Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Interacting effects of climate change and disturbance on grassland plants and plant communities Carlyle, Cameron Norman 2012

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2012_fall_cameron_carlyle.pdf [ 1.68MB ]
Metadata
JSON: 24-1.0072773.json
JSON-LD: 24-1.0072773-ld.json
RDF/XML (Pretty): 24-1.0072773-rdf.xml
RDF/JSON: 24-1.0072773-rdf.json
Turtle: 24-1.0072773-turtle.txt
N-Triples: 24-1.0072773-rdf-ntriples.txt
Original Record: 24-1.0072773-source.json
Full Text
24-1.0072773-fulltext.txt
Citation
24-1.0072773.ris

Full Text

     INTERACTING EFFECTS OF CLIMATE CHANGE AND DISTURBANCE ON GRASSLAND PLANTS AND PLANT COMMUNITIES   by   CAMERON NORMAN CARLYLE  B.Sc., The University of British Columbia, 1999 B.Sc., Royal Roads University, 2000 M.Sc., The University of Akron, 2004    A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF     DOCTOR OF PHILOSOPHY  in  THE FACULTY OF GRADUATE STUDIES  (Botany)     THE UNIVERSITY OF BRITISH COLUMBIA  (Vancouver)    April 2012   © Cameron Norman Carlyle, 2012 Abstract Grasslands are threatened by urbanization, agricultural conversion, over-grazing, tree- encroachment, and invasive plants. Simultaneously, climate change acts on all levels of biological organization, from entire communities to the physiology of individuals. The environmental stresses induced by climate change have the potential to interact with human- caused disturbance, but the response of plants to these stresses and disturbances, and how they may interact, are not well known. To conserve grasslands it is critical to know which types of grassland and which plant species will be most affected. To understand the mechanisms of change at the ecosystem level it is necessary to study the response at lower levels of biological organization. Using a variety of approaches I studied the potentially interacting effects of stress (primarily reduced water availability) and disturbance (plant biomass removal) on different levels of biological organization.  I ran a 4-year field experiment in which I manipulated water availability, temperature and clipping in three different grassland types. I found complex plant community structure and biomass response; treatments often interacted but the different grassland types had their own particular responses.  As part of this experiment I monitored the effects of treatments on soil moisture and temperature and found that the effects are generally consistent with expectations, but treatments do not act exclusively or independently on target variables. In addition to stress and disturbance, competition is a key process structuring grasslands. In the greenhouse, I examined how plant competition is affected by stress and disturbance. I found that the interpretation of how competition is affected is dependent on the way competition is measured.  Some measures of competition showed reduced competition across stress and disturbance gradients, but other measures showed no change. Finally, I examined the root traits of 18 species of grass in the greenhouse in response to reduced water availability.  I found significant variation in traits among species, maintenance of trait hierarchies across environments and little evidence of plasticity, except for root: shoot ratio. Overall, stress, disturbance and their interactions are important in influencing individual plant performance, competition, structuring plant communities, and ecosystem function.  ii Preface Chapters 2, 3, 4 and 5 were written by me, and edited by Lauchlan H. Fraser and Roy Turkington. The direction of research was a result of discussions with LHF and RT; chapter 4 is the result of discussions on an essay assignment in RT’s plant ecology course. I designed the experiments with input from LHF and RT. I conducted the research, collected the data, performed the analyses and wrote the manuscripts.  LHF and RT advised on data analysis, and assisted with manuscript editing and revision.  A version of  chapter 3 was published as: Carlyle CN, Fraser LH, Turkington R. 2011. Tracking soil temperature and moisture in a multi– factor climate experiment in temperate grassland: do climate manipulation methods produce their intended effects? Ecosystems 14: 489 – 502.  A version of chapter 4 was published as: Carlyle CN, Fraser LH, Turkington R. 2010. Using three pairs of competitive indices to test for changes in plant competition under different resource and disturbance levels. Journal of Vegetation Science 21: 1025 – 1034.         iii Table of contents  Abstract ......................................................................................................................................... ii  Preface .......................................................................................................................................... iii  Table of contents .......................................................................................................................... iv  List of tables ................................................................................................................................ vii  List of figures ............................................................................................................................... ix  List of abbreviations .................................................................................................................... xi  Acknowledgements ..................................................................................................................... xii  Defining stress and disturbance.............................................................................................................1 Chapter 1: Introduction............................................................................................................... 1  The role of interactions in ecology.........................................................................................................2   Stress – increased temperature and reduced precipitation.................................................................2   Disturbance – grazing.............................................................................................................................3   The influence of productivity on competition and community structure ..........................................3   Resistance and resilience ........................................................................................................................4   Grasslands and gradients.......................................................................................................................5   Hierarchy of response variables examined...........................................................................................5     Thesis overview .......................................................................................................................................6  Chapter 2: Interacting effects of stress and disturbance on grassland community structure and biomass production along a natural productivity gradient............................................... 8  Methods .................................................................................................................................................10 Introduction.............................................................................................................................................8  Site description ...................................................................................................................................10   Experimental design ...........................................................................................................................11   Indices used ........................................................................................................................................12   Data analyses ......................................................................................................................................13   Results ....................................................................................................................................................14   Sites ....................................................................................................................................................14   Response of standing biomass and cover to treatments .....................................................................14   Response of groups: dominants, subordinate graminoids and forbs ..................................................15   Response of litter to treatments ..........................................................................................................16   Plant community response..................................................................................................................16   Discussion ..............................................................................................................................................17   Biomass response ...............................................................................................................................18   Unexpected biomass response to water treatments ............................................................................18   Community response ..........................................................................................................................20   Interactions .........................................................................................................................................21   Dominance..........................................................................................................................................21   Change along the gradient ..................................................................................................................21     Summary ............................................................................................................................................22  Chapter 3: Tracking soil temperature and moisture in a multi-factor climate experiment in temperate grassland:  do climate manipulation methods produce their intended effects? . 33  Introduction...........................................................................................................................................33   Methods .................................................................................................................................................35  iv     Site description ...................................................................................................................................35  Experimental design ...........................................................................................................................35  Water manipulations...........................................................................................................................36 Open-top chambers.............................................................................................................................36  Soil measurements..............................................................................................................................37   Statistical analyses..............................................................................................................................37   Results ....................................................................................................................................................38   Site......................................................................................................................................................38   Soil temperature .................................................................................................................................38   Temporal effects on temperature........................................................................................................39   Weather effects on temperature..........................................................................................................39   Soil moisture.......................................................................................................................................40   Weather effects on moisture ...............................................................................................................41   Discussion ..............................................................................................................................................41   Temperature........................................................................................................................................41   Soil moisture.......................................................................................................................................43     Experimental design recommendations..............................................................................................45  Chapter 4: Using three pairs of competitive indices to test for changes in plant competition under different resource and disturbance levels...................................................................... 57  Materials and methods .........................................................................................................................59 Introduction...........................................................................................................................................57  Study species ......................................................................................................................................59   Experimental design ...........................................................................................................................59   Competitive indices ............................................................................................................................60   Data analysis.......................................................................................................................................61   Results ....................................................................................................................................................61   Biomass ..............................................................................................................................................61   Competitive indices ............................................................................................................................61   Discussion ..............................................................................................................................................62   Interacting effects on biomass ............................................................................................................62   Comparing the pairs of competitive indices .......................................................................................63   Absolute competition and relative competition..................................................................................64   Competitive effect and competitive response.....................................................................................64   Competitive intensity and competitive importance ............................................................................65   Conclusions ........................................................................................................................................67   Chapter 5: Variation in root plasticity of 18 temperate grass species to water availability 74   Introduction...........................................................................................................................................74   Methods .................................................................................................................................................75   Study species ......................................................................................................................................75   Relative growth rate ...........................................................................................................................75   Root tubes...........................................................................................................................................76   Greenhouse conditions .......................................................................................................................76   Data analysis.......................................................................................................................................76   Results ....................................................................................................................................................77   Relative growth rates..........................................................................................................................77   Species variation and rankings ...........................................................................................................77   Trait plasticity.....................................................................................................................................78   Predicting species response ................................................................................................................78   Discussion ..............................................................................................................................................78   Variation among species.....................................................................................................................79   Species rankings are maintained across treatments ............................................................................79   Plasticity .............................................................................................................................................80  v     Predicting species response ................................................................................................................81  Summary ............................................................................................................................................82  Summary of thesis.................................................................................................................................93 Chapter 6: Conclusion................................................................................................................ 93  Field experiment: the effects of stress and disturbance on grassland plant communities ..................93   Measuring competition across multiple gradients ..............................................................................96   Rooting depth .....................................................................................................................................96   Synthesis and lessons learned ..............................................................................................................96   How we measure is important ............................................................................................................96   Controlled experiments in the greenhouse can inform field experiments ..........................................97   Comments on running multi-factor experiments................................................................................98   Summary: The role of stress and disturbance on plants and plant communities...........................99     References.................................................................................................................................. 100  Appendix A: Information on common grasses in Lac du Bois Provincial Park................. 113  Introduction.........................................................................................................................................114 Appendix B: Calibration of soil moisture probes .................................................................. 114  Methods ...............................................................................................................................................114     Results ..................................................................................................................................................115    vi List of tables Table 2.1 Background data for the six experimental field sites...............................................23  Table 2.2 ANOVA results: F-values for significant factor effects on total biomass, species groups mass and litter mass in the three grassland types. .............................24  Table 2.3 ANOVA results: F-values for significant factor effects on species richness, diversity and evenness in the three grassland types ..................................................25  Table 2.4 Permutational multivariate analysis of variance results testing factor effects on community composition in the three grassland types ..............................................26  Table 3.1 Repeated measure ANOVA testing the effect of treatments on mean daily temperature ......................................................................................................................46  Table 3.2 Repeated measure ANOVA testing effects on mean hourly temperature and mean daily temperature......................................................................................................47  Table 3.3 Repeated measure ANOVA of effects on soil temperature and moisture and their variance......................................................................................................................48  Table 3.4 Repeated measure ANOVA testing the effects of treatments on mean daily soil moisture, daily maximum moisture, daily minimum moisture and variance of moisture .................................................................................................................49  Table 4.1 Indices of competition .............................................................................................68  Table 4.2 Summary of four-way ANOVAs testing treatment effects on the above ground biomass of Pseudoroegneria spicata and Festuca campestris.....................................69   vii Table 4.3  Summary of three-way ANOVAs for Pseudoroegneria spicata and Festuca campestris across treatment combinations for all 5 measures of competition ...............................................................................................................................70  Table 5.1 The relative growth rate of the 18 grass species ......................................................83  Table 5.2  Summary of ANOVA testing for differences among species on the variables total biomass, shoot biomass, root biomass, root: shoot ratio, and maximum depth of root growth ................................................................................................84  Table 5.3 Total biomass, shoot biomass, root biomass, root: shoot ratio and root depth of 18 species of grass in watered and drying tubes.........................................................85  Table 5.4  Summary of F-values from repeated measure ANOVAs testing the effect of water treatments on root biomass down the depth of the tube ...................................87  Table 5.5 The rankings of the 18 species for total biomass, shoot biomass, root biomass, root: shoot ratio and maximum rooting depth in drying and watered tubes..........................................................................................................................................88  Table A.1 Information on 18 common grass species found in Lac du Bois Provincial Park, British Columbia Canada.............................................................................113  viii List of figures Figure 2.1 Total plant biomass and biomass of three functional groups harvested at the end of the experiment, in each of the three grassland types ...........................................27  Figure 2.2 Vegetative cover, by treatment, over four years in three grassland types..........................................................................................................................................29  Figure 2.3 Species richness, diversity and evenness, based on biomass harvested in 2008, in the three grassland types.........................................................................................30  Figure 2.4 Non-metric multi-dimensional scaling plots, run separately for each grassland type ...........................................................................................................................31  Figure 2.5 Mean Bray-Curtis dissimilarity among plots, by treatment, over four years in three grassland types ...................................................................................................32  Figure 3.1 Photos of experimental site in Lac du Bois Provincial Park, Canada, an open-top chamber and a rainout shelter, and open-top chamber ..............................................50  Figure 3.2 Mean daily air temperature and daily rainfall totals...............................................51  Figure 3.3 Mean soil temperature and soil moisture in all twelve treatment combinations.............................................................................................................................52  Figure 3.4 Mean hourly soil temperatures measured at 5 cm depth in all treatments..................................................................................................................................53  Figure 3.5  Mean difference in soil temperature and moisture and their respective variances in the different treatments compared to the control plots.........................................54   ix Figure 3.6 Mean soil temperature in plots with and without open-top chambers in different ambient conditions .....................................................................................................55  Figure 3.7  Mean daily soil volumetric water content (% VWC), measured in the top 10 cm of each plot and mean daily rainfall for the same days ...........................................56  Figure 4.1 Mean biomass of Pseudoroegneria spicata and Festuca campestris under all treatment combinations..............................................................................................71  Figure 4.2 The five competition indices for Pseudoroegneria spicata and Festuca campestris for all treatment combinations................................................................................72  Figure 5.1 Shoot biomass, and depth profile of root growth for 18 grass species...................89  Figure 5.2 Shoot mass plasticity plotted against root mass plasticity and rooting depth plasticity..........................................................................................................................90  Figure 5.3 Relationship between the difference in root: shoot ratio between drying and watered tubes and the root: shoot ratio in watered tubes........................................91  Figure 5.4 Relationships between root: shoot ratio, and root mass plasticity with relative growth rate ...................................................................................................................92  Figure B.1 Actual volumetric water content plotted against the measured volumetric water content ........................................................................................................116    x List of abbreviations BC = British Columbia Cab = absolute competition Ce = competitive effect Cimp = competitive importance Cint = competitive intensity Cr = competitive response OTC = open-top chamber RGR = relative growth rate RS = rainout shelter r:s = root: shoot ratio based on biomass   xi  xii Acknowledgements This thesis represents many years of work, none of which I could have done without the help, encouragement and support of many people. My two supervisors, Lauch Fraser and Roy Turkington, have made me a better ecologist and have been instrumental in all of my academic successes. Roy’s insight, reviews and criticisms have been invaluable and he helped me become a better writer. I have worked with Lauch for 10 years (10 years!) and am grateful to have had him as a mentor. I look forward to being colleagues with both for years to come. My parents, Don and Elaine Carlyle, and sister, Melanie, have given unwavering support through every step of my education and I cannot thank them enough. When I first met Tamara McKinnon, I told her I was going to be a doctor – I don’t think she really knew what she was getting into. But, she stuck with me through the end of my thesis and I am thankful for her continuing support and encouragement. Lab-mates at TRU and UBC helped create a fantastic atmosphere for research and have become good friends and colleagues. In particular, Bill Harrower, Amber Greenall, Montana Burgess, Eleanor Bassett, Percy Folkard, Ashleigh Gilbert, Justine McCulloch, Anna Sapoznikova, Sadie Cox and Amanda Schmidt have provided encouragement and have been a pleasure to share a lab with. Brandy Ludwig, Amber Greenall, Montana Burgess, Eleanor Bassett, Lisa DeSandoli, Jessica Gosling, Amy Bitz, Anna-Marie Pellet and Tessa Jongbloets all provided invaluable assistance in the field, lab or greenhouse. My committee and proposal examiners, Gary Bradfield, Judy Myers, Greg Henry and Diane Srivastava helped start, and keep, my research on a better path. I am thankful to Veronica Oxtoby, Judy Heyes and Lebby Balakshin who all put in extra effort to help me navigate the administration of my PhD from afar. Don Thompson, at Agriculture and Agri-food Canada, provided extensive help with initial site selection and permissions to use the exclosures. Tod Haughton, at the BC Ministry of Environment, helped with permits to allow access and research in Lac du Bois Provincial Park. I was supported by an NSERC IPS scholarship in partnership with the British Columbia Grasslands Conservation Council, the BC Forest Science Program, a BC Pacific Century Scholarship and a University of British Columbia Graduate Fellowship. Research was supported with an NSERC discovery grant, a Canadian Foundation for Innovation grant, BC Knowledge Development Fund grant and BC Forest Science Program Grant to Lauch Fraser. Chapter 1: Introduction Grasslands are globally important ecosystems and cover approximately 40% of the terrestrial surface. However, grasslands are one of the most altered ecosystems as most have been converted to agricultural use. We rely on existing grassland for forage, to support livestock production, and they are vital for biodiversity. In British Columbia (BC), grasslands cover less than 1% of the land yet 9 of BC’s 28 biodiversity hotspots (Gayton 2003) are in grassland and they are habitat for over 200 of BC’s endangered or threatened species (BC Conservation Data Centre 2011). They also support BC’s cattle ranching industry. Understanding the dynamics of grassland function and biodiversity is critical to their long-term conservation and management. Grasslands are threatened by urbanization, agricultural conversion, over-grazing, tree- encroachment, and invasive plants.  Global climate change is an emerging threat to biological communities (McCarty 2001, Solomon et al. 2007); it is altering the natural distribution of species (Root et al. 2003, Parmesan and Yohe 2004) and is increasing the risk of species extinction at a global scale (Thomas et al. 2004).  Experiments have shown that simulated climate warming can reduce grassland plant species diversity with significant grazing interactions (Klein et al. 2004), and alter community composition and functional group abundance (Zavaleta et al. 2003).  Climate change can also have effects on plant populations (Bradley et al. 1999); such as flowering date and growing season (Keeling et al. 1996, Myneni et al. 1997), range effects (Grabherr et al. 1994), and sex and habitat-specific species response (Jones et al. 1999). However, the interaction of climate change with existing processes is not well studied especially in regard to plant biodiversity and community dynamics. The primary objective of the work described here is to investigate how climate change will interact with disturbance to alter grassland plants and grassland communities. Defining stress and disturbance Stress and disturbance, acting along with competition, are the most important factors influencing plant communities, populations and individuals. I will use the definitions provided by Grime (2001): stress is any process that limits the production of biomass by plants and disturbance is any process that removes or destroys plant biomass.  Stress and disturbance can take many forms. Stress can be due to toxins, low light, extreme temperature or a lack of resources. Disturbances can be the result of herbivores, wind or freezing. Here, I examine stress in the context of climate change primarily as altered water availability (but also increased  1 temperature) but apply a more general approach to disturbance in field and greenhouse experiments. Furthermore, stress and disturbance are mechanisms that can influence competition between plants so I will also examine competition in this context. The role of interactions in ecology Potential interactions between ecological processes may be important in determining the effect on the response variable of interest.  In natural systems interactions between various processes are always occurring. All possible factors are potentially present, but rarely occur in a manner useful for investigating their effects on natural systems because they are spatially and temporally unpredictable, not replicated and the magnitude of each factor is not controlled. In the context of a controlled and replicated experiment, interactions can occur when more than one treatment factor is imposed on a sample unit. Interactions between factors may increase or decrease the effect, or have no effect, and they sometimes change the direction of the effect. Yet despite their importance, interactions are commonly noted as being understudied especially across ecosystem types (Smith et al. 2009).  The reason for this is three fold: whether observational or experimental it requires more work and funds to study more factors, increasing the number of factors in a study necessitates an increase in replication due to issues of statistical power and interpreting the results of these studies is often difficult because results can be complex. This is especially true if interactions are present rather than two factors simply being additive – as is the case of some results described in later chapters. Despite these obstacles studies addressing multiple factors must be done to understand the full potential effect of climate change on ecosystems in the future and to increase our understanding of ecological systems. Stress – increased temperature and reduced precipitation As noted, climate change will have a range of effects on biological systems.  The primary mechanisms for these changes in terrestrial systems are warming and altered precipitation patterns.  Global climate models predict a mean temperature rise of between 1 and 5.5oC (Watson et al., 2001).  In the interior of BC, models predict that by the year 2100 temperature will increase by at least 1C and as high as 7C, although 4C is more likely. Over the same period it is projected that precipitation will decline 25% during the summer and increase 25% during the winter (models were run using the Canadian Institute for Climate Studies at UVIC web page, http://www.cics.uvic.ca/scenarios/index.cgi).  I have focused on these two factors as  2 stresses because of their dominance as drivers for climate change-induced alteration of ecosystems, but also because of grassland ecosystems are determined by precipitation levels. Disturbance – grazing Over half of the terrestrial land mass is grazed by livestock (Havstad et al. 2008). Herbivory by cattle is a disturbance to plant communities that can affect plant communities directly by a loss of plant biomass, selective feeding or seed dispersal.  Indirect effects on the community also occur by competition with other herbivores, effects on soils and nutrient cycling or alterations of plant competitive relationships.  Herbivory frequently promotes higher diversity in plant communities by removing palatable and dominant competitive species; however, excessive levels of herbivory can reduce diversity (Crawley 1983).  There are practical limitations of combining climate manipulation devices and live cattle, instead I used a clipping treatment throughout my research as the disturbance.  Consequently, when compared to cattle grazing, the treatments I applied are more general, lack the selectivity of cattle and are more uniformly applied across species and individuals. The influence of productivity on competition and community structure  Empirical studies have revealed two consistent, although debated, patterns that occur along productivity gradients in herbaceous communities.  First, species richness has a unimodal relationship with productivity and, second, the role of competition changes along a productivity gradient.   Consequently, productivity is a good basis to categorize grasslands to examine differential response. The humpback model of plant species richness plotted against productivity (the sum of live biomass and litter) predicts species richness is maximal at an intermediate level of productivity and less at low or high productivity (Grime 2001, but see Adler et al. 2011). In light of current global threats to grasslands and biodiversity in general (Erhlich and Wilson 1991), it is important to understand the factors that create or maintain diversity.  The primary mechanism behind this pattern are the evolved characters of plants.  At low productivity, a stressed environment, only the few plants adapted to tolerate low resource levels will survive. At high productivity, faster-growing species will outcompete other species. Thus, understanding the role of competition along productivity gradients is critical when examining the factors that may alter community composition.  3 Competition plays an important role in structuring communities (Connell 1983, Schoener 1983, Goldberg and Barton 1993, Keddy 2001). But, there has been extensive debate (Brooker et al. 2005, Craine 2005) concerning the role competition plays along productivity gradients and therefore how competition affects species diversity. Two conflicting theories in plant ecology both successfully make predictions about vegetation along productivity gradients, despite opposing views at a fundamental level.  Grime (1979) predicts that the intensity of competition increases with an increase in productivity, while Newman (1973) and Tilman (1988) predict that the intensity of competition is constant along a productivity gradient. It should be noted a third pattern of decreasing competition with increasing productivity has also been observed (Goldberg et al. 1999). Both Grime’s and Newman/Tilman’s theories attempt to identify the strategies evolved by plants to adapt to their environment and then use these evolved strategies to explain the structure and composition of plant communities.  Grime simplifies the environment to three variables, stress, competition and disturbance that together determine plant characteristics. In contrast, Tilman uses an approach where resources levels and supply rates, along with loss rate (disturbance), predict plant coexistence. It is important that competition is addressed in this research because it is the third component of Grime’s CSR theory (disturbance and stress being the other two) and it is also integral to Tilman’s theory.  Understanding of the processes controlling community composition will be enhanced if we have some knowledge of the role of competition in the system. Resistance and resilience The stability of an ecosystem can be assessed in terms of its resistance and resilience to a perturbation (either stress or disturbance).  Resistance refers to a community’s ability to withstand a perturbation, while resilience refers to the  ability of a community to recover after a perturbation (Pimm 1984, McCann 2000).  Resistance and resilience are a good measure of ecosystem health from a conservation and management viewpoint (Whitford et al.  1999). The stability of a system is often measured in terms of species diversity or biomass production. Many studies examining the relationship between diversity, stability and productivity often have productivity as the dependent variable (as an ecosystem function resulting from diversity) and report that stability increases with increasing diversity (Tilman et al. 1996, Pfisterer and Schmid 2002, Hector et al. 1999, Huston et al. 2000).  However, there has been debate as to whether diversity influences productivity or vice-versa and whether diversity increases or reduces stability (Tilman 1999).  While the relationships between diversity and stability have been  4 studied and debated the relationship between productivity and stability has received less attention.  It is predicted that productive systems will be less resistant to disturbance than unproductive systems because productive grasslands will contain fast-growing plants that are susceptible to disturbance.  However, productive grasslands should be more resilient because the community will be comprised of more fast-growing plants that can re-grow quickly (Grime 2001, page 335, but see Stone et al. 1996).  Compositional differences of the grassland result in this apparent tradeoff, between resistance and resilience, and are due to the traits of the species in the system. This research only examines resistance to environmental change because resilience in grasslands likely needs to be measured on a decadal scale. Grasslands and gradients The Lac du Bois Grasslands Provincial Park near Kamloops, BC, is an ideal location for addressing the questions posed in this research. Grasslands are relatively simple systems when compared to the structure and diversity of most other systems, and because of their relative simplicity it is easier to manipulate and measure both their biotic and abiotic components. Furthermore, the grassland field sites used in my research span a productivity gradient that correlates with elevation, temperature and precipitation. Gradients arrange environmental variation making it easier to study the relationships between biotic and abiotic factors (Whittaker 1967, Keddy 1991). When an ecosystem stretches along a gradient, one can examine the differential responses of communities and species within the system to disturbance, stress or competition at different locations along the gradient. Artificial gradients can also be created in greenhouse experiments by manipulating resource availability. Hierarchy of response variables examined Climate change acts on all levels of biological organization, from the physiology of individual plants, to ecosystem-level processes. Consequently, if a goal involves understanding an ecosystem-level response it is necessary to start at a smaller scale and link upward in order to successfully establish mechanisms and predict patterns (Levine 1992). My research examines the effects of climate change and disturbance across a connected range of variables that are intended to inform pattern at higher levels of organization and provide insight on the mechanisms of change at the level of the whole plant community. At the community level I will examine the response of community structure and ecosystem function (biomass production). Stress and disturbance also affect the way plants interact; competition will be examined in the greenhouse.  5 Finally, how this will affect individual plants is addressed through plant traits and plant performance at the species level measured on individual plants in the greenhouse.  I will address the following general questions at these different levels of organization: 1. Community and Ecosystem: How do community structure and function change in response to disturbance and stress? Which ecosystems are more resistant to the effects of stress and disturbance? 2. Plant – plant interactions: How do stress and disturbance alter the strength of competition between species? How does the way we assess competition affect our interpretation? 3. Species: Is there variation among species in their response to stress? Is a species’ response to stress predictable? Thesis overview The primary objective of the research is to investigate plant responses to the individual and interacting effects of two fundamental ecological processes: stress and disturbance. This was done in the field to examine species, functional-group and community-level response to water availability, temperature and clipping. To better understand these effects on individuals and the competition between individual plants I also did an experiment in the greenhouse that manipulated water, clipping and whether a neighboring plant was present.  Finally, rooting depth has been identified as a key plant trait affecting individual plant performance, especially in arid systems, and interactions between individuals and community processes. So I examined variation in this trait among species. Climate change has the potential to interact with other process and differentially affect ecosystems.  The objective of this research was to address the questions: what might the individual and combined effects of climate change and disturbance be on grasslands; and based on productivity, which ecosystems are most resistant to change? Chapter 2 describes a 4-year field experiment in which I manipulated temperature, water and clipping in natural grasslands and repeated this in three grasslands of differing productivity; responses focus on the both biomass production and community response. The experiment described in Chapter 2 is complex with multiple treatments.  To cause warming and change water availability I used small-scale climate manipulation devices in the field.  While the individual effects of these devices have been examined, interactions between the devices and with vegetation removal have not been addressed in the literature; in chapter 3 I  6 address this issue. Using sensors that monitored the key variables I altered, soil temperature and soil moisture, I examined the consequences of these multiple treatments and address whether the manipulations caused the intended effects.  The manner by which competition between plants changes along a productivity gradient has been debated for at least three decades (Craine 2005, Brooker et al. 2005).  Yet, both sides of the debate recognize that productivity does affect the way plants interact and that these interactions contribute to the composition of plant communities. Climate change has the potential to alter productivity levels (by increasing stress) and consequently, competition and, ultimately, the structure of communities.  Furthermore, disturbance can also affect competition and the outcomes of competitive interactions.  In Chapter 4, I report a greenhouse experiment in which I manipulated water availability, nutrient availability, clipping and whether a neighbour is present to examine the interactive effects of stress and disturbance on five indices of competition.  I also examine how the way we measure competition influences how the results are interpreted.  Predicting a species-level response to environmental change has long been a goal of ecologists and plant traits have been shown to correlate with a species’ effect on the environment and its response to the environment. Rooting depth is a key trait associated with the ability of a plant to obtain water, shown to separate the niches of species and related to ecosystem function (e.g. transpiration). In Chapter 5, I report the results of a greenhouse experiment in which I measured the plasticity of root characteristics of 18 grass species in response to reduced water availability.  Chapter 6 summarizes the results and implications of these four studies.  I discuss the importance of experimentally testing interactions, examine how the different types of experiments informed each other and the implication of not doing experiments with multiple treatments. Throughout this chapter I examine the management implications of my results and suggest future research directions.   7 Chapter 2: Interacting effects of stress and disturbance on grassland community structure and biomass production along a natural productivity gradient Introduction Climate change is influencing ecosystems globally (Walther et al. 2002, Rosenzweig et al. 2008), and increasing atmospheric CO2 concentrations will continue to alter climate into the future (Solomon et al. 2007).  Climate change will likely interact with other human-caused disturbance to alter ecosystems in unforeseen ways (Hulme 2005), but will not affect all ecosystems in the same way (Grime et al. 2000, Rustad et al. 2001, Walker et al. 2006). The interaction of climate change with other processes is not commonly studied, and changes in these interactions along gradients even less so (Smith et al. 2009), even though gradient studies are an advantageous approach to understanding the effects of climate change on plant communities (Keddy 1991, Dunne et al. 2004, Etterson 2004). Rising atmospheric CO2 levels are leading to climate changes that include increased temperature and altered precipitation patterns.  In arid and semi-arid environments, water availability, reduced through declines in precipitation and increased evapotranspiration, is likely to have the greatest impact on plant communities (Brown et al.1997, Weltzin et al. 2003). Climate change already has, or is predicted to, increase species extinction rates (Thomas et al. 2004), alter species’ distributions (Parmesan and Yohe 2003, Root et al. 2003), change food web structure and processes (Petchey et al. 1999, Hoekman 2010) and impact ecosystem function (Cramer et al. 2001).  Changes in both the amount and seasonality of water availability may alter diversity (Knapp et al. 2002) and reduce biomass production (Swemmer et al. 2007, Wu et al. 2010). Natural and experimental warming tends to increase productivity in plant communities (Hudson and Henry 2009, Wu et al. 2010), reduce diversity (Klein et al. 2004, Gedan and Bertness 2010) and can alter species interactions (Klanderud 2005, Tylianski and Didham 2008). Disturbance, the removal of plant biomass, will occur simultaneously with climate changes. One of the most wide-spread disturbances is grazing. More than 50% of the world’s terrestrial land mass is grazed and disturbed by domesticated livestock (Havstad et al. 2008), and the stresses induced by climate change will interact with disturbance from grazing. Therefore, understanding these effects is crucial for the sustainability of grazing agriculture, and the maintenance of biodiversity and ecosystem function. To manage and conserve ecosystems  8 effectively it is imperative that the least resistant systems be identified to minimize the impact of human-induced climate change and grazing disturbance. Not all systems will respond to stress or disturbance similarly (Connell 1978, Grime 2001).  Productivity has been shown to affect species richness and diversity (Al-Mufti 1977), species interactions (Choler et al. 2001), competition (Tilman 1982, Grime 2001), and trophic structure (Oksanen et al. 1981).  Generally, low-productivity communities are occupied by species that are stress-tolerant and possess mechanisms for coping with low resource availability such as storage organs and highly lignified leaves.  In contrast, high-productivity communities are composed of faster-growing species with less capacity to cope with stress (Grime 2001). The response of species richness to either stress or disturbance varies depending on productivity.  For example, an increase of stress at sites with low-productivity is predicted to lead to a decrease in species richness, while at sites with high-productivity an increase of stress is predicted to increase species richness (Grime 2001).  Interactions between stress and disturbance are also important because they can increase (Voight et al. 2007), mitigate (Klein et al. 2004, Post and Pedersen 2008) or even reverse (Suttle et al. 2007) the response to climate change. My study site was the bunchgrass grasslands of the southern interior of British Columbia, Canada. These grasslands represent the northern range of many grassland species (Tisdale 1947) and bunchgrass grasslands are listed as endangered ecosystems in British Columbia (BC Conservation Data Centre, 2011).  They are important for both biodiversity and agricultural use. The grasslands occur along an elevation gradient and are associated with changes in productivity, over relatively short distances, making them an ideal system to test how stress and disturbance differentially affect the plant community along a productivity gradient. Temperature in the region is expected to increase by up to 4°C over the next 100 years but projected changes in precipitation patterns are variable and may increase or decrease (Canadian Climate Change Scenarios Network, 2011).  Climate change may affect grassland function in terms of biomass production (forage) for both wildlife and the ranching industry (Weltzin et al. 2003).  I used open-top chambers (OTC), rainout shelters, water addition and hand clipping of vegetation, at three points along a natural productivity gradient to test the effects of stress and disturbance on grassland composition and biomass production.  The objectives of this study were 1) to test the interacting effects of climate change (warming, and water availability) and disturbance (clipping) on grassland communities, and 2) to test how these effects change along a natural productivity gradient. I predicted that stress (increased temperature and water removal) would alter community composition differentially  9 depending on what point along a productivity gradient the community initially occupies.  High- productivity communities would likely experience an increase in species richness when stress is increased, but decrease when water is added. In contrast, low-productivity sites would likely increase diversity with water addition but diversity will decline if stress increases. In addition, I predicted that stress would decrease standing biomass, while the addition of water would increase biomass. Disturbance (clipping) would likely lead to increased diversity at all points along the gradient because the system is heavily dominated by bunchgrasses; however, the change would be greatest at high-productivity sites where competitive exclusion is greatest. Disturbance would reduce biomass production at all elevations, but the impact would be greatest at high-productivity sites where vegetation is predicted to be least resistant. Methods Site description The study was done in Lac du Bois Grassland Provincial Park (UTM 10 E 0680737 N 5625980) within the bunchgrass grasslands of the interior of British Columbia, Canada, and north of the city of Kamloops. The region is semi-arid with an average annual precipitation of 279 mm, 75.5 mm of which is snowfall, the coefficient of variation for annual precipitation (1951 -2006) was 0.20; average annual temperature for the region is 8.9 °C, the warmest month is July (21.0 °C) and the coldest is January (-4.2 °C) (Environment Canada, 2009). The grasslands have been categorized as Lower grassland (LG), Middle grassland (MG) and Upper grassland (UG) and are associated with changing elevation, and also temperature, rainfall, soil type, productivity and vegetation types (van Ryswyk et al. 1964, Table 2.1). The increase of productivity is likely due to a decrease in the ratio of evapotranspiration to precipitation with increasing elevation where rainfall is higher and temperatures are cooler than at lower elevations.  The LG are dominated by Pseudoroegneria spicata (Pursh) A. Love (bluebunch wheatgrass) and the shrub Artemisia tridentata Nutt. (Big sagebrush), with Poa secunda J.S. Presl (sandberg’s bluegrass) and Vulpia octiflora (Walter) Rydb. (six-week fescue) as other common species (see Appendix A for information on common grass species). The MG are also dominated by P. spicata, but A. tridentata is less dense; other common species in the MG are Koeleria macrantha (Ledeb.) Schult. (june grass), Achillea millefolium L. (yarrow) and Astragalus collinus Douglas ex G. Don (hillside milkvetch).  The UG are mostly devoid of  10 shrubs and dominated by Festuca campestris L. (rough fescue), while Poa pratensis L. (Kentucky bluegrass) and Juncus balticus Willd. (baltic rush) are the next most common species. Standing live biomass and litter were collected at the beginning of the experiment in 2005, at the peak of vegetative growth, from six 0.25 m2 plots at each site, which were not part of the experiment, dried at 65 °C for 48 hrs then weighed. Rainfall and ambient temperature were recorded using a rain gauge and temperature logger from May to October of 2007 and 2008 (Model RG3-M, Onset Computer Corporation, Bourne, USA). Soil carbon and nitrogen were sampled in control plots in 2008 using four pooled 2 cm diameter, 10 cm deep cores from each plot. Soil carbon and nitrogen were estimated using an elemental analyzer (Model CE-440, Exeter Analytical Inc., North Chelmsford, MA). Experimental design Five 30-year-old fenced exclosures (not grazed by cattle), two each in the LG and MG and one in the UG, were selected in May 2005.  I established 36 plots (1 m2) in each of two LG and two MG exclosures, and 72 plots at each of two sites in the UG exclosure (216 plots in total).  Each plot was at least 1 meter away from adjacent plots and shrubs.  At each site, plots were arranged in a 6 x 6 grid and randomly assigned a treatment within the constraints of creating a pseudo-Latin square design such that each treatment occurred within every two rows or columns. Open-top chambers (OTC) were used to manipulate temperature, with the treatment either at ambient or manipulated  temperature (no OTC, OTC). Water was at ambient or was manipulated with rainout shelters or by hand watering (ambient, rainout shelter, and addition) and plots were either clipped or not clipped (clipped and unclipped).  Thus, using a factorial design there were twelve possible treatment combinations each of which was replicated three times at each site.  The experiment was run for four years, 2005 – 2008. Treatments were in place from April to October of each year, except 2005 when they began in May. The open-topped chambers (OTC) used in this study were similar to the plastic tent design initially described in Marion et al. (1997). These OTC have been shown to affect grassland plants (Fraser et al. 2009) and their abiotic effects are described in chapter 3. The OTC had a square base with each side 1.5 m long.  The plastic was secured to a wooden stake driven into the ground, angled so that the top opening was square, 1 m on a side and 40 cm above the soil surface. The plastic (Tufflite IV ™, 6 mil, 0.152 mm thick, Tyco Plastics and Agricultural Films, Monroe, LA, USA) has high transmission of photosynthetically active radiation and can  11 repel dust from its surface. Soil temperature was measured in a subset of plots at 5 cm depth; OTC increased temperature 1.2 °C above that of the control plots (chapter 3). Rainout shelters were constructed as described in Köchy and Wilson (2004) which comprises a 1 m2 plastic sheet, the same material used in the OTC, attached to a pole 1 m high at one corner and anchored at the remaining three corners such that they were each 30 cm above the soil surface to allow airflow. These rainout shelters have also been shown to affect grass physiology (Fraser et al. 2009) and their environmental effects are described chapter 3.  The sheet was oriented to block rain from the dominant wind directions during the growing season. Water availability in the plots was manipulated by hand watering the plots weekly from May through October of each year.  Once per week each plot received water to increase the monthly 30-year rainfall average 30% over the course of the month.  The 30-year average (1970 to 2000) monthly rainfall for the months May through September were 24.4, 35.2, 29.5, and 29.1 and 28.0 mm (Environment Canada, 2005); thus we added 1.8, 2.6, 2.2, 2.2 and 2.1 litres of water per week to the plots in each month. Plots were hand-watered slowly to ensure minimal runoff. Locally collected rainwater was used for the watering. Soil moisture was measured in the top 10 cm of soil in a subset of plots; rainout shelters decreased the mean soil moisture by 3.1 % and water additions increased moisture 2 % from a mean of 11 % VWC (volumetric water content) in control plots (chapter 3). Clipped plots were cut annually with shears in early July, at the peak of vegetative growth. The entire 1 m2 plot plus a 0.25 m wide border were clipped to a height of 5 cm. In July 2008, the vegetation in the center 0.25 m2 of each plot was clipped at ground level, sorted to species, dried for at least 48 hrs at 65 °C and then weighed. Litter from the same areas was also collected, dried and weighed in 2008. Clipping was a treatment, thus biomass was not collected from all plots in every year of the experiment. Consequently, I used visual percent cover as a proxy for biomass that was estimated for each plot in every year by a single observer; total percent cover could potentially sum to more than 100% if different species occupied different canopy layers. Indices used For each plot, I calculated Shannon’s index of species diversity, which weights uncommon species, because the system was heavily dominated by one of two bunchgrass species, Festuca campestris or Pseudoroegneria spicata.  Evar was calculated as a measure of species evenness (Smith and Wilson 1996).  12 Species functional groups One of two bunchgrasses (F. campestris and P. spicata) dominated the study sites.  To examine the response of the remainder of the community I divided species into a number of subordinate or functional groups: dominants, subordinate graminoids, and forbs. Dominants included only P. spicata and F. campestris.  Subordinate graminoids included all graminoids except the two dominants. The graminoid group was primarily grasses, but also included the rush Juncus balticus and the sedge Carex petisata. The forb group included all non-graminoid and non-woody species except for the cactus Opuntia fragilis, which occurred infrequently and uses the CAM photosynthetic pathway; thus, the only non-C3 species in the study sites. Data analyses Total biomass, litter, soil carbon, soil nitrogen and species richness of the study sites were tested using ANOVA with SITE as the factor.  Total biomass, litter and species richness were log transformed. Each test was followed with a post-hoc Tukey HSD test. I used a three-way ANOVA to test the effects of OTC, water addition and rainout shelters (WATER), and clipping (CLIP) on total plot biomass, total percent cover, litter mass, cover of biological crusts, percentage of bare ground, species richness, Shannon diversity, Evar and the response of species groups.  SITE was included as a random effect. All variables were log transformed to meet the assumption of normality. Each ANOVA was followed with a post-hoc Tukey HSD test. The data for forbs could not be normalized using a transformation, so I used a non-parametric Mann-Whitney U test. A MANOVA was used to examine changes in plot cover over time (2005 – 2008).  To compare plot biomass to percent cover I used a simple linear regression on all data collected in 2008; both variables were log transformed. To examine the multivariate response of community composition to the treatments and year I used permutational multivariate analysis of variance (PERMANOVA, Anderson 2001).  I ran separate PERMANOVAs for each grassland type using both a Euclidean and Bray-Curtis distance matrixes because results can be dependent on the underlying distance matrix and transformations. Permutations were constrained by stratifying the data by site and each test was run with 9999 permutations.  I separately ran PERMANOVAs for each grassland type for 2008 data only. For this test, I followed each significant effect with a permutational analysis of multivariate dispersions (PERMDISP, Anderson 2006) to test if groups differ in variance. PERMDISPs can only test a single factor, thus in the case of a significant interaction I separated  13 the treatment combinations of the two factors into factorial groupings and ran the test of dispersion on the full set of treatment combinations. To display the multivariate response of the 2008 data I used a non-metric multi-dimensional scaling (NMDS) of the species data, run separately for each grassland type. The NMDS was run with 100 random starts, using a Bray- Curtis distance matrix, two axes were specified for the final solution.  All analyses were done using R v. 2.11.1 (R Development Core Team, 2010).  Multivariate analyses were run in R using the vegan package v. 1.17-3 (Oksanen et al. 2010). Results I found that OTC, water availability and clipping all affected grassland composition and biomass production;  however, these effects varied among the different grassland types. Overall, the MG were most susceptible to change, while the UG showed the fewest changes in response to treatments. Sites Total plant biomass (live vegetation and litter) and soil carbon and nitrogen increased with elevation across the six experimental locations (Table 2.1); this correlates with an increase in precipitation and decrease in temperature along the elevation gradient.  Species richness was relatively low at the plot level and presented a hump-shaped relationship with the most species occurring in the middle elevation plots and fewer at the low and high elevation plots, although the total species pool, at a site, tended to decrease with elevation. Response of standing biomass and cover to treatments The response of standing live biomass to the treatments depended on the grassland type (Figure 2.1, Table 2.2).  Biomass was reduced under OTC in the MG, but had no effect in the LG or UG. Biomass increased under rainout shelters in the LG and MG while biomass in the UG was not affected by water treatments. Clipping reduced biomass in all three grassland types. Interactions were important in determining plant biomass.  While clipping always had the obvious effect of reducing biomass, the effects of temperature and water treatments tended to depend on the clipping treatment (Table 2.2).  The positive response of biomass to water reduction only occurred in unclipped plots and while OTC reduced biomass the reduction was larger in plots that were also clipped.  14 In 2008, the year all biomass was harvested at ground level from all plots, percent cover was a good predictor of biomass (simple linear regression: r2 = 0.70, F1, 214 = 510.6, p<0.001). MANOVA, examining the response of percent cover over time, concurred with univariate tests run on biomass data from 2008. Over time, clipping had the only significant effect on percentage cover in the LG (Figure 2.2, Clipped: 52 % cover, Control: 66 % cover; MANOVA: Pillai 0.227, F = 4.109 P =0.005).  Biomass in the MG was altered by both water and clipping (Clipped: 65 % cover, Unclipped: 79 % cover; MANOVA:  Pillai 0.232, F = 4.303, P = 0.004, ambient: 63 % cover, addition: 77 % cover and rainout shelter: 76% cover; MANOVA: Pillai = 0.271, F = 2.271, P = 0.027). Clipping had a significant effect in the UG (clipped: 113 % cover, unclipped: 111% cover; MANVOA: Pillai = 0.141, F = 2.32, P = 0.066). Response of groups: dominants, subordinate graminoids and forbs In the control plots, the two dominant species (P. spicata in the LG and MG, F. campestris in the UG) represented 73%, 78% and 92 % of the biomass in the LG, MG and UG respectively (Figure 2.1).  In the LG the biomass of the dominant species was changed by water treatments and clipping along with a TEMP X CLIP interaction such that the OTC only reduced biomass when the plots were not clipped.  In the MG, watering treatments changed the biomass of P. spicata, as did clipping. Only clipping reduced the biomass of the F. campestris in the UG. Subordinate graminoids responded differently than the two dominant species. In control plots, subordinate graminoids composed 7.7, 2.1 and 7.2 % of the biomass (Figure 2.1). In the LG subordinate graminoids responded to a WATER x CLIP interaction such that without clipping both watering treatments lead to a decrease in their biomass, but when plots were clipped both water treatments tended to increase biomass. In the MG, the response of this group of species was more complex.  OTC reduced their biomass while clipping increased subordinate biomass. There was a further WATER X CLIP interaction such that subordinate graminoid biomass increased in rainout shelter – unclipped plots, but was reduced in rainout shelter – clipped plots. In the UG, OTC reduced the biomass of subordinate graminoids. There was also a WATER x CLIP interaction in which both water treatments reduced subordinate graminoid biomass in unclipped plots, but rainout shelters increased biomass in clipped plots. Forbs composed 19, 19.5 and 1 % of the biomass, in control plots in, the LG, MG and UG respectively (Figure 2.1). In the LG, the biomass of forbs was reduced in the OTC (U0.1(2),36,36 = 791.5, P =0.10).  In the MG, rainout shelters (U0.1 (2), 24, 24 = 168.5, P =0.01) and  15 water addition (U0.1 (2), 24, 24 = 173, P =0.02) increased forb biomass above that of untreated plots. In the UG, clipping increased forb biomass (U0.1(2),36,36 = 503.5, P = 0.10). Response of litter to treatments The dry mass of litter was reduced by clipping in all grassland types (LG – unclipped: 18.7 g.m-2 ± 2.4 SE, clipped: 11.13 g.m-2 ± 2.4 SE; MG – unclipped: 26.5 g.m-2 ± 2.5 SE, clipped: 16.1 g.m-2 ± 2.5 SE, UG – unclipped: 197.3 g.m-2 ± 9.7 SE, clipped: 169.2 g.m-2 ± 8.5 SE).  Interactions were only significant in the LG: rainout shelters tended to increase litter amounts when unclipped and OTC tended to reduce litter when the plots were clipped. Plant community response Species richness was changed by treatments only in the LG and MG.  Clipping increased species richness in the LG and a CLIP x WATER interaction revealed that this effect was strongest when the rainout shelter was present (Figure 2.3a). In the MG, OTC reduced the mean number of species by one (Figure 2.3b).  There was no effect of treatments on species richness in the UG. After four years, the effects on plot Shannon diversity were specific to grassland type but clipping was always a component of the change as either a significant main effect or a significant interaction.  In the LG, clipping increased diversity but not when the OTC was present (Figure 2.3c).  In the MG, clipping increased diversity (Figure 2.3d).  In the UG, there was a WATER X CLIP interaction (Figure 2.3e) in which the combination of clipping and rainout shelters had the highest diversity. Plot evenness was changed in the LG (Figure 2.3f) and UG (Figure 2.3h) in WATER x CLIPPING interactions but the interactions were different.  In the LG, clipped – ambient water plots had higher evenness than unclipped – water addition plots.  In the UG, the clipped-ambient water plots had the lowest evenness and clipped-rainout shelter plots had the highest evenness. For both of these interactions the post-hoc Tukey test, which is more conservative than an ANOVA, indicated no significant pairwise differences between groups.  Evenness in the MG was increased by the OTC and clipping. Furthermore, there was an OTC x CLIP interaction in which OTC – clipped plots had significantly higher evenness than the other OTC x CLIP combinations (Figure 2.3g). The grasslands showed multivariate community response to treatments. PERMANOVAs testing these responses provided similar results regardless of whether an Euclidean or Bray-  16 Curtis distance matrix was used (Table 2.4); I focus on the results of the Bray-Curtis distances as it maintains robust relationships with the ecological data (Faith et al. 1987).  After four years, the LG was altered only by clipping (Table 2.4a), which also changed the dispersion of the two groups (Figure 2.4a, PERMDISP: df = 70, F = 9.74, P = 0.0018). The mean (± SE) distance to the centroids for the two groups were clipped = 0.38 ± 0.07 and unclipped = 0.25 ± 0.08 (Figure 2.4a).  The composition of the MG was changed by clipping, water treatments and their interaction (Table 2.4a).  However, watering treatments did not affect group dispersion in the MG (PERMDISP: df =69, F = 1.53, P = 0.226), but clipping did (PERMDISP: df = 70, F = 9.23, P = 0.003). Furthermore, a test of dispersion run on the six treatment combinations of the CLIP x WATER interaction in the MG shows that these groups also varied in dispersion (PERMDISP: df = 66, F = 2.80, P = 0.021). The mean (± SE) distances to centroids for the groups were control = 0.25 ± 0.04, ambient water – clipped = 0.36 ± 0.05, water added – unclipped = 0.28 ± 0.05, water added – clipped = 0.38 ± 0.04, rainout shelter – unclipped = 0.24 ± 0.07 and rainout shelter – clipped = 0.26 ± 0.03 (Figure 2.4b).  In both the LG and MG, clipping appears to be separating on the second MDS axis. The UG had no significant multivariate community response to treatments, although water treatments were close to significant and there were significant effects when all 4 years of data were examined together (Table 2.4). The PERMANOVAs run using all 4 years of data, with year as a factor, show that the treatment effects were dependent on the year (Table 2.4b, Figure 2.5). The clipping treatments generally lead to plots being less similar. The OTC had no effect on community composition in any of the grassland types; however, OTC treatments did interact with watering treatments and clipping suggesting that OTC effects are dependent on other conditions.  In the LG and MG, the OTC-clip treatment showed an increasing dissimilarity over time, regardless of watering treatment (Figure 2.5d –i). YEAR was significant in predicting community composition in all three grassland types. CLIP was significant in the LG and MG although the MG was subject to numerous interactions.  In the UG (Figure 2.5a – c), WATER, YEAR and a WATER x CLIP interaction were significant.  Over time, for all treatments, plots became more dissimilar, and at the end of the experiment water addition plots were more similar than ambient or water removal plots. Discussion I found that experimental manipulations of temperature, water and clipping altered grassland community structure, composition and biomass, but often in unexpected ways.  The responses varied by grassland type.  Interactions between treatments could increase, decrease or  17 reverse the direction of response to individual treatments.  Most interactions were between climate manipulations and clipping, suggesting that future alteration of plant communities due to climate change will be dependent on disturbance. While temperature did alter grassland composition and biomass, water treatments more often had significant effects suggesting that precipitation may have a greater impact in arid grasslands than changes in temperature. Theory suggested that the more productive upper grasslands would be the least resistant to change (Grime 2001); however, I found that these grasslands changed the least in both biomass production and community structure.  The two dominant species tended to respond differently than other groups suggesting that competitive interactions may play an important role along the entire gradient. Biomass response Contrary to most other experiments testing the effects of warming on plant communities, I found that warming reduced plant biomass.  I predicted this response because water, regulated in part by temperature, is likely the limiting factor in semi-arid grassland.  The majority of warming studies have occurred in the arctic and other cooler climates, where increased temperature likely increases plant metabolism and extends the growing season without limiting plant resources (Rustad et al. 2001, Walker et al. 2006). Clipping also reduced biomass, but the interactions of clipping with OTC and water treatments are important because of the implications for grazing management.  Klein et al. (2007) also reported biomass reductions due to OTC; however, they also observed increased biomass due to experimental defoliation and that defoliation mitigated the effects of the OTC.  In contrast, I found that the combination of clipping and warming tended to further reduce biomass along the entire gradient. This may be due to the combination of increased disturbance and stress, and suggests that disturbances, such as grazing, will interact with climate change to negatively influence ecosystem function – especially plant production. Unexpected biomass response to water treatments Changes in biomass due to watering treatments were complex and unexpected. Biomass in the LG and MG increased under the rainout shelters; this pattern was mostly due to the dominant species, P. spicata.  Adding water to plots had no effect on biomass production.  This is surprising because the additions did increase soil moisture during the growing season. This may occur because the plants are adapted to arid conditions, grow early in the season when water  18 is available for a short period and do not respond to increased water availability. Water addition did contribute to compositional changes in the grasslands leading to the speculation that the lack of biomass response may be transient, and I suspect that biomass may increase as composition shifts towards species adapted to use the increased resource availability. The increase in plant biomass under the rainout shelters was unexpected and could occur for a number of reasons. Unintended treatment effects from the rainshade could alter soil moisture, light or wind in a manner that led to increased growth of P. spicata. Soil moisture was measured on a regular basis (Chapter 3, Carlyle et al. 2011) and was significantly reduced by the rainout shelters. Humidity was not measured but my own observation suggests that it did not increase because of the airflow through the device.  Light was altered and slightly reduced under the shelter, which may induce leaf growth. The plastic used in the RS also makes light more diffuse which could allow light to reach lower levels of the canopy; however, P. spicata is not likely light limited due to the low amounts of standing biomass in the community. The RS may act as a wind block and lower the amount of moisture lost through evapotranspiration, but the rainout shelters were elevated higher than most of the vegetation and were open to airflow. Trophic effects have also been reported in a number of climate manipulation studies (Pendersen and Post 2007, Suttle et al. 2008).  Plant biomass could increase if herbivores were consuming biomass but avoided rainout shelter plots. However, I observed few insect or small mammalian herbivores in the low and middle grassland compared to the upper elevations and this pattern is confirmed by other researchers working in the region (Bassett 2009, Hales 2011). If trophic effects were responsible for the pattern, I would have expected them to be larger in the UG, but this was not observed. Plant biomass may also increase if top-down control increased (Barton et al. 2009, Fraser 1998). A further explanation may be a possible growth stimulation to P. spicata caused by the decline in water.  In greenhouse studies P. spicata is capable of fast and deep root growth under drought conditions (Chapter 5); the plants may be foraging into new otherwise unused water sources.  I did not trench the plots; doing so would reduce these effects if the growth were lateral but not if it were vertical. Finally, P. spicata is adapted to semi-arid climates where drought is a regular occurrence and has been shown to increase its stomatal density under RSs  when compared to ambient and water addition treatments (Fraser et al. 2009). Increased above ground biomass production may be a stress response whereby the allocation to above-ground biomass creates litter to both shade out smaller neighbouring species competing for water and act as a mechanism to trap moisture in the soil. I found that the biomass of P. spicata did not respond positively to reduced water availability when clipping occurred and OTC and clipping had  19 additive negative effects on biomass.  Reports from meta-analyses of water effects on plant biomass are mixed: one reported no effect of water addition in the arctic (Dormann and Woodin 2002) and the other reported a positive response (Wu et al. 2010) suggesting that the effects of water availability on plant systems are complex and not completely understood. Other studies have reported slight increases or no reduction in biomass due to rainout shelters.  Gilgen and Buchmann (2009) observed no significant response on biomass and slightly increased biomass at one site due to rainout shelters, despite a decrease in leaf-water potential—indicating water stress.  Lucas et al. (2008) reported an unexpected increase in population growth rate of Cryptantha flava under rainout shelters and attributed the effect to increased temperatures during a cold year. However, the response in my experiment appears to be a trend independent of inter- annual environmental conditions as the effects were apparent in all three years after the treatments began. Community response Climate change treatments and disturbance both altered community composition along the gradient, but responses were specific to each grassland type.  Clipping had a critical role in community structure at all elevations; it affected species richness, diversity, evenness, homogeneity and commonly interacted with other treatments.  OTC reduced the number of species in a plot by one in the middle grassland. While this is not a large absolute reduction compared to the substantial decline in species found in similar studies (Klein et al. 2004, Gedan and Bertness 2010), the consequences of this loss is unknown.  The mean plot species richness in this study was relatively low (6 species per plot), compared to approximately 35 species per plot in Klein et al. 2004. If an individual species contributes significantly to ecosystem function, loss of a single species in a community with low species richness could lead to functional loss, especially if there is little or no functional redundancy (Johnson et al. 1996). The loss of a species in the MG may be the result of increased water use due to functional complementarity that has been observed in higher diversity communities (van Peer et al. 2004). Although, water had few effects on species richness in this experiment; I expected that water would be the limiting resource in the system and manipulating its availability should have direct effects on species richness.  20 Interactions Interactions between treatments were important in determining community structure and composition. Interactions could prevent a response (e.g. in the LG, clipping did not lead to increased diversity when the plot had an OTC), initiate a response (e.g. evenness increased in the MG only when both clipping and the OTC were treatments) or even reverse a response (e.g. in the UG, clipping and water addition individually lowered diversity, but in combination increased diversity). Kardol et al. (2010) did not find interactions to be important in structuring plant communities. However, that experiment did not include a treatment that removed biomass, which has been shown to be an important determinant of response in some other climate change experiments (Klein et al. 2007, Post and Pedersen 2008, Suttle et al. 2007). The results partially conflict with other studies reporting that disturbance limits the amount of change induced by warming (Klein et al. 2007, Post and Pedersen 2008). Dominance Dominance by P. spicata and F. campestris is also likely a major structuring factor of these grasslands. Many of the results can be interpreted as competitive effects of dominants on subordinate species, because the positive biomass response of subordinate graminoids and forbs suggest competitive release. Dominance appears to drive vegetation response along the entire gradient, not just in the high productivity grasslands where dominance is expected to be largest (Grime 2001). While the influence of dominance on community response appears to occur universally on the gradient it has not contributed to consistent responses along the gradient, evidenced by the variability of response in the different grassland types. I expected clipping to increase diversity in the UG. This may not have happened because, while the clipping treatment affected all species, the dominant, F. campestris, still overwhelmingly dominated the plots. Water addition should shift the community toward the dominant. Only when the combined effects of stress and disturbance are present did subordinate species increase biomass in the UG. Change along the gradient I found that the MG, of intermediate productivity and highest diversity had the greatest response to the treatments, while the high productivity UG showed the least response. Grime et al. (2000) proposed four reasons why a plant community may be more resistant to climate change: 1) a history of exposure to climate extremes, 2) succession status, 3) diversity and 4) functional composition.   All of the sites were close to each other and climate monitoring  21  22 indicates that they were subject to similar climate patterns.  The LG and MG would be more likely to experience climate extremes, yet they showed the least resistance to the treatments. All of the sites were fenced from grazing, had not experienced fires or other major disturbance for at least 30 years and could all be classified as climax communities. The MG, highest in diversity, were generally most responsive to treatments and this conflicts with a long history of theory and experiments that suggest diverse systems are more stable (McCann 2000, Ives and Carpenter 2007).  Grime et al. (2000) found that low productivity grasslands were more resistant to climate manipulations than high productivity sites and functional composition was the likely explanation of resistance. The UG contained a different suite of species than the LG and MG; however, I currently do not have enough information pertaining to functional composition of these grasslands to asses this hypothesis. Summary The response of ecosystems to multiple factors are often complex. I have shown that grasslands of different productivity levels do respond individually to stress and disturbance but not as predicted.  High productivity grasslands showed the least response to treatments while medium productivity, high diversity grasslands were altered the most. Low productivity grasslands also responded less than the medium productivity grasslands. This response is counter to many predictions that diversity confers stability and my own predictions that productivity, via plant traits, lowers resistance. I demonstrated that interactions were important in determining the biomass and composition of the plant community.  Regionally, these grasslands are important for both forage and conservation of diversity.  These results indicate that disturbance in a warmer, drier environment may not support either of these goals. Table 2.1 Background data for the six experimental field sites. Standing biomass and litter were collected from 6 plots at each site that were not used in the experiment in June 2005.  Mean rainfall and temperature were recorded from May to October in 2007 and 2008, except for rainfall in the Upper grassland, which are 2007 data only (because of a logger malfunction in 2008).  Soil carbon and nitrogen were sampled in control plots in 2008.  Superscript letters indicate differences among sites (Tukey HSD, P = 0.05). Species richness Grassland type Site Elevati on (m,a.s. l) Mean rainfall (mm) Mean annual tempera ture (° C) Standing biomass (g. 0.25 m-2 ± SE) Litter (g. 0.25 m-2 ± SE) Soil Carbon (% ± SE) Soil Nitrogen (% ± SE) Mean plot SR Site Elevation Total UG1 860 62.7 ± 9.0 a 147.0 ± 27.5 a 9.09 ± 0.25 a 0.79 ± 0.01 a 5.3 ± 0.3 b 15 Upper grassland UG2 888 145 15.4 55.5 ± 5.4 b 109.2 ± 20.9 a 9.18 ± 0.58 a 0.80 ± 0.04 a 5.4 ± 0.2 b 16 18 MG1 731 29.2 ± 4.1 c 33.4 ± 3.4 b 2.70 ± 0.09 b 0.25 ± 0.01 b 5.4 ± 0.3 b 25 Middle grassland MG2 761 112.4 17.5 21.7 ± 5.4 d 24.0 ± 10.6 b 2.47 ± 0.11 b 0.23 ± 0.01 b 6.9 ± 0.3 a 26 30 LG1 581 18.3 ± 3.4 e 10.3 ± 2.6 c 2.00 ± 0.22 c 0.19 ± 0.02 c 3.2 ± 0.2 c 14 Lower grassland LG2 630 95.8 19.3 23.7 ± 3.9 d 16.7 ± 4.1 b 2.02 ± 0.12 c 0.20 ± 0.01 c 5.6 ± 0.2 b 26 34 52  23 Table 2.2 ANOVA results: F values for significant factor effects on total biomass, species groups mass and litter mass in the three grassland types. Only significant (P <0.1) effects and interactions are shown, residual degrees of freedom = 59.  Factors are open-top chambers (OTC), water addition or rainout shelters (WATER) and clipping (CLIP).  ˙ P <0.1, * P <0.05, ** P <0.01, *** P < 0.001. Response variable Factors df Lower grassland Middle grassland Upper grassland Total Biomass OTC 1  6.70*  WATER 2 2.66˙ 9.65***  CLIP 1 31.84*** 31.45*** 6.33*  WATER x CLIP 2  2.77˙  OTC x WATER x CLIP 2 2.86˙  Dominant graminoids WATER 2 5.69** 3.87*  CLIP 1 39.25*** 44.87*** 6.01*  OTC x CLIP 1 2.9˙  Subordinate graminoids OTC 1  6.04* 3.79˙  CLIP 1  8.4**  WATER x CLIP 2 2.49˙ 4.47* 3.07˙  OTC x WATER x CLIP 2  3.05˙  Litter WATER 2 2.46˙  CLIP 1 15.44*** 12.33*** 4.55*  OTC x WATER x CLIP 2 4.22*  24  25 Table 2.3 ANOVA results: F values for significant factor effects on species richness, diversity and evenness in the three grassland types. Only significant (P <0.1) effects and interactions are shown, residual degrees of freedom = 59.  Factors are open-top chambers (OTC), water addition or rainout shelters (WATER) and clipping (CLIP).  ˙ P <0.1, * P <0.5, ** P <0.01, *** P < 0.001. Response variable Factors df LG MG UG Species richness OTC 1  4.96*  CLIP 1 3.45˙  CLIP x WATER 2 2.40˙   CLIP 1  12.63** Shannon diversity OTC x CLIP 1 3.27˙  WATER x CLIP 2   4.41*   OTC 1  3.55˙ Evenness CLIP 1  10.06**  OTC x CLIP 1  3.65˙  WATER x CLIP 2 3.23*  2.55˙  Table 2.4 Permutational multivariate analysis of variance results testing factor effects on community composition in the three grassland types. Results from tests using both Euclidean and Bray-Curtis distance matrices are shown for a) 2008 only and b) 2005 – 2008 which also includes the factor YEAR. Significant effects are in bold (P < 0.05.  Factors are open-top chambers (OTC), water addition or rainout shelters (WATER) and clipping (CLIP).   Lower grassland (LG) Middle grassland (MG) Upper grassland (UG)   Euclidean Bray-Curtis Euclidean Bray-Curtis Euclidean Bray-Curtis  df F-value P F-value P F-value P F-value P F-value P F-value P a)  2008 OTC (O) 1 0.45 0.535 0.49 0.635 0.49 0.667 0.55 0.763 0.62 0.590 0.31 0.835 WATER (W) 2 2.64 0.042 0.94 0.275 3.27 0.12 2.39 0.005 1.70 0.077 1.54 0.090 CLIP I 1 19.41 0.001 7.69 0.001 15.41 0.001 7.20 0.001 1.02 0.340 1.15 0.266 O x W 2 0.34 0.781 0.69 0.537 0.360 0.904 0.69 0.740 1.02 0.346 0.90 0.43 O x C 1 2.00 0.119 1.03 0.230 2.63 0.048 1.23 0.229 1.48 0.154 1.61 0.133 W x C 2 1.01 0.336 0.61 0.611 0.63 0.665 1.86 0.039 0.90 0.428 1.15 0.246 O x W x C 2 0.414 0.705 0.22 0.980 1.23 0.276 1.15 0.233 1.07 0.288 1.14 0.254  b) All y  ears OTC (O) 1 1.15 0.254 1.30 0.194 0.68 0.513 1.36 0.180 1.29 0.218 1.76 0.143 WATER (W) 2 2.25 0.062 1.47 0.107 2.49 0.041 3.19 0.001 2.30 0.030 2.19 0.028 CLIP I 1 0.22 0.001 10.79 0.001 2.70 0.001 13.03 0.001 1.04 0.340 1.54 0.186 YEAR (Y) 2 2.60 0.022 3.81 0.001 5.57 0.001 8.55 0.001 37.11 0.001 35.93 0.001 O x W 2 1.82 0.106 2.68 0.006 3.00 0.018 3.00 0.001 0.63 0.664 1.10 0.333 O x C 1 1.07 0.290 1.62 0.125 6.57 0.002 4.40 0.001 1.05 0.316 0.72 0.482 W x C 2 2.25 0.065 1.74 0.050 1.49 0.156 2.72 0.007 1.51 0.146 2.50 0.027 O x Y 2 0.23 0.987 0.28 0.987 0.37 0.965 0.43 0.977 0.54 0.864 0.03 0.998 W x Y 4 0.67 0.736 0.29 0.999 0.89 0.519 0.72 0.851 1.20 0.208 1.22 0.183 C x Y 2 2.52 0.011 1.13 0.187 2.63 0.009 1.74 0.025 0.92 0.459 0.84 0.539 O x W x C 2 2.90 0.028 1.50 0.078 2.19 0.049 2.41 0.005 1.05 0.345 1.60 0.113 O x W x Y 4 0.67 0.735 0.51 0.949 0.74 0.718 0.62 0.951 0.95 0.465 0.83 0.591 O x C x Y 2 0.51 0.784 0.33 0.982 0.33 0.979 0.15 1.000 0.87 0.477 1.14 0.285 W x C x Y 4 0.41 0.969 0.35 0.998 0.28 1.000 0.72 0.864 0.80 0.636 0.61 0.849 O x W x C x Y 4 0.62 0.788 0.64 0.797 0.35 0.998 0.43 0.999 0.82 0.612 0.80 0.636  26  Figure 2.1 (next page) Total plant biomass and biomass of three functional groups harvested at the end of the experiment, in each of the three grassland types: a) upper grassland, b) middle grassland and c) lower grassland.  The shaded bar indicates biomass of the three functional groups: subordinate graminoids, dominant grasses and forbs. The points above the bars indicate total biomass (± SE), which also includes woody plants and cactus which were omitted from the forb group. Small letters above the bars indicate significant differences in total biomass based on a Tukey test.  Treatment combinations are labelled on the x-axis: No OTC = no open-top chamber, OTC = with open-top chamber, Ambient = ambient precipitation, W+ = water added, RS = rainout shelters, C- = unclipped and C+ = clipped.  27   28  Figure 2.2 Vegetative cover, by treatment, over four years in three grassland types: upper grassland (a – c), middle grassland (d – f) and lower grassland (g – i).  Plots were treated with open top chambers (OTC) and clipping, indicated in the legend, and received different water treatments: ambient water (a, d, g), water added (b, e, h) and rainout shelters (c, f, i).  Error bars were omitted for visual clarity.  The bold lines in a, d and g represent control plots. Note, the y- axis is different for each row of panels.   29   30  Figure 2.3 Species richness, diversity and evenness (± SE) based on biomass harvested in 2008 in the lower grassland (a, c and f), middle grassland (b, d and g) and upper grassland (e and h). There were no effects of treatments on species richness in the upper grassland. Whether a plot was not clipped (C-) or clipped (C+) is included in all figures except panel b.  Other factors are included to show the significant interactions of clipping with other treatments. The presence of an open-top chamber (OTC) is indicated in a and e, while watering treatments are indicated in c, d and f. Letters above the bars indicate significant differences based on a Tukey test.   Figure 2.4 Non-metric multi-dimensional scaling plots, run separately for each grassland type, showing (a) lower grassland, (b) middle grassland and (c) upper grassland. Open and closed symbols represent unclipped or clipped plots and small circles, squares and triangles represent the water treatments: ambient water, water added or rainout shelters respectively. In (a) and (b), the large circle and square represent the centroids of unclipped and clipped plots respectively.  In (a), the polygons represent the minimum convex hull surrounding all plots in each treatment set, unclipped or clipped plots.  In (b), the black polygon represents the minimum convex hull of control plots while the grey polygons represent the hulls of the five additional combinations of CLIP x WATER.  31  Figure 2.5 Mean Bray-Curtis dissimilarity among plots, by treatment, over four years in three grassland types: upper grassland (a – c), middle grassland (d – f) and lower grassland (g – i). Plots were treated with open top chambers (OTC) and clipping, indicated in the legend, and received different water treatments: ambient water (a, d, g), water added (b, e, h) and rainout shelters (c, f, i).  Error bars were omitted for visual clarity.  The bold lines in a, d and g represent control plots.   32  Chapter 3: Tracking soil temperature and moisture in a multi-factor climate experiment in temperate grassland:  do climate manipulation methods produce their intended effects? Introduction There is increasing evidence that global climate change is altering plant community composition and structure (Walther et al. 2002, Parmesan and Yohe 2003, Root et al. 2003, Walker et al. 2006, Hudson and Henry 2009). Ongoing greenhouse gas emissions caused by human activities will likely continue to contribute to global change in the future (Solomon et al. 2007).    Predictive climate models suggest an array of future climate conditions with the major drivers, temperature and precipitation, interacting in complex ways (Solomon et al. 2007).  To understand the potential consequences of climate change on plant communities it is necessary to manipulate climatic factors in controlled field experiments. For the most part, climate manipulations have been implemented singularly (e.g. Köchy and Wilson 2004, Gedan and Bertness 2009, but see Grime et al. 2000), but manipulations also need to be implemented factorially to understand their interactions. Furthermore, plant communities experiencing climate change will be interacting with other processes, such as disturbance, which could also be incorporated into climate experiments to understand the complexity of climate change effects on plant communities. Most climate change experiments have focused on temperature and precipitation because these are the factors most likely to be altered by global climate change.  Different methods have been used to manipulate temperature and precipitation in small-scale experiments, but do these manipulations produce a desired result when used together? Temperature manipulations have been done using heating cables (e.g. Grime 2001, Dunne et al. 2004), heaters (e.g. Norby et al. 1997), infra-red lamps (e.g. Shaw et al. 2002) and open-top chambers (OTCs, e.g. Marion et al. 1997).  In this study, I tested the performance of open-top chambers; they have been used to test the effects of warming on herbaceous plant communities in the Arctic (Henry and Molau 1997, Marion et al. 1997, Wahren et al. 2005, Rustad et al. 2001), grassland steppe in Tibet (Klein et al. 2004), in the alpine (Klanderud and Totland 2005) and in salt-marshes (Gedan and  Bertness 2009).  The design of OTCs usually consists of a translucent material walled around the experimental plot (Marion et al. 1997).  Generally, OTCs are approximately 1 m in diameter and 30 to 40 cm high, but larger OTCs (4.7 m diameter, 3.5 m high) have been used on trees (Whitehead et al. 1995). Marion et al. (1997) report temperature effects for different OTC  33  designs that range from slight reductions of soil temperature at some locations, but an average increase of 2.2 °C under optimal conditions; this is typical of other studies in the Arctic (Coulson et al. 1993) and Mongolian grasslands (Klein et al. 2005). Chambers also increased air temperature from <1 °C in 4.7 m diameter chambers (Whitehead et al. 1995) to increases as much as 6 °C in chambers with reduced top openings (Bremer et al. 1996), but a typical OTC design increases air temperature by 1 – 2 °C (Coulson et al. 1993, Klein et al. 2005). Despite the relatively small size of OTCs, they have been shown to have similar effects to natural landscape scale warming patterns (Hollister and Webber 2000) and temperature increases of less than 1°C are sufficient to induce change in soil respiration, nitrogen mineralization and aboveground plant biomass (Rustad et al. 2001). Soil moisture can be either increased by adding water, or reduced using a rainout shelter (RS). RSs have been used extensively in grassland ecosystems (Svejcar et al. 1999, Fay et al. 2000, Yahdjian and Sala 2002), but have been more variable in design than OTCs.  Typically, a translucent material is used to intercept falling rain.  These structures can either be fixed (Fay et al. 2000, Köchy and Wilson 2004) or moveable.  For example, Grime et al. (2000) used a motorized design with precipitation sensors that deploy the shelter only when it is raining to minimize unwanted shading by the shelter. The amount of rain intercepted has also been modified using slat like structures that can vary in size (Yahdjian and Sala 2002).  Regardless of design, they are effective at lowering soil moisture but have the unwanted effect of intercepting solar radiation. OTC and RS devices can be used in combination with other treatment types to understand the interactions between processes that alter plant communities. Disturbance of vegetation is a globally significant process; it is estimated that half of the terrestrial land mass is grazed by domesticated livestock (Havstad et al. 2008). Grazing can either increase (Bremer et al. 1998) or decrease (Vare et al. 1996) soil temperature depending on environment or season (Johnston et al. 1971).  Furthermore, live biomass or litter can also change soil temperature and moisture (Klein et al. 2004).  Thus, any disturbance that removes plant material may also alter soil temperature and moisture. Effects of warming, water availability and disturbance are likely to become complex when they interact. Changes in soil temperature can affect soil moisture and vice versa. Temperature increases will elevate evaporation rates while moisture, due to water’s high heat capacity, can increase the amount of energy required to raise temperature, but once heated will remain warmer for a longer period than drier soils. By measuring how soil temperature and  34  moisture are affected by climate and disturbance manipulations the validity of such experiments can be confirmed. I ran a 4-year experiment examining the interacting effects of OTCs, water availability, and the clipping of vegetation on the structure and species composition of semi-arid grasslands in the southern interior of British Columbia, Canada.  In a subset of the experimental plots, at the same location, I monitored soil temperature and moisture through two growing seasons.  Here I report the performance of OTCs and RSs in the semi-arid grassland and examine if these two climate manipulations produced their intended effects, whether clipping affected soil temperature and moisture, whether interacting treatments create unintended effects and if weather conditions affected performance. Methods Site description The study was done in Lac du Bois Grassland Provincial Park within the bunchgrass grasslands of the southern interior of British Columbia, 6 km north of Kamloops, Canada (UTM 10 E 680737  N 5625980; elevation 731 m a.s.l., Figure 3.1a).  The region is semi-arid with annual precipitation of 279 mm, 75.5 mm of which is snowfall.  Average annual daily temperature for the region is 8.9 °C, the warmest month is July (21.0 °C), and the coldest is January (-4.2 °C) (Environment Canada, 2009). The grasslands are dominated by Pseudoroegneria spicata (Pursh) A. Love (bluebunch wheatgrass) and the shrub Artemisia tridentata Nutt. (big sagebrush) (van Ryswyk et al. 1966). Other common species at the study site are Koeleria macrantha (Ledeb.) Shult. (june grass), Achillea millefolium L. (yarrow) and Astragalus collinus (Hook.) Douglas ex G. Don (hillside milkvetch). The soil is a brown Chernozem with a fine sandy loam texture (van Ryswyk et al. 1966). Air temperature and precipitation were measured at the site from May through October of each year using a rain gauge and temperature logger (Model RG3-M, Onset Computer Corporation, Bourne, USA), each of which were placed 1 meter above the ground. Experimental design Thirty-six 1 m2 experimental plots were located at a single south facing site with a slope of 13°.  While the surrounding grasslands were grazed by cattle, all of the plots were located within a fenced exclosure that had not been grazed for approximately 30 years.  All plots were  35  selected to exclude A. tridentata because the focus was on the herbaceous community.  Each plot was at least 1 m away from other plots and from any A. tridentata shrubs.  A fully factorial experiment included two warming treatments (plots were either warmed with an OTC or at ambient temperature), three precipitation treatments (plots with precipitation reduced using a RS, water increased by weekly hand watering, and natural ambient precipitation), and two clipping treatments (plots were clipped once annually in July to a height of 5 cm above the soil surface or left unclipped). The experiment was established in May 2005 and the first clipping occurred in July 2005.  OTCs and RSs were in place from April to October in 2005 through 2007 and from April to August in 2008. The factorial experiment had 12 treatment combinations, each of which was replicated three times at the site, thus I am reporting data monitored in 36 experimental plots. Open-top chambers The OTCs (Figure 3.1b) I used were similar to the plastic tent design described in Marion et al. (1997).  Each OTC had a square base, 1.5 m a side.  The plastic was secured to a wooden support driven into the ground.  The stakes were angled so that the top-opening was square, 1 m on a side and 40 cm above the soil surface. The plastic (Tufflite IV ™, 6 mil, 0.152 mm thick, Tyco Plastics and Agricultural Films, Monroe, LA, USA, 93% PAR transmission) has high transmission of photosynthetically active radiation,  repels dust from its surface (important for maintaining light transmission especially in the dust-prone study region), and is durable and economical. Water manipulations I constructed rainout shelters (Figure 3.1c) similar to the design of Köchy and Wilson (2004), which comprised a 1 m2 plastic sheet, the same material used in the OTCs, attached to a pole 1 m high at one corner and anchored at the remaining three corners such that they were each 30 cm above the soil surface to allow airflow.  The sheet was oriented to block rain from the dominant wind directions during the growing season. Some treatments had both a RS and an OTC; in this case, the RS was within the OTC but water off the RS would drain outside the border of the plot as it would in other RS plots (Figure 3.1c). Water availability in the plots was increased by hand-watering once a week from May through October of each year.  Once per week each plot received enough water to increase the monthly 30 year rainfall average by 30%. The historical (1970 – 2000) rainfall for the months  36  May through September were 24.4 mm, 35.2 mm, 29.5 mm, 29.1 mm and 28.0 mm (Environment Canada, 2009), thus I added  1.8 l, 2.6 l, 2.2 l, 2.2 l and 2.1 l each week in the respective months. Plots were hand-watered with care to ensure minimal runoff from the plot.  I used rainwater that was collected locally and stored in a black plastic cistern to minimize algal growth. Soil measurements Two response variables, soil temperature and soil moisture, were monitored in the center of each plot. A temperature probe (TMC50-HD, connected to a HOBO® U12 Data Logger, Onset Computer Corporation, Bourne, USA) was placed 5 cm deep into the soil. A soil moisture probe (10 cm long, Soil Moisture Smart Sensor, S-SMB-M005 using a ECH20® Dielectric Aquameter probe, Decagon Devices, Inc., Pullman, USA, connected to a HOBO® Micro Station data logger, Onset Computer Corporation, Bourne, USA) was placed vertically so that the top was 1 cm below the soil surface.  I installed the probes near the soil surface to measure the maximum effects of the treatments because treatment effects will dampen with depth (Marion et al. 1997). The probes were installed in May of each year and programmed to record measurements every 30 minutes.  Due to the sandy texture of the soil, data from the soil moisture probes were recalculated for the local soil type (Appendix B, Campbell 2002). Soil temperature and moisture were monitored from May to October in 2007 and from May to August in 2008. Statistical analyses Multi-factor repeated measure analysis of variance (RMANOVA) was used to test for treatment effects on mean daily temperature, data subsets based on the quartiles of the temperature in control plots, the daily minimum and maximum temperature, and daily temperature variance.  The same procedure was also used to test for treatment effect on mean daily volumetric water content (VWC), mean daily maximum VWC, mean daily minimum VWC and the daily variance of VWC. RMANOVAs that included TIME (hour, day or month) were also run to test within-subject effects of time.  In tests of month as the time factor, the response variables were mean daily temperature, daily temperature variance, mean daily VWC and daily VWC variance.  In the test of hour and day as a time factor the response variable was mean daily temperature.  A RMANOVA was run to test for differences of VWC in plots on the days after watering.  A Paired T-test was used to compare the VWC within plots on days when rain fell to  37  days when rain did not fall.  All data were approximately normal. All analyses were done using R version 2.7.0 (R Development Core Team, 2008). Results Site From May to October, the mean air temperature was 17.8 °C in 2007 and 17.1°C in 2008 (Figure 3.2). The hottest recorded air temperature was 45.8 °C at the end of June 2008. The coolest air temperature was -2.8 °C in early May 2007.  In 2007, 129 mm of rain fell; 34 mm more rain than 2008.  Most rainfall events were less than 5 mm per day, the largest daily rainfall was 15 mm in July 2007 (Figure 3.2). The average standing biomass at the site was 29.2 g / 0.25 m2 and the average amount of litter was 33.4 g / 0.25 m2. Soil temperature The mean soil temperature of control plots was 19.0 °C (range 3.1 to 49.0 °C), the mean daily maximum was 27.9 °C and the mean daily minimum was 12.6 °C (Figure 3.3a). Treatment effects on temperature Watering treatments and clipping altered soil temperature, but there was no significant effect of the OTC on mean temperature (Table 3.1; Figure 3.3a). The mean temperature in ambient water plots, water addition plots and RS plots was 19.4 °C, 18.8 °C and 20.3 °C respectively. Clipping increased temperature, clipped plots had a mean temperature of 19.8 °C and unclipped plots averaged 19.1 °C. The hottest treatment was the combination of a RS and clipping (Figure 3.3a, 3.4) that was on average 2.3 °C warmer than the control plot followed by the treatment that had an OTC and was clipped (1.9 °C warmer than control).  The coolest treatment was the water addition (1.3 °C cooler than control plots). There was a WATER x CLIPPING interaction effect on soil temperature (Table 3.1) such that clipping increased temperature in ambient water and RS plots but had no effect on temperature in water addition plots. The daily mean maximum, minimum and variance of temperature were all affected by treatments (Table 3.1).  The minimum temperature was affected by an OTC x WATER interaction where, in the RS plots, the OTC increased the mean minimum temperature from 12.3 °C to 13.5 °C. The mean daily maximum temperature of the plots was different due to water  38  treatments (ambient water = 28.6 °C, water addition = 27.4 °C and water removal plots = 30.3 °C).  Clipping increased the maximum temperature from 27.9 to 29.5 °C.  There was an OTC x WATER interaction where OTCs increased the mean maximum daily temperature in the water addition and ambient water plots, but decreased maximum temperature in the RS plots. As well, a WATER x CLIP interaction where clipping raised the maximum temperature in the water removal plots over 5 °C, from 27.7 to 33.1 °C.  The variance of temperature was increased by RS treatments (ambient water = 33.8, water addition = 30.3 and water removal plots = 42.8).  In the RS plots there was a difference of temperature variance between the OTC treatments; ambient plots had a variance of 50.2 while OTC plots had a variance of 35.2.  There was also a WATER x CLIP interaction where again, in the RS plots, clipping increased variance to 56.2 compared to 30.9 in the unclipped plots. Temporal effects on temperature The change in temperature varied hourly, daily and monthly and there were significant TREATMENT x TIME interactions (Table 3.2, 3.3), indicating that treatment effects were not consistent through time.   The effect of the water treatments and clipping follow the same general pattern as the main effects and their interactions.  On an hourly basis there was an OTC x WATER x HOUR and a WATER x CLIP x HOUR interaction. Similarly, there was an OTC x WATER x DAY and a WATER x CLIP x DAY interaction.  These analyses reveal significant time interactions with the OTCs suggesting that OTCs have an effect during some times of the day and not others. Over a daily cycle (Figure 3.4) within a treatment temperatures declined below those of the control or increase above that of the control. In the four months presented in Figure 3.5a-c, while temperature varied monthly there were no MONTH x main effect interactions (Table 3.3). There was a WATER x CLIP x MONTH interaction, where among RS plots, the ones that were clipped were warmer than the unclipped plots in June and July (Figure 3.5c). The variance of temperature varied by month, but not by treatments (Table 3.3; Figure 3.5e-g). Weather effects on temperature I separated the data based on the quartiles of the mean daily temperature in the control plots, and found that the OTCs, water and clipping treatments all affected soil temperature, but effects were dependent on the natural temperature (Table 3.1, Figure 3.6).  On the cooler days (<16.3 °C) the OTCs, water removal and clipping caused an increase in temperature.  On warm  39  days (>22.2 °C), OTCs had no effect, watering and clipping treatments responded and interacted with each other as described above.  On the warmest days (>27.2 °C, the quartile plus 5 °C) OTCs did not change temperature and there was an additional OTC x WATER interaction where in RS plots with an OTC were cooler than RS only plots, and the OTCs increased temperature in ambient water plots. Thus, OTCs significantly increased temperature on the coolest 54 days. Furthermore, on extremely cool days (<11.3 °C, the quartile minus 5 °C) the significance of the OTC effect on temperature increases, which suggests that the OTCs are more effective as ambient temperatures decline. Soil moisture The mean soil moisture of control plots was 11 % VWC, the range was 2 % to 35 % VWC, the mean daily maximum was 16 % and the mean daily minimum was 9 % VWC (Figure 3.3b). Treatment effects on moisture OTCs and RSs decreased mean, maximum and minimum soil moisture, while water addition tended to increase those three variables; clipping did not affect soil moisture (Table 3.4).  There were no significant interactions between the water treatments and either OTCs or clipping.  However, soil moisture levels in plots with both OTCs and RSs tended to be much lower than in any other plot (Figure 3.3b). Temporal effects on moisture No difference of mean daily soil moisture was observed in the water addition plots on the day water was added to the plots compared to the following days (Repeated Measure ANOVA, df =6, f-value =1.29, p = 0.257, there were no significant interactions) indicating that soil moisture levels were consistently elevated between watering events (Figure 3.7). In 2007, soil moisture in ambient water plots declined from June through August with a slight increase in September, which follows the same pattern as monthly rainfall that year. Rainout shelters kept soil moisture below control plot levels (Figure 3.5k), but the difference between these treatments and the control plots diminished from May to August. The addition of water raised the soil moisture levels early in the season, but the effect tended to decrease through the summer (Figure 3.5j).  By August there was little difference between the mean soil moisture of the three water treatments. There was an OTC x WATER x MONTH interaction and a  40  WATER x CLIP x MONTH interaction as well as a full 4-way interaction.  This pattern is likely due to the decrease in precipitation inputs; this trend is apparent in the control plots, where soil moisture decreased from June to August, but then increased in September (Figure 3.5l).  The mean daily variance of soil moisture showed similar variation by month (Figure 3.5m – o).  The variance of the control plots declined in August, the driest month.  The plots with the lowest variance were generally the RS plots. Weather effects on moisture There was a trend for precipitation events to increase soil moisture in all plot types, but the effect was not significant in the RS plots (Figure 3.3b). Discussion I have shown that OTCs, RSs, watering and clipping of vegetation all changed soil moisture and temperature.  Interactions between treatments, temporal variation and weather altered the effects of the treatments, confirming that multi-factor experiments can create complex effects within the manipulated environment. Temperature OTC effects on temperature I found variable effects of OTCs on soil temperature. Open-top chambers did not consistently increase temperature and their effect was at the lower end of reported ranges of increase. Open-top chambers in other studies of herbaceous communities have sometimes reduced temperatures but more typically raise temperature from 2.2 °C (Marion et al. 1997) to as high as 5 °C above ambient temperatures (Coulson et al. 1993, Marion et al. 1997). Plots with both OTCs and RSs had higher minimum temperatures than other plots, but also had lower maximum temperatures.  It is possible that under the hottest conditions the combination of an OTC and RS shaded the plot or prevented evaporating moisture from leaving the structure. Unintended effects on temperature Rainout shelters increased soil temperature. This is an unwanted effect but not unexpected and likely a realistic impact of future climate change because reduced precipitation will lead to declines in soil moisture that will lessen the amount of energy required to increase soil temperature.  It is also possible that the warming is a direct effect of the RS, rather than due  41  to a drop in soil water content.  Large RSs used in the Konza prairie also induced a soil temperature increase, but no air temperature increase (Fay et al. 2000).  The RSs were open to airflow so they likely did not increase air temperature. Slatted rainout shelters used in Argentina caused a reduction or an increase of soil temperature depending on natural climate conditions (Yahdian and Sala 2002); this may be due to the design that allowed more rainfall and airflow on the plot. Conversely, adding water to the plots nullified the warming effect of the chamber and reduced mean temperature below that of the control plot; however, among watered plots the ones with an OTC were warmer than those without.  Water addition has been reported to decrease soil temperature, but the effect diminishes with depth (Brown and Archer 1999). Clipping vegetation increased soil temperature.  This is likely due to an increased interception of solar radiation at the soil surface, as has been observed in other grasslands where clipping increased soil temperature up to 3°C in the Konza prairie (Bremer et al. 1998) and 1-1.7 °C in low-grazed shrubland in Mongolia (Klein et al. 2005). The 1.6 °C increase of soil temperature due to clipping that I report was likely less than that measured by Bremer et al. (1998) because of the large amount of standing biomass at their tallgrass prairie sites (700 g/m2) compared to my site (120 g/m2).  At my site, the vegetation is intercepting a smaller portion of the incoming solar radiation.  In contrast, grazing by reindeer in the Arctic removed the insulating moss layer and decreased temperature (Vare et al. 1996). In temperate grasslands, heavy grazing decreased wintertime soil temperatures, but increased summertime temperatures compared to lightly grazed sites (Johnston et al. 1971). Researchers concerned about temperature effects due to vegetation removal should be aware that responses will vary across systems and season. OTC effectiveness changes over time Temperature differences between treatments varied on an hourly, daily and monthly basis. This concurs with results from other studies that report a diurnal cycle in temperature increase due to OTCs, due to a strong relationship between solar radiation and heating (Whitehead et al. 1995, Marion et al. 1997). I also observed significant interactions across hours, days and months in the treatment effects on temperature.  This suggests that weather conditions may influence the effectiveness of OTCs in increasing temperature and the role the other treatments have influencing temperature.  42  Ambient temperature alters OTC warming effects Open-top chambers should be used with caution in studies trying to address warming effects on vegetation. OTCs have mostly been used in Arctic systems (Rustad et al. 2001, Walker et al. 2006) and were reported to be successful at increasing temperature (Marion et al. 1997).  I found that OTCs only significantly increased soil temperature when conditions were naturally cooler and the OTCs cause a slight cooling with some treatments when conditions were naturally warmer.  This suggests that OTCs may not be the ideal method of creating artificial warming in warmer climates and are better suited to cooler climates.  However, warming is expected to occur in most terrestrial systems and so investigations of warming effects should be undertaken in these systems. The use of OTCs as a warming tool should be adopted with caution especially if the system is warmer (in this study, OTCs caused no temperature increase on days when average soil temp of the control plot was >22 °C).  Soil moisture Treatment effects on moisture Open-top chambers caused a decline in soil moisture.  This could be a result of natural processes, such as evaporation or increased biological water demand; however, a decline in soil moisture may be an unwanted effect of the OTC.  Although the chambers are open on top, there is a 0.25 m border around each plot that is covered by the OTC, the chamber wall could intercept rainfall preventing water from penetrating the surrounding ground.  Personal observation and the pattern of mean soil moisture in plots suggest that this was not the case.  Only the heaviest of rainfall events wet the soil beyond a centimetre in this arid environment, indicating that water was trapped in the dry soil and did not travel far through the soil. The relative soil moisture between water addition and ambient water plots was similar in the respective plots with and without an OTC. Had OTCs been intercepting enough rainfall to lower soil moisture, I might expect the soil moisture level in water-addition plots with an OTC to be relatively higher than ambient water plots compared to plots without an OTC. On the Mongolian steppe, which receives monsoon rains, there was no observable effect on soil moisture due to the OTC (Klein et al. 2005). However, it is not unreasonable to expect changes in soil moisture due to the presence of an OTC and this variable should be monitored as a potential covariate. Rainout shelters significantly reduced soil moisture, while water addition only slightly increased soil moisture levels.  Arid and semi-arid grasslands are likely most susceptible to  43  declines in water availability because water is most likely to be a limiting resource. Consequently, rainout shelters have been used extensively in grasslands to test the effect of water reduction on plant communities (Yahdjian and Sala 2002, Fay et al. 2000, Köchy and Wilson 2004, Svejcar et al. 1999) and have all been successful at reducing soil moisture levels. Due to the variety of designs and sizes of RSs it is difficult to directly compare different methods. I added 30% more water to plots and, although there was a measureable increase in soil moisture, it was not significant.  Early in the season when temperatures are cooler, added water likely has the ability to penetrate deeper into the soil. Later in the season, however, as temperature rises and natural precipitation decreases, added water does not penetrate deeply into the soil and will be subject to rapid evaporation. Brown and Archer (1999) found that heavy watering increased soil moisture to a depth of 150 cm by more than 10 % VMC in savanna parkland.  However, they added substantially more water than in this experiment – 100 mm of water every two weeks compared to the 1.8 – 2.6 mm range I added weekly. Again, lack of treatment significance does not translate into a lack of biological significance as I did observe a positive response of vegetation to the water additions (chapter 2). Water pulsing Temporal variability of resources (pulsing) can alter the outcome of plant-plant interactions (Novoplansky and Goldberg 2001) and plant community diversity (Knapp et al. 2002).  For this reason, I was concerned that adding water may pulse soil moisture, increasing its variability, or that RSs may intercept precipitation in a way that lowers soil moisture variability. I found no difference in mean daily soil moisture averaged over the duration of the experiment following hand watering of the plots.  I did find that the daily variance of soil moisture was different among the watering treatments overall and in different months.  Soil moisture variance in the RS plots was always lower than in the control plots, this is likely due to the prevention of precipitation increasing soil moisture. This pattern raises the possibility that any observed responses in the plant community are due to temporal variance rather than resource availability. A dedicated experiment would be required to determine the cause of vegetation change because declines in resource availability, in high stress environments (Grime 2001), and declines in resource variance (Knapp et al. 2002) both predict a decline in species diversity.  44   45 Experimental design recommendations In this study I showed that OTCs and RSs were able to warm or reduce soil moisture in small scale plots in temperate grasslands.  However, the devices interacted with each other and other treatments, their effects were altered by weather, and they did not act exclusively on their target variables – OTCs altered soil moisture and RS altered soil temperature. Other treatments modified the effects of the OTC, RSs increased the temperature 0.4 °C, watering decreased the temperature 0.4 °C, and clipping raised temperature 2 °C. Similarly, changes in soil moisture due to the RSs decreased VWC by 3.1 % and increased 1 % by clipping; and soil moisture due to watering was reduced 1 % by the OTC and clipping. These observations raise a number of experimental design issues. The warming caused by OTCs may be limited to naturally cooler ambient conditions.  OTCs are passive devices, experiments requiring consistent warming or occurring in warmer climates may require other types of devices (e.g. heating cables or infra-red lamps).  Advanced testing of both OTCs and RSs in new regions and ecosystems is encouraged. Rainout shelters were effective at reducing soil moisture but they also reduced the variability of soil moisture levels; the experiment can be designed to remove these effects.  For example, Köchy and Wilson (2004) placed RSs over all experimental plots in order to control for the effects of the RS and then added different amounts of water to each plot.  This approach would also control for light intercepted by the RS. Water additions lowered soil temperature, while clipping increased temperature and lowered moisture. These effects are unavoidable without more elaborate devices, but should be considered when interpreting results even though it may be difficult to separate the intentional and unintentional treatment effects. As final recommendations, some monitoring of soil moisture and temperature is necessary in climate manipulation experiments even if one is not the target variable. While experiments with multiple factors are necessary to understand the complex interactions that alter plant communities, I urge caution when using these devices because of unintended effects.   Table 3.1 Repeated measure ANOVA testing the effect of treatments on mean daily temperature. Data were analysed separately for days of different temperatures separated on the basis of daily mean temperature in control plots (quartiles and quartiles plus 5 °C). The minimum, maximum and variance of all data were also tested. Temperature was modified using open-top chambers (OTC), water was either added, reduced with a rainout shelter or at ambient, and plots were clipped or not clipped.  Significant (P<0.1) effects are in bold. Reprinted with permission of Springer Science + Business Media, LLC.  Temperature (all data) Coolest days <11.3 °C (11 days) Cool days < 16.3 ºC (54 days) Warm days >2.2 ºC (56 days) Warmest days >27.2 °C (5 days) Min Max Variance  Df F- value P F- value P F- value P F- value P F- value P F- value P F- value p F- value P OTC 1 1.19 0.286 8.56 0.007 4.99 0.035 0.001 0.997 0.44 0.511 1.14 0.296 0.13 0.725 0.06 0.811 WATER 2 16.85 <0.001 9.90 <0.001 22.34 <0.001 7.34 0.003 4.44 0.023 1.12 0.342 5.71 0.009 3.15 0.061 CLIP 1 12.02 0.002 4.12 0.054 11.47 0.002 4.22 0.051 3.38 0.079 0.05 0.819 5.29 0.030 2.9 0.143 OTC x WATER 2 1.03 0.371 0.097 0.908 0.21 0.808 2.24 0.128 3.17 0.060 2.67 0.090 2.73 0.086 3.01 0.068 OTC x CLIP 1 0.004 0.953 0.14 0.715 0.07 0.795 0.01 0.909 0.01 0.914 0.17 0.681 0.01 0.915 0.39 0.540 WATER x CLIP 2 4.68 0.019 0.21 0.816 0.87 0.429 3.79 0.037 2.58 0.097 2.28 0.123 6.77 0.004 5.39 0.012 OTC x WATER x CLIP 2 1.19 0.323 1.31 0.287 1.14 0.336 0.73 0.492 1.09 0.351 0.37 0.692 1.51 0.240 1.53 0.237 Residuals 24    46  Table 3.2 Repeated measure ANOVA testing effects on mean hourly temperature and mean daily temperature with time of the day (measured on the hour and half hour) and day as factors (only the within-subject effects are shown).  Temperature was modified using open-top chambers (OTC), water was either added, reduced with a rainout shelter or at ambient and plots were clipped or not clipped.  Significant (P<0.1) effects are in bold. Reprinted with permission of Springer Science + Business Media, LLC.   Hour Day  df F-Value P F-value P TIME (hour or day) 1 653.59 <0.001 25.7 <0.001 OTC x TIME 1 0.634 0.974 0.32 0.569 WATER x TIME 2 1.31 0.029 14.0 <0.001 CLIP x TIME 1 1.57 0.009 0.72 0.397 OTC x WATER x TIME 2 1.75 <0.001 1.93 0.146 OTC x CLIP x TIME 1 0.25 1 4.73 0.030 WATER x CLIP x TIME 2 3.55 <0.001 5.46 0.004 OTC x WATER x CLIP x TIME 2 0.50 1 0.16 0.852 Residuals Hour =1128 Days = 7695   47  Table 3.3 Repeated measure ANOVA of effects on soil temperature and moisture and their variance with month as a factor (only within-subject effects are shown). Temperature was modified using open-top chambers (OTC), water was either added, reduced with a rainout shelter or at ambient and plots were clipped or not clipped.  Significant (P<0.1) effects are in bold. Reprinted with permission of Springer Science + Business Media, LLC.  Temperature Temp variance Moisture Moisture variance  df F – value P F – value P F – value P F – value P MONTH 3 1322.37 <0.001 12.78 <0.001 66.50 <0.001 44.65 <0.001 OTC x MONTH 3 0.17 0.913 0.40 0.752 3.62 0.017 1.49 0.225 WATER x MONTH 6 1.38 0.236 0.39 0.886 14.08 <0.001 8.44 <0.001 CLIP x MONTH 3 1.62 0.193 0.87 0.462 0.20 0.894 2.42 0.073 OTC x WATER x MONTH 6 0.84 0.542 0.40 0.874 3.54 0.004 2.32 0.042 OTC x CLIP x MONTH 3 0.34 0.795 0.05 0.982 0.48 0.700 0.29 0.836 WATER x CLIP x MONTH 6 2.11 0.062 1.64 0.148 6.73 <0.001 3.50 0.004 OTC x WATER x CLIP x MONTH 6 0.43 0.855 0.13 0.992 1.93 0.088 0.44 0.850 Residuals 72   48  Table 3.4 Repeated measure ANOVA testing the effect of treatments on mean daily soil moisture, daily maximum moisture, daily minimum moisture and variance of moisture. Temperature was modified using open-top chambers (OTC), water was either added, reduced with a rainout shelter or at ambient and plots were clipped or not clipped. Significant (P<0.1) effects are in bold. Reprinted with permission of Springer Science + Business Media, LLC.   Mavg Mmax Mmin Mvar  Df F- value P F-value P F-value P F-value P OTC 1 4.98 0.035 5.47 0.028 4.82 0.038 0.70 0.411 WATER 2 9.10 <0.001 10.54 <0.001 8.13 0.002 22.36 <0.001 CLIP 1 0.31 0.582 0.36 0.551 0.28 0.604 0.44 0.513 OTC x WATER 2 0.40 0.675 0.47 0.632 0.36 0.702 6.20 0.007 OTC x CLIP 1 0.23 0.633 0.28 0.601 0.22 0.642 0.00 0.999 WATER x CLIP 2 0.33 0.719 0.47 0.629 0.25 0.782 7.58 0.003 OTC x WATER x CLIP 2 0.29 0.752 0.22 0.797 0.32 0.727 2.68 0.088 Residuals 24   49  Figure 3.1 Photos of a) experimental site in Lac du Bois Provincial Park, Canada, b) an open-top chamber (foreground) and c) a rainout shelter and open-top chamber. Reprinted with permission of Springer Science + Business Media, LLC.  50  Figure 3.2 Mean daily air temperature (grey lines) and daily rainfall totals (black bars) at the site.  Data were collected during the growing season from May to October in 2007 and 2008. Reprinted with permission of Springer Science + Business Media, LLC.  51  Figure 3.3 a) Mean (± 1 SE) soil temperature and b) mean (± 1 SE) soil moisture in all twelve treatment combinations.  Precipitation in the plots was at ambient (Wc), increased with hand watering (W+) or reduced with rainout shelters (RS). Plots either had an open top chamber (OTC) or not (No OTC). Plots were clipped (C+) or were not clipped (C-). “Control” indicates unmanipulated plots for all treatments. In a) triangles indicate the mean daily maximum temperature and inverted triangles indicate mean daily minimum temperature.  In b) circles indicate mean daily soil moisture on days with rain (closed circle) and days without rain (open circles); “*” above the bars indicate a significant difference of mean daily soil moisture between days with (n = 67) and without (n =153) precipitation, “ns” = no significance (Paired T-test, P <0.05). Reprinted with permission of Springer Science + Business Media, LLC.  52  Figure 3.4 Mean hourly soil temperatures measured at 5 cm depth (averaged over days and replicates) in all treatments. The top right plot is soil temperature in the control plot (mean temperature 19.0 oC). All other panels show the mean temperature (black line, right axis) of treatment plots. The grey line is difference between soil temperature of the treatment plot and the control plot (left axis). The number in the top right indicates the mean temperature change relative to the control and the horizontal line indicates zero change relative to the control. Reprinted with permission of Springer Science + Business Media, LLC.  53  Figure 3.5  Mean difference in soil temperature (a-c) and moisture means (i-k) and their respective variances (e-g, m-o) in the different treatments compared to the control plots measured from June through September 2007. Also presented on the right axes are soil temperature means (d) and soil moisture means (l) and their variances (h, p) in the control plots. The treatment combinations are represented by different lines: OTC (Open-top chamber) and clipping (solid line), OTC without clipping (dashed line), no OTC with clipping (dotted line), no OTC without clipping (dash and dot line).    In panels (l to r), plots received ambient rainfall, had 30% of average monthly rainfall added weekly, or were covered with a rainout shelter. The solid horizontal line indicates zero difference from control plots. Reprinted with permission of Springer Science + Business Media, LLC.   54   55  Figure 3.6 Mean (± 1 SE) soil temperature in plots with (dark) and without (light) open-top chambers in different ambient conditions.  Ambient conditions were selected based on the quartiles of mean daily soil temperature in the control plots as well as extreme temperatures (quartiles ± 5 °C).  Numbers above the bars indicated the number of days within the range.  “*” above the bar indicates a significant difference within the pair (Repeated measures analysis of variance, P < 0.05, Table 1). Reprinted with permission of Springer Science + Business Media, LLC.  Figure 3.7  Mean daily soil volumetric water content (% VWC), measured in the top 10 cm of each plot (a-c) and mean (+1 SE) daily rainfall for the same days (d). The treatment combinations are represented by different lines: OTC (Open-top chamber) and clipping (Solid line), OTC without clipping (dashed line), no OTC with clipping (dotted line), no OTC without clipping (dash and dot line).   In the panels plots received ambient rainfall (a), had 30% of average monthly rainfall added weekly on day zero (b), or were covered with a rainout shelter (c). Reprinted with permission of Springer Science + Business Media, LLC.  56  Chapter 4: Using three pairs of competitive indices to test for changes in plant competition under different resource and disturbance levels Introduction Competition is an important factor that structures plant communities (Tilman 1988; Grime 2001; Keddy 2001), but there are conflicting predictions as to how important it is depending upon the productivity of the habitat (Grime 1977; Tilman 1985; Thompson 1987; Grace 1995; Craine 2005). The two predominant arguments are CSR Strategy theory, stating that competition increases with productivity (Grime 1977); and, R* theory (Resource-ratio hypothesis) that states competition is consistent along the productivity gradient but switches from below ground for water and nutrients to above ground for light as belowground resources become more available (Tilman 1982; 1988).  It has been suggested that the difference between these two theories is semantic rather than of disparate ecological understanding (Welden and Slausen 1986; Grace 1991; Brooker et al. 2005) and even that the debate is unnecessary because the concepts of competition intensity and competition importance distinguish the two theories (Brooker and Kikvidze 2008). Evidence to support these two theories has been divided and perhaps limited (Goldberg et al. 1999; Wilson and Lee 2000; Miller et al. 2005; Wilson et al. 2007; Miller et al. 2007); support for either theory can change depending on which index is used to measure competition (Turkington et al. 1993; Grace 1995), and there are many (Weigelt and Jolliffe, 2003). It is important to have a clear understanding of how competition might be affected by productivity, especially because both theories impact current understanding of biodiversity.  For example, CSR Strategy theory and R* theory both predict a hump-shaped relationship between species richness and productivity (Grime 2001; Tilman 1982).  Neither CSR theory nor R* theory examine competition as a process in isolation.  CSR theory examines the outcome of competition as a trade-off between competitive, ruderal and stress-tolerant plant strategies. Similarly, R* theory is based on a competition model, but resource availability is determined by other processes such as loss-rate (disturbance).   Furthermore, while CSR and R* theories predict how the level of disturbance and stress will affect competition, these are generally not included in tests of competition. There are many measures of competition (Weigelt and Jolliffe 2003), but three sets of measures have been used to explain and differentiate the sometimes conflicting predictions of CSR and R* theory: absolute and relative competition (Grace 1995; Campbell and Grime 1992;  57  Turkington et al. 1993; Wilson and Tilman 1995), effect and response competition (Goldberg 1990, 1996; Goldberg and Landa 1991), and the importance and intensity of competition (Welden and Slauson 1986; Brooker et al. 2005).  I will use the terminology above in an effort to reduce confusion resulting from the multiple names that have been associated with similar formulas (Table 4.1) (Weigelt and Jolliffe 2003). Absolute and relative competitive ability have been shown to support both the theories of Grime and Tilman respectively (Campbell and Grime 1992; Grace 1993; Turkington et al. 1993). Absolute competition is estimated as reduction in biomass due to neighbours and should therefore decline with productivity, while relative competition is standardized by biomass production in the absence of competition and is more likely to remain constant along a productivity gradient (Grace 1995; Goldberg and Scheiner 1993). Competitive effect and response have also been suggested to differentiate the two theories.  Competitive effect is the ability of a plant to suppress its neighbours and correlates with plant traits such as growth rate which concurs with Grime’s definition of competition (Goldberg 1990; 1996).  Competitive response is the ability of a plant to withstand the effects of its neighbours, but it  has not been so clearly linked to plant traits (Keddy et al. 1998; Cahill et al. 2005; Carlyle and Fraser 2006) and  reflects Tilman’s definition of a competitor – the species that can survive at the lowest resource availability (Goldberg 1990). Finally, competitive importance and competitive intensity have been used to distinguish the two theories (Welden and Slausen 1986; Brooker and Kikvidze 2008). The importance of competition is a relative measure of the effect of competition at a point along the gradient relative to other processes and tends to support CSR theory; the intensity of competition is a relative measure of competition at a single point along the gradient and has been associated with R* theory (Brooker et al. 2005; Brooker and Kikvidze 2008). Only importance of competition can incorporate the role of other processes in describing the impact of competition – the importance of competition should decrease when other processes such as disturbance or stress are present. Given the association of absolute competition, competitive effect and competitive importance with CSR theory, these indices should predict an increase with increased productivity.  In contrast, R* theory is based exclusively as a competitive model and thus under all conditions R* theory predicts that these measures will remain constant (Tilman 1982; 1988). I conducted a factorial experiment to assess the role of competition between two grass species under combinations of stress (nutrient, water) and disturbance (clipping) treatments.  I assessed the three sets of competitive indices under each combination of treatments to test if 1)  58  the pairs of competitive indices differentiated the predictions of CSR and R* theory and 2) do the indices measure a decrease in the level of competition due to disturbance (loss-rate)?  In addition, I tested if the results provided by different indices are consistent with predictions regarding how disturbance alone, and interacting with resources, affects competitive outcomes. Materials and methods Study species Pseudoroegneria spicata (Pursh) A. Love (bluebunch wheatgrass) and Festuca campestris Rydb. (rough fescue) are temperate bunchgrasses native to the southern interior of British Columbia, Canada.  Both species occur along an elevation gradient with P. spicata tending to dominate in drier less productive elevations and F. campestris tending to dominate at wetter more productive elevations. Experimental design The experiment was done in the Research Greenhouse at the Thompson Rivers University Campus, Kamloops, British Columbia, Canada.  Greenhouse conditions were electronically controlled for the duration of the experiment to maintain day-time conditions at 22 °C and 60% relative humidity, and night-time conditions at 15 °C and 85% relative humidity, which are within the range of local, natural growing season conditions.  Supplemental lighting was supplied by three 1000 W halogen sulphide lamps in a 14:10 hour day: night cycle. Seeds of P. spicata and F. campestris were collected from multiple locations in Lac du Bois Provincial Park, British Columbia, Canada (UTM 10 North 680737 5625980; 4 to 10 km north of Kamloops).  After three months of cold storage, the seeds were placed in Petri dishes on a bed of damp sand to germinate.  Seedlings with a radical at least 30 mm long were transplanted into 240 ml pots (6.4 x 6.4 cm opening, 4 x 4 cm base and 8.9 cm tall) containing clean sand saturated with 70 mL of Rorison’s solution (Hendry and Grime 1993).   After one week, dead seedlings were replaced with seedlings that had been planted individually in identical pots at the same time.  All pots received a top watering of 50 mL Rorison’s solution every five days and a bottom watering of distilled water when needed to maintain 5 mm of standing water for the first 21 days of the experiment.  After 21 days the pots were subjected to their respective nutrient and watering treatments.  59  The four factors examined in this study (water, nutrients, clipping and competition) each had two levels. Each species was planted in either competition or not in competition, so there were 3 planting combinations: P. spicata alone, R. fescue alone, and both species together. All factors were combined for a total of 16 pot combinations for each species. Each combination was replicated 20 times in 4 blocks (5 replicates per block). Thus, this experiment had 480 pots (planting combinations (3) x water (2) x nutrients (2) x clipping (2) x replicates (20) = 480). All pots received 30 ml of Rorison’s solution every 5 days. High nutrient treatments received regular strength Rorison’s solution; low nutrient treatments received a 1/10 dilution.  High water treatments received a bottom watering of distilled water as needed to maintain the pot sitting in 5 mm of water; low water treatment pots received the top watering of nutrients but did not receive any bottom watering.  Plants that received the clipping treatment were clipped after 45 days to reduce their leaves by 75% length (Hendry and Grime 1993). Competition was created by placing an individual of both species of grass together in a pot; no competition treatments had a single individual in the pot. All plants were clipped at ground level 90 days after planting; the biomass was dried for 48 hours at 65oC and then weighed. Attempts to separate roots of the two species could not be done with certainty, so only aboveground measures of biomass were used to estimate the competition indices. Competitive indices The five competitive indices compared in this study were: absolute competition (Cab), competitive intensity (Cint), competitive effect (Ce), competitive response (Cr) and competitive importance (Cimp) (Table 4.1).  Ce, Cr, and Cint are essentially equivalent calculations. Ce is the effect of the target species on its neighbour while C  e.  int is the effect of competition on the target species from its neighbour, and because I am dealing with only two species, the calculation of Cint for P. spicata is identical information to the calculation of Ce for F. campestris, which is why only five indices are examined rather than six.  Competitive response is equivalent to 1- C I have included both calculations to help illustrate the different competition indices.  These calculations would not necessarily be identical if a larger species set or community set were used.  Each index was calculated for the each plant grown with a neighbour compared to the mean dry biomass of that species grown alone. Each index of competition was applied to all pots.  60  Data analysis The effect of water treatments (high, low), nutrient treatments (high, low), clipping (clipped, unclipped) and block on the biomass was analysed separately for each species with an analysis of variance (ANOVA). Significant blocking effects in the final plant biomass caused us to calculate competitive measures on a block by block basis (i.e. competitive indices were calculated with target plants and plants grown without neighbours from within the same block). The effect of water treatments, nutrient treatments, clipping and block on each competitive index was analysed separately for each species. Block was not significant in the ANOVA of competitive indices and was not included in the final analyses; including block in the analyses did not change the interpretation of the analyses. All analyses were done using R version 2.7.0, R Development Core Team, R Foundation for Statistical Computing, Vienna 2008. Results Biomass The biomass of both species was reduced by four of the treatments: low water, low nutrients, clipping and competition; however, the two species responded differently to treatment interactions (Table 4.2). Pseudoroegneria spicata had a decrease in biomass with decreasing resource availability but there was no difference in the final biomass between the high water – low nutrient and low water – low nutrient treatments (Figure 4.1a).  Under the high water – high nutrient conditions, clipping lowered biomass of P. spicata, and there was a clipping x competition interaction.   Festuca campestris did not respond to any treatments when nutrients were low.  Under high nutrient conditions the presence of the competitor had the largest effect. Clipping reduced the biomass of F. campestris only under the high water – high nutrient conditions (Figure 4.1b). Competitive indices Absolute competition (Figure 4.2e, j) declined with decreasing resource availability for both species.  There was a significant three-way interaction between water, nutrients and clipping for Cab for F. campestris; under the high-water high nutrient conditions clipping decreased Cab (Table 4.3). Competitive intensity (or relative competition) was constant across all treatment combinations for P. spicata (Figure 4.2a, Table 4.3). Nutrients, water and clipping all had  61  significant effects on competitive intensity for F. campestris but not in a consistent manner (Figure 4.2f, Table 4.3). Similarly, the measures of Cr for P. spicata (Figure 4.2d) and Ce for F. campestris (Figure 4.2h) (both based on identical calculations) did not vary across treatments. However, Ce for P. spicata (Figure 4.2c) and Cr of F. campestris (Figure 4.2i) (also identical calculations), responded to water, nutrient and clipping treatments but these were not consistent across treatments. Competitive importance declined for both species with declining resources, although the trend was more pronounced for F. campestris (Figure 4.2b, g). Additionally, Cimp was reduced for F. campestris under the high water treatment x clipping, and all possible interactions were significant (Table 4.3). Discussion I have shown that resource availability interacts with disturbance to affect competition but it depends on the index used.  Others have shown that the index used to assess competition can influence how we interpret the effects of competition under changing resource availability (Campbell and Grime 1992; Turkington et al. 1993; Grace 1995; Wilson and Tilman 1995; Brooker et al. 2005).  Two pairs of indices, competitive importance – competitive intensity, and absolute competition – relative competition, distinguished CSR and R* theory along resource and disturbance gradients, while the competitive effect – competitive response pair of indices did not. Interacting effects on biomass Effects of treatments on biomass production were as expected; clipping, low water, low nutrients and competition individually lowered the biomass of each species.  However, additive effects of combinations of treatments tended to decrease biomass production more than any single treatment.  Furthermore, the biomass response by the two species varied due to treatment interactions. Dominance is important when interpreting the results of competition experiments because the effects of competition may be less apparent in dominant species than in subordinates (Goldberg 1987; Schwinning and Weiner 1998).  Competitive ability has been positively correlated with plant size (Gaudet and Keddy 1988) and growth rate (Grime 2001). Pseudoroegneria spicata is the larger species (Figure 4.1) and also has a faster relative growth  62  rate than F. campestris, (from 2 to 21 days: 0.49 g.g-1.day-1 and 0.22 g.g-1.day-1 respectively, C. N. Carlyle unpublished data), suggesting that P.  spicata is the better competitor and dominant species.  This is an important consideration when interpreting the results of this experiment because P. spicata was less affected by treatments than F. campestris (Table 4.2). Pseudoroegneria spicata only experienced significant biomass reductions when both clipping and competition were present, and when in high resource conditions; similar effects were apparent in a field study examining both drought and clipping on P. spicata (Busso et al. 1990). However, when grown in competition with Centaurea maculosa, P. spicata exhibited facilitative effects from the presence of the neighbour, but total foliage was reduced by increasing clipping severity (Kennett et al. 1992).  At the end of this study, the effects of clipping were not apparent on the biomass production of P. spicata, except in the high water- high nutrient treatment, but it is likely that a more severe clipping treatment, created by either increasing the amount of biomass removed or subjecting the plant to multiple clippings, would have produced more significant reductions in plant biomass as has been observed in a field study with this grass (Arredondo and Johnson 1998). The resistance of F. campestris to clipping, in all but the high resource, competition-free case, is inconsistent with field experiments (Willms and Fraser 1992; Vujnovic et al. 2000), but again the clipping regime on comparatively younger plants may not have been severe enough to induce a response.  However, this does suggest ability by both of these species to compensate for the effects of clipping and that the effect of clipping is reduced as conditions become more stressful. Comparing the pairs of competitive indices Resource availability (i.e. productivity or stress) and disturbance influence measures of competition and structure communities but, with only a few exceptions, resources have dominated research questions and the interaction of the two processes examined less frequently (but see Campbell and Grime 1992; Turkington et al. 1993). However, because indices of competition have generally been designed to examine the influence of a neighbour they can still be applied to a range of environmental conditions that includes interactions (but see Goldberg and Scheiner 1993; Grace 1995) to test the consistency of methods examining changes in competition.  63  Absolute competition and relative competition In general, Cab supports the predictions of CSR theory, and Cint supports the predictions of R* theory.  Absolute competitive indices and relative competitive indices (Cint) have been extensively compared and reviewed in the long debate examining how competition changes along productivity gradients and their advantages and disadvantages have been discussed (see Goldberg and Scheiner 1993; Grace 1995; Miller 1996; Oksanen et al.  2006). These indices have also been criticised for their inability to account for changes in plant size that will occur in different environments (Wilson 2007). Absolute competition is expected to show a decline with resource availability because the actual amount of biomass produced under low resource conditions is less; this is consistent with previous studies examining these indexes along resource gradients (Campbell and Grime 1992; Turkington et al. 1993; Wilson and Tilman 1995). These observations are consistent with CSR theory, where one process should decline when other processes are also present. Clipping did not have a consistent effect on Cint, but it did lower the measure of absolute competition for F. campestris under the high resource treatment, which suggests that under some circumstances clipping can influence Cint. My results concur with field studies that examined these two indices of competition. Cab increased with increasing resource availability and declining disturbance (Campbell and Grime 1992), while Cint did not change along a productivity disturbance gradient (Wilson and Tilman 1995).  Turkington et al. (1993) applied both Cab and Cint along a productivity-disturbance gradient and showed that Cab supported CSR theory while relative measures of competition supported R* theory. The application of Cint and Cab in tundra communities with different productivity levels showed that competition decreased when herbivory was present in lower productivity sites but with no difference in competition between sites of differing productivity(Olofsson et al. 2002).  Cab responded to both disturbance and productivity while Cint was generally consistent along both gradients in a manner similar to Cr and Ce which is du to the identical calculations used to measure these concepts of competition. e  Competitive effect and competitive response Contrary to Goldberg’s (1990) assertion, Ce and Cr did not differentiate the two theories; neither measure decreased with lowered resource availability although the Ce of P. spicata on F. campestris and the Cr of F. campestris showed some variation among the treatment combinations.  64  Diffuse competition (a variant of competitive intensity and competitive effect) measured along a lake shore increased with both lower disturbance (wave action) and higher productivity (Wilson and Keddy 1986). A negative relationship between Ce and stress and a positive relationship between Cr and stress was found when eight species of wetland plants were examined along a water depth gradient, suggesting support of CSR theory for both measures of competition (Fraser and Miletti 2008). When root and shoot competition were measured separately using a natural log of competitive response no change in competition was observed above or below ground (Cahill 2002).  Defining competitive response has been especially problematic but the solution may require a trait-based approach (Keddy et al. 1998; Cahill et al. 2005; Carlyle and Fraser 2006). Conflicting findings for the response of these measures suggest that further investigation is required to determine the relationship between Ce and Cr. There is a logical inconsistency in the way different indices have been used to describe the two theories.  Ce has been associated with CSR theory, and Cint with R* theory, but both measures of competition use the same formula in their calculation. I found that neither Ce nor Cint declined with decreasing resource availability, hence these two measures should both concur with R* theory.  Additionally, the formula commonly used to calculate Cr (Table 4.1 equation 3) is the inverse of Ce (Table 4.1 equation 1), so it is impossible for these two equations to show different trends in the strength of competition.  I used calculations of competitive effect and response that differed from the original calculations using regression methods applied over a range of competitor densities (Goldberg and Werner 1983, Goldberg and Fleetwood 1987). The equations I used in this study have been commonly used to calculate Ce and Cr (Table 4.1) and while these are still useful concepts these equations should be used with caution when investigating competitive effect and response relationships. Competitive intensity and competitive importance Competitive importance and competitive intensity differentiated the two theories; representing CSR and R* theories respectively.  The pattern of Cint along the gradients has already been discussed because, as pointed out earlier, Cint is identical to Ce and it was already paired with Cab as “relative competitive intensity”. Cimp showed a distinct decline with reduced resource availability for F. campestris; the importance of competition was greatest when resources were most available.  The trend is also apparent for P. spicata but only significant between the extremes of resource availability when unclipped.  This was not surprising because biomass production by P. spicata was not reduced by the presence of a competitor.  65  These results agree with those reported by Brooker et al. (2005), where Cimp declined with standing biomass but Cint did not. In a comparison of three plant species Cimp decreased with fertility but it was suggested that competitive importance was dependent on the ability of the species to tolerate low resource availability (Gaucherand et al. 2006).  Sammul et al. (2000 using a different calculation of competitive importance, reported a correlation between both C  ),  if  int and Cimp when competition was calculated using the number of shoots rather than biomass,  and Welden et al. (1988) detected no relationship between importance or intensity and a water stress gradient for desert shrubs.  I observed a reduction in competitive importance due to clipping under the high resource conditions for F. campestris, which further supports arguments that competition in CSR theory is best measured with this index because the measure of competition declines when additional processes are in place.  I are unaware of any other study that explicitly tests changes in Cimp due to disturbance; however, the findings suggest that Cimp concurs with the predictions of CSR strategy theory.  Recalculating Cimp on existing data sets might confirm this result is more common. I have also provided further evidence that a contributing factor to the long debate of how competition changes along a productivity gradient is the use of competitive indices.  In this study, two of the pairs of indices can lead to different conclusions depending on which index is selected.  Studies of plant competition should use more than one index when reporting results. In this study Ce and Cr lead to the same conclusion, thus these indexes should not be used in contrast.  Cint pairs against both Cab and Cimp, but Cab is more likely to show changes along a productivity gradient (Grace 1995).  Thus, I suggest reporting both Cimp and Cint together. My study used a simple system, only one or two plants growing in a single pot, but ultimately ecologists are interested in the plant-plant interactions that structure natural communities. Indexes of competition have been criticised for not addressing community-level responses over longer periods of time (Frekelton et al. 2009).  Pairwise competition experiments provide a controlled setting in which to investigate the mechanism of plant interactions even though they do not always reflect field distributions (Engel and Weltzin 2008). This problem is particularly evident in the competitive effect and response pair of indices which, in a pairwise experiment, will provide identical results. Disturbance decreased Cab and Cimp under high resource conditions. While this result has ecological significance it also creates an important consideration when testing for changes in competition along productivity gradients.  If disturbance is not controlled it would be possible to reject the hypothesis that competition declines with productivity if disturbance processes reduce  66  competition at high productivity while having no impact at low productivity, essentially creating equal measures of competition at both high and low productivity. Disturbance effects on competition should be accounted for, especially in field experiments, where it may be difficult to determine the extent of disturbance. Conclusions Levels of stress and disturbance can alter competition indices.  Cimp and Cab declined with decreasing resource availability and were also reduced by disturbance thus supporting their association with CSR theory.  Cr and Cint did not decline with resources or clipping in a manner consistent with CSR theory and were generally unchanged along both gradients, confirming their association with R* theory.  Ce, which has been associated with CSR theory did not decline along either gradient, and given that it is obtained with a calculation similar to Cr and Cint suggests that the calculation of Ce used here does not correspond to CSR theory. Relative measures of competition, such as Ce, Cr and Cint, are indexed and thus reduce the likelihood of observing change along stress or disturbance gradients.  I encourage careful consideration when selecting and using competition indices and suggest the use of multiple indices when testing for differences in competition in different environments.  67  Table 4.1 Indices of competition compared in this experiment, including some citations of each indices’ use and other names that have been used for the indices.  In equations 1 -5, T is the biomass of the target species and N is the biomass of the neighbour species. Competitors are either present (+) or absent (-).  X and y are the greater and lesser, respectively, of T+ and T-. Max T- for both species always occurred when the plant was grown under the high-water, high- nutrient condition without clipping and x was always equal to T- (i.e., plants always had higher biomass when grown alone).  Reprinted with permission of Wiley-Blackwell and the International Association of Vegetation Science. Index Other names Abbreviation, Equation  Citations Absolute competition Absolute competitive index (ACI) Cab = T-  - T+  (1) Campbell and Grime 1992, Turkington et al. 1993, Grace 1995, Wilson and Tilman 1995 Competitive intensity Competitive effect Relative competitive index (RCI), relative neighbour effect Cint = T- - T+/ x Ce = 1 – N+/N- (2) (3) Brooker et al. 2005, Grace 1995, Turkington et al. 1993, Campbell and Grime 1992, Wilson and Tilman 1995, Fraser and Miletti 2008, Markham and Chanway 1996 Competitive response Relative yield Cr = T+ / T-  (4) Goldberg and Fleetwood 1987, Keddy et al. 1994, Fraser and Miletti 2008 Competitive importance  Cimp = (T- - T+)/ (Max T- - y) (5) Brooker et al. 2005    68  Table 4.2 Summary of four-way ANOVAs testing treatment effects on the above ground biomass of Pseudoroegneria spicata and Festuca campestris.  Values in bold are significant (P < 0.05). Reprinted with permission of Wiley-Blackwell and the International Association of Vegetation Science.   P. spicata biomass F. campestris biomass  df F-value P F-value P Block 3 14.100 <0.001 2.181 0.090 Water 1 224.687 <0.001 54.564 <0.001 Nutrient 1 1012.28 <0.001 619.849 <0.001 Competition 1 15.084 <0.001 261.512 <0.001 Clipping 1 19.870 <0.001 33.291 <0.001 Water:Nutrient 1 150.567 <0.001 48.090 <0.001 Water:Competition 1 1.780 0.183 6.551 0.011 Water:Clipping 1 1.618 0.204 3.030 0.083 Nutrient:Competition 1 4.929 0.027 158.070 <0.001 Nutrient:Clipping 1 13.165 <0.001 14.907 <0.001 Competition:Clipping 1 0.099 0.753 0.645 0.423 Water:Nutrient:Competition 1 0.458 0.499 7.218 0.008 Water:Nutrient:Clipping 1 3.056 0.081 6.223 0.013 Water:Competition:Clipping 1 0.010 0.921 1.134 0.288 Nutrient:Competition:Clipping 1 0.1234 0.725 0.680 0.410 Water:Nutrient:Competition:Clipping 1 0.070 0.791 2.390 0.123 Residuals  271  267  69  Table 4.3  Summary of three-way ANOVAs for Pseudoroegneria spicata and Festuca campestris across treatment combinations for all 5 measures of competition: Competitive intensity (Cint), competitive importance (Cimp), competitive effect (Ce), competitive response (Cr) and absolute competition (C  ab).  Values in bold are significant (P < 0.05). Reprinted with permission of Wiley-Blackwell and the International Association of Vegetation Science.   Cab Cint Ce Cr Cimp  df F- value P F- value P F- value P F- value P F- value P P. spicata Water 1 0.543 0.035 1.456 0.230 9.594 0.002 1.456 0.230 4.758 0.031 Nutrient 1 15.576 <0.001 0.136 0.713 11.512 <0.001 0.136 0.713 13.175 <0.001 Clipping 1 0.003 0.955 0.123 0.726 8.194 0.005 0.123 0.726 0.030 0.864 Water:Nutrient 1 0.409 0.523 3.449 0.066 0.446 0.506 3.449 0.066 2.137 0.146 Water:Clipping 1 0.042 0.838 1.061 0.305 0.722 0.397 1.061 0.305 0.128 0.722 Nutrient:Clipping 1 0.036 0.851 0.683 0.410 0.007 0.936 0.683 0.410 0.011 0.916 Water:Nutrient: Clipping 1 3.2e-5 0.995 0.460 0.499 0.326 0.569 0.460 0.499 0.076 0.783 Residuals 122  F. campestris Water 1 11.63 0.009 9.594 0.002 1.456 0.230 9.594 0.002 56.569 <0.001 Nutrient 1 385.72 <0.001 11.512 <0.001 0.1364 0.713 11.512 <0.001 1046.44 <0.001 Clipping 1 1.275 0.261 8.194 0.005 0.1233 0.726 8.194 0.005 15.573 <0.001 Water:Nutrient 1 14.75 <0.001 0.446 0.506 3.4490 0.066 0.446 0.506 75.841 <0.001 Water:Clipping 1 2.960 0.088 0.722 0.397 1.0610 0.305 0.722 0.397 18.326 <0.001 Nutrient:Clipping 1 1.473 0.227 0.007 0.936 0.6829 0.410 0.007 0.936 16.547 <0.001 Water:Nutrient: Clipping 1 6.314 0.013 0.326 0.569 0.4599 0.499 0.326 0.569 30.416 <0.001 Residuals 122   70  Figure 4.1 Mean (± 1 SE) biomass of a) Pseudoroegneria spicata and b) Festuca campestris under all treatment combinations.  On the x-axis, H and L refer to high and low water (W) and nutrients (N). Plants were either grown with (+N) or without (-N) a neighbour, clipped (+C) or unclipped (-C).  Small case letters above the bars represent significant differences within water and nutrient treatment groups due to competition and clipping effects (Tukey HSD). Note the different range on the Y-axis for the two figures. Reprinted with permission of Wiley-Blackwell and the International Association of Vegetation Science.  71  Figure 4.2 (next page) The five competition indices (± 1 SE) for Pseudoroegneria spicata (a – e) and Festuca campestris (f –j) for all treatment combinations.  Treatments are arranged from left to right in order of increasing nutrient and water stress. H and L refer to high and low water (W) and nutrients (N).  Plants were clipped (open bars) or not clipped (shaded bars).  Bars sharing the same letter (Tukey HSD) are not significantly different (P = 0.05); in figures with no letters there was no significant difference.  Note the different range on the Y-axis for absolute competition (e and j). Reprinted with permission of Wiley-Blackwell and the International Association of Vegetation Science.      72  Pseudoroegneria spicata Festuca campestris Treatment combinations  73   Chapter 5: Variation in root plasticity of 18 temperate grass species to water availability Introduction Species composition and biomass production in grasslands are driven primarily by the amount and timing of precipitation events (Knapp and Smith 2001, Knapp et al. 2002).  Water availability is expected to change as the frequency and duration of drought increases in the future due to climate change (Solomon et al. 2007). The ability of plants to respond to drying conditions will therefore be a critical determinant of a plant’s performance, distribution and abundance. Water is generally more available at depth because evaporation and plant uptake both decline with depth. Consequently, rooting depth is important in all aspects of ecosystem ecology.  Rooting depth has the potential to determine individual plant success through resource acquisition and competitive interactions. Root depth also affects ecosystem processes; for example, the water cycle is affected because deeper-rooted plants can access more water and open their stomata more often which leads to an increase of transpiration (Canadell et al. 1996). Knowledge of plant root traits, their response to changing moisture conditions, and variation among and within species is crucial to our understanding of both ecosystem function and plant community dynamics. Variation in rooting depth among species promotes plant diversity and increased productivity in grassland and arid ecosystems – both diversity and species growth rates are higher when niche differences are present (Fargione and Tilman 2005, Levine and HilleRisLambers 2009).  Plasticity enables plants to respond to variation in resource availability (Alpert and Simms 2002). However, if many species exhibit a high degree of plasticity in rooting depth by foraging deeper, effectively reducing niche differences, rooting depth would contribute less to promoting or maintaining diversity. Thus, determining the range of variation in rooting depth among different species is important for understanding patterns of plant diversity. Climate change will likely impact plant species differently and studies to assist with the prediction of species response to environmental change are required. Plant traits can be good predictors of species distributions and abundances (Fargione and Tilman 2005), restoration potential (Pywell et al. 2003), and grazing resistance (Díaz et al. 2001).  While roots play a key role in a plants success, sampling plant roots is difficult and often destructive. However, it may be possible to predict the response of a plant roots to environmental change because traits are often correlated (Lavorel et al. 2002). For example, plants have adapted to different resource  74  availabilities and, in general, fast-growing species are adapted to higher levels of resources than slow-growing species (Grime 2001). Relative growth rate (RGR) may predict below-ground traits as it has been shown to correlate with a suite of important above-ground traits (Diaz et al. 2004). Additionally, amongst grasses, fast-growing species are likely to be more plastic, will generally have more roots because of their overall larger size and quick resource acquisition will enable them to grow deeper. The objective of my study was to examine variation in rooting ability of 18 grass species, in watered and drying conditions. It is difficult to non-destructively access below-ground biomass and, while photo techniques have been developed to monitor in situ root growth, they cannot resolve individual species. Comparative studies in the greenhouse allow an examination of rooting ability of many species in different environmental conditions. Using comparative trait screening I addressed four hypotheses. 1) Species vary in root traits. 2) Species rankings based on traits will change in different watering conditions to reflect species’ adaptations to different environmental conditions. 3) Species will exhibit plasticity in root traits across environmental conditions. 4) Faster-growing species will be more plastic in their response than slower-growing species. A knowledge of the variation in grass root traits and how they respond to different and changing moisture conditions will help to understand the distribution of individual species, patterns of diversity, and  aid in predicting grass responses to future climate change scenarios. Methods Study species Eighteen grass species that commonly occur in the semi-arid grasslands of the Pacific Northwest of North America were selected for this experiment (Table 5.1, for more information see Appendix A). Seeds of sixteen species were collected locally from multiple locations and pooled. Seeds of L. perenne and D. glomerata were obtained from an agricultural supplier.  All seeds were stored in a freezer for at least two months then transferred to Petri dishes containing clean, moistened sand for germination.  Once germinated, seeds were transferred by hand to pots or tubes. Relative growth rate Relative growth rate was measured, on plants grown in a greenhouse, using a protocol similar to Hendry and Grime (1993). Twelve individuals of each species were sown into 250 ml  75  pots saturated with Rorison’s solution (Hendry and Grime 1993). Six randomly chosen individuals of each species were harvested after 7 days; the remaining individuals were harvested after 42 days. During growth all pots were bottom-watered to maintain soil moisture and top-watered weekly with 30 ml of Rorison’s solution. At harvest, plants were separated to roots and shoots, cleaned of sand, dried at 65° C for 48 hrs and weighed. RGR was calculated on biomass using a log-linear model (Fisher 1921). Root tubes Using a methodology similar to Reader et al. (1992), root tubes were constructed of black plastic, 7 cm in diameter and 1 meter long.   Root tubes were filled with clean builders’ sand and suspended in a rack.  All root tubes were initially saturated with 2 litres of Rorison’s solution with a pinhole at the bottom of the tubes to allow excess water to drain.  Seedlings were planted in the tubes.  Control and drying treatment tubes received 20 ml of Rorison’s solution daily for the first 28 days.  After 28 days the drying treatment tubes stopped receiving water and control tubes continued to receive 20 ml of distilled water per day.  After 56 days the shoots were harvested and the tubes were cut into 10 cm sections.  The roots from each section were removed and thoroughly washed of sand. All plant material was dried at 65 °C for 48 hrs and weighed. Rooting depth was assessed as the deepest section in which roots were found. Greenhouse conditions Plants were grown in the Research Greenhouse at Thompson Rivers University campus, Kamloops, British Columbia, Canada. Throughout the experiment, the greenhouse environment was computer controlled to maintain conditions at 22° C and 60 % relative humidity, during the day, and 15 °C and 85 % relative humidity at night.  Three 1000 W halogen sulphide lamps supplemented light in a 14: 10 hour day: night cycle. Data analysis I used a two-way ANOVA to test for differences among species in the two watering treatments for five response variables; total biomass, shoot biomass, root biomass, root: shoot ratio and rooting depth. To meet assumptions of normality the three biomasses were log transformed. Depth was normalized by an arcsine-square root transformation and B. inermis, B. tectorum and L. perenne, the three species with the deepest roots, were removed from the analysis because they grew to the bottom of the tubes and likely could have grown deeper. To  76  compare rankings in watered and drying tubes I used a Spearman rank correlation on the five response variables.  T-tests were used to compare the total biomass, shoot biomass, root biomass, root: shoot ratio and rooting depth in watered and drying tubes for each species. A repeated measure ANOVA was used to test the effect of the water treatment on root biomass across the entire root depth for each species individually. Comparisons could not be made on the rooting depth of B. tectorum and L. perenne because all individuals in both treatments grew to the bottom of the tubes and there is no variation between groups. Plasticity was assessed as the quotient of performance in drying tubes divided by performance in watered tubes. Simple linear regressions were used to examine the relationship between the plasticity of shoot weight to the plasticity of root weight and the plasticity of root depth, the difference of root: shoot ratio in drying to watered tubes to the root: shoot ratio in watered tubes, and RGR to root traits. All analyses were done using R v. 2.11.1 (R Development Core Team, 2010). Results Relative growth rates There was an almost 2-fold difference in RGR of the test species ranging from 0.097 d-1 for P. spicata to 0.18 d-1 for the fastest growing species, V. octiflora (Table 5.1). Species variation and rankings There was significant variation among species for all five plant-traits and the water treatment had significant effects on total biomass, root biomass and r:s (Table 5.2). There were significant Species x Water interactions for all response variables except maximum rooting depth (Table 5.2). Lolium perenne produced the most below and above ground biomass, while P. secunda produced the least, in both watered and drying tubes (Table 5.3). Poa secunda had the highest r:s of all species and P. pratense had the lowest.  Bromus inermis, B. tectorum and L. perenne all produced roots that reached the bottom of the tube, while P. secunda had the shortest roots (Table 5.3). Furthermore, species vary in the way they distribute biomass vertically throughout the tubes (Figure 5.1). Results of the two-way repeated measure ANOVA (Table 5.4) show that all species had different root mass at different depths, but additionally some species showed an interaction between depth and watering treatment. For example, B. squarosus grew equally deep in both water treatments but the root mass in the drying tube declined rapidly in the deeper parts of the tube. On the other hand, K. macrantha produced equal biomass in both water  77  treatments at all depths. The effect of water on biomass in the repeated measure reflected the pattern in the T-tests of biomass (Table 5.3). Rankings among species (Table 5.5a), for each of the five traits, were similar  in  both water treatments (Table 5.5b). Although rankings remained generally consistent, a few species had rank changes of three or more (while not statistically tested I believe a change of this size is noteworthy). Agropyron cristatum showed large changes in ranking across all variables. Koeleria macrantha and E. cinerius decreased in ranking in unwatered tubes, while P. pratense and P. spicata had lower rankings in watered tubes across most variables.  Festuca campestris had comparatively deeper roots in unwatered tubes. Trait plasticity Five species had less total biomass when grown in drying tubes compared to watered tubes (Table 5.3, Figure 5.1).  Two species, F. campestris and K. macrantha had higher shoot biomass in drying tubes, while L. perenne and P. pratense had reduced shoot biomass; eight species had reduced root biomass.  Twelve species had lower r:s in drying tubes, but only P. pratense had reduced rooting depth in drying tubes. Predicting species response Species able to maintain more root biomass (Figure 5.2a) and root depth (Figure 5.2b) in drying tubes had higher shoot biomass. Species that had a high r:s in watered tubes showed larger reductions in r:s in drying tubes (Figure 5.3). The plasticity of rooting depth did not predict above or below ground biomass. Plant species with higher RGR had lower r:s (Figure 5.4a) in both drying and watered tubes and more plasticity in root biomass (a value of 1 indicates no change, Figure 5.4b), producing less root biomass in drying conditions. No significant relationships existed between RGR and total biomass, shoot biomass, rooting depth, in either watering treatments, or with the plasticity of these traits. Discussion The first hypothesis, that species vary in their rooting traits was supported. The second hypothesis, was not supported; species rankings were not different in watered and drying tubes. Support for the third hypothesis is mixed – while r:s is widely plastic, rooting depth was only plastic for one species.  Finally, the fourth hypothesis was partially supported; RGR predicted r:s in drying tubes and the plasticity of root mass, but not rooting depth.  78  Variation among species It is not surprising that species vary in their biomass production; I selected species for this study to represent a range of sizes.  Variation in rooting depth is important especially when considering which species tend to root the deepest. Introduced or invasive species such as B. tectorum and L. perenne tended to have the deepest roots  and this may account for their dominance in many plant communities (Martin and Field, 1984, Knapp 1996). In contrast, two native species, K. macrantha and P. secunda, had the shortest rooting depth; the deepest growing native species was P. spicata which ranked seventh in the drying tubes. Change in community composition from shallow to deep-rooted species could alter some ecosystem functions affected by rooting depth, for example, transpiration rates (Canadell et al. 1996). While root biomass may not be related to competitive ability (Cahill 2003), my results suggest that root depth may be related to dominance. Further examination of rooting depth as a trait related to invasiveness and the consequences for ecosystem function is required. Species rankings are maintained across treatments For most species the rankings of biomass production, r:s ratio and rooting depth were consistent in the watered and drying treatments.  Some species showed changes in rank which could be biologically significant and may help to explain their distribution in different environments. For example, Agropyron cristatum is a problematic invader in arid grasslands (Henderson and Naeth 2005) and has long been used for bank stabilization and forage in these environments (Rogler and Lorenz 1983); its ability to increase its root depth may contribute to its success provided that the deeper roots allow it to access additional water. In contrast, P. pratense is generally found in more mesic habitats and it performed relatively better in the watered tubes when compared to other species. So, while this and a similar study (Reader et al. 1992) find that most species are unresponsive, the species that are responsive to water availability have the potential to play an important functional role, particularly in arid and semi- arid ecosystems. Other suites of traits, such as leaf traits across seasons (Garnier et al. 2001) and competitive effect across flood and fertility treatments (Keddy et al. 1994, Keddy et al. 2000), also maintain rankings in different environments; suggesting that the results of this experiment are part of a broader trend.  79  Plasticity Overall I found that grasses are not variable in their rooting depth, but are variable in r:s between the two watering environments. These results are generally consistent with other studies and even those that examined the same species of grass. Reader et al. (1992) reported that only 7 of 42 species increased rooting depth in a similar experiment and only two of those were grasses. In a study examining the response of three species in drying soil, P. pratense and L. perenne did not differ in root length in watered and drought conditions but D. glomerata seedlings grew deeper (Molyneux and Davies 1983). Despite the general lack of plasticity in rooting depth in response to water availability, root foraging and lengthening has been observed to vary with nitrogen availability (Levang-Brilz and Biondini 2002) and root length may be important to uptake and competition for potassium (Mengel and Steffens 1985). Rather than extending roots to obtain water the majority of grasses in this study exhibited altered r:s ratio. Some models (Tilman 1988, Bloom et al. 1985) suggest that plants will allocate available resources to promote obtaining the limiting resource; I found the opposite, most species reduced r:s in drying tubes. Increased r:s would allow a plant to supply sufficient water to its leaves to maintain transpiration (Schulze et al. 1987). This study is not the first to report a pattern of plants not allocating tissue growth to obtain a limiting resource that may otherwise be made accessible by that tissue. Cahill (2003) found that plants did not increase r:s when experiencing only root competition. Gedroc et al. (1996) reported that optimal resource allocation by seedlings for nutrients may be developmentally limited and that older plants are less capable of adjusting. The patterns I detected for depth and r:s may apply to grasses and not to other plant functional groups. In a study examining the role of roots in seedling survival, Padilla et al. (2007) reported that Mediterranean woody species increased root elongation rate and root surface, but did not change r:s in response to reduced water availability – the opposite of what I observed for grasses. This raises an obvious question of why grasses do not increase r:s or elongate roots in drying soils? First, there must be an evolutionary advantage for plasticity of any trait to evolve (Alpert and Simms 2002). Grasses may not have evolved plasticity in these root traits because environmental conditions have historically been consistent. Grasslands are typically drier systems; the 30-year average rainfall for the region is 279 mm (cv 0.19) thus rainfall does not vary much and the amount it varies by is small thus plasticity may not be advantageous. Secondly, grasses may be able to tolerate low water levels through plastic response in other traits. P. spicata  increased stomatal density in response to reduced water availability in a field  80  experiment (Fraser et al. 2009), a variety of turfgrasses had plastic responses in root regrowth as well as deeper growth after a drying treatment (Huang et al. 1997) and Holcus lanatus responded at the molecular level to extreme water deficits (Pedrol et al. 2000). Finally, Reader et al. (1992) suggest that fast-growing, deep-rooting species may be able to escape the effects of drying soil. It is possible that fast growing species, such as B. tectorum, did not show reduced biomass in the drying tubes because they did most of their growth before the water was reduced and used the wet bottom end of the tubes rather than the drying top part of the tubes. This strategy would be congruent with the natural history of B. tectorum which gets its common name “cheatgrass” from its early season growth that allows it to avoid water stress.  However, species with the least rooting depth, P. secunda and V. octiflora, are small-statured plants that may be limited to the upper parts of the tubes and soils and yet these species, too, showed no change in biomass or rooting depth. Both of these observations can be predicted from observed differences in foraging strategy between fast and slow-growing species. Predicting species response Slow-growing species are adapted to low-resource environments and maintain root structures in order to access short resource pulses (Campbell and Grime 1989). I found that the slow-growing F. campestris had most of its root mass in the top 20 cm of the tube and root biomass declined rapidly below that depth. In contrast, fast-growing species generally allocate less biomass to root structures and have reduced growth rates when resources are limited (Campbell and Grime 1989), but through root placement and growth, they can take rapid advantage of resource patches (Fransen et al. 1999). In this study, B. inermis grew to the bottom of the tubes and maintained relatively equal amounts of biomass in all sections of the tube below 20 cm. It is likely that, B. inermis, a species with relatively high biomass, grew faster and established roots deep in the tube where water was still available. In arid regions where the normal growing season coincides with spring precipitation, rapid acquisition of water may be a key adaptation for weedy species while more stress-tolerant species are able to use shorter pulses throughout the season leading to competitive success (Goldberg and Novoplansky 1997).  The analyses show that species able to maintain deeper roots in drying compared to watered soil were better able to maintain shoot mass, likely because they accessed water in the deeper parts of the tubes. Slow growing species are less likely to be plastic because they take longer to respond to environmental change; as a species can respond quicker to change, it is more likely to be advantageous to have trait plasticity (Alpert and Simms 2002).  81  Summary I found that only one species had different rooting depths in the two water treatments and that the rooting depth hierarchy did not change.  Overall the finding of variation between species and lack of plasticity within a species support the idea that rooting depth is a mechanism of niche differentiation for grassland plant species (Sydes and Grime 1984). However, because rankings did not differ in different environments, variation in community composition between sites is likely not due strictly to rooting depths. Other environmental factors and interactions likely play a stronger role. Overall, rooting depth may separate species’ niches but it is not responsive to environmental change and therefore does not entirely explain species distributions across environments.  82  Table 5.1 Relative growth rates (RGR) of the 18 grass species used in the experiment. Species Abb. Common name RGR ± SE (d -1) Agropyron cristatum (L.) Gaertn. ac Crested wheatgrass 0.116 ± 0.0048 Agrostis gantea L. ag Redtop 0.162 ± 0.0014 Bromus inermis Leyss. bi Smooth brome 0.134 ± 0.0017 Bromus squarrosus L. bj Corn brome 0.133 ± 0.0026 Bromus tectorum L. bt Cheatgrass 0.142 ± 0.0017 Dactylis glomerata L. dg Orchardgrass 0.148 ± 0.0018 Elymus cinereus (Scribn. & Merr.) A. Love ec Giant wild rye 0.112 ± 0.0040 Elymus elemoides (Raf.) Swezey ee Squirrel-tail 0.111 ± 0.0014 Festuca campestris L. fc Rough fescue 0.107 ± 0.0021 Koeleria macrantha (Ledeb.) km Junegrass 0.112 ± 0.0035 Lolium perenne L. lp Perennial ryegrass Not measured Phleum pretense L. ph Timothy 0.150 ± 0.0021 Poa compressa L. pc Canada bluegrass 0.125 ± 0.0018 Poa pratensis L. pp Kentucky bluegrass 0.134 ± 0.0029 Poa secunda J.S. Presl  ps Sandberg’s bluegrass 0.122 ± 0.0046 Pseudoroegneria spicata (Pursh) A. Love bb Bluebunch wheatgrass 0.097 ± 0.0017 Puccinellia nutalliana (Schult.) Hitchc. pn Nuttall’s alkaligrass 0.127 ± 0.0033 Vulpia octoflora (Walter) Rybd. vo Six week fescue 0.176 ± 0.0013   83   84 Table 5.2  F-values from ANOVAs testing for differences among species based five growth variables.  157 residual df, *** P<0.001, ** P<0.01.  df Total biomass Shoot biomass Root biomass R:S ratio Root depth Species (S) 17 48.37*** 41.60*** 52.82*** 16.61** 14.17*** Water (W) 1 12.92*** 1.43 37.20*** 56.50*** 0.006 S x W 17 2.38** 2.47** 2.86*** 2.56** 0.671    Table 5.3 Summary of growth comparisons among 18 species of grass in watered (W+) and drying (W-) tubes. Significant t-test comparisons (T) are in bold * p < 0.1, ** P<0.05, *** P<0.001. T-tests with an “NA” could not be done as all plants in at least one treatment reached the maximum depth of 100 cm.  n Total biomass Shoot biomass Root biomass R:S ratio Root depth Species W- W + W- W+ T W- W+ T W- W+ T W- W+ T W- W+ T Agropyron cristatum  3 3 2.29 ± 0.91 1.01 ± 0.22 1.36 1.59 ± 0.71 0.60± 0.15 1.37 0.70 ± 0.21 0.41 ± 0.07 1.31 0.53 ± 0.11 0.71 ± 0.07 -1.35 97 ± 3 73 ± 17 1.4 Agrostis gigantea  6 6 0.40 ± 0.02 0.49 ± 0.09 -1.00 0.26 ± 0.02 0.26± 0.05 0.02 0.14 ± 0.01 0.23 ± 0.04 -2.29 * 0.52 ± 0.04 0.88 ± 0.05 -5.64 *** 72 ±3 65 ± 6 0.96 Bromus inermis  6 6 2.11 ± 0.43 2.24 ± 0.13 -0.28 0.92 ± 0.21 0.83± 0.07 0.38 1.19 ± 0.22 1.40 ± 0.08 -0.90 1.36 ± 0.06 1.71 ± 0.09 -3.37 ** 100 ± 0 98 ± 2 1 Bromus squarrosu s 3 5 0.86 ± 0.13 1.60 ± 0.19 -3.29 ** 0.59 ± 0.09 0.87± 0.12 -1.9 0.27 ± 0.05 0.72 ± 0.07 -5.41 ** 0.46 ± 0.04 0.85 ± 0.06 -5.44 ** 90 ± 6 92 ± 5 - 0.26 Bromus tectorum  6 5 1.29 ± 0.06 1.69 ± 0.17 -2.25 * 0.76 ± 0.03 0.87± 0.07 -1.35 0.53 ± 0.04 0.82 ± 0.10 -2.82 ** 0.70 ± 0.04 0.94 ± 0.06 -3.56 ** 100 ± 0 100 ± 0 NA Dactylis glomerata  6 5 1.05 ± 0.13 1.46 ± 0.20 -1.71 0.68 ± 0.08 0.78 ±0.12 -0.71 0.37 ± 0.05 0.68 ± 0.10 -2.86 ** 0.54 ± 0.03 0.88 ± 0.08 -3.88 ** 90 ± 4 88 ± 5 0.3 Elymus cinereus  4 5 0.54 ± 0.15 0.44 ± 0.08 0.62 0.25 ± 0.07 0.21 ± 0.04 0.44 0.29 ± 0.08 0.23 ± 0.04 0.748 1.20 ± 0.09 1.06 ± 0.06 1.30 82 ± 9 84 ± 7 - 0.13 Elymus elemoides  6 4 0.56 ± 0.15 0.78 ± 0.30 0.39 0.35 ± 0.10 0.43 ± 0.18 0.39 0.21 ± 0.05 0.36 ± 0.12 -1.09 0.69 ± 0.07 0.91 ± 0.07 -2.33 ** 83 ± 10 82 ± 12 0.06 Festuca campestris  5 5 0.29 ± 0.04 0.27 ± 0.04 1.73 0.14 ± 0.01 0.09 ± 0.01 2.89 ** 0.15 ± 0.03 0.18 ± 0.03 -0.66 1.10 ± 0.13 1.96 ± 0.10 -5.27 ** 54 ± 4 44 ± 4 1.8 Koeleria macrantha 6 5 0.56 ± 0.05 0.44 ± 0.05 -3.38 0.28 ± 0.03 0.13 ±0.02 4.14 ** 0.29 ± 0.02 0.31 ± 0.05 -0.47 1.11 ± 0.14 2.42 ± 0.28 -4.24 ** 55 ± 2 50 ± 3 1.3 Lolium perenne  5 4 2.67 ± 0.40 4.39 ± 0.31 - 2.64* * 1.56 ± 0.26 2.25 ± 0.21 - 2.05 * 1.11 ± 0.17 2.14 ± 0.35 - 2.67* 0.73 ± 0.06 1.00 ± 0.21 -1.25 100 ± 0 100 ± 0 NA Phleum pretense  6 6 0.42 ± 0.09 0.82 ± 0.12 - 1.03* * 0.33 ± 0.07 0.59 ± 0.09 - 2.41 ** 0.10 ± 0.03 0.24 ± 0.04 - 3.05* * 0.28 ± 0.03 0.40 ± 0.03 - 2.74* * 57 ± 10 77 ± 2 - 1.9*               Continued on next page   85   86 Table 5.3 (continued)  n Total biomass Shoot biomass Root biomass R:S ratio Root depth Species W- W + W- W+ T W- W+ T W- W+ T W- W+ T W- W+ T Poa compressa  6 4 0.19 ± 0.03 0.25 ± 0.06 -.316 0.13 ± 0.02 0.14 ± 0.04 -0.38 0.06 ± 0.1 0.11 ± 0.02 - 2.53* * 0.50 ± 0.05 0.84 ±0.10 - 3.04* * 48 ± 6 48 ± 2 0.13 Poa pratensis  6 6 0.50 ± 0.13 0.55 ± 0.11 -0.19 0.22 ± 0.05 0.20 ± 0.03 0.39 0.28 ± 0.08 0.35 ± 0.08 -0.68 1.04 ± 0.23 1.59 ± 0.29 -1.49 53 ± 11 60 ± 9 - 0.47 Poa secunda  6 6 0.17 ± 0.04 0.18 ± 0.03 -0.32 0.07 ± 0.02 0.04 ± 0.01 1.64 0.09 ± 0.02 0.13 ± 0.03 -1.15 1.36 ± 0.06 3.65 ± 0.95 - 2.42* 30 ± 2 30 ± 3 0 Pseudoroe gneria spicata 6 6 0.41 ± 0.06 0.71 ± 0.15 -1.9* 0.22 ± 0.03 0.36 ± 0.07 -1.85 0.19 ± 0.04 0.35 ± 0.08 - 1.89* 0.87 ± 0.07 1.01 ± 0.07 -1.42 84 ± 4 89 ± 5 - 0.77 Puccinelli a nutalliana 4 4 0.21 ± 0.04 0.18 ± 0.01 0.72 0.16 ± 0.03 0.12 ± 0.02 1.20 0.05 ± 0.01 0.06 ± 0.01 -0.97 0.31 ± 0.05 0.56 ± 0.13 -1.77 48 ± 11 42 ± 6 0.39 Vulpia octoflora  4 3 0.28 ± 0.07 0.39 ± 0.09 -1.01 0.21 ± 0.06 0.23 ± 0.05 -0.22 0.07 ± 0.01 0.16 ± 0.03 -2.67 0.39 ± 0.09 0.70 ± 0.03 -3.40 52 ± 5 53 ± 3 - 0.14   Table 5.4  Summary of F-values from repeated measures ANOVAs testing the effect of water treatments on root biomass among 10 cm segments of root tubes. . p < 0.1, * P<0.05,** P <0.01, *** P<0.001. Significant values are in bold. Species Water Depth Water x Depth Agropyron cristatum  1.19 58.04*** 0.6272 Agrostis gigantea 5.10* 212.17*** 25.18** Bromus inermis  0.74 36.22*** 6.47* Bromus squarrosus  23.29** 106.81*** 20.72*** Bromus tectorum  8.63* 71.36*** 13.118*** Dactylis glomerata  8.76* 350.12*** 89.04*** Elymus cinereus  0.45 274.83*** 0.50 Elymus elemoides  0.26 115.48*** 28.71*** Festuca campestris  0.53 146.23*** 1.66 Koeleria macrantha  0.17 192.50*** 1.94 Lolium perenne  7.90* 42.23*** 1.98 Phleum pretense  9.63** 210.18*** 47.00*** Poa compressa  6.61* 85.96*** 11.79*** Poa pratensis  0.52 170.53*** 3.499. Poa secunda  1.23 79.30*** 2.76. Pseudoroegneria spicata  4.35. 335.05*** 61.28*** Puccinellia nutalliana  0.89 81.44*** 2.82. Vulpia octoflora  9.93* 62.60*** 16.27***   87  Table 5.5 a) Rankings among the 18 species for five growth variables in the drying (W-) and watered (W+) tubes. Ties in rankings were recorded as averages.   Changes of rank greater than 3 by a species within a variable are highlighted in bold. The highest value is represented by a ranking of 1. B) Results of Spearman rank correlation tests comparing species rankings for each variable.  Total biomass Shoot biomass Root biomass R:S ratio Root Depth a) Species rankings Species W- W+ W- W+ W- W+ W- W+ W- W+ Lolium perenne 1 1 1 1 1 1 8 8 2 1.5 Agropyron cristatum 2 6 2 6 2 6 12 15 4 10 Bromus inermis 3 2 3 2 3 2 2 4 2 3 Bromus tectorum 4 3 4 3 4 3 9 9 2 1.5 Dactylis glomerata 5 5 5 5 5 5 11 12 5.5 6 Bromus squarrosus 6 4 6 4 6 4 15 13 5.5 4 Koeleria macrantha 7 12 7 12 7 12 4 2 12 14 Elymus  elymoides 8 8 8 8 8 8 10 10 8 8 Elymus cinereus 9 13 9 13 9 13 3 6 9 7 Poa pratensis 10 10 10 10 10 10 6 5 14 13 Phleum pratense 11 7 11 7 11 7 18 18 11 9 Pseudoroegneria spicata 12 9 12 9 12 9 7 7 7 5 Agrostis gigantea 13 11 13 11 13 11 13 11 10 11 Festuca campestris 14 15 14 15 14 15 5 3 13 16 Vulpia octiflora 15 14 15 14 15 14 16 16 15 13 Puccinellia nuttaliana 16 17 16 17 16 17 17 17 17 17 Poa compressa 17 16 17 16 17 16 14 14 16 15 Poa secunda 18 18 18 18 18 18 1 1 18 18  b) Spearman rank correlation tests  Total biomass Shoot biomass Root biomass R:S ratio Root Depth  W- W+ W- W+ W- W+ W- W+ W- W+ rho 0.900 0.876 0.872 0.960 0.922 S 96 120 124 40 75.23 P <0.001 <0.001 <0.001 <0.001 <0.001   88  Agropyron cristatum -100 -80 -60 -40 -20 0 Agrostis gigantea Bromus inermis Bromus squarrosus -100 -80 -60 -40 -20 0 Bromus tectorum Dactylis glomerata Elymus cinereus -100 -80 -60 -40 -20 0 Elymus elemoides Festuca campestris Koeleria macrantha -100 -80 -60 -40 -20 0 Lolium perenne Phleum pratense Poa compressa -100 -80 -60 -40 -20 0 Poa pratensis Poa sandbergii Pseudoroegneria spicata -100 -80 -60 -40 -20 0 -10 -8 -6 -4 -2 0 2 Puccinellia nutalliana -10 -8 -6 -4 -2 0 2 Vulpia octoflora -10 -8 -6 -4 -2 0 2 Log mass (g) D ep th  (c m )  Figure 5.1 Shoot biomass, and depth profile of root growth for 18 grass species in two watering treatments. Open symbols denote drying treatments that received water for the first 28 days but not for the remaining 28 days; closed symbols denote treatments that received water daily until harvest (56 days). Points above dashed lines are total shoot biomass; circles below the line are root biomass in each 10 cm segment of the depth profile.  Square points indicate the total root biomass and mean maximum root depth (the maximum possible depth was 100 cm) for each treatment.  89    Figure 5.2 Shoot mass plasticity plotted against a) root mass plasticity and b) rooting depth plasticity. R2 and P values for simple linear regressions (line) are included in each figure.  Letters indicate species according to abbreviations in Table 5.1.    90  Figure 5.3 Relationship between the difference in root: shoot ratio between drying and watered tubes and the root: shoot ratio in watered tubes. Letters indicate species according to abbreviations in Table 5.1.  91  Figure 5.4 Relationships between a) root shoot ratio and b) root mass plasticity with relative growth rate (d-1).  In a) both drying (open symbol, solid regression line) and watered (closed symbol) are shown.  In b) values larger than one indicate higher root production in drying tubes, a value of 1 indicates no plasticity. Letters indicate species according to abbreviations in Table 1.  92  Chapter 6: Conclusion Summary of thesis I used a variety of experimental approaches to examine the role of stress and disturbance in structuring plant communities across a range of response variables – from entire communities to individual species.  By examining similar processes across a range of response variables it is possible to gain insight and mechanism for the patterns observed at the higher levels of organization. The implications from each experiment vary; for example, Chapter 2 provides important data towards understanding the theory predicting responses of grasslands to stress and disturbance but also strong applied implications for grassland management. Chapter 4 is more narrowly focused on plant-plant interactions and the way competition is assessed. Consequently, I will deal with advances in knowledge and theory, and implications from each experiment, then finish by synthesizing patterns and lessons that span experiments. Field experiment: the effects of stress and disturbance on grassland plant communities Chapter 2 focused on the effects of stress (water availability, warming) and disturbance (clipping) on different grassland types located along a productivity gradient.  Chapter 3 examined the effectiveness of the treatments (singularly and in combination) for manipulating water availability and temperature. Chapter 3 informs our interpretation of chapter 2, so I will examine it first. One concern, of using RSs and OTCs to help predict the future effects of climate change on grasslands is that they are not realistic and may interfere with natural processes. It is important that the changes caused by these devices are understood so that response by the vegetation can be put in context. OTCs and RSs have been independently evaluated; however, as experiments become more complex and multiple factors are evaluated the potential for unknown or undesirable treatment effects increases.  As expected, I found that OTCs increased temperature, RSs reduced soil moisture and watering increased soil moisture. However, clipping vegetation, treatment interactions and weather conditions also affected soil temperature and moisture. The devices generally affected soil variables in predicable ways. For example, RSs increased soil temperature, which is not the variable they are intended to alter, but the increase is what one would expect when water is removed from the soil – due to the high heat capacity of water.  Similarly, clipping  reduced soil moisture and increased soil temperature. Clipping the plants is not intended to affect either variable but the removal of vegetation can increase the  93  amount of solar radiation reaching the soil surface leading to increased temperature and evaporation. I concluded that these climate manipulations are useful because they do affect temperature and moisture in realistic ways, but the interpretation of their effects should be done with caution as they do not act independently or exclusively on intended target variables. The overarching hypothesis in chapter 2 was that the response of the grassland type would be predictable based on its underlying productivity – high-productivity grasslands will be less resistant to change than low-productivity grasslands because they are composed of comparatively fast-growing species that allocate resources to rapid growth rather than traits that would allow them to tolerate stress (Grime 2001).  I examined two sets of variables: biomass and community structure. The results did not support this hypothesis. The relatively high- productivity grasslands I examined were the most resistant to change, while the comparatively medium-productivity grasslands showed most change followed by the lowest-productivity grassland; this was true for both biomass produced and change in community structure. This was also surprising because it not only rejects Grime’s theory, it rejects other theories, such as the diversity- stability hypothesis (McCann 2000) that suggest high diversity communities will be more resistant than low-diversity communities. In my study, the high-diversity communities were the least resistant to disturbance and stress.  The second objective was to examine how stress interacted with disturbance to alter the grasslands.  I demonstrated that stress and disturbance individually affect the grassland, but interactions between the two processes were very important and could prevent, initiate or reverse a response and, making the results even more complex, these interactions varied across grassland types. Chapter 2 has implications at theoretical and applied levels. The major hypothesis, that high productivity grasslands would be least resistant to change, was not supported.  This suggests that other factors contribute to community stability, at least in the short term.  As discussed, Grime et al. (2008) suggested four reasons (site history, succession status, diversity and functional composition) why vegetation may be more resistant to change. The most likely, but unverified conclusion for my study, is that functional composition contributes to stability. The dominant plants in the system are stress-tolerant species that tend to be long-lived.  The design of the experiment may also contribute to the patterns observed. The experiment may have been too short or treatments not severe enough to rapidly affect the large dominant grasses.  It may be that extreme weather events, that are likely to be more common in the future, may cause major shifts (Smith 2011).  This study also highlights the value of replicating studies in different  94  environments. Although I did not observe a clear trend in grassland response along the productivity gradient, the different grasslands responded differently. Had I only done the experiment in one grassland type the more complex pattern of response would not have been detected. Another possible reason for not observing the predicted pattern may be the range of productivity represented on the gradient used in the experiment; Grime’s (2001) productivity (biomass and litter) range for medium productivity is approximately 300 – 700 g/m2.  The maximum productivity of research sites in this research averaged approximately 800 g/m2, just beyond the medium productivity range. The range of biomass found in other herbaceous communities can exceed 1400 g/m2. I may not have tested sites of high enough productivity to see the reduced resistance that is predicted.  However, if this pattern was found it would indicate a unimodal response with stability highest at intermediate levels of productivity rather than the linear decline I’ve predicted.  Consequently, these results have generated new hypotheses about the patterns of ecosystem stability. Management of grasslands as rangelands will likely need to account for a warmer, drier climate in the future.  My results show that while disturbance or stress alone affected the grasslands, the combination of stress and disturbance often caused new or larger changes: biomass production and plant diversity were affected. Consequently, there may be more conflict between grazing and conservation of biodiversity as forage declines because climate change is inescapable while grazing can be managed to mitigate the consequences. There are caveats on extrapolating the results of this experiment to the specific situation of cattle grazing interacting with climate change.  First, the treatments were applied in a generic fashion. The timing, intensity and frequency of all treatments do not necessarily match the natural or managed occurrences of climate or cattle.  For instance, I clipped plots at the peak of the vegetative growing season, but cattle are generally on the land through more of the year.  Further, I clipped all grassland types at close to the same time, whereas cattle are moved to the different grasslands at different times of year (generally in lower grasslands early in the season and then to upper elevation grasslands later where growth is delayed due to a cooler climate). Similarly, the warming and water addition and removal regimes imposed do not match predicted future patterns. I chose this approach to simplify the experiment, knowingly trading off realism for power in my experimental design.  None-the-less, my results show interactions between climate change and disturbance and are a good starting point for understanding the potential implications of grazing livestock in a warmer, drier future.  95  Measuring competition across multiple gradients In chapter 4, I addressed the role of competition along nutrient, water and disturbance gradients. I asked two questions: 1) How do different resource and disturbance levels interact to affect competition?  2) How do different indices of competition change the interpretation of how competition changes under different resource and disturbance conditions? These questions are important because competition is a key factor structuring plant communities. Under climate change scenarios the availability of the key limiting resource in grasslands, water, will be reduced. Consequently, the strength of the interactions between plants will be altered.  My results show that some indices of competition show a decline with increased stress and disturbance, while other indices do not. Therefore, it is necessary to choose a competition index appropriate to the question being asked. Regardless of the index used, plants will compete for limiting resources because climate change will alter the availability of water and it is critical that we understand the role of competition in structuring plant communities. Rooting depth Chapter 5 discusses the variation among grass species’ rooting depth, a key plant trait related to ecosystem function and community structure. I showed that this variation is generally not plastic across different environments.  Furthermore, there is little change in the ranking of species in root traits in different environments.  I was also able to show that some root traits, which are generally not practical to measure in the field, are predicted by relative growth rate. The results support the idea that root depth is a mechanism of plant diversity in grasslands. There appears to be a trend for invasive species to be capable of producing deeper roots than native species in the pool of species I used. Whether this is a more general trend or a mechanism of invasion should be further investigated.  Contrary to expectations, grasses did not root deeper to obtain water in drying soil, but they did show reductions in r:s. The response of some species and their relative abilities, at least anecdotally, agree with patterns observed in the field. Further studies that manipulate the root environment are needed to establish these mechanisms. Synthesis and lessons learned How we measure is important The indexes and measures used by ecologists can greatly influence the interpretation of results, but also open new insight to some problems.  This is most apparent in chapter 4 on  96  competition, where depending on the index used one could infer decreasing competition with increasing stress or no change in competition along the gradient. I am not the first to make this observation (Grace 1995, Brooker et al. 2005); but I am one of few to test this in a controlled environment while incorporating a disturbance. In chapter 2, the use of an alternative measure of diversity (dissimilarity) led to new insight on the response of grassland plant communities to stress and disturbance.  I had initially tried to identify the species that were lost when changes in species richness and diversity occurred, however; there was no apparent pattern. An ordination and multivariate test of beta-diversity revealed that plots within a treatment became less similar; suggesting changes are not predictable at the species level. Instead response may be happening at the level of functional groups and may be predictable with functional response traits (Lavorel and Garnier 2002). This provides direction for future analysis such as more detailed examination of functional group response or species turnover. If these methods were not used, it would have been very difficult to detect this pattern. Controlled experiments in the greenhouse can inform field experiments An anonymous reviewer’s leading comment to reject chapter 4, the competition experiment, for publication was “The situation is completely contrived and has nothing to do with real communities.”  I disagree and argue that the two greenhouse experiments presented in this thesis provide insight into communities, help inform the field experiment and provide opportunities to develop new hypotheses. By reducing a community to its simplest form, two plants in a pot under highly controlled conditions, I was able to show that stress and disturbance can change the way two plants interact. While a sand-filled pot lacks the complexity and realism of field conditions it provides evidence that stress and disturbance will alter plant interactions and investigating similar conditions in a more complex system is warranted.  It also provides a mechanism for some of the responses detected in the field.  First, P. spicata was the better competitor and that may explain its dominance and distribution in the field.  Second, disturbance reduced the strength of competition, possibly explaining the positive response of some subordinate species in clipped plots.  I attempted a field competition experiment using transplants. Had it been successful (it was not because of high transplant mortality), it would have provided a more direct link between the greenhouse and field experiment and possibly further strengthened our understanding of competition in structuring these grasslands. Chapter 5’s root trait and RGR results also provide insight into the results from the field experiment. First, the two dominant plants in the field had the slowest growth rates of the 18 species tested.  97  This suggests that slow growth rather than rapid resource acquisition is a key trait for plants in the system.  We also learned that P. spicata is the deepest rooting native species; this may permit it to escape drying soils and explain the perplexing increased growth of P. spicata under the RSs. Also, the grasses tested in chapter 5 reduced their r:s ratio in drying soil conditions suggesting that rather than increasing roots to obtain the limiting resources they are adapted to slow their growth in an effort to tolerate drought conditions. P. spicata was one species that did not significantly alter its r:s ratio, which would also enable it to escape drought while other species declined. While all of these hypotheses require further investigation they are derived from the combination of field and greenhouse experiments. Comments on running multi-factor experiments Ecologists recognize that environmental concerns are compounding and if we hope to assist with understanding the drivers behind environmental change experiments that examine multiple issues simultaneously are necessary. Few experiments have examined multiple climate change factors across different ecosystems (Smith et al. 2009).  The reason for this is that multiple factors create larger experiments, more replication is required and the results are often more difficult to interpret. The competition experiment (chapter 4) had four factors (competition, water, nutrients and clipping) as did the field experiment (warming, water, clipping and grassland type).   I was able to provide sufficient replication (20) in the competition experiment, but I would suggest that the six replicates used in the field experiment is the minimum -- I observed a number of trends that were not statistically significant, but likely indicate a biological response.  Of course, replication trades off against the number of factors that can be tested within the constraints of manpower and funds available to perform the research. I used basic analysis of variance for many of my analyses.  Interpreting these results is straight forward for a single factor, but when three-way or four-way interactions are present single factor results are almost meaningless without understanding the interactions. However, it is possible to display and interpret these results, and it is necessary to do so to meet objectives of understanding multiple drivers of environmental change. The fact that three-way (or more) interactions occur is reason enough to do these experiments; otherwise, if experiments are limited to the typical two-factor design we are severely limiting our understanding of biological processes.   98  Summary: The role of stress and disturbance on plants and plant communities The research described, and results reported have shown that stress and disturbance can affect individual plants, their interactions with other plants and the function and composition of communities.  However, the interactions between stress and disturbance may be much more important, especially in light of climate change and our continued land use.  My results have shown that the effects on different grassland types is complex and in this system requires further investigation to understand the mechanisms of response.  99  References Adler PB, Seabloom EW, Borer ET, Hillebrand H, Hautier Y, Hector A, Harpole WS, O'Halloran LR, Grace JB, Anderson TM, Bakker JD, Biederman LA, Brown CS, Buckley YM, Calabrese LB, Chu CJ, Cleland EE, Collins SL, Cottingham KL, Crawley MJ, Damschen EI, Davies KF, DeCrappeo NM, Fay PA, Firn J, Frater P, Gasarch EI, Gruner DS, Hagenah N, Hille Ris Lambers J, Humphries H, Jin VL, Kay AD, Kirkman KP, Klein JA, Knops JM, La Pierre KJ, Lambrinos JG, Li W, MacDougall AS, McCulley RL, Melbourne BA, Mitchell CE, Moore JL, Morgan JW, Mortensen B, Orrock JL, Prober SM, Pyke DA, Risch AC, Schuetz M, Smith MD, Stevens CJ, Sullivan LL, Wang G, Wragg PD,Wright JP, Yang LH. 2011. Productivity is a poor predictor of plant species richness. Science 333: 1750 -1753. Al– Mufti MM, Sydes CL, Furness SB, Grime JP, Band SR. 1977. A quantitative analysis of shoot phenology and dominance in herbaceous vegetation. Journal of Ecology 65: 759 – 791. Alpert P, Simms EL. 2002. The relative advantages of plasticity and fixity in different environments: when is it good for a plant to adjust? Evolutionary Ecology 16: 285 – 297. Anderson MJ. 2001. A new method of non– parametric multivariate analysis of variance. Austral Ecology 26: 32 – 46. Anderson MJ. 2006. Distance– based tests of homogeneity of multivariate dispersions. Biometrics 62: 245 – 253. Arredondo JT, Johnson DA. 1998. Clipping effects on root architecture and morphology of 3 range grasses. Journal of Range Management 51: 207– 214. Bassett E. 2007. The effects of grazing and site productivity on Carabid Beetles (Coleoptera: Carabidae) in a semi– arid grassland. MSc thesis, Thompson Rivers University, Kamloops, British Columbia, Canada, 60 pp. BC Conservation Data Centre. 2011. BC Species and Ecosystems Explorer. B.C. Ministry of Environment, Victoria, B.C. http://a100.gov.bc.ca/pub/eswp/, accessed 10 January 2011. Bloom AJ, Chapin FS, Mooney HA. 1985. Resource limitation in plants – an economic analogy. Annual Reviews in Ecology and Systematics 16: 363 – 92. Bradley NL, Leopold AC, Ross J, Huffaker W. 1999. Phenological changes reflect climate change in Wisconsin. Proceedings of the National Academy of Sciences of the United States of America 96: 9701 – 9704. Bremer DJ, Ham JM, Owensby CE. 1996. Effect of elevated atmospheric carbon dioxide and open– top chambers on transpiration in a tallgrass prairie. Journal of Environmental Quality 25: 691 – 701. Bremer DJ, Ham JM, Owensby CE , Knapp AK. 1998. Response of soil respiration to clipping and grazing in a tallgrass prairie.  Journal of Environmental Quality 27: 1539 – 1548. Brooker R, Kikvidze Z, Pugnaire FI, Callaway RM, Choler P, Lortie CJ,  Michalet R. 2005. The importance of importance. Oikos 109: 73 – 70.  100  Brooker RW,  Kikvidze Z. 2008. Importance: an overlooked concept in plant interaction research.  Journal of Ecology 96: 703 – 708. Brown JH, Valone TJ, Curtin CG. 1997. Reorganization of an arid ecosystem in response to recent climate change. Proceedings of the National Academy of Sciences 94: 9729 – 9733. Brown JR, Archer S. 1999. Shrub invasion of grassland: recruitment is continuous and not regulated by herbaceous biomass or density.  Ecology 80: 2385 – 2396. Busso CA, Richards JH, Chatterton NJ. 1990. Nonstructural carbohydrates and spring regrowth of two cool – season grasses: interaction of drought and clipping. Journal of Range Management 43: 336 – 343. Cahill JF. 2002. Interactions between root and shoot competition vary among species. Oikos 99: 101– 112. Cahill JF. 2003. Lack of relationship between below– ground competition and allocation to roots in 10 grassland species. Journal of Ecology 91: 532 – 540. Cahill JF, Kembel SW, Gustafson DJ. 2005. Differential genetic influences on competitive effect and response in Arabidopsis thaliana. Journal of Ecology 93: 958 – 967. Campbell BD, Grime JP. 1992. An experimental test of plant strategy theory. Ecology 73: 15 – 29. Campbell CS. 2002. Calibrating ECH2O soil moisture probes. Application note SMR110, Decagon Devices, Inc. Canadell J, Jackson RB, Ehleringer JR, Mooney HA, Sala OE, Schulze ED. 1996. Maximum rooting depth of vegetation types at the global scale. Oecologia 108: 583 – 595. Canadian Climate Change Scenarios Network. 2011. Localizer. User location: Kamloops, BC (50.67N 120.33W). http://cccsn.ca/tools/viz– localizer/localizer.pl, accessed 10 January 2011. Carlyle CN, Fraser LH. 2006. A test of three juvenile plant competitive response strategies. Journal of Vegetation Science 17: 11 – 18. Carlyle CN, Fraser LH, Turkington R. 2011. Tracking soil temperature and moisture in a multi– factor climate experiment in temperate grassland: do climate manipulation methods produce their intended effects? Ecosystems 14: 489 – 502. Choler P, Michalet R, Callaway RM. 2001. Facilitation and competition on gradients in alpine plant communities. Ecology 82: 3295 – 3308. Coulson S, Hodkinson ID, Strathdee A, Bale JS, Block W, Worland MR, Webb NR. 1993. Simulated climate change: the interaction between vegetation type and microhabitat temperatures at Ny Ålesund, Svalbard. Polar Biology 13: 67 – 70. Craine, JM. 2005. Reconciling plant strategy theories of Grime and Tilman. Journal of Ecology 93: 1041 – 1052. Cramer W, Bondeau A, Woodward FI, Prenctice IC, Betts RA, Brovkin V, Cox PM, Fisher V, Foley JA, Friend AD, Kucharik C, Lomas MR, Ramankutty N, Sitch S, Smith B, White A,  101  Young– Molling C. 2001. Global response of terrestrial ecosystem structure and function to CO2 and climate change: results from six dynamic global vegetation models. Global Change Biology 7: 357 – 373. Crawley MJ. 1983.  Herbivory: the dynamics of animal-plant interactions.  Blackwell Scientific, Oxford. Connell JH. 1983.  On the prevalence and relative importance of interspecific competition: evidence from field experiments.  The American Naturalist 122: 661 – 696. Díaz S, Noy– Meir I, Cabido M. 2001. Can grazing response of herbaceous plants be predicted from simple vegetative traits? Journal of Applied Ecology 38: 497 – 508. Diaz S, Hodgson JG, Thompson K, Cabido M, Cornelissen JHC, Jalili A, Montserrat– Martí G, Grime JP, Zarrinkamar F, Asri Y, Band SR, Basconcelo S, Castro– Díez P, Funes G, Hamzehee B, Khoshnevi M, Pérez– Harguindeguy N, Pérez– Rontomé MC, Shirvany FA, Vendramini F, Yazdani S, Abbas– Azimi R, Bogaard A, Boustani S, Charles M, Dehghan M, de Torres– Espuny L, Falczuk V, Guerrero– Campo J, Hynd A, Jones G, Kowsary E, Kazemi– Saeed F, Maestro– Martínez M, Romo– Díez A, Shaw S, Siavash B, Villar– Salvador P, Zak MR. 2004. The plant traits that drive ecosystems: evidence from three continents. Journal of Vegetation Science 15: 295 – 304. Dormann CF, Woodin SJ. 2002. Climate change in the Arctic: using plant functional types in a meta– analysis of field experiments.  Functional Ecology 16: 4 – 17. Dunne JA, Saleska SR, Fischer ML, Harte J. 2004. Integrating experimental and gradient methods in ecological climate change research. Ecology 85: 904 –  916. Engel EC, Weltzin JF. 2008. Can community composition be predicted from pairwise species interactions? Plant Ecology 195: 77 – 85. Environment Canada. 2009. National Climate Data and Information archive. http://climate.weatheroffice.ec.gc.ca/climate_normals/index_e.html, accessed 6 October 2009. Etterson JR. 2004. Evolutionary potential of Chamaecrista fasciculate in relation to climate change. I. Clinal patterns of selection along an environmental gradient in the great plains. Evolution 58: 1446 – 1458. Erhlich PR, Wilson EO. 1991. Biodiversity studies – science and policy. Science 253: 758-762. Faith DP, Minchin PR, Belbin L. 1987. Compositional dissimilarity as a robust measure of ecological distance. Vegetatio 69: 57 – 68. Fay PA, Carlisle JD, Knapp AK, Blair JM, Collins SL. 2000. Altering rainfall timing and quantity in a mesic grassland ecosystem: design and performance of rainfall manipulation shelters. Ecosystems 3: 308 – 319. Fargione J, Tilman D. 2005. Niche differences in phenology and rooting depth promote coexistence with a dominant C4 bunchgrass. Oecologia 143: 598 – 606. Fisher RA. 1921. Some remarks on the methods formulated in a recent article on “the quantitative analysis of plant growth”. Annals of Applied Biology 7: 367 – 372.  102  Fransen B,  de Kroon H, de Kovel CGF, Van Den Bosch F. 1999. Disentangling the effects of root foraging and inherent growth rate on plant biomass accumulation in heterogeneous environments: a modeling study. Annals of Botany 84: 305 – 311. Fraser LH. 1998. Top– down vs. bottom– up control influenced by productivity in a North Derbyshire dale. Oikos 81: 99 – 108. Fraser LH,  Miletti TE. 2008. Effect of minor water depth treatments on competitive effect and response of eight wetland plants. Plant Ecology 195: 33– 43. Fraser LH, Greenall A, Carlyle CN, Turkington R, Ross Friedman C. 2009. Adaptive phenotypic plasticity of Pseudoroegneria spicata: response of stomatal density, leaf area and biomass to changes in water supply and increased temperature. Annals of Botany 103: 769– 775. Frekelton RP, Watkinson AR, Rees M. 2009. Measuring the importance of competition in plant communities. Journal of Ecology 97: 379 – 384. Garnier E, Laurent G, Bellman A, Debain S, Berthelier P, Ducout B, Roumet C, Nava M –L. 2001. Consistency of species ranking based on functional leaf traits. New Phytologist 152: 69 – 83. Gaucherand S, Liancourt P, Lavorel S. 2006. Importance and intensity of competition along a fertility gradient and across species. Journal of Vegetation Science 17: 455 – 464. Gayton D. 2003. British Columbia grasslands: monitoring vegetation change. Forest Research Extension Partnership, FORREX series 7. Kamloops, Canada. Gedan KB, Bertness MD. 2009. Experimental warming causes rapid loss of plant diversity in New England salt marshes. Ecology Letters 12: 842 – 848. Gedroc JJ, McConnaughay KDM, Coleman JS. 1996. Plasticity in root/shoot partitioning: optimal, ontogenetic, or both? Functional Ecology 10: 44 – 50. Gilgen AK, Buchmann N. 2009. Response of temperate grasslands at different altitudes to simulated summer drought differed but scaled with annual precipitation. Biogeosciences 6: 2525 – 2539. Goldberg DE, Werner PA. 1983. Equivalence of competitors in plant communities: a null hypothesis and a field experimental approach. American Journal of Botany 70: 1098 – 1104. Goldberg DE. 1987. Neighborhood competition in an old– field plant community. Ecology 68: 1211– 1223. Goldberg DE, Fleetwood L. 1987. Competitive effect and response in four annual plants. Journal of Ecology 75: 1131 – 1143. Goldberg DE. 1990. Components of resource competition in plant communities.  In Grace JB, Tilman, D. (eds.) Perspectives on plant competition, pp 27 – 49. Academic Press, San Diego, CA, US. Goldberg DE, Landa K. 1991.  Competitive effect and response: hierarchies and correlated traits in the early stages of competition. Journal of Ecology 79: 1013 – 1030.  103  Goldberg DE, Barton AM. 1992. Patterns and consequences of interspecific competition in natural communities: a review of field experiments with plants.  The American Naturalist 139: 771 – 801. Goldberg DE, Scheiner SM. 1993. ANOVA and ANCOVA: field competition experiments.   In Scheiner SM,  Gurevich J. (eds.) Design and analysis of ecological experiments, pp 77 – 98. Oxford University Press, New York, NY, US. Goldberg DE. 1996. Competitive ability: definitions, contingency and correlated traits. Philosophical Transactions of the Royal Society London B 351:1377 – 1285. Goldberg DE, Novoplansky A. 1997. On the relative importance of competition in unproductive environments.  Journal of Ecology 85: 409 – 418. Goldberg DE, Rajaniemi T, Gurevitch J,  Stewart– Oaten A. 1999. Empirical approaches to quantifying interaction intensity: competition and facilitation along productivity gradients. Ecology 80: 1118 – 1131. Grabherr G, Gottfried M, Pauli H. 1994. Climate Effects on Mountain Plants. Nature 369: 448 – 448. Grace J. 1991. A clarification of the debate between Grime and Tilman. Functional Ecology 5: 583 – 587. Grace JB. 1993. The effects of habitat productivity on competition intensity. Trends in Ecology and Evolution 8: 33 – 34. Grace JB. 1995. On the measurement of plant competition intensity. Ecology 76: 305 – 308. Grime JP. 1977. Evidence for the existence of three primary strategies in plants and its relevance to ecological and evolutionary theory.  American Naturalist 111: 1169 – 1194. Grime JP. 2001. Plant strategies, vegetation processes, and ecosystem properties, 2nd edn. John Wiley & Sons, Ltd. Chichester. Grime JP, Brown VK, Thompson K, Masters GJ, Hillier SH, Clarke IP, Askew AP, Corker D, Kielty JP. 2000. The response of two contrasting limestone grasslands to simulated climate change. Science 289: 762 – 765. Grime JP, Fridley JD, Askew AP, Thompson K, Hodgson JG, Bennet CR. 2008. Long– term resistance to simulated climate change in an infertile grassland. Proceedings of the National Academy of Sciences 105: 10028 – 10032. Hales G. 2011. Small mammal communities in two grassland ecosystems in British Columbia, Canada. MSc thesis, Thompson Rivers University, Kamloops, British Columbia, Canada, 100 pp. Havstad KM, Herrick J, Tseelei EA. 2008. Mongolia’s rangelands: is livestock production the key to the future? Frontiers in Ecology and the Environment 6: 386 – 391. Hector A, Schmid B, Beierkuhnlein C, Caldeira MC, Diemer M, Dimitrakopoulos PG, Finn JA, Freitas H, Giller PS, Good J, Harris R, Hogberg P, Huss-Danell K, Joshi J, Jumpponen A, Korner C, Leadley PW, Loreau M, Minns A, Mulder CPH, O'Donovan G, Otway SJ, Pereira JS,  104  Prinz A, Read DJ, Scherer-Lorenzen M, Schulze ED, Siamantziouras ASD, Spehn EM, Terry AC, Troumbis AY, Woodward FI, Yachi S, Lawton JH. 1999. Plant diversity and productivity experiments in European grasslands. Science 286: 1123 – 1127. Henderson DC, Naeth MA. 2005. Multi-scale impacts of crested wheatgrass invasion in mixed- grass prairie. Biological Invasions 7: 639 – 650. Hendry GAF, Grime, JP. 1993. Methods in comparative plant ecology: a laboratory manual. Chapman Hall, London, UK. Henry GHR, Molau U. 1997. Tundra plants and climate change: the International Tundra Experiment (ITEX). Global Change Biology 3: 1– 9. Hoekman D. 2010. Turning up the heat: temperature influences the relative importance of top– down and bottom– up effects. Ecology 91: 2819 – 2825. Hollister RD, Webber PJ. 2000. Biotic validation of small open– top chambers in a tundra ecosystem. Global Change Biology 6: 835 – 842. Huang B, Duncan RR, Carrow RN. 1997. Drought– resistant mechanisms of seven warm– season turfgrasses under surface soil drying: II. Root aspects. Crop Science 37: 1863 – 1869. Hudson JMG, Henry GHR. 2009. Increased plant biomass in a High Arctic heath community from 1981 to 2008. Ecology 90: 2657 – 2663. Hulme PE. 2005. Adapting to climate change: is there scope for ecological management in the face of a global threat? Journal of Applied Ecology 42: 784 – 794. Huston MA, Aarssen LW, Austin MP, Cade BS, Fridley JD, Garnier E, Grime JP, Hodgson J, Lauenroth WK, Thompson K, Vandermeer JH, Wardle DA. 2000. No consistent effect of plant diversity on productivity. Science 289: 1255a - 1255a. Ives AR, Carpenter SR. 2007. Stability and diversity of ecosystems. Science 317: 58 – 62. Johnston A, Dormaar JF, Smoliak S. 1971. Long– term grazing effects on fescue grassland soils. Journal of Range Management 24: 185 – 188. Johnson K, Vogt K, Clark H, Schmitz O, Vogt D. 1996. Biodiversity and the productivity and stability of ecosystems. Trends in Ecology and Evolution 11: 372 – 377. Jones MH, Macdonald SE, Henry GHR. 1999. Sex- and habitat-specific responses of a high arctic willow, Salix arctica, to experimental climate change. Oikos 87: 129-138. Kardol P, Campany CE, Souza L, Norby RJ, Weltzin JF, Classen AT. 2010. Climate change effects on plant biomass alter dominance patterns and community evenness in an experimental old– field system. Global Change Biology 16: 2676 – 2687. Keddy P. 1991. Working with heterogeneity: an operator’s guide to environmental gradients. In Kolasa J, Pickett STA (eds.) Ecological Heterogeneity, pp. 181 – 200. Springer – Verlag, New York. Keddy PA, Twolan-Strutt L, Wisheu IC. 1994. Competitive effect and response rankings in 20 wetland plants: are they consistent across three environments? Journal of Ecology 82: 635 – 643.  105  Keddy PA, Fraser LH, Wisheu IC. 1998. A comparative approach to examine competitive response of 48 wetland plant species. Journal of Vegetation Science 9: 777– 786. Keddy PA, Gaudet C, Fraser LH. 2000. Effects of low and high nutrients on the competitive hierarchy of 26 shoreline plants. Journal of Ecology 88: 413 – 423. Keddy PA. 2001. Competition. 2nd edn. Kluwer Academic Publishers, Dordrecht, NL. Keeling CD, Chin JFS, Whorf TP. 1996. Increased activity of northern vegetation inferred from atmospheric CO2 measurements. Nature 382: 146 – 149. Kennett GA, Lacey JR, Butt CA, Olson– Rutz K,  Haferkamp MR. 1992. Effects of defoliation, shading and competition on spotted knapweed and bluebunch wheatgrass.  Journal of Range Management 45: 363 – 369. Klanderud K. 2005. Climate change effects on species interactions in an alpine plant community. Journal of Ecology 93: 127 – 137. Klanderud K, Totland Ø. 2005. Simulated climate change altered dominance hierarchies and diversity of an alpine biodiversity hotspot. Ecology 86: 2047 – 2054. Klein JA, Harte J, Zhao XQ. 2004. Experimental warming causes large and rapid species loss, dampened by simulated grazing, on the Tibetan Plateau. Ecology Letters 7: 1170 – 1179. Klein JA, Harte J, Zhao XQ. 2005. Dynamic and complex microclimate response to warming and grazing manipulations.  Global Change Biology 11: 1440 – 1451. Klein JA, Harte J, Zhao X. 2007. Experimental warming, not grazing, decreases rangeland quality on the Tibetan Plateau. Ecological Applications 17: 541– 557. Klinkenberg B. (Ed.) 2010. E-Flora BC: Electronic Atlas of the Plants of British Columbia. Lab for Advanced Spatial Analysis, Department of Geography, University of British Columbia, Vancouver. http://www.geog.ubc.ca/biodiversity/eflora/, Accessed August 2011. Knapp PA. 1996. Cheatgrass (Bromus tectorum L) dominance in the Great Basin Desert. Global Environmental Change 6: 37 – 52. Knapp AK, Smith MD. 2001. Variation among biomes in temporal dynamics of aboveground primary production. Science 291: 481 – 484. Knapp AK, Fay PA, Blair JM, Collins SL, Smith MD, Carlisle JD, Harper CW, Danner BT, Lett MS, McCarron JK. 2002. Rainfall variability, carbon cycling, and plant species diversity in a mesic grassland. Science 298: 2202 – 2205. Köchy M, Wilson D. 2004. Semiarid grassland responses to short– term variation in water availability. Plant Ecology 174: 197 – 203. Lavorel S, Garnier E. 2002. Predicting changes in community composition and ecosystem functioning from plant traits: revisiting the Holy Grail. Functional Ecology 16: 545 – 556. Levang– Brilz N, Biondini ME. 2002. Growth rate, root development and nutrient uptake of 55 plant species from the Great Plains Grasslands, USA. Plant Ecology 165: 117 – 144. Levine SA. 1992. The problem of pattern and scale in ecology.  Ecology 73: 1943 – 1967.  106  Levine JM, HilleRisLambers J. 2009. The importance of niches for the maintenance of species diversity. Nature 461: 254 – 258. Lucas RW, Forseth IN, Casper BB. 2008. Using rainout shelters to evaluate climate change effects on the demography of Cryptantha flava. Journal of Ecology 96: 514 – 522. Marion GM, Henry GHR, Freckman DW, Johnstone J, Jones G, Jones MH, Lévesque E, Molau U, Mølgaard P, Parsons AN, Svoboda J, Virginia RA. 1997. Open– top designs for manipulating field temperature in high– latitude ecosystems. Global Change Biology 3: 20 – 32. Markham JH, Chanway CP. 1996. Measuring plant neighbour effects. Functional Ecology 10: 548 – 549. Martin MPLD, Field RJ. 1984. The nature of competition between perennial ryegrass and white clover. Grass and Forage Science 39: 247 – 253. McCann KS. 2000. The diversity– stability debate. Nature 405: 208 – 233. McCarty JP. 2001. Ecological consequences of recent climate change. Conservation Biology 15: 320 – 331. Mengel K, Steffens D. 1985. Potassium uptake of ryegrass (Lolium perenne) and red clover (Trifolium pratense) as related to root parameters. Biology and Fertility of Soils 1: 53 – 58. Miller TE. 1996. On quantifying the intensity of competition across gradients. Ecology 77: 978 – 981. Miller TE, Burns JH, Munguia P, Walters EL, Kneitel JM, Richards PM, Mouquet N,  Buckley HL. 2005. A critical review of twenty years’ use of the resource– ratio theory. American Naturalist 165: 439 – 448. Miller TE, Burns JH, Munguia P, Walters EL, Kneitel JM, Richards PM, Mouquet N, Buckley, H.L. 2007. Evaluating support for the Resource– ratio hypothesis: A reply to Wilson et al. American Naturalist 169: 707– 708. Molyneux DE, WJ Davies. 1983. Rooting pattern and water relations of three pasture grasses growing in drying soil. Oecologia 58: 220 – 224. Myneni RB, Keeling CD, Tucker CJ, Asrar G, Nemani RR. 1997. Increased plant growth in the northern high latitudes from 1981 to 1991. Nature 386: 698 – 702. Newman E. 1973. Competition and Diversity in Herbaceous Vegetation. Nature 244: 310. Norby RJ, Edwards NT, Riggs JS, Abner CH, Wullschleger SD, Gunderson CA. 1997. Temperature– controlled open– top chambers for global change research. Global Change Biology 3: 259 – 267. Novoplansky A, Goldberg DE. 2001. Effects of water pulsing on individual performance and competitive hierarchies in plants. Journal of Vegetation Science 12: 199 – 208. Oksanen L, Fretwell SD, Arruda J, Niemela P. 1981. Exploitation ecosystems in gradients of primary productivity. American Naturalist 118: 240 – 261.  107  Oksanen L, Sammul M, Mägi M. 2006. On the indices of plant– plant competition and their pitfalls. Oikos 112: 149 – 155. Oksanen J, Blanchet FG, Kindt P, Legendre P, O’Hara RB, Simpson GL, Solymos P, Stevens MHH, Wagner H. 2010. vegan: Community Ecology Package. R package version 1.17 – 3. Olofsson J, Moen J, Oksanen L. 2002. Effects of herbivory on competition intensity in two arctic– alpine tundra communities with different productivity. Oikos 96: 265 – 272. Padilla FM, de Dios Miranda J, Pugnaire FI. 2007. Early root growth plasticity in seedlings of three Mediterranean woody species. Plant and Soil 296: 103 – 113. Parmesan C, Yohe G. 2003. A globally coherent fingerprint of climate change impacts across natural systems. Nature 421: 37 – 42. Pedrol N, Ramos P, Reigosa MJ. 2000. Phenotypic plasticity and acclimation to water deficits in velvet– grass: a long– term greenhouse experiment. Changes in Leaf morphology, photosynthesis and stress– induced metabolites. Journal of Plant Physiology 157: 383 – 393. Petchey OL, McPhearson PT, Casey TM, Morin PJ. 1999. Environmental warming alters food– web structure and ecosystem function. Nature 402: 69 – 72. Pfisterer AB,  Schmid B. 2002. Diversity-dependent production can decrease the stability of ecosystem functioning. Nature 416: 84 – 86. Pimm SL. 1984. The complexity and stability of ecosystems. Nature 307: 321 – 326. Post E, Pedersen C. 2008. Opposing plant community responses to warming with and without herbivores. Proceedings of the National Academy of Sciences 105: 12353 – 12358. Pywell RF, Bullock JM, Roy DB, Warman L, Walker KJ, Rothery P. 2003. Plant traits as predictors of performance in ecological restoration. Journal of Applied Ecology 40: 65 – 77. R Development Core Team. 2008. R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing. Reader RG, Jalili A, Grime JP, Spencer RE, Matthews N. 1992. A comparative study of plasticity in seedling rooting depth in drying soil. Journal of Ecology 81: 543 – 550. Rogler GA, Lorenz RT. 1983. Crested wheatgrass: early history in the United States. Journal of Rangemanagement 36: 91 – 93. Root TL, Price JT, Hall KR, Schneider SH, Rosenzweig C, Pounds JA. 2003. Fingerprints of global warming on wild animals and plants. Nature 421: 57 – 60. Rosenzweig C, Karoly D, Vicarelli M, Neofotis P, Wu Q, Casassa G, Menzel A, Root TL, Estrella N, Seguin B, Tryjanowski P, Liu C, Rawlins S, Imeson A. 2008. Attributing physical and biological impacts to anthropogenic climate change. Nature 453: 353 – 358. Rustad LE, Campbell JL, Marion GM, Norby RJ, Mitchell MJ, Hartley AE, Cornelissen JHC, Gurevitch J, GCTE– NEWS. 2001. A meta– analysis of the response of soil respiration, net nitrogen mineralization, and aboveground plant growth to experimental ecosystem warming. Oecologia 126: 543 – 562.  108  Sammul M, Kull K, Oksanen L, Veromann P. 2000. Competition intensity and its importance: results of field experiments with Anthoxanthum odoratum. Oecologia 125: 18– 25. Schulze ED, Robichaux Rh, Grace J, Rundel PW, Ehleringer JR. 1987. Plant Water Balance. BioScience 37: 30 – 37. Schwinning S, Weiner J. 1998. Mechanisms determining the degree of size asymmetry in competition among plants. Oecologia 113: 447 –  455. Shaw MR, Zavaleta ES, Chiariello NR, Cleland EE, Mooney HA, Field CB. 2002. Grassland responses to global environmental changes suppressed by elevated CO2. Science 298: 1987 – 1990. Schoener TW.  1983.  Field experiments on interspecific competition.  The American Naturalist 122: 240 – 285. Smith B, Wilson JB. 1996. A consumer’s guide to evenness indices. Oikos 76: 70 – 82. Smith MD, Knapp AK, Collins SL. 2009. A framework for assessing ecosystem dynamics in response to chronic resource alterations induced by global change. Ecology 90: 3279 – 3289. Smith MD. 2011. The ecological role of climate extremes: current understanding and future prospects. Journal of Ecology 99: 651 – 655. Solomon SD, Qin M, Manning Z, Chen M, Marquis KB, Averyt M, Tignor, Miller HL (eds.). 2007. Climate change 2007: the physical science basis. Cambridge University Press, Cambridge UK.  Stone L, Gabric A,  Berman T. 1996. Ecosystem resilience, stability, and productivity; seeking a relationship. The American Naturalist 148: 892 – 903. Suttle KB, Thomsen MA, Power ME. 2007. Species interactions reverse grassland responses to changing climate. Science 315: 640 – 642.  Svejcar T, Angell R, Miller R. 1999. Fixed location rain shelters for studying precipitation effects on rangelands. Journal of Arid Environments 42: 187 – 193.  Swemmer AM, Knapp AK, Snyman HA. 2007. Intra– seasonal precipitation patterns and above– ground productivity in three perennial grasslands. Journal of Ecology 95: 780 – 788.  Sydes CL, Grime JP. 1984. A comparative study of root development using a simulated rock crevice. Journal of Ecology 72: 937 – 946.  Thomas CD, Cameron A, Green RE, Bakkenes M, Beaumont LJ, Collingham YC, Erasmus BFN, Ferreira de Siqueira M, Grainger A, Hannah L, Hughes L, Huntley B, van Jaarsveld AS, Midgley GF, Miles L, Ortega– Huerta MA, Peterson AT, Phillips OL, Williams SE. 2004. Extinction risk from climate change. Nature 427: 145 – 148.  Thompson K. 1987. The resource ratio hypothesis and the meaning of competition. Functional Ecology 1: 297–  303.  109  Tilman, D. 1982. Resource competition and community structure. Princeton University Press, Princeton. Tilman D. 1985. The resource– ratio hypothesis of plant succession. American Naturalist 125: 827 – 852. Tilman D. 1988. Plant strategies and the dynamics and structure of plant communities. Princeton University Press, Princeton. Tilman D, Wedin D, Knops J. 1996. Productivity and sustainability influenced by biodiversity in grassland ecosystems. Nature 379: 718-720. Tilman D. 1999. The ecological consequences of changes in biodiversity: A search for general principles. Ecology 80: 1455-1474. Tisdale EW. 1947. The grasslands of the southern interior of British Columbia. Ecology 28: 346 – 382. Turkington R, Klein E, Chanway CP. 1993. Interaction effects of nutrients and disturbance: an experimental test of plant strategy theory. Ecology 74: 863 – 878. Tylianski JM, Didham RK. 2008. Global change and species interactions in terrestrial ecosystems.  Ecology Letters 11: 1351 – 1363. van Peer L, Nijs I, Reheul D, De Cauwer B. 2004. Species richness and susceptibility to heat and drought extremes in synthesized grassland ecosystems: compositional vs physiological effects. Functional Ecology 18: 769 – 778. van Ryswyk AL, Mclean A, Marchand LS. 1966. The climate, native vegetation and soils of some grasslands at different elevations in British Columbia. Canadian Journal of Plant Science 46: 35 – 50. Vare H, Ohtonen R, Mikkola K. 1996. The effect and extent of heavy grazing by reindeer in oligotrophic pine heaths in northeastern Fennoscandia. Ecography 19: 245 – 253. Voight W, Perner J, Jones TH. 2007. Using functional groups to investigate community response to environmental changes: two grassland case studies. Global Change Biology 13: 1720 – 1721. Vujnovic K, Wein RW, Dale MRT. 2000. Factors determining the centrifugal organization of remnant Festuca grassland communities in Alberta. Journal of Vegetation Science 11: 127 – 134. Wahren CHA, Walker MD, Bret– Harte MS. 2005. Vegetation responses in Alaskan arctic tundra after 8 years of a summer warming and winter snow manipulation experiment.  Global Change Biology 11: 537– 552. Walther G, Post E, Convey P, Menzel A, Parmesan C, Beebee TJC, Fromentin J, Hoegh– Guldberg O, Bairlein F. 2002. Ecological responses to recent climate change. Nature 416: 389 – 395. Walker MD, Wahren CH, Hollister RD, Henry GHR, Ahlquist LE, Alatalo JM, Bret– Harte MS, Calef MP, Callaghan TV, Carroll AB, Epstein HE, Jönsdöttir IS, Klein JA, Magnusson B, Molau U, Oberbauer SF, Rewa SP, Robinson CH, Shaver GR, Suding KN, Thompson CC, Tolvanen A, Totland O, Turner PL, Tweedie CE, Webber PJ, Wookey PA. 2006. Plant community responses  110  to experimental warming across the tundra biome. Proceedings of the National Academy of Sciences 103: 1342 – 1346. Watson RT, the Core Writing Team (eds.) 2001.  Climate Change 2001: Synthesis Report. Cambridge University Press, Cambridge.  397 pages. Weltzin JF, Loik ME, Schwinning S, Williams DG, Fay PA, Haddad BM, Harte J, Huxman TE, Knapp AK, Lin G, Pockman WT, Shaw MR, Small EE, Smith MD, Smith SD, Tissue DT, Zak JC. 2003. Assessing the response of terrestrial ecosystems to potential changes in precipitation. BioScience 53: 941 – 952. Weigelt A, Jolliffe P. 2003. Indices of plant competition. Journal of Ecology 91: 707– 720. Welden CW,  Slauson WL. 1986. The intensity of competition versus its importance: an overlooked distinction and some implications. The Quarterly Review of Biology 61: 23–  44. Welden CW, Slauson WL, Ward RT. 1988. Competition and abiotic stress among trees and shrubs in northwest Colorado. Ecology 69: 1566–  1577. Whitehead D, Hogan KP, Rogers GND, Byers JN, Hunt JE, McSeveny TM, Hollinger DY, Dungan RJ, Earl WB, Bourke MP. 1995. Performance of large open– top chambers for long– term field investigations of tree response to elevated carbon dioxide concentration. Journal of Biogeography 22: 307 – 313. Whitford WG, Rapport DJ, de Soyza AG. 1999. Using resistance and resilience measurements for ‘fitness’ tests in ecosystem health.  Journal of Environmental Management 57: 21 – 29. Whittaker RH. 1967. Gradient analysis of vegetation. Biological Reviews 49: 207 – 264. Willms WD,  Fraser J. 1992. Growth characteristics of rough fescue (Festuca scabrella var. campestris) after three years of repeated harvesting at scheduled frequencies and heights. Canadian Journal of Botany 70: 2125 – 2129. Wilson SD,  Keddy PA. 1986. Measuring diffuse competition along an environmental gradient: results from a shoreline plant community. American Naturalist 127: 862– 869. Wilson SD, Tilman D. 1995. Competitive responses of eight old– field plant species in four environments. Ecology 76: 1169 – 1180. Wilson JB, Lee WG. 2000. C– S– R triangle theory: community– level predictions, tests, evaluation of criticisms, and relation to other theories. Oikos 91: 77– 96. Wilson JB, Spijkerman E, Huisman J. 2007. Is there really insufficient support for Tilman’s R* concept? A comment on Miller et al.  American Naturalist 169 700 – 706. Wilson MV. 2007. Measuring the components of competition along productivity gradients. Journal of Ecology 95: 301– 308. Wu Z, Dijkstra P, Koch GW, Penuelas J, Hungate BA. 2010. Responses of terrestrial ecosystems to temperature and precipitation change: a meta– analysis of experimental manipulation. Global Change Biology 17: 927 – 942.  111   112 Yahdjian L, Sala OE. 2002. A rainout shelter design for intercepting different amounts of rainfall. Oecologia 133: 95 – 101. Zavaleta ES, Shaw MR, Chiariello NR, Thomas BD, Cleland EE, Field CB, Mooney HA. 2003. Grassland responses to three years of elevated temperature, CO2, precipitation, and N deposition. Ecological Monographs 73: 585 – 604.    Appendix A: Information on common grasses in Lac du Bois Provincial Park Table A.1 Information on 18 common grass species found in the study region, Lac du Bois Provincial Park, British Columbia Canada. Whether a species is annual or perennial, tufted, rhizomatous, exotic status and height were obtained from Klinkenberg (2010). Ten seeds of each species were individually weighed on a microbalance to determine seed mass. A = annual, P = perennial, N = no, Y = yes. Species Common name Annual/ perennial Tufted Rhizomatous Exotic Height (cm) Seed mass ± SE (mg) Agropyron cristatum (L.) Gaertn. Crested wheatgrass P Y N Y 50 -100 0.996 ± 0.056 Agrostis gigantea L. Redtop P Y Y Y 20 – 120 2.351 ± 0.088 Bromus inermis Leyss. Smooth brome P N Y Y 20-120 4.99 ± 0.072 Bromus squarrosus L. Corn brome A N N Y 20 - 60 2.078 ± 0.055 Bromus tectorum L. Cheatgrass A N N Y 20-50 2.150 ± 0.048 Dactylis glomerata L. Orchardgrass P Y N Y 150 1.193 ± 0.011 Elymus cinereus (Scribn. & Merr.) A. Love Giant wild rye P Y N N 70 - 220 2.354 ± 0.100 Elymus elemoides (Raf.) Swezey Squirrel-tail P Y N N 10-50 2.237 ± 0.038 Festuca campestris L. Rough fescue P Y N N 40-90 1.598 ± 0.027 Koeleria macrantha (Ledeb.) Junegrass P Y N N 30 -60 0.077 ± 0.001 Lolium perenne L. Perennial ryegrass P Y Y Y 100 1.919 ± 0.065 Phleum pretense L. Timothy P Y N Y 100 0.041 ± 0.0008 Poa compressa L. Canada bluegrass P Y Y Y 15 – 60 0.273 ± 0.018 Poa pratensis L. Kentucky bluegrass P Y Y N 30 – 70 0.216 ± 0.013 Poa secunda J.S. Presl  Sandberg's bluegrass P Y N N 15 - 120 0.214 ± 0.008 Pseudoroegneria spicata (Pursh) A. Love Bluebunch wheatgrass P Y N N 60 – 150 2.99 ± 0.074 Puccinellia nutalliana (Schult.) Hitchc. Nuttall's alkaligrass P Y N N 40 -80 0.145 ± 0.003 Vulpia octoflora (Walter) Rybd. Six week fescue A Y N N 6 – 30 0.300 ± 0.007  113  Appendix B: Calibration of soil moisture probes Introduction In the field experiments that manipulated water, temperature and clipping it was necessary to use probes to measure the effect of my treatments on soil moisture.  However, data from the probes contained a large number of negative values for volumetric water content (VWC) that, by definition, should not be less than zero.  An investigation of this problem suggested that this issue was related to the soil type at my experimental site.  A calibration procedure was found to correct the problem (Campbell 2002).  This appendix details the slight modification to this procedure that was used and provides the correction necessary to properly report the soil moisture data obtained from the experimental site. Methods The same soil moisture probes (10 cm long, Soil Moisture Smart Sensor, S-SMB-M005 using a ECH20® Dielectric Aquameter probe, Decagon Devices, Inc. connected to a HOBO® Micro Station data logger, Onset Computer Corporation) that were used in the field study were used in this calibration. Approximately 20 litres of soil was collected from the field site where the experiment was done (Lac du Bois Grassland Provincial park within the bunchgrass grasslands of the interior British Columbia, 6 km north of Kamloops, Canada (UTM E 680737  N 5625980; elevation 731 m a.s.l.). Soil was taken from eight separate soil pits 10 cm deep with the litter layer removed. This was the same depth that the soil probes were installed during the experiment.  The soil samples were sifted through a coarse, 5 cm x 2 cm, screen to remove large rocks and break up soil clumps, and then pooled. The calibration procedure followed Campbell (2002) with a few modifications.  The soil was equally divided among 6 containers (40 cm x 27 cm x 17 cm) to fill the container with approximately 10 cm of soil.  Water was added to each of the containers to create a range of soil moisture levels, 675 mL, 1150 mL or 1625 mL; two containers received each level of water addition.  The water was thoroughly mixed into the soil.  Next, two soil moisture probes were placed horizontally 3 cm below the soil surface in each container and the soil was pressed firmly and evenly down.  The soil moisture readings were recorded and two soil cores (2 cm radius, 3.5 cm deep) were removed from each container.  114  The soil cores were placed in jars with lids that had been weighed.  The wet soil samples were weighed, then dried for 1 week at 65ºC and then weighed again. With this information the soil weight, water content, bulk density and VWC can all be determined.   Dry soil weight = dry weight – jar weight    (1)   Wet soil weight = wet weight – jar weight   (2)   Water weight = wet soil weight – dry soil weight  (3)   Bulk density = soil weight / volume of soil core  (4)   VWC = water weight/ soil weight x bulk density  (5) To calculate the calibration equation, the actual VWC of the samples determined through the above procedure was plotted against the observed values from the soil moisture probes (using the means of probes in each container). In Excel, a linear trend line was plotted through the data; the equation of this line is the calibration equation. Results The calibration procedure shows that the probes underestimate the actual VWC of the soil. Field readings should be adjusted using the equation: Actual VWC = 1.3885 x Observed VWC + 0.0798 An R2 value of 0.915 for this regression suggests that a linear calibration is sufficient.   115   116 y = 1.3885x + 0.079 .9158R 2 = 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 Measured volumetric water content A ct ua l v ol um et ric  w at er  c on te nt  Figure B.1 Actual volumetric water content plotted against the measured volumetric water content.  The equation in the figure is the formula required to adjust field readings. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0072773/manifest

Comment

Related Items