cayan, D. R. (2006). Trends in Snowfall versus Rainfall in the Western United States. Journal of Climate; Boston , 19 (18), 4545 4559. Kylander, M. E., Ampel, L., Wohlfarth, B., & Veres, D. (2011). High resolution X ray fluor escence core scanning analysis of Les Echets (France) sedimentary sequence: new insights from chemical proxies. Journal of Quaternary Science , 26 (1), 109 117. https://doi.org/10.1002/jqs.1438 Libby, W. (1972). Dating by Radiocarbon. Accounts of Chemical Re search , 5 (9), 289 295. Libby, W. F., Anderson, E. C., & Arnold, J. R. (1949). Age Determination by Radiocarbon Content: World Wise Assay of Natural Radiocarbon. Science , 109 (2827), 227 228. Lukas, J., Barsugli, J., Doesken, N., Rangwala, I., & Wolter, K. ( 2014). Climate Change in Colorado: A Synthesis to Support Water Resources Management and Adaptation . Colorado Water Conservation Board. Mann, M. E., Zhang, Z., Hughes, M. K., Bradley, R. S., Miller, S. K., Rutherford, S., & Ni, F. (2008). Proxy based recon structions of hemispheric and global surface temperature variations over the past two millennia. Proceedings of the National Academy of Sciences , 105 (36), 13252 13257. https://doi.org/10.1073/pnas.0805721105 Mann, M. E., Zhang, Z., Rutherford, S., Bradley, (2009). Global Signatures and Dynamical Origins of the Little Ice Age and Medieval Climate Anomaly. Science , 326 (5957), 1256 1260. https://doi.org/10.1126/science.1177303 Marlon, J. R., Bartlein, P. J., Carcail Prentice, I. C. (2008). Climate and human influences on global biomass burning over the past two millennia. Nature Geoscience , 1 (10), 697 702. https://doi.org/10.1038/ngeo313

PAGE 98

87 Walsh, M. K. (2012). Long term perspective on wildfires in the western USA. Proceedings of the National Academy of Sciences , 109 (9), E535 E543. https://doi.o rg/10.1073/pnas.1112839109 Mayer, A. C., & St ckli, V. (2005). Long Term Impact of Cattle Grazing on Subalpine Forest Development and Efficiency of Snow Avalanche Protection. Arctic, Antarctic, and Alpine Research , 37 (4), 521 526. Williams, J. J. (2014). Reconstructing Disturbances and Their Biogeochemical Consequences over Multiple Timescales. BioScience . https://doi.org/10.1093/biosci/bi t017 Mensing, S., Korfmacher, J., Minckley, T., & Musselman, R. (2012). A 15,000 year record of vegetation and climate change from a treeline lake in the Rocky Mountains, Wyoming, USA. The Holocene; London , 22 (7), 739 748. http://dx.doi.org.aurarialibrary. idm.oclc.org/10.1177/0959683611430339 Meurant, G. (2012). The Ecology of Natural Disturbance and Patch Dynamics . Academic Press. Minckley, T. A., Shriver, R. K., & Shuman, B. (2012). Resilience and regime change in a southern Rocky Mountain ecosystem durin Ecological Monographs , 82 (1), 49 68. https://doi.org/10.1890/11 0283.1 Moser, K. A. (2004). Paleolimnology and the Frontiers of Biogeography. Physical Geography , 25 (6), 453 480. https://doi.org/10.2747/0272 3646.25.6.453 Moser, K. A., Mordecai, J. S., Reynolds, R. L., Rosenbaum, J. G., & Ketterer, M. E. (2010). Diatom changes in two Uinta mountain lakes, Utah, USA: responses to anthropogenic and natural atmospheric inputs. Hydrobiologia , 648 (1), 91 108. http://0 dx.doi.org.skyline.u cdenver.edu/10.1007/s10750 010 0145 7 Mote, P. W., Hamlet, A. F., Clark, M. P., & Lettenmaier, D. P. (2005). Declining mountain snowpack in western north america*. Bulletin of the American Meteorological Society , 86 (1), 39 50. https://doi.org/10.1175/BAMS 86 1 39 National Oceanic and Atmospheric Administration (NOAA). (2014). ENSO Impacts on United States Winter Precipitation and Temperatures. Retrieved November 29, 2018, from http://www.cpc.ncep.noaa.gov/

PAGE 99

88 Patrick, R. (1977). Ecology of Freshwater Diatoms a nd Diatom Communities. In D. Werner (Ed.), The Biology of Diatoms (pp. 284 332). Los Angeles, California: University of California Press. Penniman, C. (n.d.). Atmospheric Fertilization of Estuaries -Watersheds and Airsheds. Retrieved January 17, 2019, from http://www.prenhall.com/esm/app/nebel/html/updates/update20.html Plicht, J. (2013). IntCal13 and Marine13 Radiocarbon Age Calibration Curves 0 50,000 Years cal B P. Radiocarbon , 55 (04), 1869 1887. https://doi.org/10.2458/azu_js_rc.55.16947 Reynolds, R. L., Mordecai, J. S., Rosenbaum, J. G., Ketterer, M. E., Walsh, M. K., & Moser, K. A. (2010). Compositional changes in sediments of subalpine lakes, Uinta Mountains ( Utah): evidence for the effects of human activity on atmospheric dust inputs. Journal of Paleolimnology; Dordrecht , 44 (1), 161 175. http://dx.doi.org.aurarialibrary.idm.oclc.org/10.1007/s10933 009 9394 8 Saros, J. E., Clow, D. W., Blett, T., & Wolfe, A. P. (2011). Critical Nitrogen Deposition Loads in High elevation Lakes of the Western US Inferred from Paleolimnological Records. Water, Air, & Soil Pollution , 216 (1), 193 202. https://doi.org/10.1007/s11270 010 0526 6 Saros, J. E., Interlandi, S. J., Wolfe, A. P., & Engstrom, D. R. (2003). Recent Changes in the Diatom Community Structure of Lakes in the Beartooth Mountain Range, U.S.A. Arctic, Antarctic, and Alpine Research , 35 (1), 18 23. Saros, J. E., Michel, T. J., Interlandi, S. J., & Wolfe, A. P. (2005). Resource requirements of Asterionella formosa and Fragilaria crotonensis in oligotrophic alpine lakes: implications for recent phytoplankton community reorganizations. Canadian Journal of Fisheries and Aquatic Sciences; Ottawa , 62 (7), 1681 1689. Savage, M. , & Swetnam, T. W. (1990). Early 19th Century Fire Decline Following Sheep Pasturing in a Navajo Ponderosa Pine Forest. Ecology , 71 (6), 2374 2378. https://doi.org/10.2307/1938649 Science Museum of Minnesota. (2018). Laboratories & Equipment. Retrieved from https://www.smm.org/scwrs/facilities/laboratories Sherriff, R. L., Veblen, T., & Sibold, J. (2001). Fire history in high elevation subalpine forests in the Colorado Front Range. Écoscience , 8 (3), 369 380. https://doi.org/10.1080/11956860.2001.11682665

PAGE 100

89 Shu Whitlock, C. (2009). Holocene lake level trends in the Rocky Mountains, U.S.A. Quaternary Science Reviews , 28 (19), 1861 1879. https://doi.org/10.1016/j.quascirev.2009.0 3.003 Jacques, J. M. (2018). Placing the Common Era in a Holocene context: millennial to centennial patterns and trends in the hydroclimate of North America over the past 200 0 years. Climate of the Past , 14 , 665 686. Shuman, Bryan N., Carter, G. E., Hougardy, D. D., Powers, K., & Shinker, J. J. (2014). A north south moisture dipole at multi century scales in the Central and Southern Rocky Mountains during the late Holocene. Ro cky Mountain Geology , 49 (1), 33 49. Spaulding, S. A., Otu, M. K., Wolfe, A. P., & Baron, J. S. (2015). Paleolimnological Records of Nitrogen Deposition in Shallow, High Elevation Lakes of Grand Teton National Park, Wyoming, U.S.A. Arctic, Antarctic, and Al pine Research , 47 (4), 703 717. http://0 dx.doi.org.skyline.ucdenver.edu/10.1657/AAAR0015 008 Steelman, J. (Unpublished). A Dendrochronology Study of Mining & Logging Impacts in a Small Mountainous Watershed. Sugita, S. (1993). A Model of Pollen Source Area for an Entire Lake Surface. Quaternary Research , 39 (2), 239 244. https://doi.org/10.1006/qres.1993.1027 Sugita, S. (1994). Pollen Representation of Vegetation in Quaternary Sediments: Theory and Method in Patchy Vegetation. Journal of Ecology , 82 (4), 881 897. https://doi.org/10.2307/2261452 Toney, J. L., & Anderson, R. S. (2006). A postglacial palaeoecological record from the San Juan Mountains of Colorado USA: fire, climate and vegetation history. The Holocene , 16 (4), 505 517. https://doi.org/10.1191/0959 683606hl946rp United States Forest Service. (2007). Grand Mesa, Uncompahgre, Gunnison National Forest Proposed Land Management Plan: Program Priorities and Proposed Possible Actions (Appendix C). University of Georgia: Center for Applied Isotope Studies. (2017). Radiocarbon: Background and History. University of Georgia: Center for Applied Isotope Studies. Retrieved from https://cais.uga.edu/submit/Radiocarbon_intro.pdf US Geological Survey. (2017). Geologic Provinces of the United States: Rocky M ountains. Retrieved from https://geomaps.wr.usgs.gov/parks/province/rockymtn.html

PAGE 101

90 Werner, D. (1977). Silicate Metabolism. In D. Werner (Ed.), The Biology of Diatoms (pp. 110 149). Los Angeles, California: University of California Press. Whitlock, C., Higue ra, P. E., McWethy, D. B., & Briles, C. E. (2010). Paleoecological Perspectives on Fire Ecology: Revisiting the Fire Regime Concept~!2009 09 02~!2009 11 09~!2010 03 05~! The Open Ecology Journal , 3 (2), 6 23. https://doi.org/10.2174/1874213001003020006 Whit lock, C., & Larsen, C. (2001a). Charcoal as a fire proxy. In J. P. Smol, H. J. B. Birks, & W. M. Last (Eds.), Tracking Environmental Change Using Lake Sediments. Volume 3: Terrestrial, Algal, and Siliceous Indicators. (Vol. 3). Dordrecht, The Netherlands: Kluwer Academic Publishers. Whitlock, C., & Larsen, C. (2001b). Charcoal as a fire proxy. In J. P. Smol, H. J. B. Birks, & W. M. Last (Eds.), Tracking Environmental Change Using Lake Sediments. Volume 3: Terrestrial, Algal, and Siliceous Indicators. (Vol. 3). Dordrecht, The Netherlands: Kluwer Academic Publishers. Wolfe, A. P., Baron, J. S., & Cornett, R. J. (2001). Anthropogenic nitrogen deposition induces rapid ecological changes in alpine lakes of the Colorado Front Range (USA). Journal of Paleolimnology , 25 (1), 1 7. https://doi.org/10.1023/A:1008129509322 Wollmann, K., Deneke, R., Nixdorf, B., & Packroff, G. (2000). The dynamics of planktonic food webs in 3 mining lakes across a pH gradient (pH 2 4). Hydrobiology , 4 33 , 3 14.

PAGE 102

91 Appendix A: Forest Hill Mining Complex Adapted from 2018 dendrochronology report by J. Steelman (unpublished) This figure shows the proposed area of the Forest Hill Mining Complex and the surrounding timber stand. The red circles represen t major mines in the area and the yellow X represent other exploratory mines. These were identified by J. Steelman during an extensive foot survey of the area done during July 2018. Eight major mines and over 40 exploratory mines were

PAGE 103

92 discovered, along wit h a series of roads and trails that connected the area. Ten additional buildings were found and indicate that people were living in the area during the mining period. These findings support the theory that Lily Pond was impacted by anthropogenic activity related to mining over the last ~200 years. A preliminary dendrochronology study of the forest surrounding Lily Pond. identified three forest stands . The stands were defined preliminarily using aerial photography and further evaluated by collecting tree rings in July 2018. The three forest stands were based on current tree age and using evidence of anthropogenic activity (i.e. timber harvest, burning) . Descriptions of each forest stand are included below. There is also evidence of a recent fire event ( ~ 50 35 cal yr BP; ~1900 1915 CE) on the slope southwest of Lily Pond. There is evidence of a fire event at the top of the Lily Pond fire record ( ~ 2 cal yr BP; 1948 CE ), however the timing between these two events does not line up. Additional tree ring analysis is needed in this area to make any further conclusions about this possible fire event. Forest Stand 03 : This stand group shows remnants of harvest activity . Releas e dates range from 1840 1870 CE , indicating these stands were likely logged around the establishment of FHM. H owever harvest may not have been used for mine operations due to distance from FHM location . Forest Stand 02: This stand type is the oldest in su rrounding area, with release dates between 1777 1871 CE. These stands show very limited anthropogenic activity, aside from clearing close to road locations. These stands were likely not harvested . Forest Stand 01 : This stand type is defined as having the y oungest trees in the surrounding area , with release between 1876 1918 CE. The timing of release indicates timber harvest in these areas during the mining period. There were also landscape indicators that suggest heavy anthropogenic activity and clear cutting.

PAGE 104

93 Appendix B : Lead 210 Dating Results

PAGE 105

94 Additional figures related to establishment of 210 Pb chronology provided by St. Croix Watershed Research Station.

PAGE 106

95 Appendix C : Supplemental Age Depth Model Material s

PAGE 107

96 Appendix D : Diatom Taxa Identified Taxa Used in Analyses Aulacoseira X Cavinula X Cyclotella X Cymbella X Cymbopleura X Encyonema X Eunotia X Fragilariaceae X Frustulia X Gomphoneis X Navicula X Neidium X Pinnularia X Sellaphora X Stauroneis X Surirella Tabellaria X

PAGE 108

97 Appendix E: P aleoclimate and paleoecology studies referenced Map of paleoecological and paleoclimatic records referenced in this thesis . The table below lists the timing of environmental changes identified at these locations, organized from north to south . Location Year (cal yr BP) Increase in Pinus Increase in fire activity Decrease in fire activity