An Examination of Stream Fishes at 3 Levels: Species-Area and Environmental Interactive Influences on Fish Species Richness, Stream Type Influences on Assemblage Membership, and Variation in Weight-Length Parameters by Stream Type & Fish Trait

Abstract

Obtaining an understanding of the multi-scale influences on species richness is of concern not only to theoretical development but to building an understanding of potential mechanisms of rapid species loss. To address modern conservation and management concerns we must further our understanding of fundamental ecological patterns. One of the most basic ecological phenomenon is the increase in the number of species with an increase in area. Though an oft-studied phenomenon the mechanisms that underlie this basic ecological pattern are not well understood. Though contributions to our understanding of this phenomenon have been gleaned from research conducted in many systems the most well-known of the species-area theories, island biogeography, credits lotic systems with being an especially useful ecotype to study the underlying mechanisms of these phenomena. Within streams many environmental variables influence fish assemblage membership across regions and within basins at the stream level. Current stream theoretical approaches explaining the hierarchical influences of fish assemblage membership within a basin lack the precision to predict differing assemblages at a relatively close spatial scale, although this is known to occur. To better develop our understanding of what dictates fish assemblage further research must address what the fundamental environmental differences are between streams in close proximity, even within a basin, with different assemblages. Indeed, because closely spaced streams within the same or nearby basins presumably share a similar regional species pool identifying differences in the fish assemblage membership and correlated environmental parameters would provide greater insight into the determinants of assemblage composition. The 1st chapter of my dissertation discusses the levels of investigation I conducted in subsequent chapters to examine influences on stream fish richness and assemblage membership at multiple scales. In the 2nd chapter I used multiple linear regression and 3 model selection techniques (stepwise, all-subsets, and AICc) and 1 model shrinkage technique (Lasso) to determine which environmental parameters were most important predictors of native fish species richness as singular and interactive terms with stream flow. I found that most of the environmental predictors investigated influenced fish species richness independent of stream flow, while some environmental predictors were also correlated with native fish species richness as interactive terms with stream flow. These results indicated that environmental predictors act on fish species richness both independently of and in conjunction with an increase in flow. Similarly, results of similar analyses on average values of environmental predictors indicated that native fish species richness was related similarly to both measures for some environmental predictors whereas other predictors show associations unique to each measure. These results indicate that ascribing the species-area relationship to a simple linear increase in heterogeneity with an increase in area may be too simplistic a model and that similar to a patch dynamics perspective both the variety of and the variability in stream habitat are important drivers of fish species richness. In my 3rd chapter I investigated whether there were fish assemblage and environmental differences between blackwater and clearwater streams of coastal Alabama. My results suggest that streams of coastal Alabama exhibit 2 unique types, blackwater and clearwater, each with unique environmental parameters and fish assemblage compositions. In my 4th chapter, I calculated weight-length regression (WLR) parameters for 17 species (9 families) of Alabama coastal stream fishes, most of which were non-game species with previously unpublished parameters. I also investigated whether season, stream type, environmental variables, or species traits influenced WLR parameters. My results suggested WLR parameters were affected by season with higher WLR slope (b) in fall-collected fish than in spring and summer collections. My results also suggested that the slope of log length was lower in clearwater stream populations and that stream water pH and invertebrate density influenced fish weight. Species closely associated with clearwater stream assemblages and surface-water column feeders demonstrated a shift into a juvenile life stage at a smaller size. My results suggested that correlating WLR parameters with both abiotic and biotic factors and life history characteristics can provide further insight into contrasting patterns of season, stream type fish assemblage membership, environmental variables, and species traits. Overall, my dissertation explores ecological questions using stream fish from the theory level in investigating potential mechanisms underlying species-area relationships, at the landscape level addressing whether and how stream fish assemblages varied by stream type, and at the population level using WLR data to investigate ecological questions with a variety of temporal, environmental, and trait predictors. My dissertation highlights the utility of using stream fishes across a range ecological investigations. My work also adds to our understanding of the species-area relationship by expressing how the contribution of different environmental predictors may need to better understood independently or in conjunction with an increase in stream flow (as a proxy for area). Additionally, it appears that the relative magnitude of some environmental predictors are important correlates of fish species richness whereas for other predictors variability might have a greater influence on fish species richness. This dissertation also provides a more definitive identification of separate stream types in coastal Alabama complete with unique fish assemblages which should better aid conservation and management objectives in this region. Finally, WLR approaches are underutilized for ecological applications and this dissertation serves as a means of broadening those investigations. My work demonstrates that more ecological information can be gleaned about the effects of seasonal variation, environmental conditions, and traits on fish populations using weight-length data.