Citation
Colorado prairie grassland ecological restoration : an analysis of land management on the Rocky Mountain Arsenal National Wildlife Refuge

Material Information

Title:
Colorado prairie grassland ecological restoration : an analysis of land management on the Rocky Mountain Arsenal National Wildlife Refuge
Creator:
Mueller, Jennifer
Place of Publication:
Denver, Colo.
Publisher:
Metropolitan State University of Denver
Publication Date:
Language:
English

Record Information

Source Institution:
Metropolitan State University of Denver
Holding Location:
Auraria Library
Rights Management:
Copyright [name of copyright holder or Creator or Publisher as appropriate]. Permission granted to University of Colorado Denver to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.

Downloads

This item has the following downloads:


Full Text
Colorado Prairie Grassland Ecological Restoration: An Analysis of Land Management on the Rocky Mountain Arsenal National Wildlife Refuge by Jennifer Mueller
An undergraduate thesis submitted in partial completion of the Metropolitan State University of Denver Honors Program
May 2015
Ronald Baxendale II
Dr. Christy Carello
Dr. Megan Hughes-Zarzo
Primary Advisor
Second Reader
Honors Program Director




Colorado Prairie Grassland Ecological Restoration: an Analysis of Land Management on the Rocky Mountain Arsenal National Wildlife Refuge
Jennifer M. Mueller
Metropolitan State University of Denver


Disclaimer: The findings and conclusions in this article are those of the author(s) and do not necessarily represent the views of the U.S. Fish and Wildlife Service.
2


Table of Contents
Abstract________________________________________________________________________4
Introduction____________________________________________________________________5
Background______________________________________________________________________5
Materials and Methods___________________________________________________________7
Study Site_______________________________________________________________7
Data Collection Methods__________________________________________________10
Statistical Analysis_____________________________________________________12
Results_________________________________________________________________________13
Percent Cover: Native and Invasive Plant Species Analysis________________17
Richness, Diversity, and Relative Abundance______________________________20
Discussion______________________________________________________________________21
Conclusions and Future Suggestions_______________________________________23
Acknowledgements________________________________________________________________24
Works Cited_____________________________________________________________________26
List of Tables and Figures
Figure 1- Map of study location at RMANWR________________________________________8
Figure 2- View of Control Plot____________________________________________________9
Figure 3- View of Treated Plot____________________________________________________9
Figure 4- Study Design___________________________________________________________11
Figure 5- Sample Quadrat_________________________________________________________11
Figure 6- Total Percent Cover____________________________________________________13
Figure 7- Plant Species Composition______________________________________________13
Figure 8- Native Species Composition_____________________________________________14
Table 1- Native Species__________________________________________________________14
Figure 9- Native Species Percent Cover___________________________________________15
Figure 10- Invasive Species Composition__________________________________________15
Table 2- Invasive Species________________________________________________________16
Figure 11- Invasive Species Percent Cover________________________________________16
Figure 12- Slender russian thistle Percent Cover_________________________________17
Figure 13- Black bindweed Percent Cover__________________________________________17
Figure 14- Species Richness______________________________________________________19
Figure 15- Species Diversity_____________________________________________________19
Figure 16- Relative Abundance____________________________________________________20
3


Abstract
The Rocky Mountain Arsenal National Wildlife Refuge (RMANWR) is located in a grassland habitat that has been highly altered as a result of military activities. In the last few decades efforts have been made to restore portions of the RMANWR to a healthier grassland, facing the challenge of fighting invasive plant species. Parcels of habitat have been delineated based on the location of prairie dog (Cynomys ludovicianus) colonies. Controlled burning and herbicides have been used in an effort to control cheatgrass (Bromus tectorum) on some plots on RMANWR. I compared two colonies on the Arsenal: one that has received little to no management (control) and one that has received high levels of management to control the invasive cheatgrass. The vegetation present on each colony was analyzed in order to determine if the management applied to the treated colony has led succession in a more native direction. Eight to ten one-meter squared quadrats were randomly sampled for species richness and abundance at each location. Statistical analyses determined the presence or absence of a statistically significant difference in various categories between the two colonies. My findings suggest that the treated colony has not begun to progress to a more natural state but instead showed statistically more black bindweed (Polygonum convolvulus) and slender russian thistle (,Salsola collina), both an invasive species. This implies the management applied has not had the desired effect on the habitat, and that further management techniques will be required to restore the site to a more native state.
Keywords: Prairie grassland, restoration, invasive, cheatgrass, bindweed, thistle, succession
4


Introduction
Parcels of land that were historically natural prairie but modified in some way are of great concern now for restoration and land management practices. The United States has experienced the loss of prairie habitat due to changes in land use, fire suppression, and other practices (Olechnoivski et al. 2009; Rudolph et al. 2014). These changes to the habitat allow invasive species to move in and possibly overtake the environment, reducing the success of native species, and threatening biodiversity (Weiner 2004). In Colorado grassland habitat, invasive species provide a significant challenge for land managers (Stohlgren et al. 1999) who are attempting to encourage the surrounding environment to progress towards or maintain a native state.
Part of prairie restoration is controlling invasive species during replanting, when land managers are attempting to encourage native plants to grow. Up for debate, is exactly which pathway enables introduced species to take over, offering interspecific competition and assembly as two possible controlling pathways (Weiner 2004; Martin and Wilsey 2012); this further increases the difficulty of the complex problem of how best to control invasive species. In the end, attempting to manage the various prairie habitats present at a variety of stages will serve to boost and maintain diversity and abundance of wildlife in the area (Olechnoivski et al. 2009), on which the restoration of prairie ecosystems depends (Station et al. 2008).
Background
The Rocky Mountain Arsenal National Wildlife Refuge (RMANWR) is a nearly 17,000 acre superfund site located in Commerce City, Colorado once used for agricultural purposes, chemical weapon synthesis and testing by the U.S. Army, and pesticide synthesis by Shell Chemical Company following army use (Cohn 1999). The Army and Shell Chemical Company
5


have been assigned responsibility for covering the costs for the clean-up of the arsenal. The habitats within the arsenal include open lakes, woodlands, wetlands, and prairie grasslands (Rocky Mountain Arsenal National Wildlife Refuge: History). The Black-tailed Prairie Dog (Cynomys ludovicianus) is present on RMANWR on many parcels of land delineated by land managers outlined by the 2013 Prairie Dog Management Plan (U.S Fish and Wildlife Service 2013).
Due to the nature of land use RMANWR has historically been subjected to, the natural habitats that once existed are highly anthropogenically disturbed and modified, requiring land management to mitigate the effects. In 1996 remediation efforts were launched on the Rocky Mountain Arsenal and were considered to be finished in the fall of 2011 with continuing native plant restoration activities (U.S Fish and Wildlife Service 2013). Managers on the arsenal have been tasked with not only clean-up or control of chemical contaminants present, but also restoring the landscape to its original state. Plants and animals are carefully monitored and various forms of control are utilized to help the habitats present progress towards a more natural state.
On RMANWR, cheatgrass (Bromus tectorum) is of great concern as it had been close to producing a monoculture on a parcel of land that managers were working on restoring to native prairie habitat. The initial invasion of species is often linked to the use or management of the land that has disturbed native vegetation in some way (Weiner 2004; Chambers et al. 2014), and the habitats within RMANWR have been extremely disturbed in the past. Invasive grasses such as cheatgrass have become increasingly dominant and can have major effects on plant communities and ecosystem services; changing the physical and chemical properties of soil (Chambers et al. 2014; Jones et al. 2015). Land managers on RMANWR attempted to recover
6


the habitat that was heading towards a cheatgrass monoculture through use of herbicides and controlled burns in order to encourage native species to return.
The two plots chosen for the study were based on prairie dog colonies outlined in the 2013 Prairie Dog Management Plan (U.S Fish and Wildlife Service 2013). Land managers applied herbicides and used controlled burns in an attempt to remove the cheatgrass and reestablish native grasses on the treated plot. The control plot has been relatively untouched, having no herbicides or burns applied to the plants, allowing natural processes such as herbivory, competition, and weather to control plant species composition. The goal of this study was to compare the two plots on RMANWR to evaluate whether the control exercised on the treated plot led succession towards a more native state. By evaluating the direction succession has headed on the treated plot, land managers can gain an understanding of the efficacy of their restoration plan and determine whether or not further action will be required.
Materials and Methods
Study Site
The two plots I selected for the study occupy two distinct parcels of land that were both once native prairie grassland (Fig 1). I labeled one plot control to signify that little to no control has been utilized by land managers on the plot. Vegetation and wildlife on the control plot have remained relatively untouched by arsenal managers. The other plot was labeled treated to signify that some amount of control has been utilized by land managers on the site. It should be noted that both plots housed prairie dogs. Treatment on the treated plot is estimated to have happened at the end of the growing season in 2013 with controlled burning and herbicides applied, the vegetation was then allowed to grow for the duration of the 2014 growing season.
7


U.S. Fish & Wildlife Service
Rocky Mountain Arsenal
National Wildlife Refuge Environmental Assessment
EZ2 Prairie Zee (2515 Aoes|
Historic Bunnieg Owl Nest leutiees'
BBB Cmptstte Pi m le Deg Celenies (1334 to 200J| KrkitatZineBwndjiy Arny Juic4cMn l~lBMANWBBui4an
lend Cunently en EPA
Nitiinil Priorities List (NPl|
White tores stow (visiter |
Bus *4 Auto Tor Rotes.
-Tht lletcre Rjrurnu fieri K'allai^lTCT iccodtd in yean IS Kft.20002002
I min 1
Figure 2
Proposed Zones for Black-tailed Prairie Dog Occupation at RMANWR
Fig. 1- Map of the Rocky Mountain Arsenal National Wildlife Refuge in Commerce City, CO
8


Fig. 2- Overall landscape view of the control plot on RMANWR (Emily Schenderlein)
Fig. 3- Overall landscape view of the treated plot (Emily Schenderlein)
9


Data Collection Methods
Data were collected August-October 2014. I measured out transect lines running North to South, 200m from each other (placing as many as could fit on each plot), and I designated five meter squared macroplots every 200m along each transect line, which I then broke down into one meter squared quadrats and labeled one through twenty five. Using a random number generator I selected two one meter squared quadrats in which to conduct a vegetation analysis. Quadrats made of PVC pipe were used to define the boundaries of the selected quadrats and I recorded the coordinate of the NW comer of each quadrat. Fig. 4 is a pictorial representation of the study layout. The yellow area is the plot, dashed lines signal where I placed transects lines on the plot. The small black boxes on the plot are the macroplots, with an enlarged view of how I broke down the macroplot for random sampling. I took a picture of each quadrat and plant species for further analysis and identification. A sample quadrat is pictured in Fig 5. To complete vegetation analysis I estimated percent cover of each plant species present in the quadrat and recorded it in a spreadsheet. Due to time and vegetation constraints, I only completed two to three transects on each plot with eight quadrats on the control plot and ten quadrats on the treated plot being analyzed.
10


I I I I I
I I I I I I I I I
A B
Fig. 4- Depiction of study layout.
Fig. 5- Sample quadrat from the control plot
For each quadrat I placed two additional one meter PVC pipes over the square to assist with visualizing what a quarter of the quadrat would look like (Fig. 5). I either identified all species of plants found or assigned a code to the plant for later identification. In order to define
200m
11


percent cover I visualized every part of the species originating within the quadrat set out in such a way that I took plant height into consideration. The taller the individual of the species, the higher the percent cover. I also estimated bare ground and dead matter for percent cover, but did not include in statistical analysis. The plants I observed were identified using a combination of RMANWRs species list and consulting a botanist familiar with the local plant life in the area. Once identified, I classified each species as native or invasive using the USD A plant database (Plants Database). Due to the time of year data were collected, I was unable to identify a total of four plants due to lack of floral structures.
Statistical Analysis
After all data were collected I used Microsoft Excel to calculate richness, diversity (using Simpsons Index), and percent cover. Then I used Excel to run a two sample t-test test assuming equal variance for each value, as well as the percent cover of native v.s non-native species. I used one tailed p-values to evaluate statistical significance in the difference between mean values on each plot, setting statistical significance to p<0.05.
12


Results
Fig. 6- Mean total percent cover and SE for control and treated plot
Control Plot Plant Treated Plot Plant
Species Composition Species Composition
Native Native 24%
39%
Invasive
61% * Invasive y
76%
Fig. 7- Plant species composition on each plot
13


Control Plot Native Plant
Treated Plot Native
Species Composition
Plant Species
Blue
Sage Composition
Fig 8- Native species representation on the control and treated plots
Table 1- Native species observed
Plant species Common name {Scientific name) Control Treated Percent cover of native species
Sunflower spp. X 7
Ribseed Sandmat {Chamaesyce glyptosperma) X 61
Little Milkweed {Asclepias pnmila) X 18
Cowpen Daisy {Verbesina encelioides) X 30
Blue Sage {Salvia Aznrea) X 9
PI U3 X 14
P4U2 X 22
P4 U6 X 34
P4 U7 X 5
14


Native Plant Species % Cover on Both
Plots
40
35
30 a>
g 25 u
S? 20 § 15 2 10 5 0
Control Treated
Fig. 9- Mean native percent cover and SE for both plots
Control Plot Invasive Plant Species Black Composition
Treated Plot Invasive Plant Species Composition
Spotted Burflin§
Fig. 10- Invasive species composition on the control and treated plots
15


Table 2- Invasive species observed
Plant Species Common name {Scientific name) Control Treated Percent cover of invasive species (Control) Percent Cover of Invasive Species (Treated)
cheatgrass (Bromus tectornm) X 19
Slender russian thistle {Salsola collina) X X 29 66
Common Mullein {Verbascnm thapsus) X 2
Five-hook Bassia {Bassia hyssopifolia) X 37
Black bindweed {Polygonum convolvulus) X X 13 27
Spotted Knapweed {Centaurea maculosa) X 2
Burning Bush {Kochia scoparia) X 5
Invasive Plant Species % Cover on Both Plots
Fig. 11- Mean invasive percent cover and SE for both plots
16


Fig. 12- Mean percent cover and SE of Slender russian thistle for both plots
% Bindweed/Total % cover
60 -i---------------------------
Control Treated
Fig. 13- Mean percent Bindweed cover and SE for both plots
Percent cover: Native and Invasive Plant Species Analysis
In order to get a surface view of the total cover of plant species on each plot, the total percent vegetation cover for each plot (excluding classifiers such as bare ground and dead matter) was calculated. Mean total percent cover for the control plot (x=106.50 + 18.55) was
17


determined to be statistically greater than that of the treated plot (x=47.60 + 31.08, t=4.71, p=. 00012) (Fig. 6). Fig. 7 provides a break down for each plot of the total percent cover of native and invasive plant species.
Of all the plant species present between both colonies, one species on the control plot and three species on the treated plot could not be identified (refer to Table 1). Statistical analysis of the percent cover the unidentified species contributed to the total percent cover of each plot was determined to be relatively low (x Control=0.11 + 0.11, x treated=0.16 + 0.13) with no statistically significant difference between the two colonies (t=-0.55, p=0.30). Thus, unknowns were accepted and classified as native species. On each plot, the percent cover each native plant species contributed to the total native percent cover was evaluated (Fig. 8 and Table 1). The control plot had a statistically significant higher mean percent cover of native species (x =29.58 + 21.08) than that of the treated plot (x= 10.33+ 14.96, t-2.67, p=0.019) (Fig. 9).
For the invasive species, the percent cover each invasive plant species contributed to the total invasive percent cover on each plot was also evaluated (Fig. 10 and Table 2). The control plot had a lower percent cover of invasive plant species (x= 71.59 + 22.20) than that of the treated plot (x= 89.00 + 14.92) and the null hypothesis of no significant difference was rejected (t=l .99, p=0.03) (Fig. 11).
Of the seven total invasive species, only two were present on both the control and the treated plot, slender russian thistle (Salsola collina) and black bindweed {Polygonum convolvulus). The control plot had a statistically significant lower mean percent cover of slender russian thistle (x=33.44 + 26.74) than that of the treated plot (x=64.94 + 15.45, t=-2.43, p=0.03) (Fig. 12). Similarly to the russian thistle, the control plot had a statistically lower mean percent
18


cover of black bindweed (x= 12.63 + 12.76) than the treated plot (x = 40.34 + 27.54, t=-2.36, p=0.03) (Fig. 13).
Plant Species Richness on both Plots
Control Treated
Fig. 14- Mean plant species richness and SE calculated for control and treated plot at RMANWR
Diversity of Plant Species on both Plots
0.25 0.2
>
\ft
£ 0.15
0.05
0
Fig. 15- Mean diversity (determined by Simpsons Index) and SE calculated for control and treated plot
19


Relative Abundance of Plant Species on Both Plots
Fig. 16- The mean relative abundance and SE of plant species on both plots calculated using total percent cover and richness
Richness, Diversity, and Relative Abundance
The mean richness of the control plot (x = 4.75 + 1.58) was calculated to be statistically greater than that of the treated plot (x = 3.30 + 0.67, t=2.63, p=0.0090) (Fig. 14). The mean diversity of the control plot (x =0.13 + 0.093) was calculated to be higher than that of the treated plot (x = 0.19 + 0.011) but the null hypothesis that they did not statistically differ failed to be rejected (t=-1.31, p=0.21) (Fig. 15). The relative abundance of plant species on both colonies using the species richness and total percent cover was calculated. The control plot had a statistically significant higher mean relative abundance (x=23.91 + 6.29) than the treated plot (x=13.67 27.66, t=3.04, p=0.003) (Fig 16).
20


Discussion
The goal of the study was to compare the control plot that had been left untouched by land managers, with the treated plot that had been treated with herbicides and controlled burns, to determine if the treatment applied by land managers was effective in encouraging succession towards a more native state. Overall, the control plot experienced a higher total percent cover of plant species (Fig. 6). These findings suggest that the control plot experienced less bare ground than the treated plot, but a consideration to be made when interpreting these results, is that the treated plot had only been allowed to grow for one year post treatment, so plants had less height and bulk to them. Nonetheless, a successful restoration as defined by the U.S. Army and Shell Oil Company must fit certain criteria that includes a minimum 70% total ground cover (U.S Fish and Wildlife Service 2013). The treated colony is well under this benchmark, sitting at 47.6% total percent cover.
Another facet of determining the level of restoration a plot has achieved, is to consider the native plant species present and their abundance. Between the two plots, the treated plot experienced higher native species richness, but lower native species composition than the control plot (Fig. 7). This was surprising, as if the treatment were to be considered effective; I would expect to see a higher abundance and cover of native plant species on the treated plot than was observed (or at least values similar to that on the control plot), yet this was not the case (Fig. 9). Also, of the native species observed overall, there was not a single native species that was present on both plots (Fig. 8 and Table 1). Possibly, this could be due to the extreme ends of the arsenal each plot resided in, allowing for exposure to different native seeds. The data suggest that the control plot is an environment more conducive to native plant species than the treated colony despite the treatments applied to discourage invasive species and encourage natives. Previous
21


research suggests that if native plant species are not established early in succession before the invasive plant species have a chance to take over, restoration will not be successful and there may be a need to restart the process (Martin & Wilsey 2012, Cohn 2009). This does not bode well for the treated plot, as my data lead to the conclusion the native plant species have not become established early in succession.
Quite possibly, of more note than the native species composition and percent cover, was that of the invasive species observed. As the control applied to the treated plot was applied to control invasive plant species (mainly to reduce the chance of a cheatgrass monoculture), it is of great interest to land managers to understand where the treated plot is headed in terms of other invasive species. The restoration cannot be considered successful if other invasive species just replace the cheatgrass and reduce the success of native species in the area (Molano-flores 2014). By evaluating the invasive species as I evaluated the native species, a pattern of invasive species dominating the native plot emerged (Fig. 10 and Table 2). On a surface level, the treated plot had a higher cover of invasive species (Fig. 11), but as I was greatly interested in determining if there was a significant difference between the two plots, a comparison of invasive plant species present on both plots provided great insight.
Slender russian thistle and black bindweed were observed and identified on both plots, providing the opportunity for this comparison. In this comparison, the truth of the efficacy of the control applied to the treated plot was revealed, as both species were significantly higher on the treated plot (Fig. 12 and Fig. 13). Both invasive species are common where there is disturbance or the area is very dry and exposed (Park et al. 2012), making RMANWR habitats the perfect place for an invasion. Added to that, bindweed has such an array of survival techniques to
22


survive many forms of control, that it has become one of the worlds ten worst weeds and requires very aggressive and careful control (Mitich 1991).
Conclusions and Future Suggestions
After observing each plot and evaluating the data, I have concluded that the treated plot has not managed to succeed towards a more native state, and is indeed progressing on a path with invasive plant species that will make achieving a native state close to impossible without major intervention. My study was limited by the amount of data I was able to collect and the length of time each plot was observed. In order to fully understand the path the treated plot is progressing towards, a long term study will be required, but my data suggest action is required now to correct the succession the treated plot demonstrates. The Arsenal recognizes that current management practices are not providing the services needed to restore the habitats successfully (U.S Fish and Wildlife Service 2013), thus I suggest a few steps to assist the treated plot.
First and foremost, I believe that for any restoration of the natural prairie to happen on the treated plot, it is in the best interest of land managers to consider removing or relocating the prairie dog colony present on the treated plot. The less competition native plants have while trying to re-establish the better chance they have of succeeding. The absence of the clipping and burrowing encouraged by the prairie dogs for a few years would provide some relief for the native plant species who are already in high competition with the slender russian thistle and black bindweed present. Although an integral part to the prairie ecosystem, herbivores such as the Black-tailed Prairie Dog can reduce diversity in arid grassland environments, and management of herbivores in these arid environments is crucial to restoration efforts (Olff and
23


Ritchie 1998). Although this will assist restoration efforts, reducing herbivory and the burrowing effects of prairie dogs alone will not be enough (Park et al. 2012).
Previous restoration sites have seen success in a variety of methods that were not limited to herbivore control. Limited reintroduction of herbaceous species native to the prairie grasslands using a mix of sod and seed could be successful in driving succession to resemble what the historic vegetation may have been on the Arsenal (Rudolph et al. 2014). Historically, prairie grasslands experienced fires periodically, which are suppressed when the land is taken over for anthropogenic uses. Potentially, fire could continue to be used as a control measure for certain species to simulate the historic disturbances and help control such species as Cheatgrass (Augustine et al. 2009; Chambers et al. 2014), but would need to be coupled with other practices to reach maximum effect. As suggested by Martin and Wilsey (2012), it may be of value to evaluate the historic order of assembly of species that could generate greater diversity by producing alternate states of native habitat that would have once existed (Martin and Wilsey 2012). Land managers could use historical knowledge to amplify the outcome of treating for invasive species added with seeding for native species and restoring natural fire regimes. If practices and knowledge are expanded and combined, I believe the treated plot will reach the native state the land managers strive for and sustainably provide the ecosystem services historically provided by the habitat.
Acknowledgments
I would like to thank the Rocky Mountain Arsenal National Wildlife Refuge for allowing me to conduct my research on their property, Emily Schednerlein for her assistance with data
24


collection and photography, Dr. Carello for her constant advising and support, Dr. Baxendale for his advice and support, Dr. Meloche for his assistance in identifying the plants, and Metropolitan State University of Denver.
25


Works Cited
Augustine DJ, Milchunas DG, Ecology SR, Jan N, Milchunas G. 2009. Society for Range Management Vegetation Responses to Prescribed Burning of Grazed Shortgrass Steppe Linked references are available on JSTOR for this article : Vegetation Responses to Prescribed Burning of Grazed Shortgrass Steppe. Soc. Range Manag. 62:89-97.
Chambers JC, Bradley B a., Brown CS, DAntonio C, Germino MJ, Grace JB, Hardegree SP, Miller RF, Pyke D a. 2014. Resilience to Stress and Disturbance, and Resistance to Bromus tectorum L. Invasion in Cold Desert Shrublands of Western North America. Ecosystems 17:360-375.
Cohn JP. 1999. A Makeover for Rocky Mountain Arsenal. Bioscience 49:273-277.
Jones R, Chambers JC, Johnson DW, Blank RR, Board DI. 2015. Effect of repeated burning on plant and soil carbon and nitrogen in cheatgrass (Bromus tectorum) dominated ecosystems. Plant Soil 386:47-64. Martin LM, Wilsey BJ. 2012. Assembly history alters alpha and beta diversity exotic native proportions and functioning of restored prairie plant communities. J. Appl. Ecol. 49:1436-1445.
MitichLW. 1991. Intriguing World of Weeds-. Weed Technol. 5:913-915.
Molano-flores B. 2014. An Invasive Plant Species Decreases Native Plant Reproductive Success. Nat. Areas J. 34:465-469.
Olechnoivski BEM, Debinski DM, Drobney P, Viste-sparkman K, Reed WT. 2009. Changes in Vegetation Structure through Time in a Restored Tallgrass Prairie Ecosystem and Implications for Avian Diversity and Community Composition. Ecol. Restor. 27:449-458.
Olff H, Ritchie ME. 1998. Effects of herbivores on grassland plant diversity. TREE 13:261-265.
Park JN, Antill TM, Naeth MA, Bork EW, Westhaver AL. 2012. Russian Thistle ( Salsola tragus L .) Control on Bighorn Sheep Winter Ranges in Jasper National Park. Nat. Areas J. 32:391-397.
Rudolph DC, Plair DE, Jones D, Williamson JH. 2014. Restoration and Winter Avian Else of Isolated Prairies in Eastern Texas. Southeast. Nat. 13:52-64.
Station BH, Park N, Springs H, America N. 2008. Does Biodiversity Ecosystem Function Science Apply to Prairie Restoration ? Ecol. Restor. 26:100-101.
Stohlgren TJ, Binkley D, Chong GW, Kalkhan MA, Lisa D, Bull KA, Otsuki Y, Newman G, Bashkin M, Son Y. 1999. Exotic Plant Species Invade FJot Spots of Native Plant Diversity. Ecol. Monogr. 69:25-46.
26


United States Department of Agriculture Natural Resources Conservation Services Plants Database. [2015 April 19] Available from: http://plants.usda.gov/java/
United States Fish and Wildlife Service Rocky Mountain Arsenal National Wildlife Refuge: History. Commerce City (Colorado): Available from: http://www.fws.gov/refuges/profiles/History.cfnUniU61170
U.S Fish and Wildlife Service. 2013. U.S. Fish & Wildlife Service Management of black-tailed prairie dog (Cynomys ludoviciamus) populations on the Rocky Mountain Arsenal National Wildlife Refuge. Commerce City, Colorado.
Weiner J. 2004. Are Invasive Plant Species Better Competitors Than Native Plant Species? Evidence from Pair-Wise Experiments. OIKOS 105:229-238.
27


Full Text

PAGE 1

Colorado Prairie Grassland Ecological Restoration: An Analysis of Land Management on the Rocky Mountain Arsenal National Wildlife Refuge by Jennifer Mueller An undergraduate thesis submitted in partial completion of the M etropolitan State University of D enver Honors Program May 2015 Ronald Baxendale II Dr. Christy Carello Dr. Megan Hughes Zarzo Primary Advisor Second Reader Honors Program Director

PAGE 3

Colorado Prairie Grassl and Ecological Restoration: an Analysis of Land M anagement on the Rocky Mountain Arsenal National Wildlife Refuge Jennifer M. Mueller Metropolitan State University of Denver

PAGE 4

2 Disclaimer: The findings and conclusions in this article are those of the author(s) and do not necessarily represent the views of the U.S. Fish and Wildlife Service.

PAGE 5

3 Table of Contents Abstract 4 Introduction 5 Background 5 Materials and Methods 7 Study Site 7 Data Collection Methods 10 Statistical Analysis 12 Results 13 Percent Cover: Native and Invasive Plant Species Analysis 17 Richness, Dive rsity, and Relative Abundance 20 Discussion 21 Conclusions and Future Suggestions 23 Acknowledgements 24 Works Cited 26 List of Tables and Figures Figure 1 Map of study location at RMANWR 8 Figure 2 View of Control Plot 9 Figure 3 View of Treated Plot 9 Figure 4 Study Design 11 Figure 5 Sample Quadrat 11 Figure 6 Total Percent Cover 13 Figure 7 Plant Species Composition 13 Figure 8 Native Species Composition 14 Table 1 Native Species 14 Figure 9 Native Species Percent Cover 15 Figure 10 Invasive Species Composition 15 Table 2 Invasive Species 16 Figure 11 Invasive Species Percent Cover 16 Figure 12 Slender russian t histle Percent Cover 17 Figure 13 Black bindweed Percent Cover 17 Figure 14 Species Richness 19 Figure 15 Species Diversity 19 Figure 16 Relative Abundance 20

PAGE 6

4 Abstract The Rocky Mountain Arsenal National Wildlife Refuge (RMANWR) is located in a grassland habitat that has been highly altered as a result of military activities. In the last few decades efforts have been made to restore po r tions of the RMANWR to a healthier grassland, facing the challenge of fighting invasive plant species. Parcels of habitat have been delineated based on the location of prairie dog ( Cynomys ludovicianus ) colonies. Controlled burning and herbicides have been used in an effort to control cheatgrass ( Bromus tectorum ) on some plo ts on RMANWR. I compared two colonies on the Arsenal: one that has received little to no management (control) and one that has received high levels of mana gement to control the invasive cheatgrass The vegetation present on each colony was analyzed in ord er to determine if Eight to ten one meter squared quadrats were randomly sampled for species richness and abundance at each location. Statistical analyses determi ned the presence or absence of a statistically significant difference in various categories between the two colonies. My findings showed statistically more blac k bindweed ( Polygonum convolvulus ) and slender russian thistle ( Salsola collina ), both an invasive species. This implies the management applied has not had the desired effect on the habitat, and that further management techniques will be required to restor e Keywords: Prairie grassland, restoration, invasive, cheatgrass bindweed, thistle, succession

PAGE 7

5 Introduction Parcels of land that were historically natural prairie but modified in some way are of great concern now for restoration and land management practices. The United States has experienced the loss of prairie habitat due to changes in land use, fire suppressi on, and other practices (Olechnoivski et al. 2 009; Rudolph et al. 2014) These changes to the habitat allow invasive species to move in and possibly overtake the environment, reducing the success of native species, and threatening biodiversity (Weiner 2004) In Colorado grassland habitat, invasive species provide a significant challenge for land managers (Stohlgren et al. 1999) who are attempting to encourage the surrounding environment to progress towards or maintain a native state. Part of prairie restoration is controlling invasive species during replanting, when land m anagers are attempting to encourage native plants to grow. Up for debate, is exactly which pathway enables introduced species to take over, offering interspecific competition and assembly as tw o possible controlling pathways (Weiner 2004; Martin and Wilsey 2012) ; this further increases the difficulty of the complex problem of how best to control invasive species. In the end, attempting to manage the various prairie habitats present at a variety of stages will serve to boost and maintain diversity and abundance of wildlife in the area (Olechnoivski et al. 2009) on which t he restoration o f prairie ecosystems depends (Station et al. 2008) Background The Rocky Mountain Arsena l National Wildlife Refuge (RMANWR) is a nearly 17,000 acre superfund site located in Commerce City, Colorado once used for agricultural purposes, chemical weapon synthesis and testing by the U.S. A rmy and pesticide synthesis by Shell Chemical Company fol lowing army use (Cohn 1999) The Army and Shell Chemical Company

PAGE 8

6 have been assigned responsibility for covering the costs for the clean up of the arsenal. The habitats within the arsenal include open lakes, woodlands, wetlands, and prairie grasslands ( Rocky Mountain Ar senal National Wildlife Refuge: History ). The Black tailed Prairie D og ( Cynomys ludovicianus ) is present on RMANWR on many parcels of land delineated by land managers outlin ed by the 2013 Prairie Dog Management Plan (U.S Fish and Wildlife Service 2013) Due to the nature of land use RMANWR has historically been subjected to, the natural habitats that once existed are highly anthropogenically disturbed and modified, requiring land management to mitigate the effects. In 1996 remediation efforts were launched on the Rocky Mountain Arsenal and e plant restoration activities (U.S Fish and Wildlife Service 2013) Managers on the arsenal have been tasked with not only clean up or control of chemical contaminants present, but also restoring the landscape to its original state Plants and animals are carefully monitored and various forms of control are utilized to hel p the habitats present progress towards a more On RMANWR, cheatgrass ( Bromus tectorum ) is of great concern as it had been close to producing a monoculture on a parcel of land that managers were working on restoring to native prairie habit at. The initial invasion of species is often linked to the use or management of the land that has disturbed native vegetation in some way (Weiner 2004; Chambers et al. 2014) and the habitats within RMANWR have been extremely disturbed in the past. Invasive grass es such as cheatgrass have become increasingly dominant and can have major effects on plant communities and ecosystem services; changing the physical and chemical properties of soil (Chambers et al. 2014; Jones et al. 2015) Land managers on RMANWR attempted to recover

PAGE 9

7 the habitat that was heading towards a cheatgrass monoculture through use of herbicides and controlled burns in order to encourage native species to return. The two plots chosen for the study were based on prairie dog colonies outlined in the 2013 Prairie Dog Management Plan (U.S Fish and Wildlife Service 2013) Land managers applied herbicides and used controlled burns in an attempt to remove the cheatgrass and re establish native grasses on the treated plot. The control plot has been relatively untouched, having no herbicides or burns applied to the plants, al lowing natural processes such as herbivory, competition, and weather to control plant species compos i tion. The goal of this study was to compare the two plots on RMANWR to evaluate whether the control exercised on the treated plot led succession towards a more native state. By evaluating the direction succession has headed on the treated plot, land managers can gain an understanding of the efficacy of their restoration plan and determine whether or not further action will be required. Materials and Metho ds Study Site The two plots I selected for the study occupy two disti nct parcels of land that were bo th once native prairie grassland (Fig 1) I labeled one plot control has been utilized by land managers on the plot Vegetation and wildlife on the control plot have remained relatively untouched by arsenal managers. The other plot was labeled It should be noted that both pl ots housed prairie dogs Treatment on the treated plot is estimated to have happened at the end of the growing season in 2013 with controlled burning and herbicides applied the vegetation was then allowed to grow for the duration of the 2014 growing seaso n.

PAGE 10

8 Fig. 1 Map of the Rocky Mountain Arsenal National Wildlife Refuge in Commerce City, CO Control Treated

PAGE 11

9 Fig. 2 Overall landscape view of the control plot on RMANWR (Emily Schenderlein) Fig. 3 Overall landscape view of the treated plot (Emily Schenderlein)

PAGE 12

10 Data Collection Methods Data were collected August October 2014. I measured out t ransect lines running North to South 200m from each other (placing as many as could fit on each plot ), and I designated five meter squared macro plots e ver y 200m along each transect line, which I then broke down into one meter squared quadrats and labeled one through twenty five. Using a random number generator I select ed two one meter squared quadrats in which to c onduct a vegetation analysis. Quadrats made of PVC pip e were used to define the boundaries of the selected quadrats and I recorded the coordinate of the NW corner of each quadrat. Fig. 4 is a pictorial representation of the study layout. The yellow area is the p lot, dashed lines signal where I placed transects lines on the plot. The small black boxes on the plot are the macroplots, with an enlarged view of how I broke down the macroplot for random sampling. I took a picture of each quadrat and plant species for f urther analysis and identification A sample quadrat is pictured in Fig 5. To complete v egetation analysis I est imated percent cover of each plant species present in the quadrat and recorded it in a spreadsheet Due to time and vegetation constraints I on ly completed two to three transects on each plot with eight quadrats on the control plot and ten quadrats on the treated plot being analyzed.

PAGE 13

11 Fig. 4 Depiction of study layout. Fig. 5 Sample q uadrat from the control plot For each quadrat I placed two additional one meter PVC pipes over the square to assist with visualizing what a quarter of the quadrat would look like (Fig. 5 ) I either identified a ll species of plant s found or assigned a code to the plant for later identification. In order to define 200m 200 m A B Plot

PAGE 14

12 p ercent cover I visualized every part of the species originating within the quadrat set out in such a way that I took plant height into consideration. The taller the individual of the species, the h igher the percent cover I also estimated b are ground and dead matter for percent cover, but did not include in statistical analysis. The plants I observed were identified using a combination of the local plan t life in the area. Once i dentified I classified each species as native or invasive using the USDA plant database ( Plants Database ). Due to the time of year data were collected, I was unable to identify a total of four plants due to lack of floral structures. Statistical Analysis After all data were collected I used Microsoft Excel to calculate richness, diversity (using and percent cover. Then I used Excel to run a two sample t test test assuming equal variance for each value, as well as the percent cover of native v.s non native species. I used o ne tailed p values to evaluate statistical significance in the difference between mean values on each plot setting statistical significance to p<0.05.

PAGE 15

13 Results Fig. 6 Mean total percent cover and SE for control and treated plot Fig. 7 Plant species composition on each plot 0 20 40 60 80 100 120 Control Treated Mean % cover Total % Cover on Both Plots Mean Invasive 61% Native 39% Control Plot Plant Species Composition Invasive 76% Native 24% Treated Plot Plant Species Composition

PAGE 16

14 Fig 8 Native species representation on the control and treated plots Table 1 Native species observed Plant species Common name ( Scientific name ) Control Treated Percent cover of native species Sunflower s p p. X 7 Ribseed Sandmat ( Chamaesyce glyptosperma) X 61 Little Milkweed ( Asclepias pumila ) X 18 Cow pen Daisy ( Verbesina encelioides ) X 30 Blue Sage ( Salvia Azurea ) X 9 P1 U3 X 14 P4 U2 X 22 P4 U6 X 34 P4 U7 X 5 Sunflower Sp. 7% P1 U3 14% Ribseed Sandmat 61% Little Milkweed 18% Control Plot Native Plant Species Composition Cowpen Daisy 30% P4 U2 22% P4 U6 34% P4 U7 5% Blue Sage 9% Treated Plot Native Plant Species Composition

PAGE 17

15 Fig. 9 Mean native percent cover and SE for both plots Fig. 10 Invasive species composition on the control and treated plots 0 5 10 15 20 25 30 35 40 Control Treated Mean % Cover Native Plant Species % Cover on Both Plots Mean cheatgras s 19% Russian Thistle 29% Common Mullein 2% Five hook Bassia 37% Black Bindweed 13% Control Plot Invasive Plant Species Composition Russian Thistle 66% Black Bindwee d 27% Spotted Knapwe ed 2% Burning Bush 5% Treated Plot Invasive Plant Species Composition

PAGE 18

16 Table 2 Invasive species observed Plant Species Common name ( Scientific name ) Control Treated Percent cover of invasive species (Control) Percent Cover of Invasive Species (Treated) c heatgrass ( Bromus tectorum ) X 19 Slender russian thistle ( Salsola collina ) X X 29 66 Common Mullein ( Verbascum thapsus ) X 2 Five hook Bassia ( Bassia hyssopifolia ) X 37 Black bindweed ( Polygonum convolvulus ) X X 13 27 Spotted Knapweed ( Centaurea maculosa ) X 2 Burning Bush ( Kochia scoparia ) X 5 Fig. 11 Mean invasive percent cover and SE for both plots 0 20 40 60 80 100 Control Treated Mean % Cover Invasive Plant Species % Cover on Both Plots Mean

PAGE 19

17 Fig. 12 Mean percent cover and SE of Slender russian thistle for both plots Fig. 13 Mean percent Bindweed cover and SE for both plots Percent cover: Native and Invasive Plant Species Analysis I n order to get a surface view of the total cover of pl ant species on each plot, the total percent vegetation cover for each plot (excluding classifiers such a s bare ground and dead matter) was calculated. Mean total percent cover for the control plot (x =106.5 0 + 18.55 ) was 0 10 20 30 40 50 60 70 80 Control Treated Mean %Russian Thistle/Total % cover % Russian Thistle/Total % Cover Mean 0 10 20 30 40 50 60 Control Treated Mean Bindweed/total % cover % Bindweed/Total % cover Mean

PAGE 20

18 determined to be statistically gr eater than that of the treated plot (x =47.6 0 + 31.0 8 t=4.71 p=.0001 2 ) (Fig. 6). Fig. 7 provides a break down for each plot of the total percent cover of native and invasive plant species. Of all the plant species present between both colonies, o ne species on the control plot and three species on the treated plot could not be identified (refer to Table 1). Statistical an alysis of the percent cover the unidentified species contributed to the total percent cover of each plot was determined to be rel atively low (x Control=0.1 1 + 0.11 x treated= 0 .1 6 + 0.1 3 ) with no statistically significant difference between the two colonies (t= 0 .55 p= 0 .30 ). Thus, unknowns were accepted and classified as native species. On each plot, the percent cover each native plant species contributed to the total native percent cover was evaluated (Fig. 8 and Table 1). The control plot had a statistically significant higher mean percent cover of native species (x =29.58 + 21.08 ) than that of the treated plot (x = 10.3 3 + 14.9 6 t 2.67 p= 0 .01 9 ) (Fig. 9). For the invasive species, the percent cover each invasive plant species contributed to the total invasive percent cover on each plot was also evaluated (Fig. 10 and Table 2). Th e control plot had a lower percent cover of invasive plant species (x = 71.59 + 22.20 ) than that of the treated plot (x = 89.00 + 14.92 ) and the null hypothesis of no significant difference was rejected (t=1.9 9 p=0 .03 ) (Fig. 11 ). Of the seven total invasive species, only two were present on both the control and the treated plot, s lender russian thistle ( Salsola collina ) and b lack bindweed ( Pol ygonum convolvulus ). The control plot had a statistically significant lower mean percent c over of s lender russian thistle (x =33.44 + 26.7 4 ) than that of the treated plot (x =64.94 + 15.4 5 t= 2.4 3 p=0.0 3 ) (Fig. 12). Similarly to the russian t histle the control plot had a statistically lower mean percent

PAGE 21

19 cover of b lack bindweed (x = 12.63 + 12.76 ) than the treated plot (x = 40.34 + 27.54 t= 2.3 6 p=0.03 ) (Fig.13). Fig. 14 Mean plant species richness and SE calculated for control and treated plot at RMANWR Fig. 15 dex) and SE calculated for control and treated plot 0 1 2 3 4 5 6 Control Treated Mean Richness Plant Species Richness on both Plots Mean 0 0.05 0.1 0.15 0.2 0.25 Control Treated Mean Diversity Diversity of Plant Species on both Plots Mean

PAGE 22

20 Fig. 16 The mean relative abundance and SE of plant species on both plots calculated using total percent cover and richness Richness, Diversity and Relative Abundance T he mean richness of the contro l plot (x = 4.75 + 1.58) was calculated to be statistically greater than that of the treated plot (x = 3.3 0 + 0.67 t=2.63 p=0.009 0 ) (Fig. 14). The mean diversity of the control plot (x = 0 .13 + 0.093 ) was calculated to be higher than that of the treated plot (x = 0.19 + 0.01 1 ) but the null hypothesis that they did not statistically differ failed to be rejected (t= 1.3 1 p= 0 .21 ) (Fig. 15). T he relative abundance of plant species on both colonies using the species richness and total percent cover was calculated The control plot had a statistically significant higher mean relative abundance (x =23.91 + 6.29 ) than the treated plot (x =13.6 7 + 27.66 t=3.04 p= 0 .003 ) (F ig 16). 0 5 10 15 20 25 30 Control Treated Mean Relative Abundance Relative Abundance of Plant Species on Both Plots Mean

PAGE 23

21 Discussion The goal of the study was to compare the control plot that had been left untouched by land managers, with the treated plot that had been treated with herbicides and controlled burns, to determine if the treatment applied by land mana gers was effective in encouraging succession towards a more native state. Overall, the control plot experienced a higher total percent cover of plant species (Fig. 6 ) These findings suggest that the control plot experienced less bare ground than the treat ed plot, but a consideration to be made when interpreting these results, is that the treated plot had only been allowed to grow for one year post treatment, so plants had less height and bulk to them. Nonetheless, a successful restoration as defined by the U.S Army and Shell Oil Company must fit certain criteria that includes a minimum 70% total ground cover (U.S Fish and Wildlife Service 2013) The treated colony is well under this benchmark, sitting at 47.6% total percent cover. Another facet of determining the level of res toration a plot has achieved, is to consider the native plant species present and their abundance. Between the two plots, the treated plot experienced higher native species richness, but lower native species composition than the control plot (Fig. 7). This was surprising, as if the treatment were to be considered effective; I would expect to see a higher abundance and cover of native plant species on the treated plot than was observed (or at least values similar to that on the control plot), yet this wa s not the case (Fig. 9). Also, o f the native species observed overall, there was not a single native species that was present on both plots (Fig. 8 and Table 1). Possibly, this could be due to the extreme ends of the arsenal each plot resided in, allowing for exposure to different native seeds. The data suggest that the control plot is an environment more conducive to native plant species than the treated colony despite the treatments applied to discourage invasive species and encourage native s Previous

PAGE 24

22 r esearch suggests that if native plant species are not established early in succession before the invasive plant species have a chance to take over, restoration will not be successful and there may be a need to restart the process ( Martin & Wilsey 2012, Cohn 2009) This does not bode well for the treated plot as my data lead to the conclusion the native plant species have not become established early in succession. Quite possibly, of more note than the native species composition and percent cover, was that of the invasive species observed. As the control applied to the treated plot was applied to control invasive plant species (mainly to reduce the chance of a c heatgrass monoculture), it is of great interest to land managers to understand where the treated plot is headed in terms of other invasive species. The restor ation cannot be considered successful if other invasive species just replace the c heatgrass and reduce the success of native species in the area (Molano flores 2014) By evaluating the invasive species as I evaluated the native species, a pattern of invasive species dominating the native plot emerged (Fig. 10 and Table 2) On a surface level, the treated plot had a higher cover of invasive species (Fig. 11), but as I was greatly interested in determining if there was a significant d ifference between the two plots, a comparison of invasive plant species present on both plot s provided great insight Slender russian thistle and b lack bindweed were observed and identified on both plots, p roviding the opportunity for this comparison In this comparison, the truth of the efficacy of the control applied to the treated plot was revealed, as both species were significantly higher on the treated plot (Fig. 12 and Fig. 13) Both invasive species are common where there is disturbance or the area is very dry and exposed (Park et al. 2012) making RMANWR habitats the perfect place for an invasion. Added to that, b indweed has such an array of survival techniques to

PAGE 25

23 surviv requires very aggressive and careful control (Mitich 1991) Conclusions and Future Suggestions After ob serving each plot and evaluating the data, I have concluded that the treated plot has not managed to succeed towards a more native state, and is indeed progressing on a path with invasive plant species that will make achieving a native state close to impos sible without major intervention. My study was limited by the amount of data I was able to collect and the length of time each plot was observed. In order to fully understand the path the treated plot is progressing towards, a long term study will be requi red, but my data suggest action is required now to correct the succession the treated plot demonstrates. The A rsenal recognizes that current management practices are not providing the services needed to restore the habitats successfully (U.S Fish and Wildlife Service 2013) t hus I suggest a few steps to assist the treated plot. First and foremost, I believe that for any restoration of the natural prairie to happen on the treated plot, it is in the best interest of land managers to consider removing or relocating the prairie dog colony present on the treated plot. The less competition native plants have while trying to re establish the better chance they have of succeeding. The absence of the cl ipping and burrowing encouraged by the prairie dogs for a few years would provide some relief for the native plant species who are already in high competition with the s lender russian thistle and black bindweed present. Although an integral part to the pr airie eco system, herbivores such as the Black tailed Prairie D og can reduce diversity in arid grassland environments, and management of herbivores in these arid environments is crucial to restoration efforts (Olff and

PAGE 26

24 Ritchie 1998) Although this will assist restorat ion efforts, reducing herbivory and the burrowing effects of prairie dogs alone will not be enough (Park et al. 2012) Previous restoration sites have seen success in a variety of methods that were not limited to herbivore control. Limited reintroduction of herbaceous species native to the prairie grasslands using a mix of sod and seed could be successful in driving succession to resemble what the historic vegetation may have been on the A rsenal (Rudolph et al. 2014) Historically, prairie grasslands experienced fires periodically, which are suppressed when the land is taken over for anthropogenic uses. Potentially, fire could continue to be used as a control measure for certain species to simulate the historic disturbances and help control such speci es as Cheatgrass ( Augustine et al. 2009; Chambers et al. 2014) but would need to be coupled with other practices to reach maximum effect. As suggested by Martin and Wilsey (2012) it may be of value to evaluate the historic order of assembly of species that could genera te greater diversity by producing alternate states of native habitat that would have once existed (Martin and Wilsey 201 2) Land managers could use historical knowledge to amplify the outcome of treating for invasive species added with seeding for native species and restoring natural fire regimes. If practices and knowledge are expanded and combined, I believe the treate d plot will reach the native state the land managers strive for and sustainably provide the ecosystem services historically provided by the habitat. Acknowledgments I would like to thank the Rocky Mountain Arsenal National Wildlife Refuge for allowing me to conduct my research on their property, Emily Schednerlein for her assistance with data

PAGE 27

25 collection and photography, Dr. Carello for her constant advising and support, Dr. Baxendale for his advice and support, Dr. Meloche for his assistance in identifyin g the plants, and Metropolitan State University of Denver.

PAGE 28

26 Works Cited Augustine DJ, Milchunas DG, Ecology SR, Jan N, Milchunas G. 2009. Society for Range Management Vegetation Responses to Prescribed Burning of Grazed Shortgrass Steppe Linked ibed Burning of Grazed Shortgrass Steppe. Soc. Range Manag 62:89 97. Miller RF, Pyke D a. 2014. Resilience to Stress and Disturbance, and Resistance to Bromus tectorum L Invasion in Cold Desert Shrublands of Western North America. Ecosystems 17:360 375 Cohn JP. 1999. A Makeover for Rocky Mountain Arsenal. Bioscience 49:273 277. Jones R, Chambers JC, Johnson DW, Blank RR, Board DI. 2015. Effect of repeated burning on pla nt and soil carbon and nitrogen in cheatgrass (Bromus tectorum) dominated ecosystems. Plant Soil 386:47 64. Martin LM, Wilsey BJ. 2012. Assembly history alters alpha and beta diversity exotic native proportions and functioning of restored prairie plant communities. J. Appl. Ecol 49:1436 1445. Mitich LW. 1991. Intriguing World of Weeds Weed Technol 5:913 915. Molano flores B. 2014. An Invasive Plant Species Decreases Native Plant Reproductive Success. Nat. Areas J 34:465 469. Olechnoivski BEM, Debin ski DM, Drobney P, Viste sparkman K, Reed WT. 2009. Changes in Vegetation Structure through Time in a Restored Tallgrass Prairie Ecosystem and Implications for Avian Diversity and Community Composition. Ecol. Restor 27:449 458. Olff H, Ritchie ME. 1998. E ffects of herbivores on grassland plant diversity. TREE 13:261 265. Park JN, Antill TM, Naeth MA, Bork EW, Westhaver AL. 2012. Russian Thistle ( Salsola tragus L .) Control on Bighorn Sheep Winter Ranges in Jasper National Park. Nat. Areas J. 32:391 397. R udolph DC, Plair DE, Jones D, Williamson JH. 2014. Restoration and Winter Avian Use of Isolated Prairies in Eastern Texas. Southeast. Nat 13:52 64. Station BH, Park N, Springs H, Ame rica N. 2008. Does Biodiversity Ecosystem Function Science Apply to Prair Ecol. Restor 26:100 101. Stohlgren TJ, Binkley D, Chong GW, Kalkhan MA, Lisa D, Bull KA, Otsuki Y, Newman G, Bashkin M, Son Y. 1999. E xotic P lant S pecies I nvade Hot S pots of N ative P lant D iversity. Ecol. Monogr 69:25 46.

PAGE 29

27 United States Department of Agriculture Natural Resources Conservation Services Plants Database. [2015 April 19] Available from: http://plants.usda.gov/java/ United States Fish and Wildlife Service Rocky Mountain Arsenal Nat ional Wildlife Refuge: History Commerce City (Colorado): Available from: http://www.fws.gov/refuges/profiles/History.cfm?ID=61170 U.S Fish and Wildlife Service. 2013. U.S. Fish & Wildlife Service Management of black tailed prairie dog (Cynomys ludoviciamus) populations on the Rocky Mountain Ars enal National Wildlife Refuge Commerce City, Colorado. Weiner J. 2004. Are Invasive Plant Species Better Competitors Than Native Plant Species? Evidence from Pair Wise Experiments. OIKOS 105:229 238.