Date of Award
Doctor of Philosophy (PhD)
Peter van den Hurk
Metacommunity theory incorporates local and regional factors to understand how biotic communities are structured across the landscape. Despite established knowledge of how humans impact aquatic systems, inclusion of anthropogenic factors in metacommunity studies have been largely ignored. Additionally, alpha, beta, and gamma diversity can all be explored at the metacommunity level to investigate mechanistic drivers of community structure. Beta diversity can be further partitioned into turnover and richness difference components, each with different mechanistic drivers. Streams provide an excellent study system for metacommunity research because of the dendritic structure of watersheds and the natural delineation that watershed boundaries provide. Large-extent datasets provide the ability to create multiple metacommunities serving as replicates for robust statistical analyses. As such, the overall goal of this dissertation was to use large datasets of stream fish community structure to investigate how anthropogenic variables affect stream fish beta diversity and metacommunity structure in conjunction with ‘traditionally investigated’ factors including natural landscape features and spatial distance among communities. This research uses two large extent datasets. The first covers 13 states on the eastern coast of the United States, and the second covers 350 sites throughout South Carolina. Three different approaches were taken to understand the factors affecting stream fish metacommunities across the landscape.
First, we created a spatial scale continuum using nested watersheds identified by hydrologic unit codes (HUCs) to explore how beta diversity and its components change over three spatial scales, and (a) how land use, (b) climatic, and (c) anthropogenic factors affect beta diversity within and between spatial scales. We found increasing beta diversity with increasing spatial scale, and equal contribution between turnover and richness difference components. All three factors were related to beta diversity or its components depending on the spatial scale, but few scale-dependent relationships were found. These results suggest that while a diversity of factors affect beta diversity at a given spatial scale their effects on beta diversity do not change across spatial scales. These effects may be scale-invariant, although other cross-scale effects may arise at finer spatial scales.
Second, we investigated how environmental and anthropogenic factors and aspects of study design affect coherence, turnover, and boundary clumping—the elements of metacommunity structure (EMS) at a single spatial scale. The EMS were affected by temperature, density of dams, and percentage of developed land, but also gamma diversity and number of sites sampled in a metacommunity. These results suggest that anthropogenic factors affect the elements of metacommunity structure and thus set the context for assigning metacommunities into archetypical processes across the landscape. Moreover, the EMS were affected by study design aspects such as the number of communities sampled and the distance between them within metacommunities. These results have important implications for both existing and future studies because they show that inference on spatial processes is contextualized by aspects of datasets that are inherent to datasets and are rarely considered in analyses. Including these factors in future analyses will allow researchers to better focus on the signal of key processes by accounting for the variability caused by aspects of study design.
Third, we used variation partitioning to parse out the relative effects of anthropogenic, natural, and spatial factors on beta diversity as measured in three key dimensions: taxonomic, functional, and phylogenetic. These analyses were done at three spatial delineations representing artificial, geomorphic, and natural watershed metacommunities. These are commonly used spatial delineations in metacommunity analyses, but are rarely included in the same study. We explained 25-81% of beta diversity where different spatial, natural, and anthropogenic factors structured these metacommunities depending on the spatial delineation and diversity dimension. Geomorphic metacommunities had very different results compared to other spatial delineations suggesting that accounting for geomorphic differences leads to stronger anthropogenic signals. By conducting this work in different spatial delineations within the same dataset, we show for the first time that defining metacommunities has bearing on results of analyses—an issue that is rarely considered in metacommunity studies.
Overall, this body of work suggests that anthropogenic factors have pervasive effects on stream fish beta diversity and metacommunity structure across the landscape—an aspect of metacommunity ecology that has until recently been largely ignored. This work suggests that considering anthropogenic effects in metacommunity studies will improve inference. Researchers must also consider important aspects of study design, including how metacommunities are defined and delineated, as well as how intensely and densely communities are sampled within those metacommunities. In all, this dissertation adds an important practical dimension to the field of metacommunity ecology, which up to this point has been largely theoretical. Considering these practicalities may improve our overall understanding of metacommunities in a variety of taxa and systems.
Stoczynski, Lauren E., "Incorporating Human Effects in Quantifying Mechanisms of Stream Fish Community Structure Using Metacommunity Theory" (2022). All Dissertations. 3056.
Author ORCID Identifier