Plant Community Restoration on Mima Mounds at Turnbull National Wildlife Refuge, WA Brandy K

Eastern Washington University EWU Digital Commons

EWU Masters Thesis Collection Student Research and Creative Works

2012 Plant community restoration on mima mounds at Turnbull National Wildlife Refuge, WA Brandy K. Reynecke Eastern Washington University

Follow this and additional works at: http://dc.ewu.edu/theses Part of the Biology Commons

Recommended Citation Reynecke, Brandy K., "Plant community restoration on mima mounds at Turnbull National Wildlife Refuge, WA" (2012). EWU Masters Thesis Collection. 62. http://dc.ewu.edu/theses/62

This Thesis is brought to you for free and open access by the Student Research and Creative Works at EWU Digital Commons. It has been accepted for inclusion in EWU Masters Thesis Collection by an authorized administrator of EWU Digital Commons. For more information, please contact [email protected]. Plant Community Restoration on Mima Mounds at Turnbull National

Wildlife Refuge, WA

A Thesis

Presented to

Eastern Washington University

Cheney, Washington

In Partial Fulfillment of the Requirements

for the Degree

Master of Science

Brandy K. Reynecke

Summer 2012

Abstract:

Prevention and management of exotic plant species invasion can diminish their negative ecological and economic impacts and maintain ecological functions provided by native species.

This is particularly important in rare landscapes that provide essential habitat. The most successful management plans integrate techniques to reduce invasive species abundance, increase native species abundance and reduce invasibility. Eastern Washington’s Channeled

Scablands bioregion has highly fragmented remnant patches of grassland underlain by glacial alluvium and basalt bedrock that are dotted with Mima mounds (naturally occurring aggregations of fine particle soil). The mounds provide valuable habitat for native plants and animals and many have been heavily invaded by exotic plant species, particularly winter annual grasses such as Bromus tectorum and Ventenata dubia . This study was designed to characterize Mima mound plant communities in the Channeled Scablands and to test the effects of restoration techniques on those communities. I conducted a vegetation survey to characterize community composition and developed three experiments to: 1) test effectiveness of Imazapic, sucrose soil amendment and native seed addition ( Pseudoroegnia spicata and Festuca idahoensis ), for exotic plant reduction on two geologic substrates (basalt and alluvium), 2) test effectiveness of a biocontrol insect

(Mecinus janthiniformis) to reduce Linaria dalmatica , and 3) test effectiveness of an experimental strain of rhizobacteria, Pseudomonas fluorescens D7, for invasive annual grass reduction. I surveyed plant species composition and cover in 235 1 m 2 vegetation plots in June-

July 2009, measured environmental variables for 203 of the plots, applied treatments in summer and fall 2009, and resurveyed plots in June-July 2010. Species assembly was analyzed using

NMS ordination, indicator species and cluster analyses. Preliminary abundance comparisons between substrates were conducted using Dunnett’s T3 pair-wise comparisons for unequal

variance. Mean change in abundance over time was compared among treatments using general linear models (GLM).

Community analysis revealed ten community types and a distinct difference in species composition between substrate types. The experimental application of control techniques did not yield the expected results; treatments containing herbicide reduced all species, neither biocontrol agent reduced the abundance of target species, native seed addition had no effect and sucrose addition reduced native species cover but did not affect exotic cover. In control plots, invasive species cover increased significantly over the course of one year, suggesting that management is needed to prevent further invasion. Insight gleaned from both the community analyses and experimental results will be used to direct future study and guide adaptive management practices.

Acknowledgements

I would like to thank my advisor Dr. Rebecca L. Brown for support on this project and for all of the long hours she spent with me in the field, at school, and in her home to make this project a success. I owe a debt of gratitude to my committee members Dr. Suzanne Schwab and

Dr. Robin O’Quinn for being very candid and for imparting their expertise in the fields of botany and ecology. I would also like to thank Dr. Ross Black for his enthusiasm and sharing his statistical prowess. The advice and guidance I received from my professors has enabled me to grow professionally and personally.

I would also like to extend my gratitude to Mike Rule, TNWR biologist, for his advice and guidance on this project as well as for securing funding through a grant from the US Fish and Wildlife Service. I also received funding through a mini-grant from EWU. I want to thank

Jessica Bryant for her extensive assistance in experimental design, surveying and treatment application. Also I want to thank my field crew, Holly O’Connor, Brittany Davidson, Cara

Hulce, Jenni Grimes, Joe Miles, Trenton Reynecke, Aaron Clausen, Dash Hibbard and Mario

Lawrence. Without your assistance in the field I could not have completed this project.

I would like to thank Dr. Ann Kennedy at WSU for supplying the rhizobacteria and to

Jeremy Hensen for applying it. Thanks to Larry Skillestad of the Spokane office USDA for supplying insects and for sharing his extensive knowledge. Thank you to Dr. Julie Beckstead for supplying me with materials for insect cages and insight on sucrose addition.

Table of Contents Abstract ...... iv

Acknowledgements ...... vi

List of Figures ...... ix

List of Tables ...... x

Introduction ...... 1

Restoration approaches and techniques ...... 1

Prairie patches in the Channeled Scablands ...... 2

Mima mound prairies of eastern Washington ...... 3

Native plant community restoration approaches on Mima mounds ...... 7

Objectives …...... 9

Methods ...... 10

Study site description ...... 10

Survey of Mima mound vegetation ...... 14

Experimental design ...... 15

Experiment 1: Overall reduction of invasive species ...... 15

Experiment 2: Targeted control of Linaria dalmatica ...... 18

Experiment 3: P.f. D7 for targeted control of B. tectorum ...... 20

Data analyses ...... 22

Results ...... 23

vii

Vegetation survey ...... 23

Ordination analysis of 203 plot ...... 24

Comparison of pre-treatment IAG abundance and richness across substrate type ...... 24

Experiment 1: Change in abundance and richness of invasive and native plants due to

combinations of herbicide, native seed addition, sucrose soil amendment and geologic

substrate ...... 30

Experiments 2 and 3: Effects of biocontrol agents on L. dalmatica and B. tectorum ....35

Discussion ...... 38

Community description ...... 38

Restoration technique experiments ...... 39

Literature cited ...... 40

Appendix I: Supplemental figures from NMS ordination and community analyses ...... 53

Appendix II: ANOVA / GLM tables for combined effects of herbicide, sucrose and native seed ...... 58

Appendix III: Change in abundance of L. dalmatica due to addition of Mecinus janthiniformis ...... 65

Appendix IV: Change in IAG abundance due to Pseudomonas fluorescens D7 ...... 65

List of Figures viii

Figure 1: Study site location at TNWR ...... 12

Figure 2: Study area location at TNWR ...... 13

Figure 3: Axes 2 and 3 of the three dimensional NMS ordination of 203 plots ...... 25

Figure 4: Axes 1 and 3 of the three dimensional NMS ordination of 203 plots………………..23

Figure 5: Cluster analysis group results overlain on the ordination space ...... 24

Figure 6: Average percent cover of IAGs, exotic forbs and natives between substrates ...... 25

Figure 7: Average richness of IAGs, exotic forbs and natives between substrates ...... 26

Figure 8: Average change in a) exotic percent cover and b) native percent cover due to herbicide between substrates ...... 28

Figure 9: Average change in a) percent cover and d) stem count of IAGs due to herbicide between substrates ...... 29

Figure 10: Average change in a) native species richness and b) exotic species richness due to herbicide between substrates ...... 30

Figure 11: Average change in native species percent cover due to sucrose between substrates ...... 31

Figure 12: Change in L. dalmatica percent cover due to addition of insects, native seed, sucrose, herbicide, and control treatments ...... 34

List of Tables

Table 1: Experimental design for the fully factorial combination of restoration techniques ...... 17

Table 2: Experimental design for targeted control of L. dalmatica using M. janthiniformis .....19

Table 3: Experimental design for the targeted control of invasive winter annual grasses using

P.f. D7 ...... 21

Table 4: Significance values and F-test statistics for the change in abundance of total IAGs,

Bromus species, B. tectorum, and V. dubia ...... 36

Appendix I Tables

Appendix 1, Table 1: Detailed species lists for plots on mounds underlain by both substrates .53

Appendix I, Table 2: Comparison of exotic and native species cover and richness between substrate types using Dunnett’s T3 unequal variances pairwise comparisons ...... 55

Appendix 1, Table 3: Monte Carlo test of significance of Indicator Values (IV) ...... 56

Appendix II Tables

Appendix II, Table 1: Change in exotic species cover due to treatments. Herbicide and an interaction between herbicide and substrate type significantly reduced exotic species cover ...58

Appendix II, Table 2: Change in native species cover due to treatments. Herbicide and an interaction between herbicide and substrate significantly reduce native species cover ...... 59

Appendix II, Table 3: Change in IAG stem counts due to treatments. Herbicide and an interaction between herbicide and substrate type significantly reduced the stem counts of invasive grasses ...... 60

Appendix II, Table 4: Change in percent cover of IAGs due to treatments. Herbicide and an interaction between herbicide and substrate type significantly reduced invasive grass cover....61

Appendix II, Table 5: Change in total species richness due to treatments. Substrate type, herbicide and an interaction between the two, significantly reduced species richness ...... 62

Appendix II, Table 6: Change in exotic species richness due to treatments. Substrate type, herbicide and an interaction between the two, significantly reduced species richness ...... 63

Appendix II, Table 7: Change in native species richness due to treatments. Substrate type, herbicide and an interaction between the two, significantly reduced species richness...... 64

Appendix III tables

Appendix III, Table 1: GLM analysis of change in percent cover of Linaria dalmatica between control, insect addition (I), native seed addition (N) and sucrose addition (S) ...... 65

Appendix III, Table 2: GLM analysis of change in stem count of Linaria dalmatica between control, insect addition (I), native seed addition (N) and sucrose addition (S) ...... 65

Appendix IV Tables

Appendix IV, Table 1: Change in total abundance of IAGs due to P.f. D7. There were two levels of treatment, not added (N), and added (Y) ...... 66

Appendix IV, Table 2: Change in abundance of Bromus spp. due to P.f. D7 ...... 66

Introduction

Restoring native plant communities can reduce the negative ecological and economic impacts of exotic plant invasions. Invasions can create economic losses by reducing crop yields and forage for cattle (Duncan and others 2004), and can reduce ecological function through resource competition, altered natural disturbance regimes (Pellant 1996; Rimer and Evans 2006;

Simberloff 2005), and the reduction or replacement of native plant species which reduces forage and habitat for wildlife (Belnap and Sherrod 2009; DiTomaso 2000; Knapp 1996; Simberloff

2005; Sperry, Belnap, Evans 2006). Because of the devastation caused by exotic plant invasions, there is a need for native plant community restoration. Managing invasive plant populations, by reducing invasive species abundance, limiting disturbance intensity and frequency and controlling resource availability, is a major component of native plant community restoration

(Brown and others 2008; Dale and others 2000; Davis, Grime, Thompson 2000; DiTomaso 2000;

Grasslands and Treatment ; James and others 2010; Seabloom and others 2003; Sheley and others 2010; Sheley and Krueger-Mangold 2003). The goal of this project is to test restoration techniques that can be used to restore native plant communities in an arid prairie habitat by reducing invasive species and increasing native species.

Restoration approaches and techniques

Developing a restoration plan includes characterizing the plant community, setting measurable goals, selecting and implementing appropriate techniques, measuring success and sharing all results (Blossey 1999; Palmer and others 2005; Sheley and others 2010).

Characterizing plant community composition helps determine which factors are most influential with respect to species assembly, such as species availability, biotic interactions and

environmental factors, and is also a crucial step toward setting attainable and measurable restoration goals and selecting appropriate restoration techniques. Community data can also be used to measure the outcome of the restoration effort (whether successful, detrimental or ineffective), and guide the development of future restoration projects.

Invasion increases when factors such as resource availability, disturbance and species availability are altered (Davis, Grime, Thompson 2000; Stohlgren and others 1999). Restoration practices have generally followed one or more of three approaches: 1) reduce invasive species 2) increase native species, and 3) reduce susceptibility to invasion by managing disturbance or succession and/or by manipulating nutrients. Common techniques to reduce invasive abundance include applying biological controls, herbicide and physical removal, whereas techniques to increase native abundance include adding native seed and plants. Restricting disturbance and modifying soil nutrient levels can reduce the susceptibility of native populations to invasion

(Davis, Grime, Thompson 2000; James and others 2010; Sheley and Krueger-Mangold 2003).

Formatted: Font: Times New Roman, 12 pt, Prairie patches in the Channeled Scablands Bold

Natural prairie and grassland habitats have been almost entirely eliminated in the US.

Guidelines for ecologically based restoration emphasize the need to preserve rare landscape elements that provide essential habitat (Dale and others 2000; James and others 2011; Noss,

LaRoe, Scott 1995). Prairies are defined as rolling landscapes with deep fertile soils dominated by grasses and forbs with few trees or shrubs, and support highly productive plant communities.

As a consequence, they have been mostly converted to cropland leading to a substantial decline in prairie plant and animal biodiversity (Noss, LaRoe, Scott 1995; Samson and Knopf 1994).

Many prairie types exist throughout the US and all have been heavily impacted by anthropogenic use. 2

In southeastern Washington, less than one percent of the pre-European settlement Palouse

Prairie habitat remains, that which endures is highly fragmented (Communities 2000). The

Palouse prairie is relatively small when compared to the prairies in the central US. It occurs in eastern Washington, north Idaho and eastern Oregon, and historically also occurred in southern

Idaho and northern Utah (Stoddart 1941). Much of this prairie is located immediately adjacent to and somewhat interwoven with the Channeled Scablands of eastern Washington.

The Channeled Scabland region has an unusual topography that supports vernal wetlands, lakes, rivers, Ponderosa pine forests and prairie patches that support plant communities similar to those of the Palouse prairie. The Channeled Scablands are comprised of basalt bedrock overlain by glacial alluvium and wind deposited silt loam soils (loess) which were channelized by multiple massive glacial Lake Missoula flood events that occurred between 12 and 20 thousand years ago (Bretz 1923; Bretz 1969). The climate is semi-arid with an average annual rainfall of

15 to 20 inches and average high temperatures of 85 °F. There are four tracts of scabland, the largest being the Cheney-Palouse River tract, which is about 20 miles wide and 50 miles long

(Bretz 1928), the head of which is located at Cheney, WA. Like the Palouse prairie, much of this region’s prairie patches have been cultivated or used as rangeland resulting in very few contiguous patches of natural prairie habitat. The few remnant prairie patches tend to be dotted with geomorphic anomalies called Mima mounds, also known as prairie mounds, pimple mounds and biscuit-swale formations.

Mima mound prairies of eastern Washington

Mima mounds are aggregations of loess top soil that range in size from 0.5 to 1.5 m tall and 0.5 to 30 m 2, and provide habitat and forage for native ungulates and fossorial rodents.

Mounds are found in six regions in the United States (Washburn 1988) and also in Africa and 3

South America (Berg 1990) and their formation remains a mystery; however, hypotheses include large ancient burrowing mammals (Dalquest and Scheffer 1942), seismic activity(Berg 1990), and runoff/wind erosion combined with vegetation anchoring (Washburn 1988).

In the Channeled Scablands, mounds are underlain by either basalt bedrock or alluvial gravels. Mounds underlain by basalt tend to be larger, especially well-defined and relatively isolated from each other. Inter-mound areas on basalt have little soil development, are often comprised primarily of cryptogamic crusts, become intensely dry and hot during the summer, and are relatively impermeable, causing rainwater and snow melt to collect to form ephemeral wetlands in early spring and fall. During the hot dry summer months, mounds on the alluvial substrate may sustain higher soil moisture and contain more mineral nutrients due to the porosity of the gravel (Searcy, Wilson, Fownes 2003). Therefore, mounds underlain by basalt may endure a somewhat harsher climate which likely affects the relative abundance of native and exotic species (Daehler 2003).

Underlying substrate type appears to affect plant species composition and to interact with grazing to affect native and exotic species richness and abundance on mounds (Bryant, Reynecke and Brown in review ). Bryant, Reynecke and Brown (in review ) observed that more recently grazed basalt underlain mounds had fewer exotic and more native species than less recently grazed basalt, and more recently and less recently grazed alluvium underlain mounds. Bryant,

Reynecke and Brown ( in review ) also found that species composition differed between the tops and sides of mounds, with the tops having relatively more exotic and fewer native species.

To maintain Mima mound ecological function, their native plant communities need to be preserved. Turnbull National Wildlife Refuge (TNWR), located near the head of the Cheney-

Palouse tract of Channeled Scablands, has a large Mima mound prairie underlain by both basalt and glacial alluvium. The mounds on this prairie support both intact and degraded native plant communities making TNWR an ideal place to develop a native plant community restoration plan on Mima mounds in the Channeled Scablands.

Mima mound native plant communities have been heavily impacted by invasive winter annual grasses (IAGs) and exotic forbs. Historically populations of native perennial bunch grasses such as Pseudoroegneria spicata (Blue bunch wheatgrass), Festuca idahoensis (Idaho fescue) and Poa secunda (Sandberg’s bluegrass) (Daubenmire 1970; Del Moral and Deardorff

1976) were dominant on Mima mound and Palouse prairies. Currently Mima mounds support populations of native annual species including Clarkia pulchella (pink fairies), Collinsia parviflora (blue eyed Mary), Amsinckia menziesii (Menzie’s fiddle neck), and native perennial species like Lomatium triternatum (Nine leaf biscuit root) and ambiguum, Eriogonum heracleoides (Parsnipflower buckwheat) and Achillea millefolium (Yarrow). There are also large populations of IAGs such as Bromus tectorum (L.) and Ventenata dubia (Leers) Coss., invasive perennial grasses and forbs such as Poa bulbosa L. (bulbous bluegrass) and Linaria dalmatica

(L.) Mill. , and exotic annual forbs including Sisymbrium altissimum L. (tall tumble mustard) and

Vicia cracca L. (bird vetch). These invasive species produce large amounts of viable seed which flood the seed bank, occupy safe sites, and germinate earlier than native seeds (Mulroy and

Rundel 1977) resulting in reduced native plant abundance.

Bromus tectorum (cheat grass) is an IAG that is especially competitive in the western

United States due to its ability to 1) produce large quantities of highly viable seed that germinate in the fall as well as throughout the year and persist in the seed bank for 5 to 10 years (Mulroy and Rundel 1977; Young, Evans, Eckert Jr 1969; Young and Allen 1997); 2) reduce soil nitrogen 5

levels (Rimer and Evans 2006); and 3) alter natural fire regimes. Fall germinating B. tectorum seedlings produce a primary root that can reach a depth of 18 to 20 cm before it branches which allows the root tips to reach warmer soil and take advantage of fall precipitation (Skipper, Ogg,

Kennedy 1996). Because of the species’ rapid development and senescence results in a large amount of fine fuel, which burns easily and increases the frequency, intensity and magnitude of fires (Pellant 1996).

Due to its ability to self-perpetuate and proliferate, B. tectorum has directly replaced a variety of native plants which can lead to reduced animal abundance (Belnap and Sherrod 2009).

Belnap (2009) reported B. tectorum as comprising 50% – 85% of the vascular plant cover on rangeland in the western United States and Hall, Mull, Cavitt (2009) found that snake abundances in Antelope Island in Utah, decreased when the abundance of B. tectorum increased.

B. tectorum provides high quality forage when green but tends to be an inadequate source of nutrition for native and domestic ungulates for most of the year (Young and Allen 1997). Lack of palatability combined with the reduction of native species reduces available forage in natural prairies and rangelands.

Whereas B. tectorum poses a severe threat to plant and animal biodiversity, other IAGs such as Ventenata dubia are also having a negative impact. Ventenata dubia, also known as wire grass, North Africa grass or hair grass, is native to central and southern Europe, Asia, and Africa and was first reported in Washington State and northern Idaho in the early to mid-1950s (Prather

2009). Ventenata dubia is adapted to shallow rocky soils that experience early spring inundation and a wide range of precipitation (36- 112 cm/yr). It germinates in the fall after B. tectorum, and is increasing in areas previously dominated by B. tectorum (Forest Service (US) 2008). It can be introduced in grass seed mixes, primarily Kentucky bluegrass, and along roads (Prather 2009,

Scheinost, Stannard and Prather 2009). This species has demonstrated resistance to some common herbicides (The Nature Conservancy 2000) and is not easily mechanically removed due to the wiry quality of its culms (Lass and Prather 2007). Little research has been conducted on the biology of this plant and little is known about its potential ecological impacts.

Invasive forbs such as Linaria dalmatica (Dalmatian toadflax) also negatively impact native plant communities and the wildlife that depend on them. Linaria dalmatica is an escaped perennial forb introduced as an ornamental from Eurasia (Breiter and Seastedt 2007) that has invaded much of the United States and Canada (Wilson and others 2005). Characteristics that facilitate L. dalmatica’s ability to invade include vegetative regeneration via root buds and rhizomes (Wilson and others 2005), large production of highly viable seed (Robocker 1974,

1970), toxic foliage (Jeanneret and Schroeder 1992; Polunin 1969), and lack of natural predators.

Infestations of L. dalmatica can reduce available native winter forage for wildlife (Lajeunesse

1999) and devalue ranchland because cattle normally will not eat the toxic foliage (Anonymous

1971; Harris and Carder 1984; Lacey and Olson 1991; Polunin 1969).

Native plant community restoration approaches on Mima mounds

Native plant restoration is most successful when a combination of techniques is used

(James and others 2010; Monsen, Stevens, Shaw 2004); therefore, concurrently reducing invasive plant species with herbicides and biological control agents, increasing native perennial plant abundance through seed addition and altering soil nitrogen levels by amending the soil with carbon is likely to be an effective approach for the restoration of native plant communities on

Mima mounds.

The herbicide Plateau® can be used to treat pre- and early post-emergent plants, when applied appropriately, and is degraded by sunlight and microbial activity in three to six weeks.

Plateau® uses Imazapic, an acetolactate synthase (ALS/AHAS) enzyme inhibitor that builds up in the meristematic regions preventing growth and often causing mortality (Beran, Gaussoin,

Masters 1999; Plateau label). Although herbicide can be an effective means of invasive species control they can have unintended and undesirable ecological impacts such as reduction of non- target species.

Many biological control agents have been introduced as alternatives to herbicide as they tend to be relatively inexpensive, affect only their target species, and can have long lasting results (Breiter and Seastedt 2007). The biocontrol agent Pseudomonas fluorescens D7 ( P.f. D7) is a strain of rhizobacteria (bacteria that occupy the space immediately surrounding the roots of a plant) that was isolated from the roots of winter wheat in eastern Washington (Kennedy and

Kremer 1996) and was found to inhibit germination and growth of B. tectorum and other IAGs in laboratory (Kennedy, Johnson, Stubbs 2001) and agricultural field trials (Kennedy and others

1991; Kennedy and Kremer 1996; Kennedy unpublished data, unreferenced). The bacterium produces an unknown phytotoxin (Tranel, Gealy, Kennedy 1993) that targets roots and seeds to reduce seed production and seedling development, and is unlikely to significantly inhibit non- target species (Kennedy, Johnson, Stubbs 2001).

The biocontrol insect Mecinus janthiniformis Toˇsevski & Caldara (formerly known as

M. janthinus Germar) is a stem boring weevil that was introduced in 1995 from Eurasia (Hansen

2004), that has significantly decreased the abundance of L. dalmatica in British Columbia and

Washington State (De Clerck-Floate and Miller 2002; Jeanneret and Schroeder 1992). The

USDA conducted a five year study of M. janthiniformis and deemed it safe for release in the U.S.

(Hennessey 1996).

Another technique used as an alternative to herbicide is soil amendment with carbon, which promotes microbial growth thereby immobilizing plant available nitrogen. Carbon addition was found to reduce B. tectorum and L. dalmatica biomass and density (Beckstead and

Augspurger 2004; Paschke, McLendon, Redente 2000; Prober and others 2005; Young and Allen

1997). Reducing invasive species is commonly coupled with increasing native perennial plant abundance through seed addition. Studies have shown that intact native perennial communities tend to be less invaded by annual grasses indicating that they may be resistant to invasion

(Beckstead and Augspurger 2004; Belnap and Sherrod 2009) and, when not limited by seed availability, native perennial grasses are competitive with IAGs (Anonymous 1971; Antognini and others 1995; Corbin and D'Antonio 2004; Seabloom and others 2003). Therefore, adding native perennial grass seed has the potential to reduce invasive species abundance and community invasibility.

Objectives

The goal of this study was to develop a native plant community restoration plan for the

Mima mound prairie at TNWR by characterizing Mima mound plant communities and testing restoration techniques. I characterized Mima mound plant communities across both alluvial and basalt substrates and examined how environmental factors such as soil depth and size of mound, position of plot on mound, and location of mound on the landscape (i.e. swale, ridge, mid-slope), relate to community composition and invasive species. Based on personal observation I hypothesized that there would be a distinct difference in the species assembly on mounds underlain by different substrate types. In addition to general differences in species composition, I posited that mounds on basalt would have lower total species richness but higher native species richness and abundance due to the relative harshness of basalt prairies and lack of intermound

vegetation. I expected that the most heavily invaded mounds would be those underlain by alluvium, and that the bulk of exotic cover would be comprised of IAGs.

In a series of experiments I tested the effects of herbicide, biocontrol agents (P.f. D7 rhizobacteria and M. janthinus beetles), soil amendment with sucrose and native seed addition on the abundance of invasive and native species on the tops of Mima mounds across alluvial and basalt substrates. I hypothesized that each technique would have its respective anticipated effect but that a combination of techniques that both reduce invasives and increase natives would be the most effective. Mima mound plant community characterization and the results of the experiments will be used to determine the best approach for restoration on Mima mounds at

TNWR and throughout the Channeled Scabland region.

Methods

Study site description

Turnbull National Wildlife Refuge (TNWR) is located in Spokane County, five miles south of Cheney WA, USA (47° 24’N; 117° 31' W) on the northern edge of the Cheney-Palouse

River tract and encompasses approximately 70 km 2 of Channeled Scabland. The elevation at

TNWR is 695 m and has an average annual rainfall of ~ 40 cm and average snowfall of ~ 104 cm with December receiving the highest average precipitation. The highest temperatures occur in

July and August with an average high of 28.3° C (83°F) and low of 13.3° C (56° F), and the lowest temperatures occur in December and January with an average high of 0° C (32° F) and low of -3.9° C (25° F) (TWC 2012; USA.com 2012).

My study area included approximately 4 km 2 of semi-arid prairie underlain by basalt bedrock and alluvial gravels spanning the Stubblefield tract (mostly underlain by alluvial

gravels) and the Public Use Area. The landscape of the basalt prairie is more environmentally heterogeneous, which suggests that there may be more heterogeneity in the plant communities on mounds underlain by basalt (Dunson and Travis 1991; Ricklefs 1977) compared to mounds underlain by alluvium. In addition to the inter-mound swales on the basalt prairie, there are larger swales where rain and snow melt collect resulting in series of more mesic mounds that support moisture loving species like Rosa woodsii (wood’s rose), Crataegus douglasii (black hawthorn), Geranium viscosissimum (sticky geranium) and Gaillardia aristata (blanket flower); and series of more xeric mounds with drought tolerant vegetation like Artemisia tridentata

(sagebrush) and Eriogonum heracleoides (parsnip flower buckwheat). Such dramatic variation is not present on the prairie underlain by alluvium.

All of the mounds underlain by alluvium are located on the Stubblefield tract which is flanked by basalt underlain Mima mound prairie to the south and ranchland to the north; all of the basalt underlain mounds are on the Public Use Area which is bordered by Ponderosa pine forest on the north edge and basalt underlain Mima mound prairie on the south. Both tracts have been grazed in the past; however, the Public Use Area was last grazed over thirty years ago, whereas the Stubblefield tract was last grazed about twenty years ago. Cattle were only grazed on the land during the fall and winter and their diets were supplemented with hay (Mike Rule and Les Camp pers. comm. unreferenced).

The purpose of this study was two-fold; to describe and characterize plant communities on mounds and to experimentally test the effects of a suite of native plant restoration techniques.

Survey of Mima mound vegetation

To describe species composition and provide a baseline vegetation survey prior to experimental treatment, I established 235 1 m 2 plots. Ninety-three plots were established on

Mima mounds underlain by basalt bedrock and 142 plots were established on mounds underlain by alluvium. There was one plot per mound and, because the experimental portion of this project was aimed at reducing invasive plant species and the tops appeared to have the greatest abundance of invasive species, plots tended to be on or near the tops of mounds. Plots were marked using aluminum tags and six inch steel nails. Of the 235 vegetation plots only 203 could be relocated for measurement of disturbance level and type, position of mound in the landscape

(i.e. ridge, swale or in between), mound height, and soil depth. Two soil samples each from 10 alluvial and 10 basalt plots (40 samples total) were collected for soil nitrogen analysis. Samples were taken to the University of Idaho and tested for nitrate-nitrite and nitrogen-ammonia levels

(µg N/g soil) using calorimetric ASA 33-8.3 and ASA 33-7.3 methods respectively. Plots were re-surveyed in the summer of 2010

Species composition and estimated percent cover were recorded for each plot using a modified Braun-Blanquet (1932) cover class system (trace, 0-1%, 1-2%, 2-5%, 5-10%, 10-25%,

25-50%, 50-75%, 75-95% and 100%) to ensure a greater level of consistency and repeatability than actual percent cover. The incremental variation between classes magnifies the differences in cover across small scales. Species were identified using the technical keys provided in Hitchcock and Cronquist (1973) and botanical nomenclature was updated, where necessary, using the

Integrated Taxonomic Information System (ITIS; http://www.itis.gov; retrieved 19 September

2009). In the L. dalmatica biocontrol plots, I counted stems of each mature ramet.

The baseline vegetation survey began in mid-June, and ended in mid-August, 2009. Plots that had at least three stems of B. tectorum in a 400 cm 2 sub-plot located in a corner of the main plot were selected to receive treatments directly testing IAG reduction techniques. To ensure that the results of any one treatment were not disproportionately influenced by phenology, I randomized the sampling times across treatments. Most species were identifiable after senescence and so species richness values were not greatly affected by phenology.

Experimental design

Using subsets of the plots described above, I tested the effects of different IAG/invasive plant control and native restoration techniques in a series of three experiments.

Experiment 1: Overall reduction of invasive species

Experiment 1 used a fully factorial design to test the effects of the herbicide (none and applied), sucrose soil amendment (none and applied), and native perennial grass seed addition

(none and applied) on Mima mound plant communities underlain by both alluvial and basalt substrates for a total of 16 treatments (Table 1). There were seven replicates per treatment. Stem counts of non-native grasses were recorded in a specified 400 cm 2 corner of the main plot.

Treatments were applied to 1 m 2 plots with 0.25 m 2 buffer for a total treatment application area of 2.25 m 2. Plateau was applied by hand in late November using a long arm sprayer mounted on the back of an ATV at a rate of 140 ml/m 2or 0.75 percent concentration (the recommended spot treatment rate). Sucrose, in the form of white table sugar, was hand broadcast (to reduce disturbance) in three installments to allow a steady immobilization of nitrogen that would be manageable by the ecosystem at a rate of ~ 50 g sucrose/m 2 for a total of 150 g sucrose/m2 (21 g

C/m 2). The first application was timed to coincide with rain events in late June when microbial

activity would be highest, and again three weeks later, with a final application three weeks after that. In late November, approximately three weeks after the herbicide was applied, I hand broadcast a mix of 40% Pseudoroegnia spicata (Blue bunch wheat grass) and 60% Festuca idahoensis (Idaho fescue) with 95% PLS, each with 90% and 70% seed germination respectively at a rate of 1.36 g/m2 (rate per Newman and Redente 2001)

Table 1: Fully factorial combination of restoration techniques. Boxes marked with an X are

factors applied as treatments. All treatments were applied to mounds on both alluvial and basalt

substrate types for a total of 16 treatments (n = 7).

Factors

Native seed Herbicide Sucrose Control

1 X

2 X

3 X

4 X

Treatments 5 X X

6 X X

7 X X

8 X X X

Experiment 2: Targeted control of Linaria dalmatica

I tested the effect of the stem-boring weevil, M. janthiniformis, in combination with herbicide, sucrose soil amendment, and native seed addition (described in experiment 1) on L. dalmatica populations for a total of 5 treatments with seven replicates each (Table 2). Treatments were only applied to mounds underlain by alluvial substrate because there is little L. dalmatica on the mounds underlain by basalt. To determine the extent to which M. janthiniformis was already present, adult M. janthiniformis weevils were counted on ten 1 m 2 plots with similar community composition to treatment plots. Twenty adult M. janthiniformis weevils per plot were added in the first two weeks of June to coincide with ovaposition (Larry Skillestad ,USDA

Spokane office, personal communication, unreferenced), which approximately doubled the population present. All plots were selected based on toadflax presence and were covered with mesh screen cages framed with wooden stakes to keep adult insects on plots. Other factors were applied as in experiment 1.

Table 2: Experimental design for targeted control of L. dalmatica using M. janthiniformis. Boxes

marked with an X indicate the factor applied as treatment (n = 7).

Factors

Native seed Herbicide Sucrose M. janthiniformis Control

1 X

2 X

3 X X Treatments

4 X X X X

Experiment 3: P.f. D7 for targeted control of B. tectorum

I tested the effect of P.f. D 7 on fourteen plots underlain by alluvium that contained B. tectorum . I focused on alluvial mounds because there tended to be more B. tectorum on them.

P.f. D7 was applied to the soil surface of seven plots in late November 2009 (after the rainy season had begun and high temperatures fell below 10 °C (Kennedy, Smith, Stubbs 1995) at a rate of 108 cells/m2 in aqueous solution using a backpack sprayer at an equivalent of 9.39 ml/m 2 as per the WSDA EUP (experimental use permit) 9025 for rangeland application. The EUP only allowed for non-forage/non-feed application, so all plots had to be fenced prior to application and had to remain so for at least 365 days. The other seven plots were also fenced and used as a control (Table 3).

Table 3: Experimental design for the targeted control of invasive winter annual grasses using P.f.

D7. Boxes marked with an X indicate factors used as treatments (n = 7)

Factors

Rhizobacteria Control

1 X

2 X Treatments

Data analyses

I examined compositional differences among the plant communities in 203 plots, in summer of 2009 with non-metric multidimensional scaling (NMS) ordination using PC-ORD software (McCune and Mefford 1999). NMS is an iterative ordination technique based on ranked distances of n entities on k axes that seeks to minimize distortions caused by reductions in dimensionality (Minchin 1987). NMS plots illustrate similarity and/or dissimilarity in the community composition among vegetation plots (entities) (Clarke 1993). For the ordination I used the relative Sørenson distance measure, one of the most robust measures for this purpose

(Faith, Minchin, Belbin 1987). The data were run 250 times using 500 iterations for up to 6 axes starting with randomly generated starting configurations and stability criteria of 0.0005. Results reported are based on a three dimensional solution (fewest number of axes needed to achieve low stress) with Varimax rotation to maximize the variation explained by each axis.

I conducted a cluster analysis using Euclidian distance measures and Ward’s linkage method (McCune and Grace 2002) to determine distinct community groups among the plots.

Ten groups were created based on ~ 67% of the information used. Communities were characterized by the species most indicative of each group in accordance with indicator species values determined by Dufrêne and Legendre’s (1997; cited in McCune and Grace 2002) indicator species analysis. Groups were then overlain onto the ordination to show group associations within the species space. For the purposes of this discussion a plant community is defined as a relatively homogeneous assembly of species and is characterized by the dominant vegetation.

Differences in the abundance and richness of total exotic species, exotic forbs, IAGs, and native species on mounds were determined using Dunnet’s T3 pairwise comparisons. Data could be transformed to achieve normality but not equal variance. Normality was tested using

Anderson-Darling tests and equal variance was tested using two-sample variance F- tests.

Some plots could not be relocated resulting in variation of sample size within treatments; therefore, I used general linear model (GLM) tests to examine the effect of treatments on exotic forb, IAG and native species abundance and richness. I used GLM and Kruskal – Wallis tests to examine the effect of treatments on change in abundance of Linaria dalmatica because normality could not be achieved with transformations . For all significant results I ran post-hoc Tukey's

Honestly-Significant-Difference tests. I used paired t-tests to examine the difference in total and species specific IAG percent cover due to Pseudomonas fluorescens D7. Normality and homogeneity of variances were tested in the same way as previously discussed. Appropriate transformations were performed when needed to meet the assumptions of normality and equal variance.

Results

Vegetation survey

There were 60 species recorded for all plots, 8 were exotic grasses, 17 were exotic forbs,

2 were native bunch grasses, and 33 were native forbs. Plots on mounds underlain by alluvium had 6 unique exotic forbs and 3 unique native forbs. Plots on mounds underlain by basalt had 3 unique exotic forb species and 6 unique native forb species. All grasses, native and exotic, were found on both substrate types (Appendix I, Table 1).

Ordination analysis of 203 plots

Species composition varied between mounds underlain by basalt and alluvium with some overlap; mounds on basalt were larger. Variation explained by axes 1 through 3 is ~16%, 33%, and ~74% respectively (r 2 = 0.156, r 2 = 0.332, r2 = 0.735; Figs. 3 and 4). Two of the ten community groups were characterized by native perennial species, E. heracleoides and A. millefolium, and both of those groups were found primarily on the alluvial substrate (Fig. 5; see appendix 1, Table 3 for indicator and p values)

Comparison of pre-treatment IAG abundance and richness across substrate types

There was no difference in the abundance of total exotic species, IAGs, exotic forbs or native species abundance between substrates (Fig. 6). Native and total species richness was greater on basalt (p = 0.002, p = 0.003 respectively; Fig. 7; Appendix 1, Table 2).

Figure 3: Axes 2 and 3 of the three dimensional NMS ordination of 203 plots. Each symbol represents the plant community in a 1 m 2 quadrat on a Mima mound and is placed along axes according to the similarity of community composition on mounds underlain by alluvial (hollow triangles) and basalt (filled triangles) substrate types. Vector lines indicate direction and strength of the correlation between mound height, length and width with the ordination axes; the longer the line the stronger the correlation.

Figure 4: Axes 1and 3 of the three dimensional NMS ordination of 203 plots. Each symbol represents the plant community in a 1 m 2 quadrat on a Mima mound and is placed along axes according to the similarity of community composition on mounds underlain by alluvial (black dots) and basalt (green triangles) substrate types. Vector line indicates direction and strength of the correlation between mound height and the ordination axes; the longer the line the stronger the correlation.

Figure 5: Cluster analysis group results overlain on the ordination space. Groups are named according to the results of the indicator species analysis of the cluster analysis.

250

200 2

150

Natives

100 IAG Exotic forbs Average percent cover per m per cover percent Average

0 Alluvium Basalt Substrate type

Figure 6: Average percent cover of IAGs, exotic forbs, and natives between substrates. Error bars = 1 SE.

12 2

10 Natives 8 IAG Exotic Forbs

Average richness perm richness Average 6

0 Alluvium Basalt Substrate type

Figure 7: Average richness of IAGs, exotic forbs, and natives between substrate. Error bars = 1 SE.

Experiment 1: Change in abundance and richness of invasive and native plants due to combinations of herbicide, native seed addition, sucrose soil amendment and geologic substrate

Herbicide reduced the abundance of both exotic and native species (p < 0.001, F =

69.222, F = 22.949) and interacted with substrate for a greater reduction in exotic abundance (p =

0.007, F = 8.448; Fig. 8a), native abundance (p = 0.045, F = 4.099; Fig. 8b), IAG percent cover

(p = 0.002, F = 9.624; Fig 9a) and IAG stem count (p = 0.004, F= 8.719; Fig. 9b) on alluvium.

Herbicide alone and in conjunction with substrate type reduced exotic species richness (p =

0.000, F = 77.792, F = 38.925) and native species richness (p = 0.000, F = 41.924, F = 13.537;

Fig. 10). Also, native and exotic species richness significantly declined on plots on basalt as compared to those on alluvium in control plots ( p < 0.001, F = 165.993).

Native seed addition did not significantly affect exotic (p = 0.463, F = 0.542) or native abundance (p = 0.054, F = 3.789), nor did it reduce total (p = 0.880, F = 0.023), exotic (p =

0.840, F = 0.041), or native richness (p = 0.726, F = 0.123), IAG percent cover (p = 0.530, F =

0.396), or IAG stem count (p = 0.550, F = 0.360).

Soil amendment with sucrose reduced only native species cover (p = 0.031, F = 1.25) and richness (p = 0.031, F = 4,759; Fig. 11; See Appendix II for all p values). There was a dramatic reduction in species richness from 2009 to 2010 in control plots on basalt as compared to alluvium (p < 0.001, F = 165.993).

10 2 5

-5 Alluvium Basalt -10 percent cover per m coverper percent -15 Average change in exotic species exotic in change Average -20 Control Herbicide Treatment a)

-1 2 -2

-3 Alluvium -4 Basalt

percent cover per m per cover percent -5

Average change in native species native in change Average -6 Control Herbicide Treatment b)

Figure 8: Average change in a) exotic percent cover and b) native percent cover due to herbicide between substrates. Error bars = 1 SE.

5 2 0

Alluvium -5 IAGs per m per IAGs Basalt

-10

Average change in percent cover of coverof percent inchange Average -15 Control Herbicide Treatment a)

150

100

50 2 Alluvium 0 per m per Basalt -50

-100 Average change in IAG stem count count stemIAG in change Average Control Herbcide Treatment b)

Fig. 9: Average change in a) percent cover and b) stem count of IAGs due to herbicide between substrates. Error bars represent 1 SE.

2 -2

-4 Alluvium -6 Basalt richness per m per richness -8

Average change in native species nativein change Average -10 Control Herbicide Treatment a)

0 2

-2 Alluvium -4 Basalt

richness per m per richness -6

-8 Average change in exotic species exotic in change Average Control Herbicide Treatment b)

Figure 10: Average change in a) native species richness and b) exotic species richness due to herbicide between substrates. Error bars = 1 SE.

-1

-2

2 -3 m AL

-4 BA

-5

Average change in native percent cover per per cover percent native in change Average -6 Control Sucrose Treatment

Figure 11: Average change in native species percent cover due to sucrose between substrates

(AL = alluvium, BA = basalt). Error bars = 1 SE.

Experiments 2 and 3: Effects of biocontrol agents on L. dalmatica and B. tectorum

The addition of M. janthiniformis alone and in conjunction with herbicide, sucrose addition and native seeding did not significantly reduce the percent cover of L. dalmatica (p =

0.1661, F = 0.794). Only the herbicide treatment reduced the percent cover of L. dalmatica

(p=0.036, F=5.12; Fig 12). There was also a reduction in stem count of L. dalmatica due to treatments containing herbicide (p = 0.002, F = 13.576).

The addition of Pseudomonas fluorescens D7 did not significantly reduce the total percent cover or stem count of IAGs, Bromus spp., B. tectorum, or Ventenata dubia (Table 4).

Table 4. P values and F-test statistics for the change in abundance of total IAGs, Bromus species,

B. tectorum, and V. dubia.

P value F-test statistic

Total IAG percent cover p = 0.937 F = 0.006

IAG stem count p = 0.532 F = 0.395

Bromus spp. percent cover p = 0.818 F = 0.054

Bromus spp. stem count p = 0.572 F = 0.324

Bromus tectorum percent cover p = 0.919 F = 0.011

Bromus tectorum stem count p = 0.404 F = 0.754

Ventenata dubia percent cover p = 0.383 F = 0.825

Ventenata dubia stem count p = 0.900 F = 0.016

30 20 10 0

2 -10 -20 -30 cover per m per cover -40 -50 -60

Average Chnge in Linaria dalmatica percent percent dalmatica Linaria in Chnge Average -70 Control Insects I,S I,S,H I,S,H,N Treatment

Figure 12: Change in L. dalmatica percent cover due to addition of insects (I),native seed (N), sucrose (S), herbicide (H), and control treatments. Error bars = 1 SE.

Discussion

This study was designed to characterize Mima mound plant communities in the

Channeled Scablands and to test the effects of restoration techniques on those communities. Prior to this study little was known about the nature of native plant communities on Mima mounds or about how best to approach restoring native species in this highly heterogeneous landscape.

Community analysis revealed ten community types and a distinct difference in species composition between substrate types (alluvium and basalt). The experimental application of control techniques did not yield the expected results and in some cases had the opposite effect, however our results from both the community analysis and experimental results provide needed insights that will direct future studies and potentially guide adaptive management practices.

Community description

My results suggest that plant community assembly on the Mima mound prairie at TNWR can be characterized by the type of underlying substrate; in this case either basalt bedrock or glacial alluvium. However, a suite of factors may influence the perceived effect of substrate including grazing history, mound size and soil depth, pocket gopher activity and soil moisture and nutrient levels. Each substrate type has a different grazing history, which may also be a contributing factor. The area underlain by alluvium has been more recently grazed (about 20 years ago), and the area underlain by basalt has been less recently grazed (about 30 years ago).

Though neither substrate type, nor grazing history was replicated in this study, in 2009, concurrent with this study’s vegetation survey, Bryant, Reynecke and Brown ( in review ) compared plots on alluvium and basalt that had opposite grazing histories to plots on the corresponding substrate type sampled in this study; they found that grazing and substrate

interacted to affect species composition. However, the less recently grazed alluvial and more recently grazed basalt areas were relatively small and geographically distinct, which made it difficult to distinguish the effects of grazing from other geographic differences. Overall, characteristics of substrate undoubtedly serve as a significant modifier of plant community assemblies as geologic substrate has been shown to affect plant community structure (Schneck et al. 2003), vegetation transition rates and response to disturbance (Callaway 1993).

Mounds underlain by alluvium had both the only community types indicated by native perennial forbs and greatest abundance of IAGs. Mounds underlain by alluvium may have environmental conditions (i.e., soil moisture, disturbance regime, nutrient availability, etc.) better suited for native perennial establishment and, either the perennial species are slowly being replaced by IAGs, or communities dominated by native perennial forb species are resistant to invasion by IAGs (Connell and Slatyer 1977). It is possible that underlying hydrology is affecting the relative invasibility of the alluvial site as there is a small lake to the east. Future studies should consider effects of water table levels on species assembly.

Because the mounds underlain by basalt are larger and have deeper soil they may experience higher levels of pocket gopher disturbance, which could be keeping the plant communities in an earlier successional stage. Though I did record disturbance in my plots I did not see a correlation with species composition. However, long term examination of animal activity disturbance may correlate with changes in species composition.

Soil nutrient and moisture levels may also vary between substrates. There was no difference in soil nitrogen levels between substrates; however, my sample size of ten mounds per substrate may not have been large enough to show variation. Also, nitrogen levels may vary

depending on time of year due to fluctuation in weather conditions. Soil moisture was not measured in my study but I recommend that both soil nitrogen and moisture be measured throughout the growing season in future studies.

Restoration technique experiments

Despite the fact that neither combining techniques from each of the three general restoration approaches nor the individual treatments themselves affected plant abundance as expected, significant insight was gained into the practice of restoration that will be useful for modifying subsequent studies. Plateau (Imazapic) is successful as a pre-emergent/early post- emergent herbicide for controlling IAGs (Masters and Sheley 2001; Nyamai, Prather, Wallace

2011; Scheinost , Stannard and Prather 2009) and the November application date was appropriate to target winter annual grasses; however, the recommended spot treatment application rate was too strong and resulted in high mortality among native species. Using the recommended broadcast rate and/or trying a different herbicide may provide better results.

I observed an interaction between substrate type and herbicide treatments in that overall percent cover was more reduced on alluvium than on basalt. One explanation could be that prior to treatment there was approximately 30% more Lactuca serriola, and approximately 70% more

Vicia cracca in plots on basalt as compared to alluvium ; both of which are known to be tolerant to acetolactate synthase (ALS) inhibiting herbicides such as Imazapic (Lass and Prather 2007;

Mikulka and Chodová 2003; Vollmer and Vollmer 2006). The residual presence of these plants on basalt plots could account for the difference in the reduction of percent cover on alluvium as compared to basalt. To avoid the persistence of herbicide tolerant plants a cocktail of herbicides that kill plants using different pathways, such as Imazapic and glyphosate may be preferred. 40

However, presence of L. serriola and V. cracca do not account for the greater reduction of IAGs on alluvium from herbicide treatment. The observed difference in herbicide effectiveness across substrates may have been caused by some basalt plot corners or edges not being treated, which I personally observed for a few plots. The inaccuracy may be due to the application method.

Herbicide was applied from an ATV mounted sprayer and perhaps identifying the exact location of plots on large, steep, basalt mounds was difficult. In the future, I recommend that herbicide be applied with backpack sprayers rather than ATV mounted sprayers for both accuracy and to reduce vehicle disturbance.

The addition of biocontrol agents did not reduce the abundance of their target species .

Past studies have shown that the addition of P. f. D7 reduces seedling vigor and plant growth of

Bromus tectorum and other winter annual grasses in both field and lab trials (Kennedy and others

1991; Kennedy, Smith, Stubbs 1995; Kennedy, Johnson, Stubbs 2001; Skipper, Ogg, Kennedy

1996; Tranel, Gealy, Kennedy 1993; Weddell and United States. Bureau of Land Management.

Idaho State Office 2001). However, the P. f. D7 bioagent did not reduce the abundance of IAGs on my study site. The inoculum was applied in early-December which is the appropriate time for this cold loving bacterium (Skipper, Ogg, Kennedy 1996) but the temperature on the day of application was approximately -6.7 °C. It is possible that the ambient temperature was too low for the bacteria to survive application. Adding the bacteria on a day when the ambient temperature is above freezing may yield better results and I suggest follow up soil surveys to determine bacteria survival. Another possible reason there was no significant reduction in IAG cover could be because the density of the B. tectorum infestation was so great that one growing season was not long enough to see a reduction. Often P. f. D7 takes three to five years before a reduction of B. tectorum is achieved (Kennedy 2010 personal communication, unreferenced),

especially if the density of the target grass is high (Weddell and United States. Bureau of Land

Management. Idaho State Office 2001). Future surveys of my plots that received the P. f. D7 treatment should be conducted to determine if there was a reduction in IAG cover over time.

There was also no reduction in the abundance of Linaria dalmatica due to the addition of the biocontrol agent M. janthiniformis, which was surprising because this particular biocontrol has consistently successfully controlled L. dalmatica (Hennessey 1996; Schat and others 2011).

One growing season might not have been enough time to see a significant reduction in abundance of the target species (De Clerck-Floate and Miller 2002; Wilson and others 2005).

Also, there are other ways to measure the effects of insects on plant productivity such as counting number of larvae per-stem, and determining flower or seed production as well as biomass. Furthermore, the insect was already present and it could be that adding more does not increase its effectiveness. I suggest future studies examine the effects of this insect on seed production over the course of several years.

Amendment of the soil with sucrose did not reduce exotic species abundance but instead reduced native species abundance. There is some evidence suggesting invasive annual grasses can outperform native perennials under N-poor conditions (Monaco and others 2003; Young and

Mangold 2008) and invasive winter annual grasses’ competitive ability at the seedling stage is not greatly altered by managing for low N because annual grasses have a relatively high leaf nitrogen production rate (rate of dry matter gain per unit leaf N per unit time) as compared to native perennials (James and others 2011). An alternative explanation may be that many of the native species on the Mima mounds were actively growing during application; L. dalmatica seeds had most likely germinated prior to application and IAGs, which comprised the bulk of the exotic species, had senesced (Paschke personal communication, unreferenced; Beckstead and 42

Augspurger 2004; Paschke, McLendon, Redente 2000). I am confident that the rate of carbon addition was sufficient because I did see a reduction in species cover (though it was not the desired species). Future studies testing the reduction of soil nitrogen should apply carbon in early spring and/or fall to coincide with the initial growth of invasive species.

There was no evidence that the native grass seed established, or that the seeded plants reduced exotic species. Perennial grasses are slow to develop so the one year time frame of my research may have been too little to see an effect. Therefore, I recommend resurveying the plots in future years. Krueger-Mangold, Sheley and Svejcar (2006) also suggested that seeding should be done multiple times to ensure that native seeds are present when environmental conditions are just right for germination. It is also possible that none of the added seed germinated because of seed predation. Many studies have suggested that the use of mulch or some other covering increases the chance of germination of added seeds by protecting them from predation by birds, rodents and other seed eaters (Brown and others 2008; Chambers 2000; Daehler 2003).

There is also the possibility that the Mima mounds, which are used heavily by pocket gophers and other burrowing mammals, are in a continuous state of early succession caused by high levels of soil disturbance such that perennial bunchgrass species will not establish or persist.

The pocket gopher, Thomomys taploides, is present on my study site and was found by Andersen and MacMahon (1985) to alter plant community composition, dynamics, and successional pathways. In a habitat that has high levels of natural disturbance, seeding with a mix that includes native ruderal species may be more effective than seeding with later successional perennials. It would also be worthwhile to examine if naturally occurring disturbance due to animal activity is keeping the mounds in an early successional stage and affecting mound

invasibility. If it is determined that the mound communities need to be moved to a later successional state, seeding with native annuals may facilitate natural succession (Zolkewitz

2010).

There was also a dramatic decrease in native and exotic species richness in control plots indicating that if no steps are taken to increase and maintain native plant communities they may be lost. The preliminary vegetation survey did not show a difference in overall species richness between substrates; however, overall species richness decreased significantly in control plots in one growing season. This is very concerning because it reveals rapid deterioration of native plant communities on basalt underlain mounds.

The results of this study were intended to create a restoration/management plan that could be used by TNWR and private landowners in the Channeled Scablands region. Further examination of abiotic factors, such as soil moisture and soil nutrient levels, drainage, soil type, as well as distance between mounds, biotic factors such as animal and human activity and continued vegetation surveys will allow for a better understanding of the mechanisms that drive plant species composition in this unusual landscape. This study demonstrates that community composition varies with substrate type. Because of the dramatic decrease in native species richness in basalt control plots on basalt, these communities will most likely not move towards a more native condition without management.

Literature Cited

Andersen DC and MacMahon JA. 1985. Plant succession following the Mount St. Helens volcanic eruption: Facilitation by a burrowing rodent, thomomys talpoides. Am Midl Nat :62-9.

Antognini J, Quimby PC,Jr., Turner CE, Young JA. 1995. Implementing effective noxious range weed control on rangelands. Rangelands 17(5):158-63.

Beckstead J and Augspurger CK. 2004. An experimental test of resistance to cheatgrass invasion: Limiting resources at different life stages. Biol Invasions 6(4):417-32.

Belnap J and Sherrod SK. 2009. Soil amendment effects on the exotic annual grass Bromus tectorum L. and facilitation of its growth by the native perennial grass Hilaria jamesii (Torr.) Benth. Plant Ecol 201(2):709-21.

Beran DD, Gaussoin RE, Masters RA. 1999. Native wildflower establishment with imidazolinone herbicides. HortScience 34(2):283-6.

Berg AW. 1990. Formation of Mima mounds: A seismic hypothesis. Geology 18(3):281.

Blossey B. 1999. Before, during and after: The need for long-term monitoring in invasive plant species management. Biol Invasions 1(2):301-11.

Braun-Blanquet J. 1932. Plant sociology. the study of plant communities. Plant Sociology.the Study of Plant Communities.First Ed.

Breiter NC and Seastedt TR. 2007. Postrelease evaluation of Mecinus janthinus host specificity, A biological control agent for invasive toadflax ( Linaria spp ). Weed Sci 55(2):164-8.

Bretz JH. 1969. The Lake Missoula floods and the Channeled Scabland. J Geol :505-43.

Bretz JH. 1928. The Channeled Scabland of eastern Washington. Geogr Rev 18(3):446-77.

Bretz JH. 1923. The Channeled Scablands of the Columbia plateau. J Geol 31(8):617-49.

Brown CS, Anderson VJ, Claassen VP, Stannard ME, Wilson LM, Atkinson SY, Bromberg JE, Grant III TA, Munis MD. 2008. Restoration ecology and invasive plants in the semiarid west. Invasive Plant Science and Management 1(4):399-413.

Bryant JB, Reynecke BK, Brown RL. In review. Geology and topography effects on exotic plant distribution in a semi-arid Mima mound prairie in eastern Washington.

Chambers JC. 2000. Seed movements and seedling fates in disturbed sagebrush steppe ecosystems: Implications for restoration. Ecol Appl 10(5):1400-13. 45

Clarke KR. 1993. Non ‐parametric multivariate analyses of changes in community structure. Aust J Ecol 18(1):117-43.

Communities RE. 2000. Ecological communities: Terrestrial vegetation of the United States. volume II: List of vegetation units. Arlington, Va.: The nature conservancy. Precious Heritage: The Status of Biodiversity in the United States 93:347.

Connell JH and Slatyer RO. 1977. Mechanisms of succession in natural communities and their role in community stability and organization. Am Nat :1119-44.

Corbin JD and D'Antonio CM. 2004. Competition between native perennial and exotic annual grasses: Implications for an historical invasion. Ecology 85(5):1273-83.

Daehler CC. 2003. Performance comparisons of co-occurring native and alien invasive plants: Implications for conservation and restoration. Annual Review of Ecology, Evolution, and Systematics :183-211.

Dale VH, Brown S, Haeuber RA, Hobbs NT, Huntly N, Naiman RJ, Riebsame WE, Turner MG, Valone TJ. 2000. Ecological principles and guidelines for managing the use of land. Ecol Appl 10(3):639-70.

Dalquest WW and Scheffer VB. 1942. The origin of the Mima mounds of western Washington. J Geol :68-84.

Daubenmire R. 1970. Steppe vegetation of Washington. Technical Bulletin.Washington Agricultural Experiment Station (62).

Davis MA, Grime JP, Thompson K. 2000. Fluctuating resources in plant communities: A general theory of invasibility. J Ecol 88(3):528-34.

De Clerck-Floate R and Miller V. 2002. Overwintering mortality of and host attack by the stem- boring weevil, Mecinus janthinus Germar, on Dalmatian toadflax ( Linaria dalmatica (L.) mill.) in western Canada. Biological Control 24(1):65-74.

Del Moral R and Deardorff DC. 1976. Vegetation of the Mima mounds, Washington State. Ecology :520-30.

DiTomaso JM. 2000. Invasive weeds in rangelands: Species, impacts, and management. Weed Sci 48(2):255-65.

Duncan CA, Jachetta JJ, Brown ML, Carrithers VF, Clark JK, Ditomaso JM, Lym RG, McDaniel KC, Renz MJ, Rice PM. 2004. Assessing the economic, environmental, and societal losses from invasive plants on rangeland and wildlands. Weed Technol 18(, Invasive Weed Symposium):1411-6.

Dunson WA and Travis J. 1991. The role of abiotic factors in community organization. Am Nat :1067-91.

Faith DP, Minchin PR, Belbin L. 1987. Compositional dissimilarity as a robust measure of ecological distance. Plant Ecol 69(1):57-68.

Forest Service (US). 2008. Wildland fire in ecosystems: Fire and nonnative invasive plants. Forest Service.

Grasslands DP and Treatment NW. Projects occurring nationwide. .

Hall LK, Mull JF, Cavitt JF. 2009. Relationship between cheatgrass coverage and the relative abundance of snakes on Antelope Island, Utah. West N Am Nat 69(1):88-95.

Hansen RW. 2004. Biological control of Dalmatian toadflax 2 nd ed. Washington, DC: USDA- APHIS-PPQ-CPHST. [pamphlet]

Harris P and Carder A. 1984. Linaria vulgaris Miller, yellow toadflax and L. dalmatica (L.) Mill., broad-leaved toadflax (Scrophulariaceae). Biological Control Programmes Against Insects and Weeds in Canada 1959–1968.Commonwealth Institute of Biological Control Technical Communication 4:179-82.

Hennessey RD. 1996. Field release of the exotic weevil Mecinus janthinus (Coleoptera: Curculionidae), for biological control of the weeds Dalmatian toadflax, Linaria dalmatica, and Yellow toadflax, Linaria vulgaris (Scrophulariaceae).

Hitchcock CL and Cronquist A. 1973. Flora of the pacific northwest: An illustrated manual. Seattle: London, Washington University Press. xx, 730p.. Illustrations.

James JJ, Smith BS, Vasquez EA, Sheley RL. 2010. Principles for ecologically based invasive plant management. Invasive Plant Science and Management 3(3):229-39.

James J, Drenovsky R, Monaco T, Rinella M. 2011. Managing soil nitrogen to restore annual grass-infested plant communities: Effective strategy or incomplete framework? Ecol Appl 21(2):490-502.

Jeanneret P and Schroeder D. 1992. Biology and host specificity of Mecinus janthinus Germar (Col.: Curculionidae), a candidate for the biological control of yellow and Dalmatian toadflax, Linaria vulgaris (L.) Mill, and Linaria dalmatica (L.) Mill. (Scrophulariaceae) in North America. Biocontrol Sci Technol 2(1):25-34.

Kennedy A and Kremer R. 1996. Microorganisms in weed control strategies. J Prod Agric 9(4):480-4.

Kennedy A, Young F, Elliott L, Douglas C. 1991. Rhizobacteria suppressive to the weed downy brome. Soil Sci Soc Am J 55(3):722.

Kennedy AC, Johnson BN, Stubbs TL. 2001. Host range of a deleterious rhizobacterium for biological control of downy brome. Weed Sci 49(6):792-7.

Kennedy AC, Smith KL, Stubbs TL. 1995. Investigating soil microorganisms for biological weed control. The American Biology Teacher 57(8):526-30.

Knapp PA. 1996. Cheatgrass ( Bromus tectorum L.) dominance in the great basin desert: History, persistence, and influences to human activities. Global Environ Change 6(1):37-52.

Krueger-Mangold JM, Sheley RL, Svejcar TJ. 2006. Toward ecologically-based invasive plant management on rangeland. Weed Sci 54(3):597-605.

Lacey JR and Olson BE. 1991. Environmental and economic impacts of noxious range weeds. Noxious Range Weeds.Boulder, CO: Westview :5-16.

Lajeunesse S. 1999. Dalmatian and yellow toadflax. Biology and Management of Noxious Rangeland Weeds.Oregon State University Press, Corvallis :202-16.

Lass L and Prather T. 2007. A scientific evaluation for noxious and invasive weeds of the highway 95 construction project between the Uniontown cutoff and Moscow. QuilaVision Inc., Missoula, MT.Online.Availble: Http://northwestmedia.net/us95/pdf-final/Final-Weed- Report.Pdf.(Accessed 5/27/2008) .

Masters RA and Sheley RL. 2001. Principles and practices for managing rangeland invasive plants. J Range Manage :502-17.

McCune B and Grace JB. 2002. Analysis of ecological communities. MjM Software Design.

McCune B and Mefford MJ. 1999. PC-ORD: Multivariate analysis of ecological data; version 4 for windows;[user's guide]. MjM Software Design.

Mikulka J and Chodová D. 2003. Germination and emergence of prickly lettuce (lactuca serriola L.) and its susceptibility to selected herbicides. Plant Soil and Environment 49(2):89-94.

Minchin PR. 1987. An evaluation of the relative robustness of techniques for ecological ordination. Plant Ecol 69(1):89-107.

Monaco TA, Johnson DA, Norton JM, Jones TA, Connors KJ, Norton JB, Redinbaugh MB. 2003. Contrasting responses of intermountain west grasses to soil nitrogen. J Range Manage :282-90.

Monsen SB, Stevens R, Shaw NL. 2004. Restoring western ranges and wildlands. United States Dept. of Agriculture, Forest Service, Rocky Mountain Research Station.

Mulroy TW and Rundel PW. 1977. Annual plants: Adaptations to desert environments. Bioscience :109-14.

Nature Conservancy. 2000. Weed Alert–Ventenata dubia. http://tncinvasives.ucdavis.edu/alert/alrtvent.html.

Newman GJ and Redente EF. 2001. Long-term plant community development as influenced by revegetation techniques. J Range Manage :717-24.

Noss RF, LaRoe ET, Scott JM. 1995. Endangered ecosystems of the united states: A preliminary assessment of loss and degradation. US Department of the Interior, National Biological Service Washington, DC, USA.

Nyamai PA, Prather TS, Wallace JM. 2011. Evaluating restoration methods across a range of plant communities dominated by invasive annual grasses to native perennial grasses. Invasive Plant Science and Management 4(3):306-16.

Palmer M, Bernhardt E, Allan J, Lake P, Alexander G, Brooks S, Carr J, Clayton S, Dahm C, Follstad Shah J. 2005. Standards for ecologically successful river restoration. J Appl Ecol 42(2):208-17.

Paschke MW, McLendon T, Redente EF. 2000. Original articles: Nitrogen availability and old- field succession in a shortgrass steppe. Ecosystems 3(2):144-58.

Pellant M. 1996. Cheatgrass: The invader that won the west. Unpublished Report.on File with: US Department of Agriculture, Forest Service .

PLATEAU® label. 2002. BASF Corporation, Research Triangle Park, NC,USA

Polunin O. 1969. Flowers of europe. A field guide. Flowers of Europe.A Field Guide. .

Prober SM, Thiele KR, Lunt ID, Koen T. 2005. Restoring ecological function in temperate grassy woodlands: Manipulating soil nutrients, exotic annuals and native perennial grasses through carbon supplements and spring burns. J Appl Ecol 42(6):1073-85.

Prather Tim. 2009. Ventenata dubia : An increasing concern to the inland northwest region. Center for Invasive Plant Management news letter.

Ricklefs RE. 1977. Environmental heterogeneity and plant species diversity: A hypothesis. Am Nat 111(978):376-81.

Rimer R and Evans R. 2006. Invasion of downy brome ( Bromus tectorum L.) causes rapid changes in the nitrogen cycle. The American Midland Naturalist 156(2):252-8.

Robocker W. 1970. Seed characteristics and seedling emergence of Dalmatian toadflax. Weed Sci :720-5.

Robocker WC. 1974. Life history, ecology, and control of Dalmatian toadflax. Washington Agr. Exp. Sta. Tech. Bull.

Samson F and Knopf F. 1994. Prairie conservation in North America. Bioscience 44(6):418-21.

Schat M, Sing SE, Peterson RKD, Menalled FD, Weaver DK. 2011. Growth inhibition of Dalmatian toadflax, Linaria dalmatica (L.) Miller, in response to herbivory by the biological control agent Mecinus janthinus Germar. J Entomol Sci 46(3):232-46.

Scheinost P, Stannard M, Prather T. 2009. USDA NRCS Plant Guide: Ventenata dubia . Pullman, WA: Pullman Plant MaterialsCenter.

Seabloom EW, Harpole WS, Reichman O, Tilman D. 2003. Invasion, competitive dominance, and resource use by exotic and native California grassland species. Proceedings of the National Academy of Sciences 100(23):13384.

Searcy KB, Wilson BF, Fownes JH. 2003. Influence of bedrock and aspect on soils and plant distribution in the Holyoke range, Massachusetts. J Torrey Bot Soc :158-69.

Sheley R, James J, Smith B, Vasquez E. 2010. Applying ecologically based invasive-plant management. Rangeland Ecology & Management 63(6):605-13.

Sheley RL and Krueger-Mangold J. 2003. Principles for restoring invasive plant-infested rangeland. Weed Sci 51(2):260-5.

Simberloff D. 2005. Non-native species do threaten the natural environment! Journal of Agricultural and Environmental Ethics 18(6):595-607.

Skipper HD, Ogg AG,Jr., Kennedy AC. 1996. Root biology of grasses and ecology of rhizobacteria for biological control. Weed Technol 10(3):610-20.

Sperry L, Belnap J, Evans R. 2006. Bromus tectorum invasion alters nitrogen dynamics in an undisturbed arid grassland ecosystem. Ecology 87(3):603-15.

Stoddart L. 1941. The Palouse grassland association in northern Utah. Ecology 22(2):158-63.

Stohlgren TJ, Binkley D, Chong GW, Kalkhan MA, Schell LD, Bull KA, Otsuki Y, Newman G, Bashkin M, Son Y. 1999. Exotic plant species invade hot spots of native plant diversity. Ecol Monogr 69(1):25-46.

The Weather Channel, 1995-2012. The Weather Channel: Cheney monthly averages page. http://www.weather.com/weather/wxclimatology/monthly/99004 .(Accessed 6/10/2012).

Tranel PJ, Gealy DR, Kennedy AC. 1993. Inhibition of downy brome ( Bromus tectorum ) root growth by a phytotoxin from Pseudomonas fluorescens strain D7. Weed Technol 7(1):134- 9.

USA.com, 2012. USA.com: Cheney, WA. http://usa.com/cheney-wa.htm . (Accessed 6/10/2012).

Vollmer JL and Vollmer JG. 2006. Controlling cheatgrass in winter range to restore habitat and endemic fire. SG Kitchen, RL Pendleton, TA Monaco, J.Vernon [COMPS.].Proceedings-- Shrublands Under Fire: Disturbance and Recovery in a Changing World :6-8.

Washburn A. 1988. Mima mounds. Report of Investigations 29:I98S.

Weddell BJ and United States. Bureau of Land Management. Idaho State Office. 2001. Restoring Palouse and canyon grasslands: Putting back the missing pieces. Bureau of Land Management, Idaho State Office.

Wilson LM, Sing SE, Piper GL, Hansen RW, De Clerck-Floate R, MacKinnon DK, Randall CB. 2005. Biology and biological control of Dalmatian and yellow toadflax. USDA Forest Service :05-13.

Young JA and Allen FL. 1997. Cheatgrass and range science: 1930-1950. J Range Manage :530- 5.

Young JA, Evans RA, Eckert Jr RE. 1969. Population dynamics of downy brome. Weed Sci :20- 6.

Young K and Mangold J. 2008. Medusahead ( Taeniatherum caput-medusae ) outperforms squirreltail ( Elymus elymoides ) through interference and growth rate. Invasive Plant Science and Management 1(1):73-81.

Young JA and Allen FL. 1997. Cheatgrass and range science: 1930-1950. J Range Manage 50(5):530-5.

Zolkewitz MA. 2010. The use of assisted succession to restore the globally imperiled dwarf Pine-Oak forests of the New Jersey Pinelands: An investigation of the functional and structural recovery of an abandoned gravel pit.

Appendix I: Supplemental figures from NMS ordination and community analyses

Appendix 1, Table 1: Detailed species lists for plots on mounds underlain by both substrates.

BASALT ALLUVIUM Exotic Forbs x x Buglossoides arvensis (L.) I.M. Johnst x Camelina microcarpa Andrz. ex DC x Centaurea solstitialis L. x Convolvulus ssp. L. x Draba verna L. x x Holosteum umbellatum L. x Hypericum perforatum L. x Lactuca biennis (Moench) Fernald x x Lactuca serriola L. x x Linaria dalmatica (L.) Mill. x Medicago lupulina L. x x Myosotis stricta Link ex Roem. & Schult x x Potentilla argentea L. x Senecio sylvaticus L. x x Sisymbrium atlissimum L. x Taraxacum laevigatum (Willd.) DC. x x Tragopogon dubius Scop. x x Vicia cracca L.

Exotic Grasses x x Apera interrupta (L.) P. Beauv. x x Bromus arenarius Labill. x x Bromus arvensis L. x x Bromus racemosus L. x x Bromus secalinus L. x x Bromus tectorum L. x x Poa bulbosa L. x x Ventenata dubia (Leers) Coss.

Native Forbs x x Achillea millefolium L. x x Agoseris grandiflora (Nutt.) Greene x x Agoseris heterophylla (Nutt.) Greene

BASALT ALLUVIUM

Allium columbianum (Ownbey & Mingrone) x P. Peterson, Annable & Rieseberg x x Amsinckia menziesii (Lehm.) A. Nelson & J.F. Macbr. x Belpharipappus scaber Hook. x x Clarkia pulchella Pursh x x Collinsia parviflora Lindl. x x Collomia linearis Nutt. x x Delphinium nuttallianum Pritz. ex Walp x x Descurainia incana (Bernh. ex Fisch. & C.A. Mey.) Dorn x x Epilobium brachycarpum C. Presl x x Eriogonum heracleoides Nutt. x x Frittillaria pudica (Pursh) Spreng. x Gaillardia aristata Pursh x x Galium aparine L. x x Geranium viscosissimum Fisch. & C.A. Mey. ex C.A. Mey. x Geum triflorum Pursh x x Lagophylla ramosissima Nutt. x x Lithophragma parviflorum ( Hook.) Nutt. ex Torr. & A. Gray x x Lomatium ambiguum (Nutt.) J.M. Coult. & Rose x x Lomatium triternatum (Pursh) J.M. Coult. & Rose x x Lotus unifoliolatus (Hook.) Benth . x x Lupinus sericeus Pursh x x Madia gracilis (Sm.) D.D. Keck x x Microseris nutans (Hook.) Sch. Bip. x x Microsteris gracilis (Hook.) Greene x Montia linearis (Douglas ex Hook.) Greene x Perideridia gairdneri (Hook. & Arn.) Mathias x x Polemonium micranthum Benth. x x Polygonum douglasii Greene x Rosa woodsii Lindl. x Triteleia grandiflora Lindl.

Native Grasses x x Festuca idahoensis Elmer x x Pseudoroegneria spicata (Pursh) Á. Löve

Appendix I, Table 2: Comparison of exotic and native species cover and richness between substrate types using Dunnett’s T3 unequal variances pairwise comparisons.

95.0% Confidence Substrate Difference p-value Interval Lower Upper Native Cover al ba -0.242 0.098 -0.529 0.045 Total Exotic Cover al ba 0.183 0.188 -0.091 0.457 IAG Cover al ba -0.359 0.127 -0.822 0.103 Exotic Forb Cover al ba -0.065 0.772 -0.504 0.374 Native Richness al ba -0.931 0.002 -1.53 -0.333 Total Exotic Richness al ba -0.281 0.22 -0.731 0.169 IAG Richness al ba -0.068 0.581 -0.309 0.174 Exotic Forb Richness al ba -0.189 0.274 -0.529 0.15 Total Richness al ba -1.264 0.003 -2.09 -0.439

Appendix 1, Table 3: Monte Carlo test of significance of Indicator Values (IV)

Monte Carlo test of significance of observed maximum indicator value for species code 4999 permutations Random number seed: 3814 species code Max. group Value (IV) Mean S. Dev p * 1 ACHIMIL 0 26.7 11.2 2.11 0.0002 2 AGOSHET 1 14.7 8.9 1.86 0.0134 3 AMSIMEN 1 47.1 48.6 1.51 0.8894 4 BROMCOM 1 43.1 36 2.4 0.011 5 BROMTEC 1 51 48.9 1.59 0.0868 6 CLARPUL 0 31.7 14.5 2.15 0.0002 7 COLLLIN 0 66.8 36 2.42 0.0002 8 COLLPAR 0 43.6 43.8 2.15 0.4731 9 EPILBRA 1 50.7 24.2 2.62 0.0002 10 ERIOHER 0 10.1 4.6 1.39 0.0038 11 GALIAPA 1 21.2 19 2.39 0.165 12 HOLOUMB 1 31.5 31.1 2.5 0.3607 13 LACTSER 1 53.2 42.8 2.15 0.0002 14 LINADAL 0 53.3 18.9 2.32 0.0002 15 LITHPAR 1 56.7 30.4 2.47 0.0002 16 LOMAAMB 1 25.6 8 1.72 0.0002 17 LOTUNNI 1 1 1.4 0.63 0.7644 18 LUPISER 0 5.2 4 1.28 0.1824 19 MADIGRA 1 41.7 35.7 2.39 0.0214 20 MICRGRA 1 49.7 28.8 2.56 0.0002 21 MYOSSTR 0 17.8 14.6 2.26 0.0944 22 TRAGDUB 1 31.1 24.3 2.5 0.0162 23 VENTDUB 0 64.1 42.8 2.18 0.0002 24 APERINT 0 11.6 5.7 1.46 0.0044 25 BROMARV 1 13.8 10.6 2.01 0.0794 26 BUGLARV 0 11.4 8 1.72 0.0506 27 FRITPUD 1 2.1 2.7 1 0.6997 28 POLEMIC 1 18.5 18.7 2.4 0.4415 29 POLYDOU 1 37.8 26 2.49 0.0004 30 BROMARE 0 10.7 13 2.16 0.918 31 SISYALT 1 38.7 39.9 2.35 0.6221 32 DELPNUT 0 23.4 10.8 2 0.0002 33 AGOSGRA 1 13.1 11.4 2.07 0.1858 34 ALLICOL 0 3.4 2.1 0.88 0.1464 35 FESTIDA 0 5.9 4.9 1.44 0.214 55

36 PSEUSPI 1 2.7 3.1 1.11 0.5673 37 CENTSOL 0 0.8 1 0.17 1 38 POTEARG 1 1.8 1.7 0.8 0.4609 39 LAGORAM 1 35.6 25.1 2.53 0.002 40 LOMATRI 0 4 5.5 1.5 0.8796 41 LOTUUNI 1 7.7 4.2 1.27 0.0266 42 MICRNUT 0 0.8 1 0.17 1 43 SENESYL 0 3.4 2.1 0.88 0.1384 44 POA_BUL 1 15.8 6.4 1.61 0.0002 45 MEDILUP 0 0.8 1 0.17 1 46 VICICRA 1 84.4 24.5 2.44 0.0002 47 CAMEMIC 0 0.8 1 0.17 1 48 CONV-SP 0 0.8 1 0.17 1 49 BROMSEC 1 36.4 14.3 2.16 0.0002 50 DESCINC 1 0.8 1.3 0.65 0.7582 51 MONTLIN 0 12.6 5.4 1.44 0.0002 52 PERIGAI 1 4.8 2.1 0.9 0.0302 53 GAILARI 1 2.4 1.3 0.67 0.1724 54 LACTBIE 1 1.2 1 0.17 0.4031 55 TRITGRA 0 1.7 1.4 0.66 0.5373 56 TARALAE 0 0.8 1 0.17 1 57 DRABVER 0 1.7 1.3 0.65 0.5255 58 BELPSCA 0 0.8 1 0.17 1

Appendix II: ANOVA/GLM tables for combined effects of herbicide, sucrose and native seed.

Appendix II, Table 1: Change in exotic species cover due to treatments. Herbicide and an interaction between herbicide and substrate type significantly reduced exotic species cover.

Change in Dependent Variable invasive cover N 147 Multiple R 0.637 Squared Multiple R 0.406

Analysis of Variance Source Type III SS df Mean Squares F-ratio p-value SUBSTRATE 3.639 1 3.639 0.092 0.762

SUCROSE 0.142 1 0.142 0.004 0.952

HERBICIDE 2,736.52 1 2,736.52 69.222 0.000

NATIVES 21.435 1 21.435 0.542 0.463

SUBSTRATE*SUCROSE 208.073 1 208.073 5.263 0.023

SUBSTRATE*HERBICIDE 301.725 1 301.725 7.632 0.007

SUBSTRATE*NATIVES 6.678 1 6.678 0.169 0.682

SUCROSE*HERBICIDE 37.816 1 37.816 0.957 0.330

SUCROSE*NATIVES 0.02 1 0.02 0.001 0.982

HERBICIDE*NATIVES 36.639 1 36.639 0.927 0.337 SUBSTRATE*SUCROSE*HERBICIDE 115.856 1 115.856 2.931 0.089 SUBSTRATE*SUCROSE*NATIVES 0.2 1 0.2 0.005 0.943 SUBSTRATE*HERBICIDE*NATIVES 107.187 1 107.187 2.711 0.102 SUCROSE*HERBICIDE*NATIVES 14.205 1 14.205 0.359 0.550 SUBSTRATE*SUCROSE*HERBICIDE*NATIVES 95.866 1 95.866 2.425 0.122 Error 5,178.73 131 39.532

Appendix II, Table 2: Change in native species cover due to treatments. Herbicide and an interaction between herbicide and substrate significantly reduce native species cover.

Dependent Variable Change in Native cover N 147 Multiple R 0.503 Squared Multiple R 0.253

Analysis of Variance Type III Mean Source df F-ratio p-value SS Squares SUBSTRATE 1.713 1 1.713 0.104 0.747 SUCROSE 78.052 1 78.052 4.759 0.031 HERBICIDE 376.371 1 376.371 22.949 0.000 NATIVES 62.137 1 62.137 3.789 0.054 SUBSTRATE*SUCROSE 0.504 1 0.504 0.031 0.861 SUBSTRATE*HERBICIDE 67.22 1 67.22 4.099 0.045 SUBSTRATE*NATIVES 0.535 1 0.535 0.033 0.857 SUCROSE*HERBICIDE 58.563 1 58.563 3.571 0.061 SUCROSE*NATIVES 1.582 1 1.582 0.096 0.757 HERBICIDE*NATIVES 0.978 1 0.978 0.060 0.807 SUBSTRATE*SUCROSE*HERBICIDE 9.891 1 9.891 0.603 0.439 SUBSTRATE*SUCROSE*NATIVES 32.69 1 32.690 1.993 0.160 SUBSTRATE*HERBICIDE*NATIVES 12.411 1 12.411 0.757 0.386 SUCROSE*HERBICIDE*NATIVES 6.077 1 6.077 0.371 0.544 SUBSTRATE*SUCROSE*HERBICIDE*NATIVES 1.936 1 1.936 0.118 0.732 Error 2,148.46 131 16.400

Appendix II, Table 3: Change in IAG stem counts due to treatments. Herbicide and an interaction between herbicide and substrate type significantly reduced the stem counts of invasive grasses.

Change in total stem Dependent Variable count N 138 Multiple R 0.602 Squared Multiple R 0.363

Analysis of Variance Mean Source Type III SS df F-ratio p-value Squares SUBSTRATE 5,979.48 1 5,979.48 1.904 0.170 SUCROSE 66.762 1 66.762 0.021 0.884 HERBICIDE 204,403.93 1 204,403.93 65.094 0.000 NATIVES 1,129.76 1 1,129.76 0.360 0.550 SUCROSE*SUBSTRATE 235.104 1 235.104 0.075 0.785 HERBICIDE*SUBSTRATE 13,698.73 1 13,698.73 4.362 0.039 NATIVES*SUBSTRATE 610.323 1 610.323 0.194 0.660 HERBICIDE*SUCROSE*SUBSTRATE 7.069 1 7.069 0.002 0.962 NATIVES*HERBICIDE*SUBSTRATE 183.723 1 183 .723 0.059 0.809 NATIVES*HERBICIDE*SUCROSE*SUBSTRATE 2,114.63 1 2,114.63 0.673 0.413 Error 398,800.03 127 3,140.16

Appendix II, Table 4: Change in percent cover of IAGs due to treatments. Herbicide and an interaction between herbicide and substrate type significantly reduced invasive grass cover.

Dependent Variable Change in IAG percent cover N 520 Multiple R 0.339 Squared Multiple R 0.115

Analysis of Variance Source Type III SS df Mean F-ratio p- Squares value SUBSTRATE 521.176 1 521.176 1.682 0.195 SUCROSE 217.899 1 217.899 0.703 0.402 HERBICDE 16,826.81 1 16,826.81 54.315 0.000 NATIVES 122.578 1 122.578 0.396 0.530 SUCROSE*SUBSTRATE 119.586 1 119.586 0.386 0.535 HERBICDE*SUBSTRATE 1,667.20 1 1,667.20 5.381 0.021 NATIVES*SUBSTRATE 4.298 1 4.298 0.014 0.906 HERBICDE*SUCROSE*SUBSTRATE 366.085 1 366.085 1.182 0.278 NATIVES*HERBICDE*SUBSTRATE 67.726 1 67.726 0.219 0.640 NATIVES*HERBICDE*SUCROSE* 357.058 1 357.058 1.153 0.284 SUBSTRATE Error 157,689.13 509 309.802

Appendix II, Table 5: Change in total species richness due to treatments. Substrate type, herbicide and an interaction between the two, significantly reduced species richness.

Total Change in species Dependent Variable richness N 136 Multiple R 0.832 Squared Multiple R 0.692

Analysis of Variance Mean p- Source Type III SS df F-ratio Squares value 143.95 SUBSTRATE 1,646.81 1 1,646.81 0.000 3 SUCROSE 0.527 1 0.527 0.046 0.830 HERBICIDE 904.941 1 904.941 79.103 0.000 NATIVES 0.263 1 0.263 0.023 0.880 SUCROSE*SUBSTRATE 12.504 1 12.504 1.093 0.298 SUBSTRATE*HERBICIDE 376.996 1 376.996 32.954 0.000 SUBSTRATE*NATIVES 8.247 1 8.247 0.721 0.397 HERBICIDE*SUCROSE*SUBSTRATE 0.99 1 0.99 0.087 0.769 NATIVES*HERBICIDE*SUBSTRATE 0.473 1 0.473 0.041 0.839 SUBSTRATE*NATIVES*HERBICIDE 0.848 1 0.848 0.074 0.786 *SUCROSE Error 1,430.00 125 11.44

Appendix II, Table 6: Change in exotic species richness due to treatments. Substrate type, herbicide and an interaction between the two, significantly reduced species richness.

Dependent Variable Change in exotic richness N 137 Multiple R 0.774 Squared Multiple R 0.599

Analysis of Variance p- Source Type III SS df Mean Squares F-ratio value SUBSTRATE 183.42 1 183.42 50.562 0.000 SUCROSE 0.02 1 0.02 0.005 0.941 HERBICIDE 282.202 1 282.202 77.792 0.000 NATIVES 0.001 1 0.001 0.000 0.985 SUCROSE*SUBSTRATE 3.016 1 3.016 0.831 0.364 SUBSTRATE*HERBICIDE 141.205 1 141.205 38.925 0.000 SUBSTRATE*NATIVES 1.4 1 1.400 0.386 0.536 HERBICIDE*SUCROSE*SUBSTRATE 0.348 1 0.348 0.096 0.757 NATIVES*HERBICIDE*SUBSTRATE 0.139 1 0.139 0.038 0.845 SUBSTRATE*NATIVES*HERBICIDE 0.07 1 0.070 0.019 0.890 *SUCROSE Error 457.084 126 3.628

Appendix II, Table 7: Change in native species richness due to treatments. Substrate type, herbicide and an interaction between the two, significantly reduced species richness.

Change in native Dependent Variable richness N 137 Multiple R 0.813 Squared Multiple R 0.662

Analysis of Variance Mean Source Type III SS df F-ratio p-value Squares SUBSTRATE 740.384 1 740.384 169.587 0.000 SUCROSE 0.234 1 0.234 0.053 0.817 HERBICIDE 183.03 1 183.03 41.924 0.000 NATIVES 0.197 1 0.197 0.045 0.832 SUCROSE*SUBSTRATE 2.942 1 2.942 0.674 0.413 SUBSTRATE*HERBICIDE 59.098 1 59.098 13.537 0.000 SUBSTRATE*NATIVES 2.541 1 2.541 0.582 0.447 HERBICIDE*SUCROSE*SUBSTRATE 2.248 1 2.248 0.515 0.474 NATIVES*HERBICIDE*SUBSTRATE 1.296 1 1.296 0.297 0.587 SUBSTRATE*NATIVES*HERBICIDE* 0.588 1 0.588 0.135 0.714 SUCROSE 12 Error 550.09 4.366 6

Appendix III: Change in abundance of L. dalmatica due to addition of

Mecinus janthiniformis

Appendix III, Table 1: GLM analysis of change in percent cover of Linaria dalmatica between control, insect addition (I), native seed addition (N) and sucrose addition (S).

TREATMENT (5 levels) Control Insect (I) I, Natives(N) I,S,H I,S,H,N Dependent Variable = Change in% cover N = 27 Multiple R = 0.496 Squared Multiple R = 0.246 Analysis of Variance F- Source Type III SS df Mean Squares ratio p-value TREATMENT 10,415.04 4 2,603.76 1.794 0.166 Error 31,922.98 22 1,451.04

Appendix III, Table 2: GLM analysis of change in stem count of Linaria dalmatica between control, insect addition (I), native seed addition (N) and sucrose addition (S).

TREATMENT (5 Levels) Control I I,N I,S,H I,S,H,N

4 Cases are deleted due to missing data

Dependent Variable = Change in Stem Count N = 27 Multiple R Squared 0.609 Multiple R Squared 0.371

Analysis of Variance Source Type III SS df Mean Squares F-ratio p-value TREATMENT 16,143.67 4 4,035.92 3.242 0.031 Error 27,385.00 22 1,244.77

Appendix IV: Change in IAG abundance due to Pseudomonas fluorescens D7

Appendix IV, Table 1: Change in total abundance of IAGs due to P.f. D7. There were two levels of treatment, not added (N), and added (Y).

Variables Levels P.f. D7 (2 levels) N Y

Dependent Change in percent Variable cover N 61 Multiple R 0.005 Squared Multiple R 0

Analysis of Variance Mean Source Type III SS df F-ratio p-value Squares P.f D7 2.14 1 2.14 0.001 0.972 Error 99,116.12 59 1,679.93

Appendix IV, Table 2: Change in abundance of Bromus spp. due to P.f. D7.

Variables Levels P.f. D7 (2 levels) N Y

Dependent Change in cover Variable N 61 Multiple R 0.005 Squared Multiple R 0

Analysis of Variance Mean Source Type III SS df F-ratio p-value Squares RHIZO 2.14 1 2.14 0.001 0.972 Error 99,116.12 59 1,679.93

Brandy K. Reynecke Graduate Student, Plant Ecology Lab, Biology Department, Eastern Washington University, Cheney, Washington

VITA

Author: Brandy Kathleen Reynecke

Place of Birth: Fontana, California Undergraduate Schools Attended: North Idaho Community College University of Idaho Eastern Washington University Graduate Schools Attended: Eastern Washington University

Degrees Awarded 2012 M.S . Eastern Washington University, Plant Ecology Lab, Department of Biology, Thesis: Plant community restoration at Turnbull National Wildlife Refuge. (Advisor: Dr. Rebecca L. Brown)

2009 B.S. Eastern Washington University, Biology with an emphasis in Botany, cum laude

Professional Experience 2009-2012 Graduate Fellowship, Department of Biology, Eastern Washington University, Cheney, WA

Teaching Experience 2012 spring Teaching Assistantship, Honors Biology Lab (HONS104) (with Dr. Justin Bastow), Eastern Washington University, Cheney, WA

2009-2011 Teaching Assistantship, Botany (BIOL302) (with Dr. Rebecca Brown and Dr. Suzanne Schwab), Eastern Washington University, WA

2009-2011 Teaching Assistantship, Field Botany (BIOL411) (with Dr. Rebecca Brown and Dr. Suzanne Schwab), Eastern Washington University, Cheney, WA

2010 fall Teaching Assistantship, Biology I - Discussion Sessions 01 and 03 (BIOL171) (With Dr. Robin O’Quinn) Eastern Washington University, Cheney, WA

2011 winter Teaching Assistantship, Biology II- Discussion Sessions 01 and 02 (BIOL172) (With Dr. Robin O’Quinn) Eastern Washington University, Cheney, WA

Research Experience 2009—2012 “Plant community restoration at Turnbull National Wildlife Refuge”, Plant Ecology Lab, Department of Biology, Eastern Washington University, Cheney, WA

Publications Bryant JAM, Reynecke BK , and Brown RL. Grazing history influences exotic plant distribution differently depending on soil parent material in a semi-arid Mima mound prairie. Journal of Arid Environments. Submitted August 2011 based on undergraduate research Comment [R1]: Would I put scholarships and things like pell grants? Funding and Grants 2009 Eastern Washington University mini grant (granted $400)

Presentations Reynecke BK and Brown RL. 2012. Approaches for Restoration of Native Plant Communities on Mima Mounds at Turnbull National Wildlife Refuge (TNWR), Society for Range Management 46 th Annual Youth Forum Range and Ecology Tour, Turnbull National Wildlife Refuge, WA

Reynecke BK and Brown RL. 2011. (Poster) Control methods for invasive grasses on Mima mounds at Turnbull National Wildlife Refuge, WA, Ecological Society of America 96 th Annual Meeting, Austin, TX

Reynecke BK. 2011. Approaches for Restoration of Native Plant Communities on Mima Mounds at Turnbull National Wildlife Refuge (TNWR), North-East Chapter Washington Native Plant Society, Spokane, WA

Reynecke BK, Brown RL. 2011. (Poster) Examining the effects of invasive annual grass control methods on Mima mounds across two geological substrates at Turnbull National Wildlife Refuge, WA Columbia Basin Landscapes: Linking Science and Management to Improve Restoration Success in the Shrub Steppe Workshop, Kennewick, WA

Reynecke BK. 2011. Examining the effects of invasive annual grass control methods on Mima mounds at Turnbull National Wildlife Refuge, WA, Student Research and Creative Works Symposium, Eastern Washington University, Cheney, WA

Bryant J, Brown RL , Reynecke B. 2010. Exotic and native plant distribution on the Mima Mounds of eastern Washington , Society for Ecological Restoration Northwest, Marysville, WA

Professional Memberships Washington Native Plant Society: Northeast Chapter Society for Range Management Ecological Society of America