There is a pressing need for research into the impacts on organisms of climate change and human land-use change due to the increasingly urgent threat to biodiversity they impose. Since the start of the industrial revolution, anthropogenic greenhouse gas emissions have caused a marked increase in global mean temperatures as well as shifts in temperature variances. Conversion of natural land for human use has resulted in the destruction of vital habitat and has altered the spatial structure of terrestrial landscapes. As conservation is intimately linked with spatial context, it is vital to develop a thorough understanding of the consequences of environmental spatial structure for organisms, particularly in regard to how the spatial structure might influence responses to climate change. Studies investigating the effects of spatial structure on organisms often adopt a simulation modeling approach due to the difficulty of surveying and tracking organisms over large spatial scales and long time periods. Such modeling studies often explore the impact of spatial structure through the lens of metapopulation dynamics. A major drawback of this approach is that metapopulation studies only consider single species, making extrapolations to the community level problematic. Metacommunity studies in contrast consider multiple metapopulations of potentially interacting species, but metacommunity modeling studies are often not spatially explicit at all or only consider space in a highly simplified manner. In this thesis, I attempt to bridge this research gap using a spatially explicit individual-based simulation model to systematically explore how compositional heterogeneity, the magnitude of spatial environmental variation, and spatial autocorrelation affect the way diversity patterns and adaptation patterns at the landscape level respond to an environmental shift mimicking climate change. My model simulates guild-like communities of annual, asexual organisms living, reproducing, and dispersing in continuous fractal landscapes. Landscapes in this model consist of grids of patches with two environmental attributes, both of which vary independently over space, and one of which fluctuates over time. Organisms possess niche optima and tolerances for the respective patch attributes which determine expected reproductive output and two traits governing dispersal probability and dispersal distance respectively. Communities are additionally subject to immigration from external sources and extinctions if a species fails to successfully produce offspring, allowing for temporal community turnover. To acquire a baseline for comparison against, I first ran the model over a series of scenarios in which the environmental mean remained constant, and analyzed the richness, diversity and adaptational patterns that arose under different landscape structures. Regarding richness and diversity patterns, I was able to demonstrate that compositional heterogeneity and spatial autocorrelation had different effects at the landscape and patch levels due to different community assembly processes dominating. Landscape diversity was driven by patch-level community dissimilarity and increased with greater compositional heterogeneity, while landscape richness was largely driven by the accumulation of immigrant species and decreased with increasing compositional heterogeneity. Greater compositional heterogeneity resulted in a reduction in the temporal variance of landscape richness and diversity, indicating reduced temporal community turnover. Patch-level richness and diversity patterns were driven by overall landscape richness and by mass effects, leading to maximum richness and diversity at intermediate heterogeneity. Spatial autocorrelation affected landscape richness and diversity through its effects on beta diversity, with low autocorrelation producing greater landscape richness and diversity. At the patch level, high spatial autocorrelation promoted greater richness and diversity due to greater spillover between more similar neighboring patches. For adaptation patterns, I could show that compositional heterogeneity and spatial autocorrelation had varying selective effects on niche traits and dispersal strategy. These selective effects depended not only on the overall spatial structure of the landscape, but also on temporal environmental variation and the amount of suitable habitat available to a species. Greater compositional heterogeneity produced a greater variety of environmental niches in organisms. Tolerance traits behaved differently depending on whether the respective environmental attribute only varied over space or fluctuated over time as well, indicating that tolerances were driven by a balance between pressure to avoid risk imposed by the spatial environment and pressure to hedge against temporal fluctuation in the environment. Dispersal probability and dispersal distance responded to compositional heterogeneity and spatial autocorrelation in different manners and to different degrees, underscoring the importance of considering the two as distinct facets of dispersal strategy. Organism traits exhibited syndromes in relation to patch environments, with varying relationships depending on the degree of compositional heterogeneity, with an increasing tendency towards greater dispersal and tolerance in rare patch habitats as compositional heterogeneity increased. Finally, I ran a series of scenarios in which the landscapes were subjected to a sustained global shift in one of the patch attributes mimicking climate change. Here, my goal was to examine how the patterns that arose under a stable environmental mean changed when that mean shifted. I was able to show that the effect of landscape structure influenced diverse responses to an environmental shift, primarily through its effects on temporal turnover of species. Low compositional heterogeneity led to rapid extinction and replacement of pre-existing communities with recent immigrants, leading to trait patterns defined by the external species pool rather than by selection imposed by the landscape. Higher compositional heterogeneity slowed the rate of extinctions leading to slower turn-over by permitting species to track environmental changes by shifting their distribution within the landscape. This resulted in selective effects having a more prominent role in shaping niche and dispersal traits and allowed more of the pre-shift community to survive. As with before the environmental shift, I found that diversity patterns differed between landscape and patch levels. Regarding traits, post-shift communities were characterized by shifted niche optima for the environmental attributes which changed, greater dispersal frequency, reduced dispersal distance, as well as increased tolerances, including to environmental attributes which did not shift. The greater tolerances were attributable to the increase in dispersal, which results in species being exposed to a greater variety of patch conditions, thereby selecting for greater tolerances. Pre-shift species which survived the shift showed traits similar to the larger post-shift communities.
Joseph Paul Tardanico (Wed,) studied this question.