Freshwater biodiversity is at once the most diverse and the most imperiled among the world's ecosystems. In the southwest, regional biodiversity and endemism face challenges imposed by declining water availability and widespread nonnative species proliferation. In this dissertation, I explore how these challenges affect fish community dynamics and native species persistence in dryland rivers, and explore the effectiveness of nonnative removal programs toward native fish conservation. The overarching questions motivating my research are: (1) How are fish communities responding to a changing climate? (2) How does flow intermittence and species origin shape freshwater fish beta diversity across dryland riverscapes? (3) Can we restore native species food-web dynamics through invasive species management? (4) Do strategic and opportunistic removal programs result in measurable, and if so comparable, benefits to native species conservation? Demographic models linking native and nonnative populations to flow dynamics predicted that contemporary declines in the frequency of peak flows, and increases in drought frequency are likely to result in nonnative dominant fish assemblages and diminished native fish populations. I found that intermittent and perennial streams play complementary roles in supporting fish beta diversity, and that contributions of intermittent streams to overall beta diversity were relatively consistent through time, primarily supporting a unique composition of native fishes. Although nonnative species control and removal programs are a common management strategy they have not always been successful. However, I found that nonnative removal efforts allowed native species to recover in their food-web dynamics, by returning to higher trophic levels and isotopic niches comparable to individuals that did not co-occur with nonnative fishes. In a model informed by long-term monitoring programs, I also found that both opportunistic and strategic removal strategies were predicted to decrease native fish extinction probabilities. These results were encouraging, and demonstrated that removal programs can meet recovery goals even over large areas and long after nonnative species are established.

The most comprehensive synthesis of stream fish community research ever produced. Winner of the CHOICE Outstanding Academic Title of the Choice ACRL Ecologists have long struggled to understand community dynamics. In this groundbreaking book, leading fish ecologists William Matthews and Edie Marsh-Matthews apply long-term studies of stream fish communities to several enduring questions. This critical synthesis reaches to the heart of ecological theory, testing concepts against the four decades of data the authors have collected from numerous warm-water stream fish communities in the central and eastern United States. Stream Fish Community Dynamics draws together the work of a single research team to provide fresh analyses of the short- and long-term dynamics of numerous streams, each with multiple sampling sites. Conducting repeated surveys of fish communities at temporal scales from months to decades, the authors' research findings will fascinate anyone searching for a deeper understanding of community ecology. The study sites covered by this book range from small headwater creeks to large prairie rivers in Oklahoma and from Ozark and Ouachita mountain streams in Arkansas to the upland Roanoke River in Virginia. The book includes • A comparison of all global and local communities with respect to community composition at the species and family level, emergent community properties, and the relationship between those emergent properties and the environments of the study sites • Analyses of traits of individual species that are important to their distribution or success in harsh environments • A review of evidence for the importance of interactions—including competition and predation—in community dynamics of stream fishes • An assessment of disturbance effects in fish community dynamics • New analysis of the short- and long-term dynamics of variation in stream fish communities, illustrating the applicability and importance of the "loose equilibrium concept" • New analyses and comparisons of spatiotemporal variation in community dynamics and beta diversity partitioning • An overview of the effects of fish in ecosystems in the central and eastern United States The book ends with a summary chapter that places the authors' findings in broader contexts and describes how the "loose equilibrium concept"—which may be the most appropriate default assumption for dynamics of stream fishes in the changing climate of the future—applies to many kinds of stream fish communities.

The upper Gila River basin in southwest New Mexico, USA is one of the few unimpounded drainage basins in North America and is a stronghold for the unique and endemic fishes west of the Continental Divide. Multiple non-indigenous fishes have been introduced to the Gila River and are a potential threat to native fishes, yet very little is known of the trophic ecology of the native and nonnative fishes. We used diet and stable isotopes collected from native and nonnative fishes to identify their trophic relationships and evaluate potential interactions in the upper Gila River basin during June-July, 2007 and 2008. Diet and stable isotope data indicated aquatic invertebrates were the primary food for both native and nonnative fishes. Native large-bodied fishes were mainly algivore/detritivores and native small-bodied fishes were primarily insectivores. Small-bodied nonnative fishes fed on detritus and aquatic invertebrates. Nonnative predators preyed on small-bodied fishes and predaceous aquatic invertebrates and had higher trophic positions than all native fishes. Although nonnative predators did not rely exclusively on native fishes as prey, their presence extended community food-chain lengths, and the combined predation on juvenile native fishes by multiple apex predators may threaten persistence of native fishes. The lack of concise evidence for negative effects suggested that impacts of nonnative predators were more subtle and confirmed the underlying complexity of a relatively simple community The extensive database on feeding relations of Gila River fishes allowed us to further understand how energy moves through ecosystems. Specifically, the goal of chapter two was to characterize variation in fish-community food web structure within and among study reaches on the Gila River using [superscript]13C and [superscript]15N stable isotopes. We hypothesized that food web structure would reflect variation in fish community structure, resource availability and environmental conditions across habitats. Food web structure in isotope bi-plot space was estimated using community-wide measures of trophic structure, mean trophic position, and food-chain length. Permutational multivariate analysis of variance indicated that indices of food web structure were more variable among than within reaches and this pattern was primarily associated with variation in trophicl area occupied by taxa in isotope bi-plot space and mean trophic position of those taxa. Variation in food web structure was significantly associated with fish species richness across macrohabitats but was weakly associated with abiotic reach-scale factors. Variation in food web structure was concordant with variation in fish community composition and suggested that factors influencing the distribution of fishes also influence food web structure.

PUBLIC ABSTRACT: Native fishes of the Colorado River Basin have experienced dramatic reductions in range and abundance as a result of extensive human alterations to the basin's waterways. Many of these native fishes are federally listed under the Endangered Species Act, while several others are subject to range-wide conservation agreements between state and federal management agencies. Three of the native species subject to range-wide conservation agreements are the flannelmouth sucker, bluehead sucker, and roundtail chub (hereafter, the "three species"). Each of the "three species" is still found in the San Rafael River of southeastern Utah, which has experienced habitat degradation and non-native species establishment representative of many desert streams. In this study, I examined the effects of non-native species on the food web structure and relative growth rates of the "three species" (Chapter 2) in two sections of the river characterized by the presence and absence of non-native fish. I found that the presence of non-native species lengthens the food chain, presenting new predators and competitors to the 'three species.' However, I found no evidence of reduced growth in the presence of these non-native fishes, likely due to movement of individuals of the "three species" between the two sections of river. Secondly, I developed a model to identify and rank limiting factors to the "three species" along the continuum of the lower river. Finally, I used this model to simulate and predict the relative effect of different restoration actions at different locations along the river on the abundance of the 'three species.' These models predicted that removal of non-native fishes and the restoration of long stretches of suitable habitat would be most beneficial to the 'three species.' Models such as those developed in this study can be useful for management agencies to prioritize restoration efforts to ensure the persistence of the "three species" both in the San Rafael River and throughout their historic range.

Freshwaters ecosystems support extraordinary biodiversity relative to their extent and provide important societal benefits. As such, freshwater environments and biota are often heavily impacted by anthropogenic activities. Freshwater fishes in Mediterranean-climate regions are especially impacted because of large human populations in these regions and extensive agricultural production, extensive river modification for flood control and to meet societal demands, and because these systems are heavily invaded by non-native organisms. The distribution and ecology of freshwater fishes in Mediterranean-climate regions are also influenced by the distinct wet and dry periods and the high inter-annual variability in precipitation. Thus, efforts to manage and conserve native fishes in Mediterranean-climate regions require understanding both the effects of human disturbance and the strong seasonality that characterizes these regions. In this dissertation, I examine the relationship between land use and Mediterranean seasonality on freshwater fishes in streams within the greater San Francisco Bay region in California, USA. In my second chapter, I use a multivariate approach to explore variability among fish communities in 25 Bay Area watersheds. I found that a combination of local (water conductivity) and watershed-scale factors (percent forested watershed, watershed area, elevation) were important predictors of fish communities across sites. Furthermore, watershed-scale factors had indirect effects on fish communities through their influence on a local-scale factor, water conductivity. The results of this chapter highlight the importance of considering both the direct and indirect effects of watershed-scale factors on freshwater fish communities. In my third chapter, I continued my analysis of land use and fish communities with a focus on contemporary land change. For this chapter, I performed a resurvey study, surveying the habitat and fish communities in 32 sites in the Alameda Creek Watershed that had been surveyed by Dr. Robert Leidy in the mid-1990s, including sites in the rapidly urbanizing Livermore Valley region. Again using a multivariate approach, I found that the increase in urbanization across an approximately 16-year period was related to change in fish community composition, a decline in native species richness, and a decline in a common native cyprinid - changes that were not observed in another part of the watershed that has experienced little land use change in the last 16 years. The relationship between land use change and fish community change was strongest when considering land use change at a local scale. These results suggest that ongoing land change alters fish communities and that contemporary resurveys are an important tool for examining how freshwater taxa respond to recent and ongoing environmental change. In my final chapter, I assessed how seasonal drought, a characteristic feature of Mediterranean-climate systems, influenced food webs in a small intermittent stream in Marin County, CA that provides rearing habitat for threatened steelhead trout (Oncorhynchus mykiss). I used stable isotopes of carbon and nitrogen to characterize food web structure and the trophic position of a suite of predators in this system, including O. mykiss and several macroinvertebrate predators. I compared food web snapshots across time, as well as between permanent and temporary pools. I found that the intermittent stream food web remained relatively stable across time and did not differ between pool types. However, I also found significant changes in the trophic position, niche width, and mean [delta]13C values for aquatic predators. This study provides an important first look at the trophic ecology of an imperiled fish species in intermittent streams during the summer drought season, and emphasizes that food chain length increases across the drought season, possibly because invertebrate prey are concentrated with declining water levels. In conclusion, my research shows that anthropogenic factors at the watershed scale influence instream conditions and freshwater fish communities, and emphasizes that contemporary changes in land use can have subtle changes on fish community structure, which may be indicative of future declines of extirpations of native fishes. Finally, my research shows that changing conditions across the summer drought season lead to shifts in the trophic ecology of some, but not all, aquatic predators, including threatened steelhead trout. Overall my research contributes to a growing body of research that demonstrates how multi-scale natural and anthropogenic factors influence freshwater fishes in Mediterranean-climate region.

Non-native species introductions are a ubiquitous form of environmental change. However, the role of introductions in ecosystem functioning is still poorly understood, especially in highly invaded systems with multiple non-native species. This thesis employs a functional trait framework-- in which the net ecosystem effect of changing community structure is evaluated by quantifying change in the community-level distribution of biological traits-- to assess the potential effects of multi-species introductions on ecosystem functioning in the Verde River, Arizona, USA. Specifically, I assess changes in body size distributions associated with non-native fish and crayfish introductions to estimate change in consumer-mediated nutrient recycling-- an important size-scaling function in which aquatic consumers can control the rates and ratios of inorganic nutrient availability. In chapter 1, I compare the central moments of individual size distributions of basin-wide native and non-native fish species pools. Native and non-native species pools were characterized by significantly different mean, coefficient of variation, and skewness moments of individual size distributions. These differences were more pronounced within trophic guilds than across the whole community, and notably differences in higher order moments (CV, skewness) were relatively greater than differences in the mean of size distributions. In Chapter 2 I evaluated whether such changes in the variance of size distributions affected nutrient recycling independent of the mean. I coupled nutrient recycling incubations and field sampling of habitat-specific size distributions of a non-native crayfish (Orconectes virilis) to scale up nutrient recycling from individuals to aggregate ecosystem functioning using mean-only or variance-incorporating approaches. A mean-only approach overestimated true rates of aggregate nutrient recycling by as much as 20% in habitats with low mean body size, but the bias induced by ignoring variance declined with the mean of the distribution. Given that the relationship between individual body size and nutrient recycling is a general representation of many size-functioning relationships, these qualitative results likely hold for other consumer-mediated functions. In Chapter 3 I determined whether body size provided a common functional currency to predict individual nutrient recycling across multiple fish species, and in turn whether native to non-native fish turnover was expected to generate large differences in aggregate nutrient recycling in the Verde River. Body size-recycling models with species-specific parameters were more parsimonious than global models with parameters shared across species. Using species-specific models I found that non-native dominated communities excreted ammonium at similar rates as native-dominated communities, but phosphate at significantly lower rates. The resultant difference in the N:P ratio was surprisingly large, generally independent of body size, and potentially important for aquatic microbial communities in this system. By contrast, the global model incorporating body size but not taxonomy did not capture this significant N:P difference. Together these chapters suggest that quantifying change in body size distributions yields important ecological insights, but taxonomic identity or additional traits are still necessary for a full evaluation of the ecosystem-level effects of multi-species introductions.