Homogenizing Effects of Historic Lake Bonneville: Can Phylogeography Inform Community Assembly?

Authors: Trevor Williams, Jerald B. Johnson

Year: 2018

Abstract

What structures ecological communities? This is one of the predominant questions asked in community ecology. Traditionally, ecologists have tackled this question by investigating the contemporary biotic and abiotic factors that govern community assembly, such as predation, competition, and environmental filtering. Though informative, this approach neglects historical factors and may be overlooking important assembly drivers such as shared dispersal histories or climatic alterations that can have lasting effects on community pattern. However, modern phylogeographic techniques are now able to estimate and model species responses to historical phenomena and therefore can be used to examine historical community assembly. We investigated the contemporary and historical factors governing community assembly of freshwater fishes in the Bonneville Basin in Utah. The Bonneville Basin is composed of several isolated, closed drainage basins that over the past 30,000 years have been temporarily connected by large lakes due to climatic change during the last glacial maximum. Approximately 14,500 to 16,000 years ago, the entire Bonneville Basin was covered by Lake Bonneville, thereby connecting all of the separate basins and presumably allowing aquatic species to migrate throughout the Bonneville Basin. Because its geologic history, as well as its fish taxa, are well studied, Lake Bonneville is an ideal system for studying both the historic and contemporary factors that govern community assembly. To investigate the historic factors governing community assembly, we gathered sequence and allozyme data from Genbank and the literature and used them in coalescent simulations to test whether fish species were able to migrate through Lake Bonneville. To determine the contemporary factors working on community assembly, we used co-occurrence null models to analyze whether species are randomly or non-randomly structured in respect to each other. Co-occurrence null models were run on historic occurrence data collected before 1940 in order to minimize the effects of non-native introductions on communities. These methods allowed us to analyze two sets of hypotheses. If Lake Bonneville allowed migration to occur, we hypothesized that communities would be structured randomly due to ecological drift. If Lake Bonneville did not allow migration, we hypothesized that communities would be structured non-randomly, likely due to biotic and abiotic habitat requirements. We found that fish species had independent responses to the rise of Lake Bonneville. Some species were able to migrate while others did not. Likewise, the co-occurrence null models indicated that fish communities are structured randomly. These results support our hypothesis that Lake Bonneville allowed homogenization of communities across the Bonneville Basin. Current membership in communities seems to be structured due to ecological drift as lake levels dropped and communities became isolated. Our results further show how phylogeographic methods can be implemented to investigate the historical aspects of community assembly.