Citation
Water diversions contributing to mountain pine beetle infestation and subsequent tree mortality

Material Information

Title:
Water diversions contributing to mountain pine beetle infestation and subsequent tree mortality
Creator:
Smolinski, Sharon Louise
Publication Date:
Language:
English
Physical Description:
xiv, 111 leaves : color illustrations, color maps ; 28 cm

Subjects

Subjects / Keywords:
Water diversion -- Colorado ( lcsh )
Mountain pine beetle -- Colorado ( lcsh )
Trees -- Mortality -- Colorado ( lcsh )
Mountain pine beetle ( fast )
Trees -- Mortality ( fast )
Water diversion ( fast )
Colorado ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 101-110).
Statement of Responsibility:
by Sharon Louise Smolinski.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
786195637 ( OCLC )
ocn786195637
Classification:
LD1193.L547 2011M S56 ( lcc )

Downloads

This item has the following downloads:


Full Text
WATER DIVERSIONS CONTRIBUTING TO MOUNTAIN PINE BEETLE
INFESTATION AND SUBSEQUENT TREE MORTALITY
by
Sharon Louise Smolinski B.S., University of Colorado Denver, 1998 B.F.A., University of Colorado Denver, 1998
A thesis submitted to the University of Colorado Denver in partial fulfillment of the requirements for the degree of Master of Science Environmental Sciences
2011


This thesis for the Master of Science
degree by
Sharon Louise Smolinski has been approved by
Sa AhJ Z&ll


Smolinski, Sharon Louise (M.S., Environmental Sciences)
Water Diversions Contributing to Mountain Pine Beetle Infestation and Subsequent Tree Mortality
Thesis directed by Professor Frederick Chambers
ABSTRACT
This novel study explored the hypothesis that water diversions cause an increased incidence of mountain pine beetle infestation and subsequent increased tree mortality in areas downstream of or below diversion points. This study included the use of field surveys to determine the locations and types of diversion present in the areas of interest. Field data was collected from a total of eighteen plots over four study areas. The study areas were established around two different types of diversion structures, with plots above and below the diversion. Sites were characterized by tree count, diameter and species, in addition to visual assessment of pine beetle infestation, and soil and vegetation moisture content. Additionally, analysis of Landsat data examined differences and changes in NDVI values over time. Survey data revealed extensive diversions of entire basins, pointing to the need for assessments of environmental impacts. Field assessments of beetle infestation showed increased percentages of beetle-kill trees in


habitat below diversions. These findings support the hypothesis that water diversions cause an increase in mountain pine beetle infestation and subsequent tree mortality in habitat below diversions. Other field data suggest the possibility that diversions lead to general habitat effects.
This abstract accurately represents the conte recommend its publication.
Signed
Frederick B. Chambers


DEDICATION
I dedicate this thesis to my parents, who introduced me to natural world at a young age and who taught me the value of learning. I especially want to thank my mother, who filled my crib with books.


ACKNOWLEDGEMENT
I acknowledge my advisor, Frederick B. Chambers, and all of the members of my committee for their contribution and insight.


TABLE OF CONTENTS
Figures..............................................................xii
Tables...............................................................xiv
Chapter
1. Introduction........................................................1
1.1 Purpose of the Study................................................1
1.2 Scope of the Study..................................................2
2. Literature Review...................................................4
2.1 Water Diversions....................................................4
2.1.1 Impacts of Water Diversions......................................5
2.2 Mountain Pine Beetle Infestation...................................7
2.2.1 Factors Facilitating Pine Beetle Infestation.....................8
2.2.2 Environmental Impacts of Pine Beetle Infestation................10
2.3 Field Methods.....................................................12
2.3.1 Visual Assessment of Pine Beetle Infestation..................12
2.3.2 Assessment of Moisture Conditions...............................13
2.3.3 Statistical Analysis of Field Data..............................14
2.4 Methods Using Remotely Sensed Data................................15
2.4.1 Data Sources....................................................16
2.4.2 Spectral Analysis...............................................16
viii


2.4.3 Error mitigation................................................18
3. Methods...........................................................20
3.1 Field Surveys of Water Diversions.................................20
3.1.1 Inadequate Public Data.........................................20
3.1.2 Survey Overview................................................23
3.1.3 Locations and Types of Diversions..............................24
3.1.4 Surveyed Areas.................................................31
3.2 Detailed Field Studies...........................................37
3.2.1 Study Sites....................................................37
3.2.2 Basic Site Assessment..........................................45
3.2.3 Assessment of Pine Beetle Infestation........................45
3.2.4 Assessment of Percentage of Green Canopy.......................46
3.2.5 Soil Water Content.............................................47
3.2.6 Vegetation Water Content.......................................48
3.2.6 Statistical Analysis of Field Data.............................49
3.3 Analysis of Remotely Sensed Data.................................49
3.3.1 Data and Sources................................................49
3.3.2 Data Analysis...................................................50
4. Results...........................................................55
4.1 Field Surveys of Water Diversions.................................55
ix


4.2 Detailed Field Studies..........................................55
4.2.1 Study Sites...................................................56
4.2.2 Tree Density..................................................56
4.2.3 Tree Diameter.................................................61
4.2.4 Species Composition............................................63
4.2.5 Assessment of Pine Beetle Infestation.........................64
4.2.6 Assessment of Percentage of Green Canopy......................75
4.2.7 Soil Water Content............................................76
4.2.8 Vegetation Water Content......................................77
4.3 Analysis of Remotely Sensed Data.................................78
5. Discussion.......................................................83
5.1 Field Surveys of Water Diversions................................83
5.2 Detailed Field Studies..........................................85
5.2.1 Study Areas and Plots..........................................85
5.2.2 Tree Density...................................................86
5.2.3 Tree Diameter.................................................87
5.2.4 Assessment of Beetle Infestation...............................88
5.2.5 Assessment of Percentage of Green Canopy.......................89
5.2.6 Soil Water Content.............................................90
5.3 Analysis of Remotely Sensed Data................................93


6. Conclusion.....................................................95
Appendix
A. Special-Use Permit...............................................98
Bibliography......................................................101
XI


LIST OF FIGURES
Figures
3.1 Water Diversions in Southeastern Grand County
Based on CDSS Data..................................................22
3.2 Types of Diversion Structures.........................................25
3.3 Detached Diversion Outlet Pipe in Berthoud Pass Area................29
3.4 Field Survey of Rollins Pass Area.....................................33
3.5 Field Survey of Berthoud Pass Area....................................34
3.6 Field Survey of Williams Fork Area....................................35
3.7 Study Areas in Grand County...........................................40
3.8 Rollins Pass Study Area 1.............................................41
3.9 Rollins Pass Study Area 2.............................................42
3.10 Rollins Pass Study Area 3.........................................43
3.11 Berthoud Pass Study Area 1........................................44
3.12 Spatial Resolution Limitations....................................53
4.1 Beetle Indicator Data for Rollins Pass Study Areas 1 and 2............67
4.2 Beetle Indicator Data for Rollins Pass Study Area 3...................68
4.3 Beetle-kill Trees in Rollins Pass Study Areas 1 and 2.................70
4.4 Beetle-kill Trees in Rollins Pass Study Area 3........................71
4.5 Beetle Indicator Data for Rollins Pass Study Areas 1 and 2..........73
xii


4.6 Beetle Indicator Data for Rollins Pass Study Area 3.............74
4.7 Percentage of Green Canopy by Plot Type..........................76
4.8 NDVI Change at Rollins Pass Study Area 3.........................80
4.9 NDVI Change in Jones Pass Area..................................82
5.1 Dam at Rollins Pass Study Area 3.................................84
A.1 Special-Use Permit...............................................99
xiii


LIST OF TABLES
Table
3.1 Summary of areas covered by field surveys.................24
3.2 Summary of study areas and plots..........................39
4.1 Summary of study areas and plots..........................55
4.2 Summary of tree counts for study areas....................57
4.3 Summary of plot central elevation.........................59
4.4 Summary of tree counts by plot............................60
4.5 Summary of tree diameter data.............................62
4.6 Summary of tree type percentages..........................64
4.7 Summary of pine beetle indicator data.....................66
4.8 Summary of soil moisture content data.....................77
XIV


1. Introduction
1.1 Purpose of the Study
This is a novel study the objective of which is determining if water diversions cause increased incidence of infestation by the mountain pine beetle, Dendroctonus ponderosae Hopkins, hereafter referred to as pine beetle. This study is motivated by my first-hand observations over several years in areas of Grand County, Colorado. I observed what appeared to be a marked increase in beetle kill trees below water diversions, relative to areas above diversions. Presumably, water diversions lead to reduced water availability downstream of or below diversion points. This study will examine the hypothesis that water diversions facilitate pine beetle infestation and lead to increased tree mortality in areas downstream of and below diversion points.
There has been no research examining a possible relationship between water diversion and pine beetle activity. Furthermore, there has been no research regarding the impact of water diversion on coniferous trees located within the basin of extraction. Based on the widespread extent and impacts of pine beetle infestation, research exploring additional factors facilitating pine beetle activity would be of substantial importance. If
l


a relationship is found, the findings could prompt significant changes in water policy and forest management.
1.2 Scope of the Study
This study examined the aforementioned hypothesis through field surveys of water diversions, detailed field studies examining several factors, and analysis of remotely sensed data. Field surveys determined the location and types of diversions present. These field surveys investigated three areas in the southeast quarter of Grand County, Colorado, and formed the basis of site selection for the detailed field studies.
The detailed field studies included the selection of study areas and sampling plots, followed by a variety of assessments. Based on the results of the field surveys, three study areas were established in the Rollins Pass Area, in addition to one in the Berthoud Pass area. All sampling plots within the study areas were subject to basic characterization, including determination of GPS coordinates, elevation, and aspect, in addition to tree counts, measurement of tree diameter at breast height (DBH), and identification of all trees. Lodgepole pines were visually assessed for pine beetle indicators, and all trees were assessed for basic health. Visual
2


assessments were also explored based on visual assessments of percentage of green needles. Some study areas were assessed for moisture condition, specifically soil water content and vegetation water content.
Remotely sensed data were analyzed in a geographic information system (GIS) in order to compare spectral values between diverted and non-diverted areas. Landsat data were analyzed in ArcMapIO (Esri, 2011), in order to determine and compare the change in NDVI from 2003 to 2010.
3


2. Literature Review
2.1 Water Diversions
Water diversion projects stretch across the Western United States and Canada, transporting significant amounts of water out of basins of origin. In Colorado, numerous projects transport water from the west side of the Continental Divide to the east side, servicing the developed Front Range for municipal water consumption and industrial application. Many water projects were instituted before environmental impacts were considered (Doremus & Hanemann, 2007). The field surveys for this study found numerous diversion structures in the Rollins Pass area dating to the 1950s. Only recently have some water policy decisions begun to consider environmental issues (Doremus & Hanemann, 2007). However, as climate-driven drought progresses, there may be a revived battle between water demand and environmental impacts (Doremus & Hanemann, 2007). Unfortunately, the impacts of diversions have not been adequately explored.
4


2.1.1 Impacts of Water Diversions
Diversions result in reduced water availability downstream from diversion points. Researchers have demonstrated detrimental impacts to mountain streams, as well as habitats far downstream (Carriquiry & Sanchez,1999; Rodriguez et al., 2001; Schone et al., 2003; Wohl, 2006). Researchers have also demonstrated impact to high-elevation wetlands as a result of water diversions (Chimner & Cooper, 2003). However, there has been an absence of research examining impact on coniferous habitat at the site of high-elevation water diversion projects in the basins of origin. Also, there has been an absence of research regarding a correlation between water diversion projects and mountain pine beetle activity.
Wohl (2006) discussed significant and numerous impacts to mountain streams from reduced flow, including sedimentation, water temperature, and available habitat. These effects have in turn impacted multiple species, resulting in measureable reductions in biotic integrity (Wohl, 2006). In fact, impacts to flow within mountain streams in the Rocky Mountains of the United States are rated as severe, due to the number of projects and amount of water involved (Wohl, 2006, p. 222).
5


Marine habitats far downstream from diversion points, which may be states away, are also affected. Reduced freshwater flows alter the sediment composition and salt content within estuaries (Carriquiry & Sanchez, 1999). Reduced freshwater flows have impacted the growth and distribution of some species in these same estuaries (Rodriguez et al.,
2001; Schone et al., 2003). These endpoints of river systems receive the cumulative effect of all water diversions along the contributing watersheds, producing a greater measureable effect.
Outside of mountain streams, there is little research regarding diversion impacts on high-altitude ecosystems. Notably, researchers have shown an impact to high-altitude wetlands as a result of diversion projects (Chimner & Cooper, 2003). Diversion projects result in lower water tables (Chimner & Cooper, 2003). These lower water tables caused increased CO2 emissions from subalpine fens due to increased decomposition (Chimner & Cooper, 2003).
Previous inquiries into a potential relationship between water diversion projects and pine beetle infestation have been limited to the impacts of beetle-kill trees on diversions (Rey, 2008; USFS, 2008a). Specifically, concerns only included the possibility of decreased water
6


quality and blockage of diversion channels from fallen trees (Rey, 2008; USFS, 2008a). The USDA directly pointed to a threat to the water supply for Front Range metropolitan areas (Rey, 2008).
Based on the extent of diversions and the limited research on the impacts of those operations, additional research is needed, particularly into the impacts on the ecosystems within the basins of extraction.
2.2 Mountain Pine Beetle Infestation
The mountain pine beetle has decimated pine forests throughout the western US and Canada in recent years (Colorado State University [CSU], n.d.; Gillette et al., 2009; Jones et al., 2004; Kurtz et al., 2008; McKinney and Tomback, 2007; Miller & Spoolman, 2009; Steventon & Daust, 2009).
A native species, the pine beetle can provide beneficial ecosystem services when acting on a limited number of trees, including reduction of stand density and facilitation of natural wildfire (Jones et al., 2004; Logan & Powell, 2001; USFS, n.d.). However, in recent years pine beetle activity has escalated considerably, resulting in vast regions of damaged and dead forest (CSU, n.d.; Gillette et al., 2009; Jones et al., 2004; Kurtz et al., 2008;
7


McKinney & Tomback, 2007; Miller & Spoolman, 2009; Steventon & Daust, 2009).
2.2.1 Factors Facilitating Pine Beetle Infestation
Researchers have identified a variety of factors that have facilitated the escalation of pine beetle activity. Largely anthropogenic, these factors include drought, increased temperatures, increased stand density resulting from fire suppression, and pollutants (Bleiker & Six, 2009; Breece et al., 2008; Gillette et al., 2009; Jones et al., 2004; Logan & Powell, 2001; McKinney & Tomback, 2007; Miller & Spoolman, 2009; U.S. Forest Service [USFS], n.d.).
Drought adversely impacts tree health, creating favorable conditions for the pine beetle (Bleiker & Six, 2009; Jones et al, 2004; USFS, n.d.). Tree resin concentration is altered by drought conditions, thereby reducing resistance to pine beetles (Jones et al., 2004). Drought leads to lowered moisture content in trees, increasing vulnerability to colonization by the fungi associated with the pine beetle (Bleiker & Six, 2009). Forecasts indicate that climate change has increased drought frequency and severity in the Western U.S. through increased temperatures and altered
8


precipitation, effects that are expected to intensify as climate change progresses (IPCC, 2007; Logan & Powell, 2001; Miller & Spoolman 2009). Higher temperatures enable pine beetle activity to shift to higher altitudes and latitudes, further intensifying impacts and potentially allowing pine beetles to infest species other than pines (Logan & Powell, 2001). These increased temperatures allow larvae populations to survive through the winter (Cain 2005). Researchers have determined that climate change has thereby already increased the severity and extent of pine beetle activity (Gillette et al., 2009; Logan & Powell, 2001; Miller & Spoolman, 2009).
Wildfire is a crucial natural force for modifying forest structure, but modern forest management actively suppresses fire (Breece et al., 2008; Jones et al., 2004; Logan & Powell, 2001; McKinney & Tomback, 2007). This management policy produced increased stand density, providing an increased and densely associated host supply (Breece et al., 2008; Jones et al., 2004; McKinney & Tomback, 2007). Also, increased stand density has amplified the effects of drought, forcing larger populations of trees to share increasingly limited supplies of water (Jones et al., 2004; McKinney & Tomback, 2007).
9


Researchers have determined that pollutants also facilitate infestation (Jones et al., 2004). The mechanism for impact is that anthropogenic pollutants enter the atmosphere and are deposited in remote regions, posing a significant threat to natural habitats (Jones et al., 2004). Jones et al. (2004) examined ponderosa pine already affected by drought, and found increased pine beetle attacks as well as increased tree mortality in areas with higher concentrations of nitrogen and ozone.
2.2.2 Environmental Impacts of Pine Beetle Infestation
Damage resulting from pine beetle infestation causes significant localized and global impacts (Breece et al., 2008; Jones et al., 2004; Kurtz, et al., 2008; Logan & Powell, 2001; McKinney &Tomback, 2007, Steventon & Daust, 2009). The coniferous forests of western North America support a wide range of species and provide important ecosystem services. The large-scale loss of pine trees presents a dramatic change to these ecosystems, impacting all resident species. For example, whitebark pines are a key food source for a variety of animals, including Clarks Nutcracker and grizzly bears (Logan & Powell, 2001; McKinney & Tomback, 2007; Miller & Spoolman, 2009). As whitebark pine populations decrease, those
10


species are adversely impacted (Logan & Powell, 2001; McKinney & Tomback, 2007). The loss of foliage cover can lead to increased predation upon prey species (Hoffman, 2006). Changes in prey populations in turn affect predators, such as pine martens (Steventon & Daust, 2009). Indeed, the loss of entire regions of forest will cause habitat fragmentation and edge effects, further impacting species, increasing predation, and creating opportunities for invasive species (Primack, 2006). In general, habitat loss leads to reduced populations of species, and reduced biodiversity (Miller & Spoolman, 2009; Primack, 2006).
Pine beetle infestation also has substantial global effect by contributing to the progression of climate change (Kurtz et al., 2008). Forests provide an important ecosystem service by absorbing C02, acting as carbon sinks (Kurtz et al., 2008; Miller & Spoolman, 2009; Primack, 2006). The reduction in functional forests has reduced C02 uptake, and increased C02 input through decomposition (Kurtz et al., 2008). As explained by Kurtz et al. (2008), these forests now function as carbon sources, instead of carbon sinks, thereby accelerating climate change.
li


2.3 Field Methods
2.3.1 Visual Assessment of Pine Beetle Infestation
Numerous studies have relied on visual assessment to determine pine beetle infestation in pine trees (Breece et al., 2008; Clow et al., 2011; Jones et al., 2004; Kulakowski & Jarvis, 2011; Morehouse et al., 2008; Negron et al., 2009; Sanchez-Martinez & Wagner, 2002; Waring & Six, 2005; Zausen et al., 2005). Assessments classified trees based on the presence or absence of indicators of pine beetle activity, including brown needles, faded crowns, bore holes, boring dust, pitch tubes, and actual pine beetles (Breece et al., 2008; Jones et al., 2004; Kulakowski & Jarvis, 2011; Negron et al., 2009; Waring & Six, 2005; Zausen et al., 2005). Ratings systems generally typically differentiated between live trees without indicators, live trees with indicators, dead trees with indicators, and dead from other causes (Breece et al., 2008; Negron et al., 2008; Negron et al., 2009, Waring & Six, 2005).
Some researchers assessed all trees within plots, or all trees with a specified DBH (Breece et al., 2008; Jones et al., 2004; Morehouse et al., 2008; Negron et al., 2008; Negron et al., 2009; Sanchez-Martinez & Wagner, 2002; Waring & Six, 2005). Explained by Waring and Six (2005),
12


researchers use DBH limits to exclude trees that are too small to be subject to pine beetle attack (Waring & Six, 2005). However, minimum DBH limits were not consistent across studies, including 2.54 cm (1 in) (Negron et al., 2008; Negron et al., 2009), 5 cm (2 in) (Morehouse et al., 2008), 7.6 cm (3 in) (Waring & Six, 2005), and 13 cm (5 in) (Breece et al., 2008).
Studies which have included visual assessment of pine beetle indicators used a variety of numbers and sizes of plots. Negron et al.
(2009) utilized 633 circular plots, each with an 8 m (26.2 ft) radius, spread over five national forests. Morehouse et al. (2008) used 20 circular plots, each with a 10 m (32.8 ft) radius. Waring and Six (2005) used 105 circular plots, each with an 11.28 m (37 ft) radius.
2.3.2 Assessment of Moisture Conditions
Researchers have also assessed moisture conditions in field studies examining infestation, including soil water content (Clow et al., 2011), and vegetation water content (Zausen et al., 2005). Soil water content is a common parameter in assessing water availability in forest ecosystems (Brooks & Kyker-Snowman, 2009; Clow et al., 2011; Penna et al., 2009; Salle et al., 2008; Walker et al., 2004). Traditionally, soil samples are
13


removed from varying depths and transported for lab analysis (Clow et al., 2011; Salle et al., 2008). Typically soil samples are dried at approximately 100C and weighed to determine water content (Salle et al., 2008). Soil water content is also measured directly on-site using portable soil moisture probes and sensors (Brooks & Kyker-Snowman, 2009; Penne et al., 2009; Walker et al., 2004). Assessment of water stress in coniferous trees commonly relies upon needle water content and predawn needle water potential (Salle et al., 2008; Zausen et al., 2005). Similar to soil sampling, needles are physically removed and transported for lab analysis, where they are dried and weighed to determine water content (Salle et al., 2008; Zausen et al., 2005). To determine predawn water potential, needles are collected before dawn and measured with a pressure chamber (Salle et al., 2008; Zausen et al., 2005).
2.3.3 Statistical Analysis of Field Data
Researchers have typically used non-parametric statistical tests to analyze field data, including Wilcoxon-rank-sum, chi-square, and Kruskal-Wallis tests (Breece et al., 2008; Guarin et al., 2005; Negron et al., 2008;
14


Negron et al., 2009). If parametric tests were used, the data were first log-transformed (Breece et al., 2008).
2.4 Methods Using Remotely Sensed Data
Researchers have utilized remotely sensed data and GIS to assess and analyze ecosystems in relation to a variety of issues, including pine beetle infestation (Coops et al., 2006; Coops et al., 2009; Dennison et al., 2010; Hais et al., 2009; Hatala et al., 2010; Hilker et al., 2009; Meddens et al., 2011; Vogelmann et al., 2009; White et al., 2005; Wulder et al., 2006). This is possible through the use of spectral values as measured by satellite or airborne sensor, based on wavelengths of electromagnetic radiation reflected or emitted from vegetation (Campbell, 2007). Spectral analysis allows for assessment of vegetation stress and mortality, as well as species identification and canopy characterization (Campbell, 2007). The use of this spectral data allows researchers to assess infestation in large and remote areas, as well as over time (Campbell, 2007; Coops et al., 2006; Coops et al., 2010; Hais et al., 2009; Wulder et al., 2006).
15


2.4.1 Data Sources
Researchers have examined infestation using a wide range of platforms and sensors, including airborne HyMap (Hatala et al. 2010), airborne Vexcel Utracam-D (Meddens et al., 2011); Landsat TM or ETM+ (Coops et al., 2006; Coops et al., 2010; Hais et al., 2009; Hilker et al., 2009; Vogelmann et al., 2009; Wulder et al., 2006), MODIS (Coops et al., 2009), IKONOS (White et al., 2005), and GeoEye-1 (Dennison et al., 2010). Onboard sensors provide data of diverse spatial and spectral resolution.
For example, MODIS provides a low spatial resolution of 250 m (Campbell, 2007), while GeoEyel provides 1.65 m resolution (Digital Data Services, n.d.). These variations in resolution influence data quality and analytical capabilities. While of moderate spatial resolution of 30 m, Landsat data have been used for comparisons of sizeable areas over time (Coops et al., 2006; Coops et al., 2010; Hais et al., 2009; Vogelmann et al., 2009; Wulder et al., 2006).
2.4.2 Spectral Analysis
In assessing infestation, a primary goal has been to differentiate between green canopy, representative of living trees, and red,
16


representative of dead trees with needles (Coops et al., 2010; Dennison et al., 2010; Meddens et al., 2011; White et al., 2005; Wulder et al., 2006). Some researchers also differentiated grey canopy, which is representative of dead trees that have shed all needles (Coops et al, 2006; Coops et al., 2010; Dennison et al., 2010; Wulder et al., 2006). Non-vegetation features can first be masked out in order to more accurately assess the vegetated areas (White et al., 2005; Wulder et al., 2006). Pixels in the areas of interest can then be classified using supervised (Dennison et al., 2010), or unsupervised classification (White et al., 2005). Some researchers have directly worked from spectral values to evaluate red and green pixels (Dennison et al., 2010; White et al., 2005).
Researchers have also utilized a variety of Vegetation Indices (Vis) in the analysis of spectral data (Coops et al., 2006; Coops et al., 2009; Coops et al., 2010; Hais et al., 2009; Hilker et al., 2009; Vogelmann at al., 2009; Wulder et al., 2006). Vegetation Indices calculate relationships between specific bands in order to assess some factor of vegetation (Campbell, 2007). One common VI is the Normalized Difference Vegetation Index (NDVI), which utilizes red and infrared bands, corresponding to bands 3 and 4 respectively in Landsat TM and ETM+
17


(Campbell, 2007; Esri, 2010). Studies assessing infestation have used a variety of Vis: NDVI (Hais et al., 2009; Vogelmann et al., 2009), Enhanced Difference Vegetation Index (EDVI) (Coops et al., 2006), Disturbance Index (Dl) (Vogelmann et al., 2009), Tasseled Cap Transformation (TCT) (Coops et al., 2006; Wulder et al., 2006; Hais et al., 2009), and Red Green Index (RGI) (Coops et al., 2006; Meddens et al., 2011). Some of these indices can evaluate red and green values, such as the RGI which calculates the ratio of red to green scaled values (Meddens et al., 2011). Again, classification schemes are used to process these values (Campbell, 2007).
2.4.3 Error mitigation
The use of remotely sensed data for ecological assessments carries the potential for error due to the fact that spectral data values can be attributed to a variety of factors (Campbell, 2007). Some researchers use airborne imagery to validate satellite imagery in order to verify that features are being interpreted correctly (Dennison et al., 2010; White et al., 2005). Many researchers use field data physically collected from sites on the ground to validate remotely sensed data, using protocols similar to those
18


discussed in 2.3 (Campbell, 2007; Coops et al., 2006; Hais et al., 2009 Meddens et al., 2011; White et al., 2005; Wulder et al., 2006)
19


3. Methods
3.1 Field Surveys of Water Diversions
Field surveys determined the locations and types of water diversions within the region of interest. Inadequate existing public data necessitated the use of first-hand field surveys. The finding of the surveys, specifically the locations and types of diversions, are presented in 3.1.3. Those findings are discussed here, instead of the Results section, because they determined the selection of the study areas used for detailed field study. Also, these findings characterize the nature of and the extent of diversions in the region, the basis for the study.
3.1.1 Inadequate Public Data
Field surveys were necessary due to the lack of adequate public data regarding extent, location, and types of water diversions in the region. Geospatial data regarding diversion projects in Colorado are available from Colorados Decision Support System, the CDSS, which is operated by the Colorado Division of Water Resources and the Colorado Water Conservation Board (Colorados Decision Support System, n.d.). Quality issues were apparent when the CDSS data were examined in ArcMap, and
20


compared with first-hand observations as well as GIS data from the Colorado Department of Transportation (CDOT) (Colorado Department of Transportation, 2010). Major diversions were missing from the CDSS data, such as the transbasin diversion at the headwaters of Vasquez Creek. In Figure 3.1, CDOT data show the Vasquez Tunnel extending from Vasquez Creek to the adjacent county. Furthermore, entire lengths of extensive diversions were omitted, such as those bisecting the basins of Current Creek and adjacent basins, northwest of Berthoud Pass. This diversion system is known through first-hand observation. Also, CDOT data show this diversion as a stream extending from First Creek south to Berthoud Pass, as shown in Figure 3.1. In addition to omitting diversions, CDSS data do not fully differentiate or classify the various types of diversions, compared to first-hand observations. Despite omissions, CDSS data indicate that water diversions are extensive, and that all streams are diverted and often at multiple points, as suggested by the numerous diversion points in Figure 3.1.
21


Water Diversions in Southeastern Grand County Based on CDSS Data
Sharon Smoilnski
Diversions
------ Streams
County boundaries
Data Sources:
Colorado's Decision Support System water diversion data Colorado Department of Transportation: county and stream data
Coordinate system: NAD 63 UTM Zone 13N
Figure 3.1. Water Diversions in Southeastern Grand County Based on CDSS Data.
22


3.1.2 Survey Overview
Field surveys were conducted along three areas in Grand County, with the goal of determining locations and types of diversions. A Garmin GPSmap 60CSx GPS unit, with an accuracy of approximately +/-10 feet, was used to collect coordinate data of diversion points. A digital camera was used to photograph all surveyed diversion points. Surveys were conducted from June to October in 2010. Additionally, survey areas were observed multiple times, and some areas were also visited in 2011. The area west of Rollins Pass was surveyed from the Fraser River Dam northwards for approximately 16 miles. The diversions bisecting the Current Creek basin northwest of Berthoud Pass was surveyed for approximately 1.5 miles. The area west of Jones Pass was surveyed for approximately 3.5 miles from Steelman Creek in the southwest of the basin to McQueary Creek in the north of the basin. Miles surveyed are based upon the linear distance traveled along diversion lengths. The areas covered by the field surveys are summarized in Table 3.1.
23


Table 3.1. Summary of areas covered by field surveys.
Area Distance Surveyed (miles) Elevation Range (ft) Accuracy of GPS unit (ft) Transportation
Rollins Pass 16 9358-9644 +/-9-12 car
Berthoud Pass 1.5 11264- 11352 +/-9-18 foot
Williams Fork 3.5 10284- 10519 +/-9-15 foot
3.1.3 Locations and Types of Diversions
Field surveys revealed extensive networks of varied diversion structures and thoroughly diverted basins. All streams in the surveyed basins were diverted, regardless of their size. Several structural types of diversions were identified, and can be classified into ten types, and an additional five subtypes. The ten main types are as follows: dams, canals, gates, ditches, grates, complete intercepts, pipelines, reservoirs, tunnels, and indeterminate structures. Canals consist of three subtypes based on materials: dirt, concrete, and stone. Pipelines consist of two subtypes, above-surface and buried pipelines. Figure 3.2 provides examples of some of these structures, all of which were photographed during the field surveys and are located in one of the survey areas. Some structures cannot be photographed, such as buried pipelines.
24


Figure 3.2. Types of Diversion Structures. Examples of some diversion structure types found during field surveys: A) dam, B) dam, C) gate, D) dirt canal, E) concrete canal, F) stone canal. Figure is continued on following page.
25


Figure 3.2 (Cont.). Types of Diversion Structures. Examples of some diversion structure types found during field surveys: G) grate, H) complete intercept, I) pipeline.
26


Dams were found in a range of sizes, from very large to small, based on the size of the stream being intercepted. Dams provide for the potential of some release or overflow of water with gates and spillways. Spillways only allow for overflow during high stream flow. Some spillways of small dams are very rudimentary, and functionality may be limited. Gates served as control points to allow water through the diversion point, back into the stream, or served to control water into the diversion pipe. Large dams generally had two gates, allowing for control of water through the diversion and into the diversion. Dams are shown in Figure 3.2, A and B, at Bobtail Creek and Jones Creek, respectively, in the Williams Fork area. Note both dams have a spillway and two gates.
Gate structures are also found separate from dams, and can generally be considered as control points where water can be released from the diversion system. They can provide a release point from a canal, as is shown in Figure 3.2 C. It is important to note that nearly all gates observed during all field surveys were completely closed, prohibiting any release from the diversion system. Notably, gates remained in the same position while surveyed areas were visited multiple times from June to October and in some cases over multiple years. In some cases, gates that could function
27


to conduct water back into the natural stream were not connected to any pipes that would allow that possibility. An example of this is shown in Figure 3.3, a detached outlet pipe at Berthoud Pass. The pipe labeled S conducts the water into the diversion pipe that directs the water south over Berthoud Pass and across the Divide into Clear Creek County. The pipe labeled with blue paint would release some water back into the natural stream course and towards the Colorado River to the west. However, the pipe was not connected to the gate system, preventing the possibility of any water release. Close examination indicated the pipe was never connected. The field surveys demonstrate the thorough level of diversion in the area, the implications of which will be discussed further.
28


Figure 3.3. Detached Diversion Outlet Pipe in Berthoud Pass Area. The pipe on the right would allow water out of the diversion into the natural course, but is not attached.
Canals can be classified as concrete, stone, and dirt, as depicted in Figure 3.2 D-F. This differentiation is important, as each type can capture and contain water to varying degrees. Open canals intercept surface and in some cases subsurface water. Dirt canals intercept some subsurface water, yet can potentially allow water to enter or re-enter the soil. Concrete
29


canals prevent water from re-entering the soil, and may capture some subsurface water depending on slope and amount of exposed soil above the edge of the canal. Stone canals generally have stone walls may contain mortar between stones, and have dirt bottoms. These canals are likely more permeable than concrete canals, yet less permeable than dirt canals. Canals are significant in size; the dirt canal in the Rollins Pass area depicted in Figure 3.2 D was measured at approximately 18 feet in width.
Relative to canals, ditches are smaller, shallower, and more rudimentary. Grates, depicted in Figure 3.2 G, are permanently open collection points along ditches, collecting direct precipitation as well as surface water draining from slopes and along dirt ditches. Some structures were of indeterminate function, and were classified as indeterminate structures. At some points, streams were piped directly into canals without any feature to allow for release of water past the diversion, as shown in Figure 3.2 H. These have been termed as complete intercepts for this study. Flowever, since most gates can be considered permanently closed to releasing water, this definition is not entirely adequate. Finally, all diverted water is eventually passed into pipeline. Above-surface pipeline is readily apparent, as shown in Figure 3.2 I, yet less frequently used. Most
30


pipeline is buried, and typically follows dirt roads before traveling in tunnels under the Continental Divide.
In summary, the diversity and functionality of diversion structures indicate the complexity of the diversion network and the ability of the network to remove large amounts of water from the basins of origin. Also, there is evidence that streams are entirely removed at many diversion points.
3.1.4 Surveyed Areas
The area west of Rollins Pass was surveyed from the Fraser River Dam northwards for approximately 16 miles. The main diverted portion was approximately 10.5 miles, and the northern 5.5 miles was generally buried aqueduct. The elevation of the main diverted portion ranged from 9358 ft to 9644 ft, with a GPS accuracy ranging from +/- 9 ft to +/-12 ft. The Rollins Pass survey revealed a large number and variety of diversions: dams, dirt and concrete canals, gates, grates, dirt ditches, and above-surface pipeline and buried pipeline. Surveyed points were mapped in ArcMap, with basemap topographic data from the USDA, as shown in Figure 3.4. Due to the large number and often concurrent nature of diversion structures,
31


structure types are not differentiated in the map. Also, buried pipeline were not mapped, as these are generally continuous and connect the abovesurface diversion points.
The area northwest of Berthoud Pass was surveyed from the top of Berthoud Pass west and north around Current Creek basin for approximately 1.5 miles. The elevation ranged from 11264 ft to 11352 ft, with a GPS accuracy ranging from +/- 9 ft to +/-18 ft. The Berthoud Pass survey revealed an extensive network of diversion structures: dams, gates, intercepts, and all types of canals. Surveyed points are mapped in Figure 3.5, similar to Figure 3.4.
The area west of Jones Pass was surveyed for approximately 3.5 miles from Steelman Creek in the southwest of the basin to McQueary Creek in the north of the basin. The elevation ranged from 10284 ft to 10519 ft, with a GPS accuracy ranging from +/- 9 ft to +/-15 ft. The diversions consisted of 3 major dams, and a range of other dam sizes. Surveyed points and types are shown in Figure 3.6.
32


Surveyed Diversions West of Rollins Pass Grand County, Colorado
Southern Section of Survey: Approximately 10.5 Miles
Sharon Smolinski
December 12.2010
Coordinate System:
NAD 83 UTM 2one13N
Topographic Data Source USOA/NRCS
Surveyed Diversions
^- i Miles
0 1 2
A
Figure 3.4. Field Survey of Rollins Pass Area. Shown are the survey points in the southern portion of the area.
33


Surveyed Diversions in Berthoud Pass Area (Current Creek Basin) Grand County, Colorado

Sharon Smolinski
December 12,2010
Coordinate System NAD 83 UTM Zone13N
Topographic Data Source USDA/NRCS
Figure 3.5. Field Survey of Berthoud Pass Area. Shown are the survey points in Current Creek basin.
34


Sharon Smdinski
Surveyed Diversions
0 25
Surveyed Diversions in Williams Fork Area V\fest of Jones Pass, Grand County, Colorado
December 12,2010
Coordinate System NAD 83 UTM Zone13N
Topographic Data Source: s
USDA/NRCS A
Figure 3.6. Field Survey of Williams Fork Area. Shown are the survey points in the Williams Fork Area, west of Jones Pass.
35


Networks of dirt roads follow the diversions, providing vehicle access for water agencies. The Rollins Pass diversion roads are open to public vehicles, and the entire area is subject to significant human activity, including recreational use, logging, and construction. The Berthoud Pass diversion road is closed to public vehicles, but open to non-motorized recreational use. The William Fork diversion roads provide only partial and limited access to public vehicles. Due to the remote location and vehicle restrictions, this area experiences seasonal and limited recreational impact.
The original thesis proposal planned to include study areas in non-diverted areas as controls. However, as indicated by the CDSS data, all basins in the target region west of the Divide are diverted at a minimum of one point, generally more. Bordering areas east of the Divide receive diverted water at multiple locations. As a result, those areas do not act as true controls. Areas farther west of the target region are separated from the target area by distance that is sufficient to create significant differences in precipitation, soil, and other factors. As a result, non-diverted areas were not available or feasible to use as controls in this study.
36


3.2 Detailed Field Studies
3.2.1 Study Sites
Areas for detailed field study were selected based on information gathered in the field surveys of diversions, including the location and types of diversions as well as general area characteristics such as the level of human activity. Three study areas were selected in the Rollins Pass area. Two areas were established around open dirt canal diversions, and a third around a dam diversion. One area was established in the Berthoud Pass area around an open dirt canal diversion, in addition to two areas utilized only for soil measurements. A study area was selected in the Williams Fork area, but was discontinued due to safety concerns related to large animal activity, including bears. In the Rollins Pass area, variably impacted by human disturbance, study areas were selected to exclude human disturbances such as recreational dirt roads, campsites, or logging. Human disturbance was not a site selection issue in the Berthoud Pass area due to lack of public motor vehicle access. In the Berthoud Pass area, study areas were selected to exclude terrain hazards such as cliffs or rockfall. Those terrain hazards were not present in the Rollins Pass area.
37


Due to the fact that every stream in the area of interest was diverted, a non-diverted control basin could not be established in the immediate area west of the Continental Divide. Due to the fact that basins east of the Continental Divide received the diverted water, control basins could not be established in those adjacent areas. Non-diverted control basins were not established in different regions due to the fact that would introduce other variations between basins, such as difference in geology and precipitation.
Each study area consists of sampling plots above and below the diversion structure, as shown in Table 3.2. Plots positioned around open canals were chosen by walking along the diversion and entering the habitat at regular intervals. Plot centers were approximately 18.3 m (60 ft) from the diversion or neighboring road or path. This proximity was selected so that the diversion was the most immediate factor influencing the plots. Plots below the diversion were selected first, followed by selection of the plots above the diversion and approximately directly above a lower plot. Plots positioned around the dam diversion were selected in a similar manner by walking along the stream. Plot centers were approximately 10.7 m (35 ft) from the stream edge. Plots were established at increasing distance from the dam to keep all plots on the same aspect. Sampling plots had radius of
38


9.75 m (32 ft), similar to other studies (Morehouse et al., 2008; Negron et al., 2009). Measuring tape was used to measure distance from the center outwards, and the perimeter was marked. Central plot coordinates were mapped in ArcMap, as shown in Figures 3.7-3.11, with basemap topographic data from the USDA (USDA, n.d.).
Table 3.2. Summary of study areas and plots.
Study Area Diversion Aspect Total Plots Plots Above Diversion Plots Below Diversion
Rollins Pass 1 dirt canal W 4 2 2
Rollins Pass 2 dirt canal W 6 3 3
Rollins Pass 3 dam N 4 2 2
Berthoud Pass 1 dirt canal E 4 2 2
39


Visual Assessment Study Areas
in Grand County, Colorado
Sharon Smolinski Visual Assessement Study Areas
Berthoud Pass 1
December 12,2010 Rollins Pass 1
Rollins Pass 2
Coordinate System: NAD 83 UTM Zone13N Rollins Pass 3
Topographic Data Source: USDA/NRCS
^---------------- i Miles
0 125 2.5
A
Figure 3.7. Study Areas in Grand County.
40


Rollins Pass Study Area 1
Grand County, Colorado
Sharon Smolinski
December 12,2010
Coordinate System:
NAD 83 UTM Zone 13N
Topographic Data Source USDA/NRCS
Figure 3.8. Rollins Pass Study Area 1. Plots were established above and
below a dirt canal.
41


Rollins Pass Study Area 2
Grand County, Colorado
Sharon Smolinski
December 12,2010
Coordinate System NAD 83 UTM Zone 13N
Topographic Data Source USOA/NRCS
Figure 3.9. Rollins Pass Study Area 2. Plots were established above and
below a dirt canal.
42


Rollins Pass Study Area 3
Grand County, Colorado
Sharon Smolinski
December 12,2010
Coordinate System NAD 83 UTM Zone 13N
Topographic Data Source USDA/NRCS
Figure 3.10. Rollins Pass Study Area 3. Plots were positioned above and
below a dam.
43


Berthoud Pass Study Area 1 Grand County, Colorado


Sharon Smolinski
December 12,2010
Coordinate System NAD 83 UTM Zone 13N
Topographic Data Source USDA/NRCS
Berthoud Pass Study Area 1 Sampling Rots
Q Above Diversion A Below Diversion
iMilr~
0 02 0 4
A
Figure 3.11. Berthoud Pass Study Area 1. Plots were positioned above
and below a dirt canal.
44


3.2.2 Basic Site Assessment
In each plot, elevation and coordinates of the central point, as well as aspect, were measured using the GPS unit. Accuracy of the GPS unit was recorded for each measurement. A photograph was taken from the central point. Slope was estimated with a slope meter. Every tree in the plot was counted and DBH was measured using diameter measuring tape. Every tree was identified by genus, and species if possible, following the Field Guide to Trees of North America (Kershner et al., 2008). All study areas were examined between July and October of 2010. All plots within a study area were sampled within a two-day period. After measurement, trees were marked with chalk to ensure trees were only counted once.
3.2.3 Assessment of Pine Beetle Infestation
All pine trees were visually assessed for pine beetle indicators and scored using rating scale similar to published studies (Breece et al., 2008; Jones et al., 2004; Negron et al., 2008; Negron et al., 2009; Sanchez-Martinez & Wagner, 2002; Waring & Six, 2005; Zausen et al., 2005). In this study, the only pine species in the area was lodgepole pine. Living trees without indicators were deemed healthy and rated 1. Trees with evidence
45


of holes (bore or exit), and entirely brown needles or lost needles, were deemed beetle-killed trees and rated 3. Some of these trees also exhibited extruded sap. Trees with indicators but with a majority of green needles were deemed attacked by pine beetles but living and rated 2. Trees dead from indeterminate or non-infestation-related causes were rated 4. All lodgepole pines with DBH equal to or greater than 2.54 cm (1 in) were assessed for indicators. However, only pines with DBH equal to or greater than 7.6 cm (3 in), and 15.24 (6 in) were utilized for analysis. This range of values is similar to those implemented in other studies (Breece et al., 2008, Morehouse et al., 2008; Negron et al., 2008; Negron et al., 2009; Waring & Six, 2005).
3.2.4 Assessment of Percentage of Green Canopy
Other visible factors were also noted that might impact tree health, including animal activity such as gnawing or rubbing, as well as fungal infections. Non-pine tree species were also assessed for health using a rating system. Living trees without any dead needles or leaves were considered healthy and rated 1. Living trees with some yellow or brown needles or leaves were rated 2. Dead trees were rated 4. An additional
46


assessment of percentage of green needles was utilized in some study areas. This approach aimed to estimate the percentage of green needles based on visible foliage. This was utilized to serve as an additional measure of tree health of non-pine species.
3.2.5. Soil Water Content
Soil water content was assessed using a FieldScout TDR100 Soil Moisture Meter from Spectrum Technologies. This unit measures volumetric water content, at 0.1% resolution and +/- 3.0% accuracy (Spectrum Technologies, n.d.). A similar unit has been used in mountainous terrain with publishable results (Penna et al., 2009). Measurements were taken at a total of fourteen plots located in the following study areas: Rollins Pass Areas 1 and 2, and Berthoud Pass. Twenty measurements were collected per plot, sampled randomly throughout the plots within 3 feet on the western side of a tree. In a given study area, all samples were taken within a 3 hour time period. Measurements were attempted with two different probes, 7.6 cm (3 in) and 12cm (4.75 in). However, high rock content in the soil prevented use of the 12 cm probe, resulting in use of the 7.6 cm probe.
47


3.2.5 Vegetation Water Content
Vegetation water content was assessed by harvesting needle samples from Lodgepole Pines in four sampling plots in Rollins Pass Study Area 2. Protocol was similar to that used by Salle et al. (2008) and Zausen et al. (2005). A twelve-foot Corona compound action tree pruner was used to cut branches from the maximum height possible. The maximum height achieved for sampling, based on manageable length of pruner and height of sampler, was approximately 510.5 cm (201 in). Needles from branch ends (newest growth) were removed by hand, placed in plastic bags, and stored on ice in a cooler. Two bundles of needles were collected from each tree. Samples were transported to a lab and weighed before and after drying at 105C for 72 hours, similar to previous studies. A permit was obtained from the USFS Sulphur Range District authorizing the removal of needles (Appendix A).
48


3.2.6 Statistical Analysis of Field Data
Microsoft Excel software was used to tabulate data, calculate basic statistics, and produce graphs. Minitab 16.1.1 software was used to determine normality using graphical distribution plots, perform hypothesis tests, and produce some graphs (Minitab, 2011). Selection of hypothesis tests was based on data type and distribution. A standard a-value of 0.05 was used.
3.3 Analysis of Remotely Sensed Data
3.3.1 Data and Sources
Landsat 5 TM data were obtained from the USGS GloVis website as a standard Level 1T terrain-corrected product in GeoTIFF format (USGS, June 2011). Landsat 7 ETM+ data could not be used due to data gaps over areas of interest, the result of the 2003 failure of the scan-line corrector (USGS, January 2011). The entirety of Grand County, containing the areas of interest, is covered in a single Landsat scene, path 34, row 32. Scenes were selected based on minimal cloud cover, either 0% for the entire scene or absence of cloud cover over areas of interest. Scenes were selected from September 2003 and September 2010, and obtained from GloVIS as
49


files LT50340322003265PAC02 (USGS, n.d., a), and LT50343221010268PAC01 (USGS, n.d., b).
In addition to Landsat data, additional data were obtained for reference and analysis. Basemap data, including county boundary, stream and road data, were obtained from CDOT (Colorado Department of Transportation, 2010). Landcover data (USDA, 2006) and an orthorectified aerial photo were obtained from the U.S. Department of Agriculture (U.S.D.A., 2000).
3.3.2 Data Analysis
All geospatial data were processed in ArcGIS 10 (Esri, 2011), specifically ArcCatalog and ArcMap, with the exception of the use of Minitab for statistical tests on NDVI change values. NDVI values were calculated for the 2003 and 2010 Landsat scenes, and then the 2003 raster was subtracted from the 2010 raster. The resulting raster contained values for the change in NDVI, with negative values reflecting a decrease in healthy vegetation and positive values reflecting an increase in healthy vegetation.
In this way, the change in NDVI over the seven-year period was calculated per pixel. The NDVI change values can then be compared between
50


habitats above and below diversions. By comparing the change values, instead of NDVI values for a single scene, the analysis does not need to account for slope and aspect variations between compared pixels.
Different methods to isolate areas above and below diversions were explored. First, attempts focused on selecting areas corresponding to study areas and plots. The diversion and service road were manually isolated by creating polylines in the editor extension, which were then buffered for 100 ft. Nearby habitat above and below the diversion were then manually selected against this buffered area. These areas were then converted to rasters and used to extract the NDVI change values. In order to mask out obvious non-vegetated areas as well as non-coniferous vegetated areas, landcover data from the National Landcover Dataset were limited to coniferous vegetation and used as a mask for NDVI change values. The change values were then compared between habitats.
Factors related to spatial resolution and reference data prevented the isolation and analysis of pixels corresponding to most study areas, with the exception of Rollins Pass Study Area 3. First, the Landsat raster data had 30 m pixel resolution. This pixel size generally corresponded to mixtures of land types and features, as shown in Figure 3,12. Pixels
51


containing plots often contained diversion roads and diversions as well.
This was particularly an issue in the Rollins Pass area, which contains a variety of land types and uses, including logged habitat, dirt roads, and nearby development. Furthermore, the orthorectified aerial photo that was used for reference has a processed data of 2000 (U.S.D.A., 2001). As a result, newly logged or developed areas are not included in the 2000 photo, limiting its usefulness for areas with changing land use
52


Limitations of NDVI Spatial Resolution: Pixel Coverage, Plots, and Landscape
Sharon Smolinski
Data Sources:
USDA: ortho photo USGS: landsat data
Coordinate system: NAD 83 UTM Zone 13N
Figure 3.12. Spatial Resolution Limitations. This map shows Landsat data overlaid on an aerial photo. The red box indicates one 30 m pixel which contains forest, road and the edge of a diversion. This illustrates how Landsat pixels correspond to a variety of land types and uses. The red points represent the center points of study plots.
Ortho photo
Value
- High : 255
Low 1
NDVI Change Value
High : 39
- Low:-54
Plot*
Placement O Above Diversion O Beiow Diversion
3 Feet
53


A second method utilized an entire basin bisected by a diversion road which corresponded to a series of point diversions, previously shown in Figure 3.6. Areas were selected as previously described, but included the entire basin. The change values were then compared between above-and below-diversion habitats. Statistical tests were performed on the raster data contained in the attribute tables using Minitab.
54


4. Results
4.1 Field Surveys of Water Diversions
The field surveys produced information regarding the locations and types of diversions in three areas. These were reported in detail in 3.1.3 and 3.1.4.
4.2 Detailed Field Studies
4.2.1 Study Sites
Study areas and plots are summarized in Table 4.1. All study areas and plots were subjected to basic site assessment, which included tree counts, DBH measurements, and tree identification.
Table 4.1. Summary of study areas and plots.
Study Area Diversion Total Plots Plots Above Plots Below
Rollins Pass 1 dirt canal 4 2 2
Rollins Pass 2 dirt canal 6 3 3
Rollins Pass 3 dam 4 2 2
Berthoud Pass 1 dirt canal 4 2 2
55


4.2.2 Tree Density
Tree count data are shown in Table 4.2. All trees above 2.5 cm (1 in) were counted. Only lodgepole pines with DBH greater than or equal to
7.6 cm (3 in) were assessed for pine beetle indicators. Rollins Pass 1 and 2 data can be combined, based on the similarities in the study areas including the aspect and type of diversion, a dirt canal. In fact, a continuation of the same diversion divides both areas. The Rollins Pass study areas contained aspen, fir, and spruce. The Berthoud Pass study area was comprised entirely of fir and spruce, and as a result could not be used for visual assessment of pine beetle indicators.
56


Table 4.2. Summary of tree counts for study areas. Rollins Pass 1 and 2 data can be combined, based on the similarities in the study areas including the aspect and type of diversion, a dirt canal.
Study Area Plot Type Lodgepole Pines >7.6cm DBH Lodgepole Pines > 2.5cm DBH All Trees > 2.5 cm DBH
Rollins Pass 1 Above 78 116 198
Below 44 72 131
Total 122 188 329
Rollins Pass 2 Above 294 408 411
Below 132 157 216
Total 426 565 627
Rollins Pass 1 and 2 Above 372 524 609
Below 176 229 347
Total 548 753 956
Rollins Pass 3 Above 65 95 101
Below 50 69 71
Total 115 164 172
Berthoud Pass Above 98
Below 122
Total 220
Above-diversion plots contained more trees than below-diversion plots in the Rollins Pass areas. In the Berthoud Pass area, above-diversion plots contained fewer trees than below-diversion plots. This difference in the Berthoud Pass area may be attributable to the higher elevation of those plots, ranging from 11279 to 11407 feet, as shown in Table 4.3. In
57


comparison, the elevation range in the Rollins Pass areas is 9329 to 9578 feet. In the Berthoud Pass area, the decrease in tree density over the short distance from below- to above-diversion plots may be due to the natural effect of decreasing tree density with elevation near treeline. In contrast, the Rollins Pass areas are at much lower elevation, so elevation should not have an impact on tree density over the narrow elevation distance between above- and below-diversion plots.
58


Table 4.3. Summary of plot central elevation.
Study Area Plot Plot Type Elevation (ft) Accuracy (ft)
Rollins Pass 1 3 Above 9509 9
4 Above 9513 11
1 Below 9493 11
2 Below 9489 10
Rollins Pass 2 3 Above 9420 12
4 Above 9471 10
5 Above 9421 10
1 Below 9329 11
2 Below 9340 9
6 Below 9359 10
Rollins Pass 3 3 Above 9578 9
4 Above 9546 15
1 Below 9447 9
2 Below 9479 10
Berthoud Pass 1 Above 11407 11
2 Above 11390 21
3 Below 11279 14
4 Below 11316 10
59


Table 4.4. Summary of tree counts by plot.
Study Area Plot Plot Type Lodgepole Pines > 7.6cm DBH Lodgepole Pines > 2.5cm DBH All Trees > 2.5 cm DBH
Rollins Pass 1 3 Above 51 79 82
4 Above 27 37 116
1 Below 17 28 46
2 Below 27 44 85
Rollins Pass 2 3 Above 84 108 111
4 Above 98 126 126
5 Above 106 189 189
1 Below 47 56 74
2 Below 37 51 93
6 Below 49 54 57
Rollins Pass 3 3 Above 44 50 54
4 Above 21 44 47
1 Below 26 43 44
2 Below 24 26 27
Berthoud Pass 1 Above 53
2 Above 45
3 Below 69
4 Below 53
Tree counts were also tabulated by plot, as shown in Table 4.4. Counts were compared using non-parametric tests, the Mann-Whitney and Kruskal-Wallis. There was not a statistically significant difference in tree density, likely due the small sample size which is based on the number of plots. There is also considerable variability in tree density within plot types.
60


4.2.3 Tree Diameter
Tree DBH varies by plot type, as shown in Table 4.5. Abovediversion plots had smaller average, median, and maximum tree diameters. Rollins Pass Areas 1 and 2 data were combined and tested together, and Rollins Pass Area 3 data were tested separately. Tree diameters were different with statistical significance (p = 0.000) between above- and below-diversion plots, based on the two-way difference Mann-Whitney and Kruskal-Wallis tests. Above-diversion plots contained statistically significant (p = 0.000) smaller tree diameters than below-diversion plots, based on the one-way difference Mann-Whitney test.
61


Table 4.5. Summary of tree diameter data. A minimum DBH cutoff of 7.6 cm (3 in) is applied. Data are given in inches.
Study Area Plot Type Number of trees Avg DBH Median DBH Minimum DBH Maximum DBH
Rollins Pass 1 Above 78 6.3 6.2 3.0 11.9
Below 44 7.9 6.9 3.0 14.3
Rollins Pass 2 Above 294 5.5 5.5 3.0 10.3
Below 133 7.3 7.3 3.0 13.7
Rollins Pass 1 and 2 Above 372 5.7 5.6 3.0 11.9
Below 177 7.4 7.2 3.0 14.3
Rollins Pass 3 Above 65 6.8 6.8 3.0 15.6
Below 50 9.7 10.4 3.1 15.8
Berthoud Pass Above 98 8.8 6.9 1.0 33.2
Below 122 6.7 5.1 1.0 32.0
Similar to tree density, tree diameter data for the Berthoud Pass area differed from Rollins Pass data. Above-diversion plots contained higher average, median, and maximum DBH values. Tree diameters were different between above- and below-diversion plots with statistical significance (p = 0.039, p=0.004), based on the two-way difference Mann-Whitney and Kruskal-Wallis tests respectively. Above-diversion plots contained statistically significant (p = 0.019) larger tree diameters than below-diversion plots, based on the one-way difference Mann-Whitney test.
62


Similar to tree density, this difference in tree diameter may be affected by elevation.
4.2.4 Species Composition
Species composition of the study areas can be examined in more detail, to give more detailed characterization of the plots, as shown in Table 4.6. At Rollins Pass Study Area 1, data show the species compositions are similar across plot types, with higher percentages of aspen than any other study site. At Rollins Pass 2, above-diversion plots contained over 99% lodgepole pine, while below-diversion plots contained higher amounts of aspen. At Rollins Pass 3, percentages of lodgepole pines remain consistently high, above 90%, across both plot types. At Berthoud Pass, data show the species compositions are similar between above- and below-diversion plots, with spruce dominating. Most sites, with the exception of Rollins Pass 2, show similar species compositions across plot types, which showed an increased amount of aspen in below-diversion plots.
63


Table 4.6 Summary of tree type percentages.
Study Area Plot Type Percent Tree Type
Lodgepole Pine Aspen Fir Spruce
Rollins Pass 1 Above 58.59 37.89 3.03 0.51
Below 54.96 38.17 0.76 6.11
Rollins Pass 2 Above 99.3 0.7 0 0
Below 71.88 28.12 0 0
Rollins Pass 1 and 2 Above 86.38 12.50 0.96 0.16
Below 65.64 31.83 0.28 2.25
Rollins Pass 3 Above 93.07 0 5.94 0.99
Below 97.18 1.41 1.41 0
Berthoud Pass 1 Above 0 0 31.63 68.37
Below 0 0 28.93 71.31
4.2.5 Assessment of Pine Beetle Infestation
Lodgepole pines were visually assessed for pine beetle infestation using two different minimum DBH values, greater than or equal to 7.6 cm (3 in). The data for Rollins Pass Study Areas 1 and 2 were combined based on the fact that they have the same aspect and type of diversion. Data for Rollins Pass Study Area 3 were analyzed separately, since the aspect and diversion type differed from the other two Rollins Pass areas.
A summary table of beetle indicator data using a minimum DBH of
7.6 cm (3 in) is shown in Table 4.7, and graphs are shown in Figures 4.1 and 4.2. For all Rollins Pass Study Areas, data show a decreased
64


percentage of healthy trees (1) in below-diversion areas. There are also an increased percentage of beetle-kill trees (3) in below-diversion areas. In Rollins Pass Study Areas 1 and 2, below-diversion areas also show an increased percentage of live attacked trees (2) and dead trees from other causes (4). However, this is not the case with Rollins Pass Study Area 3.
65


Table 4.7. Summary of pine beetle indicator data. All lodgepole pines with a DBH greater than or equal to 7.6 cm (3 in) were assessed for pine beetle indicators. The ratings are as follows: live trees without indicators (1), live trees with indicators (2), dead trees with indicators (3), and dead trees from
indeterminate or non-pine-beetle-related causes (4).
Study Area Plot Score Number of Trees Percentage
Rollins Pass 1 & 2 Above 1 324 87.10
2 18 4.84
3 18 4.84
4 12 3.22
Below 1 81 45.76
2 23 13.00
3 62 35.03
4 11 6.21
Rollins Pass 3 Above 1 53 81.54
2 5 7.69
3 3 4.62
4 4 6.15
Below 1 14 28.00
2 3 6.00
3 31 62.00
4 2 4.00
66


Visual Assessment of PineBeetle Indicators in Lodgepole Pines
(DBH £ 7.6 cm) in Rollins Pass Study Areas 1 and 2
Pine Beetle Indicator Rating
Figure 4.1. Pine Beetle Indicator Data for Rollins Pass Study Areas 1 and 2. All lodgepole pines with a DBH greater than or equal to 7.6 cm (3 in) were assessed for pine beetle indicators. The ratings are as follows: live trees without indicators (1), live trees with indicators (2), dead trees with indicators (3), and dead trees from indeterminate or non-pine-beetle-related causes (4).
67


100
Visual Assessment of Pine Beetle Indicators in Lodgepole
Pines (DBH > 7.6 cm) in Rollins Pass Study Area 3
Pine Beetle Indicator Rating
Figure 4.2. Pine Beetle Indicator Data for Rollins Pass Study Area 3. All lodgepole pines with a DBH greater than or equal to 7.6 cm (3 in) were assessed for pine beetle indicators. The ratings are as follows: live trees without indicators (1), live trees with indicators (2), dead trees with indicators (3), and dead trees from indeterminate or non-pine-beetle-related causes (4).
The cross-tabulation and chi-square method tested for two-way difference in proportions of pine beetle indicators between plot types, and determined statistically significant difference (p=0.000) for combined Rollins
68


Pass Study Areas 1 and 2, and Rollins Pass Study Area 3. Specifically, below-diversion areas have statistically significant higher percentages of beetle-kill trees and lower percentage of live healthy trees.
Additionally, the indicator data were reclassified as beetle-kill and non-beetle-kill, as shown in Figures 4.3 and 4.4. Here, all non-beetle-kill trees include trees previously scored 1, 2 or 4. Percentages of beetle-kill in above- and below-diversion areas were statistically significant (p = 0.000), based on cross-tabulation and chi-square tests, as well as Fishers exact test.
69


Visual Assessment of Pine Beetle Indicators in Lodgepole
Pines (DBH > 7.6 cm) in Rollins Pass Study Areas 1 and 2
dJ
c
d)
o
CL
dJ
CD
O
C
(U
o
d)
Q.
Above Diversion
Below Diversion
non-beetle-kill beetle-kill
Figure 4.3. Beetle-Kill Trees in Rollins Pass Study Areas 1 and 2. All lodgepole pines with a DBH greater than or equal to 7.6 cm (3 in) were assessed for pine beetle indicators.
70


Visual Assessment of Pine Beetle Indicators in Lodgepole Pines
(DBH £ 7.6 cm) in Rollins Pass Study Area 3
Figure 4.4. Beetle-Kill Trees in Rollins Pass Study Area 3. All lodgepole pines with a DBH greater than or equal to 7.6 cm (3 in) were assessed for pine beetle indicators.
An examination of DBH values show minimum DBH values for beetle-kill trees are 4.0 and 3.9 in, for Rollins Pass Study Areas 1 and 2, and Rollins Pass Study Area 3, respectively. However, median DBH values for beetle-kill trees were much higher, at 9.3 and 12.1 in, for Rollins Pass Study Areas 1 and 2, and Rollins Pass Study Area 3, respectively. Recall
71


above-diversion plots contained a higher number of smaller diameter trees than below-diversion plots.
In order to compensate for these factors, the data can be reexamined using a minimum DBH cutoff of 15.2 cm (6 in), as shown in Figures 4.5 and 4.6. Trends are similar to the previous analyses.
Compared to above-diversion areas, below-diversion areas had increased percentages of beetle-kill trees and decreased percentages of healthy trees. With the increased DBH cutoff value, there was an increase in the amount of beetle-kill trees and decreased percentage of healthy trees. The cross-tabulation and chi-square method tested for two-way difference in proportions of pine beetle indicators between plot types, and determined statistically significant difference (p=0.000) for combined Rollins Pass Study Areas 1 and 2, and Rollins Pass Study Area 3. Specifically, below-diversion areas have statistically significant higher percentages of beetle-kill trees and lower percentage of live healthy trees.
72


Visual Assessment of Pine Beetle Indicators in Lodgepole
Pines (DBH > 15.2 cm) in Rollins Pass Study Areas 1 and 2
Pine Beetle Indicator Rating
Figure 4.5. Pine Beetle Indicator Data for Rollins Pass Study Areas 1 and 2. All lodgepole pines with a DBH greater than or equal to 15.2 cm (6 in) were assessed for pine beetle indicators. The ratings are as follows: live trees without indicators (1), live trees with indicators (2), dead trees with indicators (3), and dead trees from indeterminate or non-pine-beetle-related causes (4).
73


Visual Assessment of Pine Beetle Indicators in Lodgepole Pines
(DBH £ 15.2 cm) in Rollins Pass Study Area 3
Above Diversion
Below Diversion
Pine Beetle Indicator Rating
Figure 4.6. Pine Beetle Indicator Data for Rollins Pass Study Area 3. All lodgepole pines with a DBH greater than or equal to 15.2 cm (6 in) were assessed for pine beetle indicators. The ratings are as follows: live trees without indicators (1), live trees with indicators (2), dead trees with indicators (3), and dead trees from indeterminate or non-pine-beetle-related causes (4).
74


4.2.6 Assessment of Percentage of Green Canopy
At the Berthoud Pass Study Area, all trees with DBH greater than or equal to 2.5 cm (1 in) were visually assessed for percentage of green canopy. There are a higher percentage of green needles in abovediversion areas, as shown in Figure 4.7. The difference in percentage of green canopy between above- and below-diversion plots is statistically significant (p = 0.000), based on the two-way Kruskal-Wallis test.
Figure 4.7. Percentage of Green Canopy by Plot Type. Graph prepared in Minitab.
75


4.2.7 Soil Water Content
Data show higher soil water content in plots below diversion, compared with plots above diversions, as shown in Table 4.8. Here, the Rollins Pass Study Area 2 and Berthoud Pass Study Area were analyzed separately, based the fact that the Berthoud Pass area did not have lodgepole pines. Rollins Pass 1 was assessed, but not analyzed due to the fact that two rounds of measurements were taken and both interrupted by rain, preventing complete sampling of all plots in the area each time.
At Rollins Pass Study Area 2, below-diversion plots contained increased soil moisture compared to above-diversion plots with statistical significance (p = 0.0004, p = 0.001), based on the two-way difference Mann-Whitney and Kruskal-Wallis tests respectively. At the Berthoud Pass Study Area, below-diversion plots contained increased soil moisture compared to above-diversion plot with statistical significance (p = 0.0327), based on the two-way Mann-Whitney test. However, the Kruskal-Wallis test showed a lack of statistical significance (p = 0.065). Rollins Pass Study Area 2 showed high standard deviation (cr) among measurements within plot types (cr= 3.9, 76


deviation (cr= 1.9, Table 4.8. Summary of soil moisture content data. Measurements are given in percent soil water content (SWC).
Study Area Plot Type Number of samples Avg SWC Median SWC Min SWC Max SWC Stnd Dev
Rollins Pass 2 Above 60 8.2 6.9 3.6 22.3 3.87
Below 60 11.5 10.5 2.2 29.2 5.72
Berthoud Pass Above 40 5.5 4.6 3.1 13.5 1.95
Below 40 6.5 6.1 3.1 13.0 2.39
4.2.8 Vegetation Water Content
The maximum height achieved for sampling, based on manageable length of the pruner combined with the height of the sampler, was approximately 510.5 cm (201 in). However, due to tree height, only a small number of trees could be sampled. In Rollins Pass Study Area 2, two above- and two below-diversion plots were sampled. Only 12 total trees could be reached in the combined above-diversion plots, and 12 in the combined below-diversion plots The small sample size precludes any
77


statistical significance to the data. Within the small number of samples, the average vegetation water content was 57.93% for above-diversion trees, and 58.12% for below-diversion trees. The median values were 58.34% for above-diversion trees, and 59.77% for below-diversion trees. Potentially due to small sample size, the data did not show a significant difference in percent vegetation water content due to diversions. Based on the difficulty in reaching a significant number of trees with this approach, vegetation sampling was not pursued further.
4.3 Analysis of Remotely Sensed Data
NDVI values reflecting change in NDVI from 2003 to 2010 were compared between above- and below-diversion areas at Rollins Pass Study Area 3, as shown in Figure 4.8. Negative values indicate loss of vegetation from 2003 to 2010, and increasingly negative values indicate increased loss of vegetation. There was a difference in NDVI change values between above- and below-diversion habitats with statistical significance (p = 0.004, p = 0.004), using the one-way Mann-Whitney and two-way Kruskal-Wallis tests respectively. There was a greater loss of healthy vegetation in below-
78


diversion plots with statistical significance (p = 0.0022), using the two-way Mann-Whitney test.
79


Change in NDVI at Rollins Pass Study Area 3
(2003 to 2010)
Sharon Smolinski
Data Sources:
USDA: ortho photo USGS: landsat data
Coordinate system:
NAD 83UTM Zone 13N
A
Change In NDVI Below Diversion
H-24--19 I 1-H9--1S I I -14 9--11
HB -10 9 -8
Change In NDVI Above Diversion
B*1*
a-iM-i*
' 1-13.9- -11
1-10 9-8
620
zzaFeet
Figure 4.8. NDVI Change at Rollins Pass Study Area 3. Points represent plots, while colored pixels represent NDVI change values in selected above- and below-diversion areas.
80


NDVI values reflecting change in NDVI from 2003 to 2010 were compared between above- and below-diversion areas in a basin in the Jones Pass area, as shown in Figure 4.9. The below-diversion area contained more negative change values than above-diversion areas with statistical significance (p = 0.000), using the one-way and two-way Mann-Whitney tests.
81


Change in NDVI in Jones Pass Area
(2003 to 2010)
Sharon Smolinski
Data Sources:
USDA: ortho photo USGS: landsat data
Coordinate system:
NAD 83UTM Zone 13N
A
Change in NOVI Value
-34 -29
-24--19
-19.9--15
-14.9--11
-10.9--7
-S.9-.2
-<9-4
40-39
1.400 2.900
5i600 Feet
Figure 4.9. NDVI Change Values at Rollins Pass Study Area 3. Points represent plots, while colored pixels represent NDVI change values in selected above- and below-diversion areas.
82


5. Discussion
5.1 Field Surveys of Water Diversions
This study raises several concerns related to water diversions, particularly in regards to management and extent. First, the data quality issues surrounding the CDSS data raise questions as to whether the data omissions are present only in public data, or in data used by water management agencies. It is critical for water agencies to maintain and utilize accurate and complete diversion databases in order to properly manage water resources.
The field survey data regarding the extent of diversion systems and the amount of water removal are startling. Every stream, no matter how small, was diverted along the surveyed areas. Additionally, nearly all features that could provide for release of water out of the diversion system were observed as closed through surveys over multiple months and in some cases over multiple years. Recall Figure 3.3, which demonstrated that some diversion points are not even properly connected to allow for this possibility. Figure 5.1 further illustrates the issue of closed gates, depicting the dam in Rollins Pass Study Area 3, and the abundant water retained.
83


Figure 5.1. Dam at Rollins Pass Study Area 3. A) Dam, B) dry conduit leading from closed dam gate, and C) dry stream bed below dam. For scale, note the items on the concrete platform in upper right of photo A, an orange hat, bag, and notebook.
84


This suggests a lack of management or any type of assessment of the diversion systems by water agencies. By demonstrating the nearly complete diversion of entire basins, it suggests the potential for large and significant environmental impacts in the diverted basins. This also alludes to the lack of and need for environmental impact assessments regarding existing water diversions.
5.2 Detailed Field Studies
5.2.1 Study Areas and Plots
In this study, three study areas and fourteen plots were established and studied, including the count of 1348 total trees and pine beetle indicator rating of 663 lodgepole pines. The coverage is less than some published studies, particularly Negron et al. (2009), which covered five national forests with 633 plots. The number of plots used in this study was sufficient to determine statistical significance in differences in proportions of pine beetle indicators, tree diameter, and soil moisture content.
Placement of the plots in proximity to diversions was designed to limit effects from other factors. Notably, roads accompany most diversions in order to provide access for water agencies. These roads can be
85


significant in size, measured at approximately eighteen feet in the Rollins Pass area. A similarly large road accompanies the Grand Ditch diversion in Rocky Mountain National Park. These roads may have effects on areas, due to transport of sediments to lower habitat, or other effects. Effects may be limited, since most diversion roads through the areas surveyed in this study were closed to public access. However, since roads are an integral part of diversions, it may not be necessary to differentiate or quantify the effect from the roads as a separate feature. Essentially, whatever effects roads may have on surrounding habitat, those effects could be considered together with the impact from water removal.
It is important to note that this study did not take into account the history of the areas, such as fire, logging, and other factors. Given the close proximity of the plot types to each other, it is presumed that historic effects would likely impact the habitats in a similar way.
5.2.2 Tree Density
Tree density appears to be decreased in above-diversion plots. However, sample size, based on plots, was not sufficient to determine statistical significance. It is important to note that increased stand density is
86


thought to increase occurrence of pine beetle infestation (Breece et al., 2008; Jones et al., 2004; McKinney & Tomback, 2007). Higher tree density in above-diversion plots would suggest increased infestation, compared to below-diversion plots with decreased density; yet the opposite has been shown. This might suggest that water diversion impacts may have greater impact than stand density on infestation. Also, it could be proposed that above-diversion habitat may be able to support more trees due to increased water availability, compared to below-diversion habitat. Additional examination of such areas is needed, and larger study areas would allow for statistical tests.
5.2.4 Tree Diameter
Above-diversion plots had statistically significant smaller tree diameters compared with below-diversion plots. Similar to the case with increased stand density, above-diversion habitat may provide more favorable conditions, compared with below-diversion plots. This could account for more new growth, evidenced through the higher amount of smaller-diameter trees. As with the case of stand density, this could
87


Full Text

PAGE 1

WATER DIVERSIONS CONTRIBUTING TO MOUNTAIN PINE BEETLE INFESTATION AND SUBSEQUENT TREE MORTALITY by Sharon Louise Smolinski B S., University of Colorado Denver 1998 B F A., University of Colorado Denver 1998 A thesis submitted to the University of Colorado Denver in partial fulfillment of the requirements for the degree of Master of Science Environmental Sciences 2011

PAGE 2

This thesis for the Master of Science degree by Sharon Louise Smolinski has been approved by Jon M. Barbour ii ... Peter An thamatten Nov 2e;11 Date

PAGE 3

Smolinski, Sharon Louise (M.S., Environmental Sciences) Water Diversions Contributing to Mountain Pine Beetle Infestation and Subsequent Tree Mortality Thesis directed by Professor Frederick Chambers ABSTRACT This novel study explored the hypothesis that water diversions cause an increased incidence of mountain pine beetle infestation and subsequent increased tree mortality in areas downstream of or below diversion points This study included the use of field surveys to determine the locations and types of diversion present in the areas of interest. Field data was collected from a total of eighteen plots over four study areas The study areas were established around two different types of diversion structures, with plots above and below the diversion. Sites were characterized by tree count, diameter and species, in addition to visual assessment of pine beetle infestation and soil and vegetation moisture content. Additionally, analysis of Landsat data examined differences and changes in NDVI values over time Survey data revealed extensive diversions of entire basins pointing to the need for assessments of environmental impacts Field assessments of beetle infestation showed increased percentages of beetle-kill trees in

PAGE 4

habitat below diversions. These findings support the hypothesis that water diversions cause an increase in mountain pine beetle infestation and subsequent tree mortality in habitat below diversions Other field data suggest the possibility that diversions lead to general habitat effects. This abstract accurately represents the conte recommend its publication.

PAGE 5

DEDICATION I dedicate this thesis to my parents who introduced me to natural world at a young age and who taught me the value of learning I especially want to thank my mother who filled my crib with books

PAGE 6

ACKNOWLEDGEMENT I acknowledge my advisor, Frederick B. Chambers and all of the members of my committee fo r their contribution and insight.

PAGE 7

TABLE OF CONTENTS Figures ............ ... ........... ..... ...... ...... ........ ... ... .......... ......... ..... xii Tables ............... ............... .... ...... ......... .. .... .... ..... ...... ... ....... xiv Chapter 1 Introduction ..... .... .... ......... .......... ... ..................... ...... .......... 1 1.1 Purpose of the Study ... ...... ........ ...... ............... ..... ...... ............. .... 1 1 2 Scope of the Study ........ ........... .... ......... ......... ........ ...... ... ......... 2 2 Literature Review ...... .................. ................................ .... .......... .4 2 1 Water Diversions ....... ............................ ..... ......................... .4 2 1 1 Impacts of Water Diversions ... ............. .... ..... .......... ............. 5 2 2 Mountain Pine Beetle Infestation .................................................. 7 2 2 1 Factors Facilitating Pine Beetle Infestation ............................ ....... 8 2.2.2 Environmental Impacts of Pine Beetle Infestation .... .................... 1 0 2.3 Field Methods ......... ..... ......... .... ......... ...... ............................... 12 2.3 1 Visual Assessment of Pine Beetle Infestation ........................... 12 2 .3.2 Assessment of Moisture Conditions . .............. .... ... .... ........... 13 2 3.3 Statistical Analysis of Field Data ...................................... ......... 14 2.4 Methods Using Remotely Sensed Data ........ ................................. 15 2.4 1 Data Sources .......... ...... ... ... .... .... ....... ............... . .......... 16 2.4.2 Spectral Analysis ............ . ......... .... ... .... .......... ... ... ........ .... 16 viii

PAGE 8

2.4 3 Error mitigation ............ . ................... ........... .. ... ........... .... 18 3 Methods .... ... ... ... .... ... ... ...... ...... ... ...................... ............ 20 3 1 Field Surveys of Water Diversions ... ............. ..... ... .. ....... .... ... .... 20 3 .1.1 Inadequate Public Data ....................................... ... ............. 20 3 1.2 Survey Overview ................ ........... ........ ........ .. ........ .......... 23 3 1 3 Locations and Types of Diversions ... .............. .... ....... ... ....... 24 3 1 4 Surveyed Areas ............ ...... ........... ... . ............. ............ 31 3 2 Detailed F i eld Studies ............. ..... ...... . ..... ....... ....... ..... .... 37 3 2 1 Study Sites .............. ... ....... ...... .......... . ... .............. ... ..... 37 3 2.2 Basic Site Assessment.. ... ...... ... .......... ............ ....... ..... ..... .45 3.2 3 Assessment of Pine Beetle Infestation ...... ... ..... .. ..... .............. .45 3.2.4 Assessment of Percentage of Green Canopy ..... ....... ..... ..... ... ... .46 3.2 5 Soil Water Content.. .... ........................ ........ .... ... ... ... ..... .47 3.2 6 Vegetation Water Content.. ... ...... ...... .... .... ... .. ., ........... .48 3.2.6 Statistical Analysis of Field Data .... .......................... . .......... .49 3 3 Analysis of Remotely Sensed Data .... ........ ........ .... .... .... ......... .49 3 3 1 Data and Sources ... ................. ... .......... ............ ... ...... .... ... .49 3 3.2 Data Analysis ...... ........ .......... .... .. ........... ...... ....... . . 50 4 Results ...... ..... .. ........ ... ..... ..... ...... ....... .. .. .......... ....... ....... 55 4 1 Field Surveys of Water Diversions ... . ......... ......... .......... ..... ..... 55 ix

PAGE 9

4.2 Detailed Field Studies .............................. ......... ... ....... ........ ...... 55 4.2 1 Study Sites ........... ................................................................ 56 4 2 2 Tree Density ......................................................... . ........... 56 4.2.3 Tree Diameter ... .................................................................... 61 4.2.4 Species Composition ....... ...................................................... 63 4.2 5 Assessment of Pine Beetle Infestation .................... .......... ..... 64 4.2 6 Assessment of Percentage of Green Canopy .... ... .. ............ ....... 75 4.2.7 Soil Water Content.. ..... ......................................................... 76 4.2.8 Vegetation Water Content.. .... ................................................ 77 4 3 Analysis of Remotely Sensed Data ............................................... 78 5 Discussion .......................................... ... ... ..... ....... .. ...... ........ 83 5.1 Field Surveys of Water Diversions ................................................ 83 5.2 Detailed Field Studies ....... ... .... ................................................. 85 5 .2.1 Study Areas and Plots ... ................. ..... ............................... 85 5.2 2 Tree Density ...... ....... ... .................................. ......... ... .... 86 5.2.3 Tree Diameter. ......... ........... ................................... ...... ....... 87 5.2.4 Assessment of Beetle Infestation ... ..... ..... ...... ............. ... ...... 88 5 2.5 Assessment of Percentage of Green Canopy ............................ 89 5 2 6 Soil Water Content. ........ .............................................. ...... 90 5.3 Analysis of Remotely Sensed Data ............................................. 93 x

PAGE 10

6 Conclusion ... .. ....... . .... ... ..... .......... ................................ 95 Appendix A. Special-Use Permit.. ....... ........... ............ ... ... ........ .. .. .. ... ..... 98 Bibliography ........... .... .. ..... ........... . ...... ............... ......... ...... ... 101 xi

PAGE 11

LIST OF FIGURES Figures 3.1 Water Diversions in Southeastern Grand County Based on CDSS Data ......... .......... .... ... .. ........ ...... ............. 22 3 2 Types of Diversion Structures ... ................. ....... ................. ....... 25 3.3 Detached Diversion Outlet Pipe in Berthoud Pass Area ................. 29 3.4 Field Survey of Rollins Pass Area ......................... ..................... 33 3.5 Field Survey of Berthoud Pass Area ............. ...... ...... ......... ........ 34 3.6 Field Survey of Williams Fork Area ...... .............. .. .............. ....... 35 3 7 Study Areas in Grand County ... ...... .................... .......... .... .. ..... .40 3 8 Rollins Pass Study Area 1 ...... ....... ..... ...... ..... ...... .................... .41 3 9 Rollins Pass Study Area 2 .............. .... ... ....... .... ....... ....... . .42 3.10 Rollins Pass Study Area 3 ... ......................... .... ............. ...... .43 3.11 Berthoud Pass Study Area 1 ...... .............................. ... .... ..... .44 3.12 Spatial Resolution Limitations .......... ..... .... .......... ......... ....... 53 4.1 Beetle Indicator Data for Rollins Pass Study Areas 1 and 2 ............. 67 4.2 Beetle Indicator Data for Rollins Pass Study Area 3 .......... ......... . 68 4.3 Beetle-kill Trees in Rollins Pass Study Areas 1 and 2 ... ......... ........ 70 4.4 Beetle-kill Trees in Rollins Pass Study Area 3 ........ ........... ..... ... .... 71 4.5 Beetle Indicator Data for Rollins Pass Study Areas 1 and 2 ... ... ... 73 xii

PAGE 12

4 6 Beetle Indicator Data for Rollins Pass Study Area 3 ....... ... ...... .... 74 4 7 Percentage of Green Canopy by Plot Type ... ................. .. ..... ..... 76 4 8 NOVI Change at Rollins Pass Study Area 3 .... .... .... ....... .. .... ... 80 4 9 NOVI Change in Jones Pass Area ... ... .... .. .... .... ..................... 82 5.1 Dam at Rollins Pass Study Area 3 ......................... .. ..... .......... 84 A.1 Special-Use Permit. .................... ..... ...... .... .................. ......... 99 xiii

PAGE 13

LIST OF TABLES Table 3 1 Summary of areas covered by field surveys ..... ...... .. .............. ....... 24 3.2 Summary of study areas and plots ...... ............................ ........... 39 4.1 Summary of study areas and plots ...... ....... .................. ..... .... 55 4.2 Summary of tree counts for study areas ...... ... ..... ... .... ............... 57 4.3 Summary of plot central elevation ........ ....... '" ...... ... ... .... .... ... 59 4.4 Summary of tree counts by plot. ...... .... .................. .......... ... ... .. 60 4.5 Summary of tree diameter data ..... ............ .. ...... .. .... ............ .... 62 4.6 Summary of tree type percentages ................... .................... ... 64 4.7 Summary of pine beetle indicator data .... .... .............................. 66 4.8 Summary of soil moisture content data ....................................... 77 xiv

PAGE 14

1. Introduction 1.1 Purpose of the Study This is a novel study the objective of which is determining if water diversions cause increased incidence of infestation by the mountain pine beetle, Dendroctonus ponderosae Hopkins hereafter referred to as pine beetle". This study is motivated by my first-hand observations over several years in areas of Grand County Colorado. I observed what appeared to be a marked increase in beetle kill trees below water diversions relative to areas above diversions. Presumably water diversions lead to reduced water availability downstream of or below diversion points. This study will examine the hypothesis that water diversions facilitate pine beetle infestation and lead to increased tree mortality in areas downstream of and below diversion points. There has been no research examining a possible relationship between water diversion and pine beetle activity Furthermore there has been no research regarding the impact of water diversion on coniferous trees located within the basin of extraction Based on the widespread extent and impacts of pine beetle infestation research exploring additional factors facilitating pine beetle activity would be of substantial importance. If 1

PAGE 15

a relationship is found the findings could prompt significant changes in water policy and forest management. 1.2 Scope of the Study This study examined the aforementioned hypothesis through field surveys of water diversions detailed field studies examining several factors and analysis of remotely sensed data. Field surveys determined the location and types of diversions present. These field surveys investigated three areas in the southeast quarter of Grand County Colorado, and formed the basis of site selection for the detailed field studies. The detailed field studies included the selection of study areas and sampling plots followed by a variety of assessments. Based on the results of the field surveys three study areas were established in the Roll ins Pass Area in addition to one in the Berthoud Pass area All sampling plots within the study areas were subject to basic characterization including determination of GPS coordinates elevation and aspect, in addition to tree counts measurement of tree diameter at breast height (DBH) and i dentification of all trees Lodgepole p i nes were visually assessed for pine beetle indicators and all trees were assessed for basic health Visual 2

PAGE 16

assessments were also explored based on visual assessments of percentage of green needles Some study areas were assessed for moisture condition, specifically soil water content and vegetation water content. Remotely sensed data were analyzed in a geographic information system (GIS) in order to compare spectral values between diverted and non-diverted areas Landsat data were analyzed in ArcMap10 (Esri 2011) in order to determine and compare the change in NOVI from 2003 to 2010. 3

PAGE 17

2. Literature Review 2.1 Water Diversions Water diversion projects stretch across the Western United States and Canada transporting significant amounts of water out of basins of origin. In Colorado numerous projects transport water from the west side of the Continental Divide to the east side servicing the developed Front Range for municipal water consumption and industrial application. Many water projects were instituted before environmental impacts we r e considered (Doremus & Hanemann 2007). The field surveys for this study found numerous diversion structures in the Rollins Pass area dating to the 1950s Only recently have some water policy decisions begun to consider environmental issues (Doremus & Hanemann 2007) However as climate driven drought progresses there may be a revived battle between water demand and environmental impacts (Doremus & Hanemann 2007). Unfortunately the impacts of diversions have not been adequately explored 4

PAGE 18

2.1.1 Impacts of Water Diversions Diversions result in reduced water availability downstream from diversion points Researchers have demonstrated detrimental impacts to mountain streams as well as habitats far downstream (Carriquiry & Sanchez 1999 ; Rodriguez et aI., 2001; Schone et aI., 2003 ; Wohl 2006). Researchers have also demonstrated impact to high-elevation wetlands as a r esult of water diversions (Chimner & Cooper 2003). However there has been an absence of research examining impact on coniferous habitat at the site of high-elevation water diversion projects in the basins of origin Also, there has been an absence of research regarding a correlation between water diversion projects and mountain pine beetle activity Wohl (2006) discussed significant and numerous impacts to mountain streams from reduced flow including sedimentation water temperature, and available habitat. These effects have in turn impacted multiple species resulting in measureable reductions in biotic integrity (Wohl, 2006) In fact impacts to flow within mountain streams in the Rocky Mountains of the United States are rated as severe", due to the number of projects and amount of water involved (Wohl 2006 p 222) 5

PAGE 19

Marine habitats far downstream from diversion points which may be states away, are also affected Reduced freshwater flows alter the sediment composition and salt content within estuaries (Carriquiry & Sanchez, 1999). Reduced freshwater flows have impacted the growth and distribution of some species in these same estuaries (Rodriguez et aI., 2001 ; Schone et aI., 2003) These endpoints of river systems receive the cumulative effect of all water diversions along the contributing watersheds producing a greater measureable effect. Outside of mountain streams, there is little research regarding diversion impacts on high-altitude ecosystems Notably researchers have shown an impact to high-altitude wetlands as a result of diversion projects (Chimner & Cooper 2003). Diversion projects result in lower water tables (Chimner & Cooper 2003) These lower water tables caused increased CO2 emissions from subalpine fens due to increased decomposition (Chimner & Cooper 2003). Previous inquiries into a potential relationship between water diversion projects and pine beetle infestation have been limited to the impacts of beetle kill trees on diversions (Rey 2008 ; USFS, 2008a). Specifically, concerns only included the possibility of decreased water 6

PAGE 20

quality and blockage of diversion channels from fallen trees (Rey 2008 ; USFS 2008a) The USDA directly pointed to a threat to the water supply for Front Range metropolitan areas (Rey, 2008) Based on the extent of diversions and the limited research on the impacts of those operations additional research is needed particularly into the impacts on the ecosystems within the basins of extraction 2.2 Mountain Pine Beetle Infestation The mountain pine beetle has decimated pine forests throughout the western US and Canada in recent years (Colorado State University [CSU] n d.; Gillette et aI., 2009 ; Jones et aI., 2004; Kurtz et aI., 2008; McKinney and Tomback 2007 ; Miller & Spoolman 2009 ; Steventon & Daust 2009) A native species, the pine beetle can provide beneficial ecosystem services when acting on a limited number of trees including reduction of stand density and facilitation of natural wildfire (Jones et aI., 2004; Logan & Powell 2001 ; USFS n.d.). However in recent years pine beetle activity has escalated considerably resulting in vast regions of damaged and dead forest (CSU n d ; Gillette et aI., 2009 ; Jones et aI., 2004 ; Kurtz et aI., 2008 ; 7

PAGE 21

McKinney & Tomback 2007; Miller & Spoolman 2009 ; Steventon & Daust 2009). 2.2.1 Factors Facilitating Pine Beetle Infestation Researchers have identified a variety of factors that have facilitated the escalation of pine beetle activity Largely anthropogenic these factors include drought increased temperatures increased stand density resulting from fire suppression and pollutants (Bleiker & Six 2009 ; Breece et aL, 2008 ; Gillette et aL, 2009 ; Jones et aL, 2004 ; Logan & Powell 2001 ; McKinney & Tomback 2007 ; Miller & Spoolman 2009 ; U.S Forest Service [USFS] n d ) Drought adversely impacts tree health, creating favorable conditions for the pine beetle (Bleiker & Six 2009 ; Jones et ai 2004; USFS n d ) Tree resin concentration is altered by drought conditions, thereby reducing resistance to pine beetles (Jones et aL, 2004) Drought leads to lowered moisture content in trees, increasing vulnerabil i ty to colonization by the fungi associated with the pine beetle (Bleiker & Six 2009) Forecasts indicate that climate change has increased drought frequency and severity in the Western U S through increased temperatures and altered 8

PAGE 22

precipitation effects that are expected to intensify as climate change progresses (IPCC 2007 ; Logan & Powell, 2001; Miller & Spoolman 2009) Higher temperatures enable pine beetle activity to shift to higher altitudes and latitudes further intensifying impacts and potentially allowing pine beetles to infest species other than pines (Logan & Powell 2001) These increased temperatures allow larvae populations to survive through the winter (Cain 2005). Researchers have determined that climate change has thereby already increased the severity and extent of pine beetle activity (Gillette et aI., 2009 ; Logan & Powell 2001 ; Miller & Spoolman 2009) Wildfire is a crucial natural force for modifying forest structure, but modern forest management actively suppresses fire (Breece et aI., 2008; Jones et aI., 2004; Logan & Powell 2001; McKinney & Tomback 2007) This management policy produced increased stand density, providing an increased and densely associated host supply (Breece et aI., 2008; Jones et aI., 2004; McKinney & Tomback, 2007) Also increased stand density has amplified the effects of drought forcing larger populations of trees to share increasingly limited supplies of water (Jones et aI., 2004; McKinney & Tomback 2007). 9

PAGE 23

Researchers have determined that pollutants also facilitate infestation (Jones et aL, 2004) The mechan i sm for impact is that anthropogenic pollutants enter the atmosphere and are deposited in remote regions posing a significant threat to natural habitats (Jones et aL, 2004) Jones et aL (2004) examined ponderosa pine already affected by drought and found increased pine beetle attacks as well as increased tree mortality in areas with higher concentrations of nitrogen and ozone. 2.2.2 Environmental Impacts of Pine Beetle Infestation Damage resulting from pine beetle infestation causes significant localized and global impacts (Breece et aL, 2008; Jones et aL, 2004; Kurtz, et aL, 2008 ; Logan & Powell 2001 ; McKinney & Tomback, 2007, Steventon & Daust 2009) The coniferous forests of western North America support a wide range of species and provide important ecosystem serv i ces. The large scale loss of pine trees presents a dramatic change to these ecosystems impacting all resident species. For example whitebark pines are a key food source for a variety of animals including Clark s Nutcracker and grizzly bears (Logan & Powell 2001 ; McKinney & Tomback, 2007 ; Miller & Spoolman, 2009) As whitebark pine populations decrease those 10

PAGE 24

species are adversely impacted (Logan & Powell 2001 ; McKinney & Tomback, 2007). The loss of foliage cover can lead to increased predation upon prey species (Hoffman, 2006). Changes in prey populations in turn affect predators such as pine martens (Steventon & Oaust 2009) Indeed the loss of entire regions of forest will cause habitat fragmentation and edge effects further impacting species increasing predation and creating opportunities for invasive species (Primack 2006) In general habitat loss leads to reduced populations of species and reduced biodiversity (Miller & Spoolman 2009 ; Primack 2006). P i ne beetle infestation also has substantial global effect by contr i buting to the progression of climate change (Kurtz et aI., 2008) Forests provide an important ecosystem service by absorbing CO2 acting as carbon sinks (Kurtz et aI., 2008; Miller & Spoolman 2009 ; Primack 2006) The reduction in functional forests has reduced CO2 uptake and increased CO2 input through decomposition (Kurtz et aI., 2008) As explained by Kurtz et al. (2008) these forests now function as carbon sources instead of carbon sinks thereby accelerating climate change 11

PAGE 25

2.3 Field Methods 2.3 1 Visual Assessment of Pine Beetle Infestation Numerous studies have relied on visual assessment to determine pine beetle infestation in pine trees (Breece et aI., 2008; Clow et aI., 2011 ; Jones et aI., 2004 ; Kulakowski & Jarvis, 2011 ; Morehouse et aI., 2008 ; Negron et aI., 2009 ; Sanchez Martinez & Wagner, 2002; Waring & Six 2005 ; Zausen et aI., 2005) Assessments classified trees based on the presence or absence of indicators of pine beetle activity including brown needles faded crowns bore holes, boring dust, pitch tubes, and actual pine beetles (Breece et aI., 2008 ; Jones et aI., 2004; Kulakowski & Jarvis 2011 ; Negron et aI., 2009 ; Waring & Six 2005; Zausen et aI., 2005) Ratings systems generally typically differentiated between live trees without indicators live trees with indicators dead trees with indicators and dead from other causes (Breece et aI., 2008 ; Negron et aI., 2008; Negron et aI., 2009 Waring & Six 2005) Some researchers assessed all trees within plots, or all trees with a specified DBH (Breece et aI., 2008 ; Jones et aI., 2004 ; Morehouse et aI., 2008 ; Negron et aI., 2008 ; Negron et aI., 2009 ; Sanchez-Martinez & Wagner, 2002 ; Waring & Six, 2005) Explained by Waring and Six (2005), 12

PAGE 26

researchers use DBH limits to exclude trees that are too small to be subject to pine beetle attack (Waring & Six 2005) However, minimum DBH limits were not consistent across studies including 2 54 cm (1 in) (Negron et aL, 2008 ; Negron et aL, 2009), 5 cm (2 in) (Morehouse et aL, 2008) 7 6 cm (3 in) (Waring & Six 2005) and 13 cm (5 in) (Breece et aL, 2008) Studies which have included visual assessment of pine beetle indicators used a variety of numbers and sizes of plots Negron et aL (2009) utilized 633 circular plots each with an 8 m (26 2 ft) radius spread over five national forests. Morehouse et aL (2008) used 20 circular plots, each with a 10m (32.8 ft) radius Waring and Six (2005) used 105 circular plots each with an 11.28 m (37 ft) radius. 2.3.2 Assessment of Moisture Conditions Researchers have also assessed moisture conditions in field studies examining infestation including soil water content (Claw et aL, 2011), and vegetation water content (Zausen et aL, 2005) Soil water content is a common parameter in assessing water availability in forest ecosystems (Brooks & Kyker-Snowman, 2009 ; Claw et aL, 2011 ; Penna et aL, 2009 ; Salle et aL, 2008; Walker et aL, 2004). Traditionally, soil samples are 13

PAGE 27

removed from varying depths and transported for lab analysis (Clow et aI., 2011 ; Salle et aI., 2008) Typically soil samples are dried at approximately 100C and weighed to determine water content (Salle et aI., 2008) Soil water content is also measured directly on-site using portable soil moisture probes and sensors (Brooks & Kyker-Snowman 2009 ; Penne et aI., 2009 ; Walker et aI., 2004) Assessment of water stress in coniferous trees commonly relies upon needle water content and predawn needle water potential (Salle et aI., 2008; Zausen et aI., 2005) Similar to soil sampling, needles are physically removed and transported for lab analysis where they are dried and weighed to determine water content (Salle et aI., 2008; Zausen et aI., 2005) To determine predawn water potential needles are collected before dawn and measured with a pressure chamber (Salle et aI., 2008 ; Zausen et aI., 2005). 2.3.3 Statistical Analysis of Field Data Researchers have typically used non-parametric statistical tests to analyze field data, including Wilcoxon-rank-sum chi-square, and Kruskal Wallis tests (Breece et aI., 2008; Guarin et aI., 2005 ; Negron et aI., 2008; 14

PAGE 28

Negron et aL, 2009). If parametric tests were used the data were first log transformed (Breece et aL, 2008). 2.4 Methods Using Remotely Sensed Data Researchers have utilized remotely sensed data and GIS to assess and analyze ecosystems in relation to a variety of issues including pine beetle infestation (Coops et aL, 2006 ; Coops et aL, 2009 ; Dennison et aL, 2010 ; Hais et aI., 2009; Hatala et aL, 2010; Hilker et aL, 2009 ; Meddens et aI., 2011; Vogelmann et aL, 2009; White et aL, 2005; Wulder et aL, 2006) This is possible through the use of spectral values as measured by satellite or airborne sensor, based on wavelengths of electromagnetic radiation reflected or emitted from vegetation (Campbell, 2007). Spectral analysis allows for assessment of vegetation stress and mortality as well as species identification and canopy characterization (Campbell 2007). The use of th i s spectral data allows researchers to assess infestation in large and remote areas, as well as over time (Campbell 2007 ; Coops et aL, 2006; Coops et aL, 2010 ; Hais et aL, 2009 ; Wulder et aL, 2006). 15

PAGE 29

2.4.1 Data Sources Researchers have examined infestation using a wide range of platforms and sensors including airborne HyMap (Hatala et aL, 2010) airborne Vexcel Utracam-D (Meddens et aL, 2011) ; Landsat TM or ETM+ (Coops et aL, 2006 ; Coops et aL, 2010 ; Hais et aL, 2009 ; Hilker et aL, 2009 ; Vogelmann et aL, 2009 ; Wulder et aL, 2006) MODIS (Coops et aL, 2009) IKONOS (White et aL, 2005) and GeoEye-1 (Dennison et aL, 2010) Onboard sensors provide data of diverse spatial and spectral resolution. For example MODIS provides a low spatial resolution of 250 m (Campbell, 2007) while GeoEye1 provides 1 65 m resolution (Digital Data Services n.d .). These variations in resolution influence data quality and analytical capabilities. While of moderate spatial resolution of 30 m, Landsat data have been used for comparisons of sizeable areas over time (Coops et aL, 2006 ; Coops et aL, 2010 ; Hais et aL, 2009; Vogelmann et aL, 2009 ; Wulder et aL, 2006) 2.4 2 Spectral Analysis In assessing infestation a primary goal has been to differentiate between green canopy, representative of living trees and red, 16

PAGE 30

representative of dead trees with needles (Coops et aI., 2010 ; Dennison et aI., 2010; Meddens et aI., 2011 ; White et aI., 2005 ; Wulder et aI., 2006) Some researchers also differentiated grey canopy which is representative of dead trees that have shed all needles (Coops et ai 2006 ; Coops et aI., 2010 ; Dennison et aI., 2010 ; Wulder et aI., 2006). Non vegetation features can first be masked out in order to more accurately assess the vegetated areas (White et aI., 2005; Wulder et aI., 2006). Pixels in the areas of interest can then be classified using supervised (Dennison et aI., 2010) or unsupervised classification (White et aI., 2005). Some researchers have directly worked from spectral values to evaluate red and green pixels (Dennison et aI., 2010 ; White et aI., 2005). Researchers have also utilized a variety of Vegetation Indices (Vis) in the analysis of spectral data (Coops et aI., 2006 ; Coops et aI., 2009 ; Coops et aI., 2010 ; Hais et aI., 2009 ; Hilker et aI., 2009; Vogelmann at aI., 2009; Wulder et aI., 2006). Vegetation Indices calculate relationships between specific bands in order to assess some factor of vegetation (Campbell, 2007) One common VI is the Normalized Difference Vegetation Index (NDVI) which utilizes red and infrared bands, corresponding to bands 3 and 4 respectively in Landsat TM and ETM+ 17

PAGE 31

(Campbell, 2007 ; Esri, 2010) Studies assessing infestation have used a variety of Vis: NDVI (Hais et aI., 2009; Vogelmann et aI., 2009) Enhanced Difference Vegetation Index (EDVI) (Coops et aI., 2006), Disturbance Index (01) (Vogelmann et aI., 2009), Tasseled Cap Transformation (TCT) (Coops et aI., 2006 ; Wulder et aI., 2006 ; Hais et aI., 2009) and Red Green Index (RGI) (Coops et aI., 2006 ; Meddens et aI., 2011) Some of these indices can evaluate red and green values such as the RGI which calculates the ratio of red to green scaled values (Meddens et aI., 2011). Again classification schemes are used to process these values (Campbell, 2007). 2.4.3 Error mitigation The use of remotely sensed data for ecological assessments carries the potential for error due to the fact that spectral data values can be attributed to a variety of factors (Campbell 2007). Some researchers use airborne imagery to validate satellite imagery in order to verify that features are being interpreted correctly (Dennison et aI., 2010 ; White et aI., 2005). Many researchers use field data physically collected from sites on the ground to validate remotely sensed data using protocols similar to those 18

PAGE 32

d i scussed in 2 3 (Campbell 2007 ; Coops et aI., 2006 ; Hais e t aI., 2009 ; Meddens et aI., 2011; White et aI., 2005 ; Wulder et aI., 2006) 19

PAGE 33

3. Methods 3.1 Field Surveys of Water Diversions Field surveys determined the locations and types of water diversions within the region of interest. Inadequate existing public data necessitated the use of first-hand field surveys The finding of the surveys specifically the locations and types of diversions are presented in 3.1.3 Those findings are discussed here instead of the Results section because they determined the selection of the study areas used for detailed field study. Also these findings characterize the nature of and the extent of diversions in the region the basis for the study. 3.1.1 Inadequate Public Data Field surveys were necessary due to the lack of adequate public data regarding extent location and types of water diversions in the region. Geospatial data regarding diversion projects in Colorado are available from Colorado s Decision Support System the CDSS which is operated by the Colorado Division of Water Resources and the Colorado Water Conservation Board (Colorado s Decision Support System n.d ) Quality issues were apparent when the CDSS data were examined in ArcMap and 20

PAGE 34

compared with first-hand observations as well as GIS data from the Colorado Department of Transportation (COOT) (Colorado Department of Transportation 2010) Major diversions were missing from the CDSS data such as the transbasin diversion at the headwaters of Vasquez Creek. In Figure 3.1, COOT data show the Vasquez Tunnel extending from Vasquez Creek to the adjacent county Furthermore, entire lengths of extensive diversions were omitted such as those bisecting the basins of Current Creek and adjacent basins, northwest of Berthoud Pass. This diversion system is known through first-hand observation. Also, COOT data show this diversion as a stream extending from First Creek south to Berthoud Pass as shown in Figure 3 .1. In addition to omitting diversions CDSS data do not fully differentiate or classify the various types of diversions compared to first-hand observations. Despite omissions, CDSS data indicate that water diversions are extensive and that all streams are diverted and often at multiple points as suggested by the numerous diversion points in Figure 3.1. 21

PAGE 35

Water Diversions in Southeastern Grand County Based on CDSS Data 9'Iaron Smoilnski DIverSIons -streams o County boundaries 0.11 Sources: Colorodo', O.cision Support System: ..-tor diversion dlta Colorado Department ofTran.portot ion: county .nd strum dati Coordinate syst.m : NAn B3 UTM Zone 13N 0;.,;;0.5;..;,1 ==2;... .... iles A Figure 3 1 Water Diversions in Southeastern Grand County Based on CDSS Data 22

PAGE 36

3.1.2 Survey Overview Field surveys were conducted along three areas in Grand County, with the goal of determining locations and types of diversions. A Garmin GPSmap 60CSx GPS unit, with an accuracy of approximately +/10 feet, was used to collect coordinate data of diversion points A digital camera was used to photograph all surveyed diversion points Surveys were conducted from June to October in 2010 Additionally, survey areas were observed multiple times, and some areas were also visited in 2011 The area west of Rollins Pass was surveyed from the Fraser River Dam northwards for approximately 16 miles The diversions bisecting the Current Creek basin northwest of Berthoud Pass was surveyed for approximately 1.5 miles. The area west of Jones Pass was surveyed for approximately 3.5 miles from Steelman Creek in the southwest of the basin to McQueary Creek in the north of the basin. Miles surveyed are based upon the linear distance traveled along diversion lengths. The areas covered by the field surveys are summarized in Table 3.1. 23

PAGE 37

Table 3 .1. Summary of areas covered by field surveys. Area Distance Elevation Accuracy of Transportation Surveyed Range (ft) GPS unit (ft) (miles) Rollins Pass 16 9358-9644 +/-9-12 car Berthoud 1 5 11264 +/ 9 18 foot Pass 11352 Williams Fork 3 5 10284 +/ 9 15 foot 10519 3.1.3 Locations and Types of Diversions Field surveys revealed extensive networks of varied d i version structures and thoroughly diverted basins All streams in the surveyed basins were diverted regardless of their size. Several structural types of diversions were identified and can be classified into ten types and an additional five subtypes. The ten main types are as follows: dams canals gates ditches, grates complete intercepts pipelines, reservoirs tunnels, and indeterminate structures. Canals consist of three subtypes based on materials : dirt concrete and stone Pipelines consist of two subtypes above-surface and buried pipelines. Figure 3.2 provides examples of some of these structures all of which were photographed during the field surveys and are located in one of the survey areas Some structures cannot be photographed such as buried pipelines. 24

PAGE 38

Figure 3.2. Types of Diversion Structures. Examples of some diversion structure types found during field surveys : A) dam 8) dam, C) gate, D) dirt canal, E) concrete canal, F) stone canal. Figure is continued on following page. 25

PAGE 39

Figure 3 2 (Con't.) Types of Diversion Structures. Examples of some diversion structure types found during field surveys: G) grate, H) complete intercept I) pipeline. 26

PAGE 40

Dams were found in a range of sizes, from very large to small based on the size of the stream being intercepted. Dams provide for the potential of some release or overflow of water with gates and spillways. Spillways only allow for overflow during high stream flow Some spillways of small dams are very rudimentary, and functionality may be limited Gates served as control points to allow water through the diversion point, back into the stream or served to control water into the diversion pipe. Large dams generally had two gates, allowing for control of water through the diversion and into the diversion Dams are shown in Figure 3.2 A and B at Bobtail Creek and Jones Creek respectively in the Williams Fork area Note both dams have a spillway and two gates. Gate structures are also found separate from dams, and can generally be considered as control points where water can be released from the diversion system. They can provide a release point from a canal as is shown in Figure 3.2 C. It is important to note that nearly all gates observed during all field surveys were completely closed, prohibiting any release from the diversion system Notably gates remained in the same position while surveyed areas were visited multiple times from June to October and in some cases over multiple years In some cases gates that could function 27

PAGE 41

to conduct water back into the natural stream were not connected to any pipes that would allow that possibility. An example of this is shown in Figure 3.3 a detached outlet pipe at Berthoud Pass The pipe labeled S conducts the water into the diversion pipe that directs the water south over Berthoud Pass and across the Divide into Clear Creek County The pipe labeled with blue paint would release some water back into the natural stream course and towards the Colorado River to the west. However the pipe was not connected to the gate system, preventing the possibility of any water release. Close examination indicated the pipe was never connected. The field surveys demonstrate the thorough level of diversion in the area, the implications of which will be discussed further. 28

PAGE 42

Figure 3.3. Detached Diversion Outlet Pipe in Berthoud Pass Area. The pipe on the right would allow water out of the diversion into the natural course but is not attached. Canals can be classified as concrete stone and dirt as depicted in Figure 3.2 D F This differentiation is important as each type can capture and contain water to varying degrees Open canals intercept surface and in some cases subsurface water Dirt canals intercept some subsurface water yet can potentially allow water to enter or re-enter the soil. Concrete 29

PAGE 43

canals prevent water from re-entering the soil and may capture some subsurface water depending on slope and amount of exposed soil above the edge of the canal. Stone canals generally have stone walls may contain mortar between stones, and have dirt bottoms These canals are likely more permeable than concrete canals yet less permeable than dirt canals Canals are significant in size ; the dirt canal in the Rollins Pass area depicted in Figure 3.2 0 was measured at approximately 18 feet in width Relative to canals ditches are smaller shallower and more rudimentary Grates depicted in Figure 3 2 G are permanently open collection points along ditches collecting direct precipitation as well as surface water draining from slopes and along dirt ditches. Some structures were of indeterminate function and were classified as indeterminate structures". At some points streams were piped directly into canals without any feature to allow for release of water past the diversion as shown in Figure 3 2 H These have been termed as complete intercepts for this study. However since most gates can be considered permanently closed to releasing water, this definition is not entirely adequate. Finally all diverted water is eventually passed into pipeline. Above-surface pipeline is readily apparent as shown in Figure 3.2 I yet less frequently used. Most 30

PAGE 44

pipeline is buried, and typically follows dirt roads before traveling in tunnels under the Continental Divide. In summary, the diversity and functionality of diversion structures indicate the complexity of the diversion network and the ability of the network to remove large amounts of water from the basins of origin. Also, there is evidence that streams are entirely removed at many diversion points. 3.1.4 Surveyed Areas The area west of Rollins Pass was surveyed from the Fraser River Dam northwards for approximately 16 miles The main diverted portion was approximately 10 5 miles, and the northern 5.5 miles was generally buried aqueduct. The elevation of the main diverted portion ranged from 9358 ft to 9644 ft, with a GPS accuracy ranging from +/9 ft to +/12 ft The Rollins Pass survey revealed a large number and variety of diversions: dams dirt and concrete canals, gates, grates, dirt ditches, and above-surface pipeline and buried pipeline. Surveyed points were mapped in ArcMap, with basemap topographic data from the USDA, as shown in Figure 3.4 Due to the large number and often concurrent nature of diversion structures 31

PAGE 45

structure types are not differentiated in the map Also buried pipeline were not mapped as these are generally continuous and connect the above surface diversion points The area northwest of Berthoud Pass was surveyed from the top of Berthoud Pass west and north around Current Creek basin for approximately 1 5 miles The elevation ranged from 11264 ft to 11352 ft, with a GPS accuracy ranging from +/9 ft to +/18 ft. The Berthoud Pass survey revealed an extensive network of diversion structures: dams gates intercepts and all types of canals. Surveyed points are mapped in Figure 3 5 similar to Figure 3.4 The area west of Jones Pass was surveyed for approximately 3 5 miles from Steelman Creek in the southwest of the basin to McQueary Creek in the north of the basin. The elevation ranged from 10284 ft to 10519 ft with a GPS accuracy ranging from + / 9 ft to +/15 ft The diversions consisted of 3 major dams and a range of other dam sizes. Surveyed points and types are shown in Figure 3 6 32

PAGE 46

Surveyed Diversions West of Rollins Pass Grand County Colorado Southern Sect ion o f Surv e y : Approximat ely 10.5 Miles Sharon Smolinski December 12, 2010 Coordinate System: NAO 83 UTM Zone 13N Tq>ographic Data Source: U SDA /NRCS I __ Surveyed Diversions 0 2 Figu r e 3.4. Field Survey of Rollins Pass Area Shown are the survey points in the southern portion of the area 33

PAGE 47

Surveyed Diversions in Berthoud Pass Area (Current Creek Basin) Grand County, Colorado ____ ===::::J,Miles Sharon Smolinski December 12, 2010 Coordinate System: NAD 83 UTM Zone 13N Tq>ographic Data Source: USDA / NRCS Surveyed Diversions o 02 0 4 Figure 3.5. Field Survey of Berthoud Pass Area Shown are the survey points in Current Creek basin. 34

PAGE 48

Surveyed Diversions in Williams Fork Area West of Jones Pass Grand County Colorado Sharon Smolinski Decernber12 2 0 1 0 Coo r dinate System : NAO 83 UTM Zone 13N D ata S ource : U SOA / NRCS o SIll. eyed Diversions I __ -====IAIII 0 25 0 5 Figure 3.6 Field Survey of Williams Fork Area Shown are the survey points in the Williams Fork Area west of Jones Pass. 35

PAGE 49

Networks of dirt roads follow the diversions providing vehicle access for water agencies. The Rollins Pass diversion roads are open to public vehicles and the entire area is subject to significant human activity including recreational use logg ing, and construction The Berthoud Pass diversion road is closed to public vehicles, but open to non motorized recreational use The William Fork diversion roads provide only partial and limited access to public vehicles. Due to the remote location and vehicle restrictions this area experiences seasonal and limited recreational impact. The original thesis proposal planned to include study areas in non diverted areas as controls. However as indicated by the CDSS data all basins in the target region west of the Divide are diverted at a minimum of one point generally more. Bordering areas east of the Divide receive diverted water at multiple locations. As a result those areas do not act as true controls. Areas farther west of the target region are separated from the target area by distance that is sufficient to create significant differences in precipitation soil and other factors As a result non diverted areas were not available or feasible to use as controls in this study 36

PAGE 50

3.2 Detailed Field Studies 3.2.1 Study Sites Areas for detailed field study were selected based on information gathered in the field surveys of diversions, including the location and types of diversions as well as general area characteristics such as the level of human activity Three study areas were selected in the Rollins Pass area Two areas were established around open dirt canal diversions, and a third around a dam diversion One area was established in the Berthoud Pass area around an open dirt canal diversion, in addition to two areas utilized only for soil measurements A study area was selected in the Williams Fork area, but was discontinued due to safety concerns related to large animal activity, including bears In the Rollins Pass area, variably impacted by human disturbance, study areas were selected to exclude human disturbances such as recreational dirt roads, campsites, or logging. Human disturbance was not a site selection issue in the Berthoud Pass area due to lack of public motor vehicle access. In the Berthoud Pass area, study areas were selected to exclude terrain hazards such as cliffs or rockfall. Those terrain hazards were not present in the Rollins Pass area. 37

PAGE 51

Due to the fact that every stream in the area of interest was diverted, a non-diverted control basin could not be established in the immediate area west of the Continental Divide Due to the fact that basins east of the Continental Divide received the diverted water control basins could not be established in those adjacent areas Non diverted control basins were not established in different regions due to the fact that would introduce other variations between basins, such as difference in geology and precipitation. Each study area consists of sampling plots above and below the diversion structure as shown in Table 3.2 Plots positioned around open canals were chosen by walking along the diversion and entering the habitat at regular intervals Plot centers were approximately 18 3 m (60 ft) from the diversion or neighboring road or path This proximity was selected so that the diversion was the most immediate factor influencing the plots. Plots below the diversion were selected first, followed by selection of the plots above the diversion and approximately directly above a lower plot. Plots positioned around the dam diversion were selected in a similar manner by walking along the stream Plot centers were approximately 10.7 m (35 ft) from the stream edge. Plots were established at increasing distance from the dam to keep all plots on the same aspect. Sampling plots had radius of 38

PAGE 52

9 75 m (32 ft) similar to other studies (Morehouse et aI., 2008 ; Negron et aI., 2009). Measuring tape was used to measure d i stance from the center outwa r ds and the perimeter was marked. Central plot coordinates were mapped in ArcMap as shown in Figures 3.7 3.11 with basemap topographic data from the USDA (USDA, n.d ) Table 3 2 Summary of study areas and plots. Study Area Diversion Aspect Total Plots Above Plots Below Plots Diversion Diversion Rollins Pass 1 dirt canal W 4 2 2 Rollins Pass 2 dirt canal W 6 3 3 Roll i ns Pass 3 dam N 4 2 2 Berthoud dirt canal E 4 2 2 Pass 1 39

PAGE 53

Visual Assessment Study A r eas in G r and County Co l orado ___ -===='Mlles Sharon Smolinski DecenAber12,2010 Coordinate S ystem: NAD 83 UTM Zone 13N Tq>ographic Data Source: USDA / NRCS VI.u", A ..... m .. Study Ar Berthoud P ... 1 RoDin. P 1 RoDin. Pass 2 Ronins PI .. 3 Figu r e 3 7. Study Areas in Grand County. 40 o 1 25 2.5

PAGE 54

Sharon Smolinski Decerroer 1 2 2010 Coordinate System: NAD 83 UTM Zone 13N Tcpographi c Data Source: USDA f NRCS Rollins Pass Study Area 1 Grand County Colorado Rolli ... P ... St udy Are. 1 Sampling Plots H:Jovt Djy ... ;on A Below DiYt" ion --=Mfles 00050. 1 Figure 3 8 Rollins Pass Study Area 1 Plots were established above and below a dirt canal. 41

PAGE 55

Sharon Smolinski December 12,201 0 Coordinate System: NAO 83 UTM lone 13N Rollins Pass Study Area 2 Grand County Colorado Rolli ... P ... Study A .... 2 Sampling Plota Ab .. e Diversion A Below Diversi o n 00. 050 1 Tq>ograp hic Data Source: USDAINRCS Figure 3 9 Rollins Pass Study Area 2 Plots were established above and below a dirt canal. 42

PAGE 56

Sha ron S moli nski December 12, 2010 Coordinate System: NAD 83 UTM Zone 13N Tcpographic Data Source: USDA / NRCS Rollins Pass Study Area 3 Grand County Colorado R olli ... P ... Study A r .. 3 Sampling Plot. Above Diver';on ... Below Diversion __ -===:J.MiIe S o 0 .15 0 3 Figure 3.10. Rollins Pass Study Area 3. Plots were positioned above and below a dam 43

PAGE 57

Sharon Smolinski December 12,2010 Coordinate System: NAD 83 UTM Zone 13N Berthoud Pass Study Area 1 Grand County. Colorado ___ -===::J.Miles Benhoud P. t Study A .... 1 0 0 2 04 SIImpllng Plot. PbrN. Divtrsion .6. Be tow Div. rsio n Topographic Data Source: l' A USDA/NRCS Figure 3.11. Berthoud Pass Study Area 1. Plots were positioned above and below a dirt canal. 44

PAGE 58

3.2.2 Basic Site Assessment In each plot elevation and coordinates of the central point as well as aspect were measured using the GPS unit. Accuracy of the GPS unit was recorded for each measurement. A photograph was taken from the central point. Slope was estimated with a slope meter. Every tree in the plot was counted and DBH was measured using diameter measuring tape. Every tree was identified by genus and species if possible following the Field Guide to Trees of North America (Kershner et aI., 2008). All study areas were examined between July and October of 2010 All plots w i thin a study area were sampled within a two day period. After measurement trees were marked with chalk to ensure trees were only counted once. 3.2.3 Assessment of Pine Beetle Infestation All pine trees were visually assessed for pine beetle indicators and scored using rating scale similar to published studies (Breece et aI., 2008; Jones et aI., 2004 ; Negron et aI., 2008 ; Negron et aI., 2009 ; Sanchez Martinez & Wagner, 2002 ; Waring & Six 2005; Zausen et aI., 2005) In this study the only pine species in the area was lodgepole pine Living trees without indicators were deemed healthy and rated 1". Trees with evidence 4S

PAGE 59

of holes (bore or exit) and entirely brown needles o r lost needles were deemed beetle killed trees and rated 3". Some of these trees also exhibited extruded sap Trees with indicators but with a majority of green needles were deemed attacked by pine beetles but living and rated 2 Trees dead from indeterminate or non-infestation related causes were rated 4". All lodgepole pines with DBH equal to or greater than 2.54 cm (1 in) were assessed for indicators. However only pines with DBH equal to or greater than 7 6 cm (3 in), and 15 24 (6 in) were utilized for analysis. This range of values is similar to those implemented in other studies (Breece et aI., 2008 Morehouse et aI., 2008 ; Negron et aI., 2008 ; Negron et aI., 2009 ; Waring & Six 2005) 3.2.4 Assessment of Percentage of Green Canopy Other visible factors were also noted that might impact tree health including animal activity such as gnawing or rubbing as well as fungal infections. Non pine tree species were also assessed for health using a rating system Living trees without any dead needles or leaves were considered healthy and rated 1". Living trees with some yellow or brown needles or leaves were rated 2". Dead trees were rated 4". An additional 46

PAGE 60

assessment of percentage of green needles was utilized in some study areas. This approach aimed to estimate the percentage of green needles based on visible foliage This was utilized to serve as an additional measure of tree health of non pine species 3.2.5. Soil Water Content Soil water content was assessed using a FieldScout TDR100 Soil Moisture Meter from Spectrum Technologies This unit measures volumetric water content at 0 1 % resolution and +/3 0% accuracy (Spectrum Technologies n d.) A similar unit has been used in mountainous terrain with publishable results (Penna et aI., 2009) Measurements were taken at a total of fourteen plots located in the following study areas: Rollins Pass Areas 1 and 2 and Berthoud Pass Twenty measurements were collected per plot, sampled randomly throughout the plots within 3 feet on the western side of a tree In a given study area, all samples were taken within a 3 hou r time period. Measurements were attempted with two different probes, 7.6 cm (3 in) and 12cm (4.75 in) However, high rock content in the soil prevented use of the 12 cm probe resulting in use of the 7.6 cm probe 47

PAGE 61

3.2.5 Vegetation Water Content Vegetation water content was assessed by harvesting needle samples from Lodgepole Pines in four sampling plots in Rollins Pass Study Area 2. Protocol was similar to that used by Salle et al. (2008) and Zausen et al. (2005) A twelve foot Corona compound action tree pruner was used to cut branches from the maximum height possible The maximum height achieved for sampling based on manageable length of pruner and height of sampler was approximately 510 5 cm (201 in) Needles from branch ends (newest growth) were removed by hand placed in plastic bags and stored on ice in a cooler. Two bundles of needles were collected from each tree Samples were transported to a lab and weighed before and after drying at 105C for 72 hours similar to previous studies A permit was obtained from the USFS Sulphur Range District authorizing the removal of needles (Appendix A). 48

PAGE 62

3.2.6 Statistical Analysis of Field Data Microsoft Excel software was used to tabulate data calculate basic statist i cs, and produce graphs. Minitab 16.1. 1 software was used to determine normality using graphical distribution plots perform hypothesis tests and produce some graphs (Minitab 2011) Selection of hypothesis tests was based on data type and d i stribution A standard a -value of 0.05 was used 3.3 Analysis of Remotely Sensed Data 3.3.1 Data and Sources Landsat 5 TM data were obtained from the USGS GloVis website as a standard Level 1T terrain-corrected product in GeoTIFF format (USGS June 2011) Landsat 7 ETM+ data could not be used due to data gaps over areas of interest the result of the 2003 failure of the scan-line corrector (USGS January 2011) The entirety of Grand County containing the areas of interest, is covered in a single Landsat scene, path 34, row 32 Scenes were selected based on minimal cloud cover either 0% for the entire scene or absence of cloud cover over areas of interest. Scenes were selected from September 2003 and September 2010 and obtained from GloVIS as 49

PAGE 63

files LT50340322003265PAC02 (USGS n.d a), and LT50343221010268PAC01 (USGS n.d., b). In addition to Landsat data additional data were obtained for reference and analysis. Basemap data including county boundary stream and road data were obtained from COOT (Colorado Department of Transportation 2010) Landcove r data (USDA 2006) and an orthorectified aerial photo were obtained from the U.S Department of Agriculture (U. S D A., 2000). 3.3.2 Data Analysis All geospatia l data were processed in ArcGIS 10 (Esri 2011) specifically ArcCatalog and ArcMap with the exception of the use of Minitab for statistical tests on NDVI change values NDVI values were calculated for the 2003 and 2010 Landsat scenes, and then the 2003 raster was subtracted from the 2010 raster. The resulting raster contained values for the change in NDVI with negative values reflecting a decrease in healthy vegetation and positive values reflecting an increase in healthy vegetation. In this way the change in NDVI over the seven-year period was calculated per pixel. The NDVI change values can then be compared between so

PAGE 64

habitats above and below diversions By comparing the change values, instead of NDVI values for a single scene, the analysis does not need account for slope and aspect variations between compared pixels. Different methods to isolate areas above and below diversions were explored First attempts focused on selecting areas corresponding to study areas and plots The diversion and service road were manually isolated by creating polylines in the editor extension which were then buffered for 100 ft. Nearby habitat above and below the diversion were then manually selected against this buffered area. These areas were then converted to rasters and used to extract the NDVI change values In order to mask out obvious non-vegetated areas as well as non coniferous vegetated areas, land cover data from the National Landcover Dataset were limited to coniferous vegetation and used as a mask for NDVI change values. The change values were then compared between habitats Factors related to spatial resolution and reference data prevented the isolation and analysis of pixels corresponding to most study areas, with the exception of Rollins Pass Study Area 3 First the Landsat raster data had 30 m pixel resolution This pixel size generally corresponded to mixtures of land types and features, as shown in Figure 3 12 Pixels 51

PAGE 65

containing plots often contained diversion roads and diversions as well This was particularly an issue in the Rollins Pass area, which contains a variety of land types and uses, including logged habitat dirt roads and nearby development. Furthermore the orthorectified aerial photo that was used for reference has a processed data of 2000 (U. S D A., 2001) As a result newly logged or developed areas are not included in the 2000 photo limiting its usefulness for areas with changing land use 52

PAGE 66

limitations of NDVI Spatial Resolution : Pixel Coverage Plots and Landscape Sharon Smolinski Data Sources : USDA : ortho phclto USGS : landsat data Coordinate system : NAD 83 UTM Zone 13N t-DVl Change v_ -High:39 -lDw: 54 Plots Abov. DN.rsion Bilow Dlv,rslon A Figure 3.12 Spatial Resolution Limitations. This map shows Landsat data overlaid on an aerial photo. The red box indicates one 30 m pixel which contains forest road and the edge of a diversion. This illustrates how Landsat pixels correspond to a variety of land types and uses. The red points represent the center points of study plots. 53

PAGE 67

A second method utilized an entire basin bisected by a diversion road which corresponded to a series of point diversions, previously shown in Figure 3 6 Areas were selected as previously described but included the entire basin The change values were then compared between above and below-diversion habitats Statistical tests were performed on the raster data contained in the attribute tables using Minitab. 54

PAGE 68

4. Results 4.1 Field Surveys of Water Diversions The field surveys produced information regarding the locations and types of diversions in three areas These were reported in detail in 3.1. 3 and 3 1.4. 4 2 Detailed Field Studies 4.2.1 Study Sites Study areas and plots are summarized in Table 4.1. All study areas and plots were subjected to basic site assessment which included tree counts DBH measurements and tree identification Table 4 1 Summary of study areas and plots Study Area D i version Total Plots Plots Plots Above Below Rollins Pass 1 dirt canal 4 2 2 Rollins Pass 2 dirt canal 6 3 3 Roll i ns Pass 3 dam 4 2 2 Berthoud Pass dirt canal 4 2 2 1 55

PAGE 69

4.2.2 Tree Density Tree count data are shown in Table 4 2 All trees above 2 5 cm (1 in) were counted Only lodgepole pines with DBH greater than or equal to 7.6 cm (3 in) were assessed for pine beetle indicators. Rollins Pass 1 and 2 data can be combined based on the similarities in the study areas including the aspect and type of diversion a dirt canal. In fact, a continuation of the same diversion divides both areas The Rollins Pass study areas contained aspen, fir, and spruce The Berthoud Pass study area was comprised entirely of fir and spruce and as a result could not be used for visual assessment of pine beetle indicators 56

PAGE 70

Table 4.2. Summary of tree counts for study areas Rollins Pass 1 and 2 data can be combined, based on the similarities in the study areas including the aspect and type of diversion, a dirt canal. Lodgepole Lodgepole All Trees Study Area Plot Type Pines Pines 2.5cm DBH 2 5cm DBH DBH Rollins Above 78 116 198 Below 44 72 131 Pass 1 Total 122 188 329 Rollins Above 294 408 411 Below 132 157 216 Pass 2 Total 426 565 627 Rollins Above 372 524 609 Pass 1 and Below 176 229 347 2 Total 548 753 956 Rollins Above 65 95 101 Below 50 69 71 Pass 3 Total 115 164 172 Berthoud Above 98 Below 122 Pass Total 220 Above-diversion plots contained more trees than below-diversion plots in the Rollins Pass areas In the Berthoud Pass area, above-diversion plots contained fewer trees than below-diversion plots. This difference in the Berthoud Pass area may be attributable to the higher elevation of those plots ranging from 11279 to 11407 feet, as shown in Table 4 3 In 57

PAGE 71

comparison the elevation range in the Rollins Pass areas is 9329 to 9578 feet. In the Berthoud Pass area the decrease in tree density over the short distance from below to above-diversion plots may be due to the natural effect of decreasing tree density with elevation near treeline. In contrast, the Rollins Pass areas are at much lower elevation, so elevation should not have an impact on tree density over the narrow elevation distance between aboveand below-diversion plots. 58

PAGE 72

Table 4.3 Summary of plo t central elevation Study A r ea Plot Plot Type Elevation Accuracy (ft) (ft) 3 Above 9509 9 Rollins 4 Above 9513 11 Pass 1 1 Below 9493 11 2 Below 9489 10 3 Above 9420 12 4 Above 9471 10 Roll i ns 5 Above 9421 10 Pass 2 1 Below 9329 11 2 Below 9340 9 6 Below 9359 10 3 Above 9578 9 Rollins 4 Above 9546 15 Pass 3 1 Below 9447 9 2 Below 9479 10 1 Above 11407 11 Berthou d 2 Above 11390 21 Pass 3 Below 11279 14 4 Below 11316 10 59

PAGE 73

Table 4.4. Summary of tree counts by plot. Lodgepole Lodgepole All Trees Plot Pines Study Area Plot Type Pines 7 6cm 2 5cm 2.5 cm DBH DBH DBH 3 Above 51 79 82 Rollins 4 Above 27 37 116 Pass 1 1 Below 17 28 46 2 Below 27 44 85 3 Above 84 108 111 4 Above 98 126 126 Rollins 5 Above 106 189 189 Pass 2 1 Below 47 56 74 2 Below 37 51 93 6 Below 49 54 57 3 Above 44 50 54 Rollins 4 Above 21 44 47 Pass 3 1 Below 26 43 44 2 Below 24 26 27 1 Above 53 Berthoud 2 Above 45 Pass 3 Below 69 4 Below 53 Tree counts were also tabulated by plot as shown in Table 4.4 Counts were compared using non parametric tests the Mann Wh i tney and Kruskal Wallis. There was not a statistically significant difference i n tree density, likely due the small sample size which is based on the number of plots There is also considerable variability in tree density within plot types. 60

PAGE 74

4.2.3 Tree Diameter Tree DBH varies by plot type as shown in Table 4.5. Above diversion plots had smaller average median and maximum tree diameters. Rollins Pass Areas 1 and 2 data were combined and tested together, and Rollins Pass Area 3 data were tested separately Tree diameters were different w ith statistical significance (p = 0 000) between above and below diversion plots based on the two way difference Mann-Whitney and Kruskal Wallis tests Above diversion plots contained statistically significant (p = 0.000) smaller tree diameters than below-diversion plots based on the one way difference Mann Whitney test. 6 1

PAGE 75

Table 4 .5. Summary of tree diameter data. A minimum DBH cutoff of 7 6 cm (3 in) is applied. Data are given in inches. Study Plot Number Avg Median Minimum Maximum Area Type of trees DBH DBH DBH DBH Rollins Above 78 6 3 6 2 3.0 11.9 Pass 1 Below 44 7.9 6.9 3.0 14.3 Rollins Above 294 5 5 5.5 3 0 10.3 Pass 2 Below 133 7 3 7.3 3 0 13 7 Rollins Above 372 5 7 5.6 3 0 11.9 Pass 1 Below 177 7.4 7 2 3 0 14.3 and 2 Rollins Above 65 6 8 6 8 3 0 15 6 Pass 3 Below 50 9 7 10.4 3 1 15.8 Berthoud Above 98 8.8 6 9 1 0 33.2 Pass Below 122 6.7 5.1 1 0 32 0 Similar to tree density tree diameter data for the Berthoud Pass area differed from Rollins Pass data Above-diversion plots contained higher average, median, and maximum DBH values. Tree diameters were different between above and below diversion plots with statistical significance (p = 0 039, p=0.004) based on the two way difference MannWhitney and Kruskal Wallis tests respectively Above diversion plots contained statistically significant (p = 0 019) larger tree diameters than below-diversion plots based on the one way difference Mann Whitney test. 62

PAGE 76

Similar to tree density, this difference in tree diameter may be affected by elevation. 4.2.4 Species Composition Species composition of the study areas can be examined in more detail, to give more detailed characterization of the plots, as shown in Table 4 .6. At Rollins Pass Study Area 1, data show the species compositions are similar across plot types, with higher percentages of aspen than any other study site. At Rollins Pass 2, above-diversion plots contained over 99% lodgepole pine, while below-diversion plots contained higher amounts of aspen At Rollins Pass 3, percentages of lodgepole pines remain consistently high, above 90%, across both plot types. At Berthoud Pass, data show the species compositions are similar between aboveand below diversion plots with spruce dominating Most sites with the exception of Rollins Pass 2, show similar species compositions across plot types, which showed an increased amount of aspen in below-diversion plots. 63

PAGE 77

Table 4.6 Summary of tree type percentages Percent Tree Type Study Area Plot Type Lodgepole Aspen Fir Spruce Pine Rollins Pass 1 Above 58.59 37 89 3 03 0.51 Below 54.96 38 17 0 76 6.11 Rollins Pass 2 Above 99.3 0.7 0 0 Below 71. 88 28 .12 0 0 Rollins Pass 1 Above 86 38 12 50 0.96 0 16 and 2 Below 65 64 31. 83 0 28 2 25 Rollins Pass 3 Above 93.07 0 5.94 0.99 Below 97.18 1.41 1.41 0 Berthoud Pass 1 Above 0 0 31. 63 68 .37 Below 0 0 28.93 71.31 4.2.5 Assessment of Pine Beetle Infestation Lodgepole pines were visually assessed for pine beetle infestation using two different minimum DBH values, greater than or equal to 7 6 cm (3 in). The data for Rollins Pass Study Areas 1 and 2 were combined based on the fact that they have the same aspect and type of diversion Data for Rollins Pass Study Area 3 were analyzed separately, since the aspect and diversion type differed from the other two Rollins Pass areas A summary table of beetle indicator data using a minimum DBH of 7.6 cm (3 in) is shown in Table 4.7 and graphs are shown in Figures 4.1 and 4.2 For all Rollins Pass Study Areas, data show a decreased 64

PAGE 78

percentage of healthy trees (1) in below-diversion areas There are also an increased percentage of beetle-kill trees (3) in below diversion areas In Rollins Pass Study Areas 1 and 2 below diversion areas also show an increased percentage of live attacked trees (2) and dead trees from other causes (4) However, this is not the case with Rollins Pass Study Area 3. 65

PAGE 79

Table 4.7. Summary of pine beetle indicator data All lodgepole pines with a DBH greater than or equal to 7 6 cm (3 in) were assessed for pine beetle indicators The ratings are as follows: live trees without indicators (1), live trees with indicators (2), dead trees with indicators (3), and dead trees from indeterminate or non-pine-beetle-related causes (4) Study Area Plot Score Number of Percentage Trees 1 324 87.10 Above 2 18 4.84 3 18 4.84 Rollins Pass 4 12 3.22 1&2 1 81 45.76 Below 2 23 13.00 3 62 35.03 4 11 6.21 1 53 81.54 Above 2 5 7.69 3 3 4.62 Rollins Pass 4 4 6.15 3 1 14 28.00 Below 2 3 6.00 3 31 62.00 4 2 4 00 66

PAGE 80

(/) 100 Q) c a:: 80 Q) (5 a. Q) 60 C> '0 0 ....J 40 -0 ...... c 20 Q) u 'Q) a... 0 Visual Assessment of PineBeetle Indicators in Lodgepole Pines (DBH 7 6 cm) in Rollins Pass Study Areas 1 and 2 1 2 3 4 Pine Beetle Indicator Rating -Above Diversion o Below Diversion Figure 4 1 Pine Beetle Indicator Data for Rollins Pass Study Areas 1 and 2. All lodgepole pines with a DBH greater than or equal to 7.6 cm (3 in) were assessed for pine beetle indicators The ratings are as follows : live trees without indicators (1), live trees with indicators (2), dead trees with indicators (3), and dead trees from indeterminate or non-pine-beetle related causes (4) 67

PAGE 81

100 (/J Q) 80 c a:: Q) 60 (5 a. OJ Q) 40 "0 0 -l -20 0 ...... c Q) 0 () .... Q) a.. Visual Assessment of Pine Beetle Indicators in Lodgepole Pines (DBH 7 6 cm) i n Rollins Pass Study Area 3 2 3 Pine Beetle Indicator Rat ing 4 -Above Dive r sion o Below Diversion Figure 4 .2. Pine Beetle Indicator Data for Rollins Pass Study Area 3. All lodgepole pines with a DBH greater than or equal to 7 6 cm (3 in) were assessed for pine beetle indicators. The ratings are as follows : live trees without indicators (1) live trees with indicators (2) dead trees with indicators (3), and dead trees from indeterminate or non-pine-beetle related causes (4) The cross tabulation and chi square method tested for two way difference in proportions of pine beetle indicators between plot types and determined statistically significant difference (p=O.OOO) for combined Rollins 68

PAGE 82

Pass Study Areas 1 and 2, and Rollins Pass Study Area 3 Specifically below-diversion areas have statistically significant higher percentages of beetle kill trees and lower percentage of live healthy trees Additionally, the indicator data were reclassified as beetle-kill and non-beetle-kill, as shown in Figures 4.3 and 4.4. Here, all non beetle-kill trees include trees previously scored 1, 2 or 4. Percentages of beetle-kill in aboveand below-diversion areas were statistically significant (p = 0 000), based on cross-tabulation and chi-square tests as well as Fisher's exact test. 69

PAGE 83

I/) 100 QJ c a:: QJ 80 (5 a. QJ 60 OJ "0 0 ....J ..... 40 0 C QJ 20 () ..... QJ (l, 0 Visual Assessment of Pine Beetle Indicators in Lodgepole Pines (DBH 7 6 cm) in Rollins Pass Study Areas 1 and 2 -Above Diversion o Below Diversion non beetle kill beetle-kill Figure 4.3. Beetle-Kill Trees in Rollins Pass Study Areas 1 and 2. All lodgepole pines with a DBH greater than or equal to 7 6 cm (3 in) were assessed for pine beetle indicators 70

PAGE 84

(/) OJ c: 0:: OJ "0 a. 0> OJ "0 0 -.J -0 C OJ U L-OJ a. Visual Assessment of Pine Beetle Indicators in Lodgepole P i nes (DBH 7 6 cm) in Rollins Pass Study Area 3 100 80 -Above 60 D i version 40 o Below D i ve r sion 20 0 non-beetle-kill beetle kill Figure 4.4 Beetle-Kill Trees in Rollins Pass Study Area 3. All lodgepole pines with a DBH greater than or equal to 7.6 cm (3 in) were assessed for pine beetle indicators An examination of DBH values show minimum DBH values for beetle-kill trees are 4 0 and 3 9 in, for Rollins Pass Study Areas 1 and 2 and Rollins Pass Study Area 3 respectively. However median DBH values for beetle-kill trees were much higher, at 9.3 and 12 1 in, for Rollins Pass Study Areas 1 and 2 and Rollins Pass Study Area 3 respectively. Recall 71

PAGE 85

above-diversion plots contained a higher number of smaller diameter trees than below-diversion plots. In order to compensate for these factors, the data can be reexamined using a minimum DBH cutoff of 15 2 cm (6 in), as shown in Figures 4.5 and 4.6. Trends are similar to the previous analyses Compared to above-diversion areas below-diversion areas had increased percentages of beetle-kill trees and decreased percentages of healthy trees. With the increased DBH cutoff value there was an increase in the amount of beetle-kill trees and decreased percentage of healthy trees The cross-tabulation and chi-square method tested for two way difference in proportions of pine beetle indicators between plot types and determined statistically significant difference (p=O.OOO) for combined Rollins Pass Study Areas 1 and 2 and Rollins Pass Study Area 3 Specifically, below diversion areas have statistically significant higher percentages of beetle-kill trees and lower percentage of live healthy trees 72

PAGE 86

Visual Assessment of Pine Beetle Indicators in Lodgepole Pines (DBH 15.2 cm) in Rollins Pass Study Areas 1 and 2 II) 100 -r----------------, Q) C 0::: Q) "0 c.. Q) OJ -0 o -I .... o c Q) Q) a.. 80 60 40 20 o 1 -Above Diversion o Below Diversion 2 3 4 Pine Beetle Indicator Rating Figure 4.5 Pine Beetle Indicator Data for Rollins Pass Study Areas 1 and 2. All lodgepole pines with a DBH greater than or equal to 15 2 cm (6 in) were assessed for pine beetle indicators. The ratings are as follows: live trees without indicators (1), live trees with indicators (2), dead trees with indicators (3), and dead trees from indeterminate or non-pine-beetle-related causes (4) 73

PAGE 87

100 If) a.> 80 c:: a::: a.> 60 "0 0.. C> a.> 40 "0 0 ....J ..... 20 0 c:: a.> 0 u .... a.> 0-Visual Assessment of Pine Beetle Indicators in Lodgepole Pines (DBH 15 2 cm) in Rollins Pass Study Area 3 1 2 3 Pine Beetle Indicator Rating 4 -Above Diversion o Below Diversion Figure 4.6. Pine Beetle Indicator Data for Rollins Pass Study Area 3. All lodgepole pines with a DBH greater than or equal to 15.2 cm (6 in) were assessed for pine beetle indicators. The ratings are as follows: live trees without indicators (1), live trees with indicators (2), dead trees with indicators (3), and dead trees from indeterminate or non-pine-beetle-related causes (4). 74

PAGE 88

4.2.6 Assessment of Percentage of Green Canopy At the Berthoud Pass Study Area, all trees with DBH greater than or equal to 2 5 cm (1 in) were visually assessed for percentage of green canopy There are a higher percentage of green needles in abovediversion areas as shown in Figure 4.7 The difference in percentage of green canopy between above and below diversion plots is statistically significant (p = 0.000) based on the two-way Kruskal-Wallis test. 100 80 60 8. 40 20 o Individual Value Plot of Percent Green Canopy III III III, 111111 --. 111111-above type IIIIIIIIIU. below Figure 4 .7. Percentage of Green Canopy by Plot Type. Graph prepared in Minitab 75

PAGE 89

4.2.7 Soil Water Content Data show higher soil water content in plots below diversion, compared with plots above diversions, as shown in Table 4.8. Here, the Rollins Pass Study Area 2 and Berthoud Pass Study Area were analyzed separately, based the fact that the Berthoud Pass area did not have lodgepole pines. Rollins Pass 1 was assessed, but not analyzed due to the fact that two rounds of measurements were taken and both interrupted by rain preventing complete sampling of all plots in the area each time At Rollins Pass Study Area 2 below-diversion plots contained increased soil moisture compared to above diversion plots with statistical significance (p = 0.0004, P = 0 001), based on the two-way difference Mann-Whitney and Kruskal-Wallis tests respectively. At the Berthoud Pass Study Area below-diversion plots contained increased soil moisture compared to above-diversion plot with statistical significance (p = 0.0327), based on the two-way Mann-Whitney test. However, the Kruskal-Wallis test showed a lack of statistical significance (p = 0 065) Rollins Pass Study Area 2 showed high standard deviation (a) among measurements within plot types (0'= 3 .9, 0'= 5.7), for above and below-diversion plots respectively. The Berthoud Pass Study Area showed lower standard 76

PAGE 90

deviation (0"= 1.9 0"= 2.4), for above and below diversion plots respectively Table 4.8 Summary of soil moisture content data. Measurements are given in percent soil water content (SWC) Study Plot Number Avg Median Min Max Stnd of Area Type samples SWC SWC SWC SWC Dev Rollins Above 60 8 2 6 9 3 6 22.3 3.87 Pass 2 Below 60 11.5 10 5 2 2 29 2 5 72 Berthoud Above 40 5.5 4.6 3 1 13 5 1.95 Pass Below 40 6.5 6.1 3 1 13.0 2.39 4.2.8 Vegetation Water Content The maximum height achieved for sampling based on manageable length of the pruner combined with the height of the sampler, was approximately 510 5 cm (201 in) However due to tree height only a small number of trees could be sampled In Rollins Pass Study Area 2, two above and two below-diversion plots were sampled. Only 12 total trees could be reached in the combined above diversion plots and 12 in the combined below-diversion plots The small sample size precludes any 77

PAGE 91

statistical significance to the data Within the small number of samples, the average vegetation water content was 57.93% for above-diversion trees, and 58 12% for below diversion trees The median values were 58 34% for above diversion trees and 59 77% for below diversion trees. Potentially due to small sample size the data did not show a significant difference in percent vegetation water content due to diversions Based on the difficulty in reaching a significant number of trees with this approach, vegetation sampling was not pursued further. 4.3 Analysis of Remotely Sensed Data NOVI values reflecting change in NOVI from 2003 to 2010 were compared between aboveand below diversion areas at Rollins Pass Study Area 3 as shown in Figure 4.8 Negative values indicate loss of vegetation from 2003 to 2010, and increasingly negative values indicate increased loss of vegetation. There was a difference in NOVI change values between above and below-diversion habitats with statist i cal significance (p = 0.004 P = 0 004), using the one-way Mann Whitney and two-way Kruskal Wall i s tests respectively. There was a greater loss of healthy vegetation in below7 8

PAGE 92

diversion plots with statistical significance (p = 0.0022), using the two-way Mann-Whitney test. 79

PAGE 93

Change in NOVI at Rollins Pass Study Area 3 (2003 to 2010) Sharon Smolinski Data Sourc.s: USDA : ortho photo USGS: landsat data Coordinat. syst.m: NAD 8 3 UTM Zon. 13N -ll. I--U 0-1 3.8--11 -10.1--8 820 __ ==:::;,feet Figure 4 .8. NOVI Change at Rollins Pass Study Area 3 Points represent plots while colored pixels represent NOVI change values in selected above and below-diversion areas. 80

PAGE 94

NOVI values reflecting change in NOVI from 2003 to 2010 were compared between above and below-diversion areas in a basin in the Jones Pass area as shown in Figure 4 .9. The below -diversion area contained more negative change values than above diversion areas with statistical significance (p = 0 000) using the one way and two-way Mann Whitney tests. 81

PAGE 95

Change in NOVI in Jones Pass Area (2003 to 2010) Sharon Smolinski Chong. In NOV I Value 0 ..... 1 .... 00 .. 1,_80:=0 Data Sources : USDA : ortho photo USGS : landsat data Coordinat e system : NAD 83 UTM Zone 13N 54 ... 19 .&.9 .. 1.9 ... 0 -39 Figure 4.9 NOVI Change Values at Rollins Pass Study Area 3. Points represent plots, while colored pixels represent NOVI change values in selected above and below diversion areas. 82

PAGE 96

5. Discussion 5.1 Field Surveys of Water Diversions This study raises several concerns related to water diversions particularly in regards to management and extent. First, the data quality issues surrounding the CDSS data raise questions as to whether the data omissions are present only in public data or in data used by water management agencies It is critical for water agencies to maintain and ut i lize accurate and complete diversion databases in order to properly manage water resources. The field survey data regarding the extent of diversion systems and the amount of water removal are startling. Every stream no matter how small was diverted along the surveyed areas Additionally nearly all features that could provide for release of water out of the diversion system were observed as closed through surveys over multiple months and in some cases over multiple years. Recall Figure 3 .3, which demonstrated that some diversion points are not even properly connected to allow for this possibility. Figure 5 1 further illustrates the issue of closed gates depicting the dam in Rollins Pass Study Area 3 and the abundant water retained 83

PAGE 97

Figure 5.1. Dam at Rollins Pass Study Area 3. A) Dam 8) dry conduit leading from closed dam gate, and C) dry stream bed below dam For scale note the items on the concrete platform in upper right of photo A an orange hat, bag, and notebook 84

PAGE 98

This suggests a lack of management or any type of assessment of the diversion systems by water agencies. By demonstrating the nearly complete diversion of entire basins it suggests the potential for large and significant environmental impacts in the diverted basins. This also alludes to the lack of and need for environmental impact assessments regarding existing water diversions 5.2 Detailed Field Studies 5.2.1 Study Areas and Plots In this study three study areas and fourteen plots were established and studied including the count of 1348 total trees and pine beetle indicator rating of 663 lodgepole pines. The coverage is less than some published studies particularly Negron et al. (2009), which covered five national forests with 633 plots. The number of plots used in this study was sufficient to determine statistical significance in differences in proportions of pine beetle indicators tree diameter and soil moisture content. Placement of the plots in proximity to diversions was designed to limit effects from other factors Notably, roads accompany most diversions in order to provide access for water agencies These roads can be 85

PAGE 99

significant in size measured at approximately eighteen feet in the Rollins Pass area. A similarly large road accompanies the Grand Ditch diversion in Rocky Mountain National Park. These roads may have effects on areas, due to transport of sediments to lower habitat, or other effects. Effects may be limited since most diversion roads through the areas surveyed in this study were closed to public access However, since roads are an integral part of diversions, it may not be necessary to differentiate or quantify the effect from the roads as a separate feature. Essentially, whatever effects roads may have on surrounding habitat, those effects could be considered together with the impact from water removal. It is important to note that this study did not take into account the history of the areas, such as fire logging, and other factors Given the close proximity of the plot types to each other, it is presumed that historic effects would likely impact the habitats in a similar way. 5.2.2 Tree Density Tree density appears to be decreased in above-diversion plots However, sample size, based on plots, was not sufficient to determine statistical significance It is important to note that increased stand density is 86

PAGE 100

thought to increase occurrence of pine beetle infestation (Breece et aI., 2008 ; Jones et aI., 2004 ; McKinney & Tomback, 2007) Higher tree density in above-diversion plots would suggest increased infestation compared to below diversion plots with decreased density ; yet the opposite has been shown. This might suggest that water diversion impacts may have greater impact than stand density on infestation. Also, it could be proposed that above diversion habitat may be able to support more trees due to increased water availability, compared to below diversion habitat. Additional examination of such areas is needed, and larger study areas would allow for statistical tests. 5.2.4 Tree Diameter Above-diversion plots had statistically significant smaller tree diameters compared with below diversion plots. Similar to the case with increased stand density, above-diversion habitat may provide more favorable conditions, compared with below-diversion plots This could account for more new growth evidenced through the highe r amount of smaller-diameter trees As with the case of stand density, this could 87

PAGE 101

potentially indicate water diversions have an overall negative impact to habitat health 5.2.4 Assessment of Pine Beetle Infestation Below-diversion sites had statistically significant lower percentages of healthy trees and higher percentages of beetle kill trees compared with above-diversion sites Two different statistical tests yielded near-zero p values (0. 000) on the percentages of scores, values well below a standard a -value of 0 05 When the data were rescored to compare beetle kill and non beetle-kill trees three different tests yielded near-zero p values The data were also examined using an additional minimum DBH cutoff value with similar results. These findings support the hypothesis that water diversions cause an increase in pine beetle infestation and subsequent tree mortality in habitat below diversions Notably the methods used in the assessment were comparable to those used in other studies (Breece et aL, 2008 Morehouse et aL, 2008; Negron et aL, 2008 Negron et aL, 2009; Waring & Six 2005) The ind i cator rating system was similar to those used by previous studies (Breece et aL, 2008 ; Negron et aL, 2008 ; Negron et aL, 2009 Waring & Six 88

PAGE 102

2005). Also the original DBH minimum limit of 7.6 cm (3 in) was equivalent to that used by Waring and Six (2005) and in the middle of the range used by other studies (Breece et aL, 2008 Morehouse et aL, 2008 ; Negron et aL, 2008 Negron et aL, 2009) The use of a second, higher, DBH minimum cutoff limit of 15.2 cm (6 in) accommodate the wide range of possible DBH cutoff values. 5.2.4 Assessment of Percentage of Green Canopy The visual assessment of percentage of green canopy is intended to provide an estimate of tree health similar to assessment of infestation indicators. Whereas the assessment of pine beetle indicator was based on detecting the presence or absence of specific factors the assessment of percentage of green canopy was based on visual estimations of quantities. As a result assessment of the percentage of green canopy a continuous variable, was subject to error. Visual estimations may vary by observer, introducing subjectivity Also there is a limited view of a tree from the ground preventing visualization of the entire tree Based on these limitations, this type of assessment has limited importance compared to visual assessment of pine beetle indicators 89

PAGE 103

However, the higher percentage of green canopy in above-diversion plots was determined to be statistically significant. As a result, the data could be considered alongside pine beetle indicator data in suggesting diversions impact habitat health for multiple tree species. 5.2.5 Soil Water Content One basic premise of this study is that water diversions cause decreased water availability in areas below diversion points. As a result, one would anticipate decreased soil moisture below diversions However, data show increased soil moisture values below diversions Notably, the findings from the Rollins Pass Study Areas 1 and 2 showed greater statistical significance across two tests compared to the findings from Berthoud Pass In fact the Kruskal-Wallis test determined there was not a statistically significant difference at the Berthoud Pass site This difference between Rollins Pass 1 and 2 and the Berthoud Pass site can be explained, primarily based on the fact that Berthoud Pass is lacking in pine trees. Observations at the Rollins Pass sites noted the ground cover of below-diversion plots was found to be covered with an increased amount of dead needle litter, compared to above-diversion plots This may serve to 90

PAGE 104

trap moisture and cause increased soil moisture values. Also, below diversion plots contained less canopy cover compared to above-diversion plots. This would allow a greater portion of the ground to directly receive precipitation, causing increased soil moisture values The shallow depth of measurement 7.6 cm, may cause soil moisture measurements to be subject to these factors Additionally, due to the decreased numbers of live trees in below-diversion plots, one could expect decreased evapotranspiration, which could lead to increased soil moisture These soil moisture findings from 2010 field measurements have similarity with a 2011 study by Clow et al. That study likewise found increased soil moisture in areas with increased beetle-kill. Clow et al. (2011) also attributed these results to ground litter and decreased evapotranspiration It is important to note that the Clow et al. (2011) study was also conducted in Grand County, with some sampling sites very close to the Rollins Pass area These published findings give credence to the soil moisture results of this study. At Berthoud Pass, observations did not note a significant difference in the amount of needle litter or ground cover. However, the decreased soil moisture values can be attributed to other factors While the site did not 91

PAGE 105

contain beetle kill trees below diversion plots contained a lower percentage of green needles Trees with decreased percentages of green needles could have some decreased health and possibly some impairment of evapotranspiration Also, the nature of the diversions themselves could contribute to soil moisture differences. Recall that the probe used for soil moisture measurements was 7.6 cm (3 in) This length did not penetrate past the subsurface soil. Some diversion structure could potentially lead to artificial recharge of subsurface soils Specifically dirt or stone canals could recharge the soil if some water held in the diversion penetrates the soil lining the canal. This water could potentially transport to sites below diversions and thereby artificially recharge the upper level of soil. This could potentially contribute to increased subsurface soil moisture at below diversion sites Overall the soil moisture data suggest impaired stand health below diversion sites through decreased canopy and decreased evapotranspiration Examination of moisture content from deeper soil strata would be worth pursuing in order to determine deeper impacts to the soil from diversions 92

PAGE 106

5.3 Analysis of Remotely Sensed Data Landsat spatial resolution limited reliable analysis of NOVI change values for the field study sites Specifically Landsat data pixels contained mixed land features including vegetation and roads Additional sources of error included the age of data Aerial imagery dated to 2000 and did not reflect recent changes in land cover or use. Similarly, landcover data used to isolate vegetated coniferous areas were dated to 2006 did not reflect recent changes in landcover. These factors limited the selection of pixels corresponding to field sites to Rollins Pass Study Area 3 It was only this area that could be reliably and manually selected based on knowledge of the area A comparison of the changes in N OVI per pixel from 2003 to 2010 in Rollins Pass Study Area 3 and the Jones Pass basin showed a greater loss of vegetation in below diversion plots compared with above diversion plots This suggests that NOVI change value can be used to compare vegetation above and below diversion plots. However, the spatial resolution of Landsat data means that individual pixels generally contain mixed vegetation type E xcept where pixels correspond to field plots with data quantifying pine species, the data analysis cannot be reliably interpreted as 93

PAGE 107

corresponding to pine trees and pine beetle activity Instead the data analysis is limited to general vegetation health. Remotely sensed data with higher spatial resolution, as well as more recent aerial data, would allow for more accurate and wider analysis. Additional statistical analysis should also be explored Increased complexity of analysis can also be explored by exploring change in NOVI values in relation to proximity to diversions Furthermore the use of NOVI change values per pixel has advantages over comparing values between separate pixels by eliminating immediate concerns regarding differences in slope and aspect. However considerations of slope and aspect can be included for a more thorough analysis Additional data can also be included such as precipitation and soil type contributing to a more detailed analysis and understanding of the differences in vegetation health between habitats. 94

PAGE 108

6. Conclusion Improved knowledge of the extent and impacts from water diversion projects is vital in order to ensure proper policy law and infrastructure With climate change projections predicting severe water shortages for the south-west Rocky Mountain Region proper allocation of water resources is vital. This begins with accurate database management from water agencies to ensure that at a minimum the extent of diversions is fully known Indeed impacts from diversions cannot be assessed if the locations of the diversions are not known or documented. These data quality issues also raise the question if the volumes of extracted water are being accurately documented, a matter not explored in this study. To this end one may further question if there is adequate oversight of water agencies. The results from this study revealed extensive diversions in three areas. Field survey findings also suggest an ongoing extraction of nearly all water from basins of extraction Furthermore this study found that habitat below water diversions contained statistically significant highe r percentages of beetle kill trees compared to habitat above diversions, using multiple statistical tests and two different minimum DBH values. These find i ngs 95

PAGE 109

support the hypothesis that water diversions cause an increase in pine beetle infestation and subsequent tree mortality in habitat below diversions This is in spite of increased tree density above diversions a factor attributed to increased infestation and mortality. While stand density differences were not statistically significant this could suggest water diversions are more important than stand density in determining beetle infestation. The results of other assessments within this study indicate diversions may impair habitat health for pine trees and other tree species. This was indicated through increased stand density and smaller tree diameters in above-diversion plots suggesting non-diverted habitat may be able to support more trees and young growth. The negative impact of diversions on habitat health is also i ndicated through decreased green canopy and increased subsurface soil moisture in below diversion habitats pointing to decreased functional canopy and decreased evapotranspiration. Overall changes in functional vegetation are supported by findings of greater loss of healthy vegetation in below diversion areas as indicated by changes in NOVI over time This study determined water d i versions adversely impact habitats below diversions causing increased beetle infestation and tree mortality as 96

PAGE 110

well as impact to other tree species and overall decreased habitat health These findings point to the urgency for thorough and large scale environmental impact assessments of habitat in areas with water diversions Measures should be taken to fully assess and quantify diversion impacts The results could impact design and implementation of new water projects as well as management of existing water diversions 97

PAGE 111

APPENDIX A SPECIAL USE PERMIT 98

PAGE 112

Autnonzauon 10' SUL4S9 RECEIVED U 0 7 Conlact 10. SMOLINSKI S EXPlrarlOO Dale 12/3112010 Use Code 422 U S DEPARTMENT OF AGRICULTURE Foresl Service TEMPORARY SPECIAL USE PERMIT (FSH 2709.11. sec 54.6) AUTHORITY : Act 01 May 26. 20()(). PL 106-206 The holdP-r. authorized conducIlhe 100IOwiI>g aclrvlllflS CollectIOn 01 conll nGedl&s In support 01 u werslty l esearch on wal ,shed TI:I11015 AND F S 27()o'25 (03106) OMS NO 0596-0082 1 Use under thIS permn shall beg", on 00.01 /2010 and end on 08.'30 2010 The permll shall nol be ox1onded 2 The fOIl f or IS US8 Is 1'<81ved II sha" paid n advance and IS nOI r lund bI 3 Th o holclot hall cooduclthe au1honzed IICtNI1re!i accord "II 10 a lached approved ptans and &peCllteahonS n accor dance wnh appllCalJOll s i gned on 05117 2010 4 T he holder shall nollnstail any impf has lhe rBSpons hlyof Inspoc1rng tho use area and ad nlOQ areas for dangerous trees. hangIng bmbS. and other elndenCe 01 halardou conditIOns whICh would poso a lis!< 01 InjUry to Ind'Ytduals Alter secu,m9 pem1'SSIOrt from the au1honzed oHrcer. the holder shall remove such hazards 9 The hotoor shall be liable lor any damage sufrered by the Ufllted States msunlng from or related to use 01 this perm" tnlCludlllg damages to N atlOMl Forest rO$ources and costs 01 tIle SUppresSIOO 10 The holder shail hold harmless the Uruted Stales lrom any fiabllo1y from damage 10 hie or property arising from the hoIder's occupancy or trSe 01 NabOnlll Foresl Ian under this permrt I I The holder agrees 10 permit the free and unreslrlclcd access 10 and upon the premises at all limos for an lawful and proper purposes not lIlConslStenl wnh the mtenl of the perm" or Wlt the reasonable exetcrse and eojoyment by the holder of the prIVIlegeS thereol 12. Th,s permll IS subject 10 ail valtd roghlS and claIms outstand,ng In Ih l rd partres Figure A.1. Special Use Permit. Permit authorizing removal of needles 99

PAGE 113

" ThIS rt IS aloc0<\$6 for use oIlodofally crNMd lalld n not graIIl any Inlerest In ",al p.-ope y ThiS IS noluOfI5, ... b The holde< shall not en\8f Into age """""I, .. rd pe"'es lor occupancy of lhe aull\oflled premlSos ar>d ITIPf0""""'" 15 Appeal 01 any PfOVISIOf\S of lib perma or any r wemeru, thel80l shaM be solJtCCl 10 appeal al36CFR 251. S\.tJpart C or rw,sons I roof 16 ThIS J*'T1' Is accepled .JIlj 110 the cono.l>OtUI lonll neroln. 8IId E.llholkU&) 811_ 10 and made a pan 01 tIlla penn" 17 Theabovec:1aus.,sha con ofM yconlct .. rthaddloOnalcla ...... orpro.1OOS YO road and .'ldI>rsta'1d \he I m and ton(! loOrIS alld agree to abodu II r 1I>em U 5 DEPARTMEtH or AGRICUlTURE F 0It!$1 SaMce By 111111111111111111111111111111111111111111 n.us "' ..... .-.. -. -----.... (" "pnhbho Figure A.1. (Con t.) Special-Use Permit. Permit authorizing removal of needles 100

PAGE 114

BIBLIOGRAPHY Bleiker, K.P., & Six D.L. (2009). Competition and coexistence in a multipartner mutualism: Interactions between two fungal symbionts of the mountain pine beetle in beetle attacked trees. Fungal Ecology 57, 191-202 doi : 1 0 1 007/s00248 008 9395-6 Breece C.R. Kolb T.E Dickson B.G., McMillin J.D & Clancy K.M. (2008) Prescribed fire effects on bark beetle activity and tree mortality in southwestern ponderosa pine forests Forest Ecology and Management, 255 119-128 doi :10. 1016/j.foreco.2007.08.026 Brooks R.T. & Kyker-Snowman T.D. (2009) Forest-floor temperatures and soil moisture across riparian zones on first to third-order headwater streams in southern New England, USA. Forest Ecology and Management, 258 2117-2126. doi:1 0.1 016/j.foreco.2009.08.007 Cain R.J. & Howell B .E. (2005) Biological evaluation of spruce beetle and mountain pine beetle activity on the French Creek Analysis Area of Brush Creek/Hayden Ranger District Medicine Bow-Routt National Forests and Thunder Basin National Grassland Biological Evaluation R205-05 USDA Forest Service. Retrieved from http : //www.fs.usda.govllnternetlFSE DOCUMENTS/fsbdev3_03945 6 pdf Campbell James B. (2007). Introduction to remote sensing (4t h ed ) New York : The Guilford Press. Carriquiry, J.D & Sanchez, A. (1999) Sedimentation in the Colorado River delta and Upper Gulf of California after nearly a century of discharge loss Marine Geology, 158 125 145 doi : 10.1016/S0025 3227(98)00189 3 101

PAGE 115

Chimner, R.A., & Cooper, D.J (2003) Influence of water table levels on CO2 emissions in a Colorado subalpine fen : an in-situ microcosm study. Soil Biology and Biochemistry, 35, 345-351. doi : 1 0.1 016/S0038 0717(02)00284-5 Clow, D W., Rhoades, C., Briggs J., Caldwell, M & Lewis Jr., W.M (2011). Responses of soil and water chemistry to mountain pine beetle induced tree mortality in Grand County, Colorado, USA. Applied Geochemistry 26, 5174-5178. doi: 1 0.1 016/j.apgeochem 2011 03 096 Colorado Department of Transportation. (2010). Grand county [Data file]. Retrieved May 20 2010, from the Colorado Department of Transportation website: http : //apps coloradodot.info/dataaccess/GeoData/index cfm?fuseacti on=GeoDataMain&MenuType=GeoData Colorado s Decision Support System (n.d .). Division 5 Structures [Data file]. Retrieved June 10, 2010, from the website Colorado's Decision Support System website: http : //cdss state co us/ Colorado State University. (n. d .). Mountain Pine Beetle. Retrieved April 4 2009, from the website Colorado State University website : csfs. colostate. ed u/pages/mou nta i n-p i ne-beetle. htm I Coops N.C., Wulder, M.A., & White, J.C. (2006) Integrating remotely sensed and ancillary data sources to characterize a mountain pine beetle infestation Remote Sensing of Environment, 105,8397 doi : 1 0 1 016/j res.2006.06.007 102

PAGE 116

Coops, N C., Wulder M .A, & Iwanicka D (2009). Large area monitoring with a MODIS-based Disturbance Index (01) sensitive to annual and seasonal variations. Remote Sensing of Environment 113 1250 1261. doi: 1 0 1 016/j rse 2009.02.015 Coops, N C., Gillanders, S.N., Wulder, M.A, Gergel, S.E. Nelson T., & Goodwin, N .R. (2010) Assessing changes in forest fragmentation following infestation using time series Landsat imagery. Forest Ecology and Management 259, 2355 2365 doi: 10.1 016/j.foreco.201 0.03.008 Dennison, P E., Brunelle, AR., & Carter, V.A (2010). Assessing canopy mortality during a mountain pine beetle outbreak using GeoEye 1 high spatial resolution satellite data Remote Sensing of Environment 114,2431-2435 doi : 10.1016/j rse 2010.05 018 Digital Data Services. (n.d.) GeoEye 1 Satellite Imagery. Retrieved June 18,2011 from the U.S. Geological Survey website: http : //www usgsquads com/index php/digital map-data/aerial and satellite-imagery/satellite-imagery/geoeye 1-satellite Doremus, H Hanemann, M. (2007) The challenges of dynamic water management in the American West. UCLA Journal of Environmental Law and Policy, 26,55-75. Retrieved from http://www2 law.ucla edu Eklundh, L., Johansson T., & Solberg, S (2009). Mapping insect defoliation in Scots pine with MODIS time-series data Remote Sensing of Environment, 113, 1566 1573. doi : 1 0 1 016/j.rse.2009 03 008 103

PAGE 117

Esri. (2010, September 14). ArcGIS Resource Center: Desktop 10. Retrieved July 24, 2011, from the Esri website: http://help arcgis com/en/arcgisdesktop/10.0/help/index.html Esri. (2011), ArcGIS1 0 [Software]. Available from http : //www esri com/software/arcgis/arcgis10/index.html Gillette N E., Erbilgin N., Webster, J N., Pederson L., Mori, S.R., Stein J.D., et al. (2009). Aerially applied verbenone-releasing laminated flakes protect Pinus contorta stand from attack by Dendroctonus ponderosae in California and Idaho Forest Ecology and Management 257 1405 1412 doi:10 016/j.foreco.2008 12 017 Guarin A, & Taylor, AH. (2011). Drought triggered tree mortality in mixed conifer forests in Yosemite National Park California, USA Forest Ecology and Management, 218, 229-244. doi:1 0.1 016/j.foreco/2005 07 014 Hais, M., Jonasova M Langhammer, J., & Kucera T. (2009). Comparison of two types of forest disturbance using multitemporal Landsat TM/ETM+ imagery and field vegetation data. Remote Sensing of Environment 113 835 845. doi: 1 0.1 016/j.rse.2008 12 012 Hatala J .A, Crabtree, R.L., Halligan K.Q & Moorcraft, P.R. (2010) Landscape-scale patterns of forest pest and pathogen damage in the Greater Yellowstone Ecosystem Remote Sensing of Environment, 114, 375 384 doi:1 0 1 016/j rse 2009 09.008 104

PAGE 118

Hilker T., Wulder, M A., Coops N .C., Linke J McDermid G., Masek J.G. et al. (2009). A new data fusion model for high spatial and temporal-resolutions mapping of forest disturbance based on Landsat and MODIS. Remote Sensing of Environment 113, 1613 1627 doi: 1 0.1 016/j.rse 009 03 007 Hoffman, R W. (2006). White-tailed Ptarmigan (Lagopus leucura): A technical conservation assessment. U.S. Department of Agriculture Forest Service, Rocky Mountain Region. Retrieved March 21,2009, from the U S Forest Service website: www fs.fed us/r2/projects/scp/assessments/whitetailedptarmigan.pdf Intergovernmental Panel on Climate Change (2007). Climate change 2007: synthesis report. Retrieved March 14, 2009, from the Intergovernmental Panel on Climate Change website: www.ipcc.ch/pdf/assessment report/ar4/syr/ar4 syr pdf Jones M .E., Paine T.D., Fenn M .E., & Poth, M .A. (2004) Influence of ozone and nitrogen deposition on bark beetle activity under drought conditions. Forest Ecology and Management 200 67 76. doi: 10.1016/j.foreco.2004.06.003 Kershner B.S., Matthews, D., Nelson G., & Spellenberg R. (2009). Field guide to trees of North America New York, NY : Sterling Publishing. Kulakowski, D., & Jarvis D (2011). The influence of mountain pine beetle outbreaks and drought on severe wildfires in northwestern Colorado and southern Wyoming : A look at the past century Forest Ecology and Management, 262, 1686 1696 doi : 10 1016 j.foreco.2011 07.016 105

PAGE 119

Kurtz W A., Stinson G., Rampley G.J., Dymond C.C., & Neilson, E.T. (2008). Risk of natural disturbances makes future contribution of Canada's forests to the global carbon cycle highly uncertain. Proceedings of the National Academy of Sciences 105 (5) 15511555 doi : 10.1073/pnas 0708133105 Logan, J.A. & Powell, J.A. (2001) Ghost forests, global warming and the mountain pine beetle (Coleoptera: Scolytidae). American Entomologist 47, 160 173. Retrieved from http://digitalcommons usu edu McKinney, S.T. & Tomback D .F. (2007) The influence of white pine blister rust on seed dispersal in whitebark pine Canadian Journal of Forest Research 37 (6) 1044 1057 doi : 1 0.1139/X06 305 Meddens, A.J H., Hicke J A., & Vierling, L.A. (2011). Evaluating the potential of multispectral imagery to map multiple stages of tree mortality Remote Sensing of Environment 115 1632 1642 doi: 1 0 1 016/j rse.2011.02 018 Miller Jr., G.T. & Spoolman S.E. (2009). Living in the environment: Concepts connections, and solutions. Belmont, California: Brooks/Cole Cengage Learning Minitab. (2011). Minitab 16 [Software]. Available from http://www.minitab com/en-US/products/minitab/default.aspx Morehouse K., Johns T. Kaye J & Kaye M (2008) Carbon and nitrogen cycling immediately following bark beetle outbreaks in southwestern ponderosa pine forests. Forest Ecology and Management 255, 2698 2708 doi : 1 0 1 016/j.foreco 2008 01.050 106

PAGE 120

Negron J F McMillin, J.D. Anhold J.A. & Coulson D. (2009). Bark beetle-caused mortality in a drought-affected ponderosa pine landscape in Arizona, USA. Forest Ecology and Management 257 1353-1362 doi :10.1 016/j.foreco.2008.12.002 Penna, D Borga, M Norbiato, D., & Fontana G.D (2009) Hillslope scale soil moisture variability in a steep alpine terrain Journal of Hydrology, 364, 311327 doi : 10.1016/j.jhydroI.2008.11 009 Primack, Richard B. (2006) Essentials of conservation biology. Sunderland Massachusetts: Sinauer Associates Rey M (2008 May 5) Statement before the Senate Committee on Appropriations Concerning impacts of the bark beetle infestation on forest health fire, and recreation on Colorado s National Forests. Retrieved April 4, 2009, from the United States Senate Committee on Appropriations website: http://appropriations.senate.gov/ht interior.cfm Rodriguez C.A., Flessa, K W & Dettman, D.L. (2001). Effects of upstream diversion of Colorado River water on the estuarine bivalve mollusk Mulinia coloradoensis. Conservation Biology 15 249-258. doi: 1 0.1111/j 1523 1739 2001 99463.x Salle, A., Ye H., Yart A., & Lieutier, F. (2008). Seasonal water stress and the resistance of Pinus yunnanensis to a bark beetle associated fungus Tree Physiology 28, 679-687. doi: 1 0 1 093/treephys/28 5 689 107

PAGE 121

Sanchez-Martinez G., & Wagner M.R (2002). Bark beetle community structure under four ponderosa pine forest stand conditions in northern Arizona Forest Ecology and Management, 170 145-160 Retrieved from http://www sciencedirect.com Schone B.R., Flessa K W., Dettman, D.L., & Goodwin, D.H. (2003) Upstream dams and downstream clams: growth rates of bivalve mollusks unveil impact of river management on estuarine ecosystems (Colorado River Delta Mexico) Estuarine Coastal and Shelf Science, 58 715-726. doi : 10 1016/S0272-7714(03)00175-6 Spectrum Technologies (n.d.). FieldScout TDR 100 Soil Moisture Meter. Retrieved April 3, 2010 from the Spectrum Technologies website : http://www.specmeters.com/Soil MoisturelTDR 100 Soil Moisture Probe.html. Steventon, J.D. & Daust D .K. (2009) Management strategies for a large scale mountain pine beetle outbreak : Modeling impacts on American martens Forest Ecology and Management 257 1976-1985 doi : 1 0 1 016/j.foreco 2009 02.013 U.S Department of Agriculture Natural Resources Conservation Service. (2000). Ortho photo [Data file]. Retrieved May 20, 2010 from the U .S. Department of Agriculture website: http://datagateway.nrcs.usda.gov/ U S Department of Agriculture Natural Resources Conservation Service. (2006). National Landcover Dataset [Data file] Retrieved May 20, 2010, from the U S Department of Agriculture website : http://datagateway.nrcs.usda gov/ 108

PAGE 122

u.s. Forest Service. (n.d.). Mountain Pine Beetle Retrieved April 4, 2009 from the U.S. Forest Service website : www.fs fed us/r6/rogue/swofidsc/beetles/mountainpine html U.S. Forest Service (200B May 5). Clean Water-Insect Outbreaks and Watersheds Retrieved February 3 2010 from the U .S. Forest Service website: http://www.fs.fed us/rm/boise/awae home.shtml U.S. Forest Service (200B May 12) District spraying for beetles ; Campground Openings Delayed Retrieved April 4 2009 from the U.S Forest Service website : www fs.fed us/r2/arnf/news/200B/05/12-clrdbeetlespraying.shtml U S Geological Survey. (2011, January 11). Landsat Enhanced Thematic Mapper Plus (ETM+). Retrieved July 2 2011, from the U .S. Geological Survey website : http : //eros usgs gov/#/Find Data/Products and Data Available/ETM U.S. Geological Survey (2011 June 22) Landsat Processing Details. Retrieved July 2 2011 from the U .S. Geological Survey website : http://landsat.usgs gov/Landsat Processing Details php U.S. Geological Survey (n. d.) L T50340322003265PAC02 [Data file]. Retrieved August 30 2011 from U S Geological Survey website : http://g lovis. usgs gov U .S. Geological Survey (n. d ) L T50340322010268PAC01 [Data file] Retrieved August 30, 2011, from U.S Geological Survey website : http : //glovis.usgs.gov 109

PAGE 123

Vogelmann, J .E., Tolk B., & Zhu Z (2009) Monitoring forest changes in the southwestern United States using multitemporal Landsat Data Remote Sensing of the Environment 113 1739 1748. doi: 1 0.1 016/j.rse.2009 04.014 Walker J.P., Willgoose, G.R., & Kalma J D (2004) In situ measurement of soil moisture : a comparison of techniques Journal of Hydrology 293 85 99. doi:10.1 016/j jhydroI.2004.01 008 Waring K .M., & Six D.L. (2005) Distribution of bark beetle attacks after whitebark pine restoration treatments: A case study Western Journal of Applied Forestry 20(2) 110 116 Retrieved from http : //saf publisher.ingentaconnect.com White J.C Wulder, M.A. Brooks D., Reich, R & Wheate, RD. (2005) Detection of red attack stage mountain pine beetle infestation with high spatial resolution satellite imagery. Remote Sensing of Environment 96 340-351 doi : 1 0 1 016/j rse 2005.03.007 Wohl E. (2006). Human impacts to mountain streams. Geomorphology 79, 217 248 doi : 1 0.1 016/j.geomorph.2006 06 020 Wulder, M.A. White J .C., Bentz B., Alvarez M F., & Coops N .C. (2006) Estimating the probability of mountain pine beetle red-attack damage Remote Sensing of Environment 101, 150 166 doi : 1 0.1 016/j rse 2005 12.01 0 Wulder M.A., White J.C Coops N C., & Butson C R (2008) Multi temporal analysis of high spatial resolution imagery for disturbance monitor i ng. Remote Sensing of Environment, 112 2729-2740 doi : 1 0.1 016/j rse.2008 01.01 0 110

PAGE 124

Zausen G.L., & Kolb T.E (2005). Long-term impact of stand management on ponderosa pine physiology and bark beetle abundance in northern Arizona : A replicated landscape Forest Ecology and Management 218, 291-305 doi:1 0 1 016/j.foreco.2005.08.023 111