Regional conservation strategies to address climate change effects on native biota should benefit from a good understanding of which species are most likely to be affected, and where on the landscape the greatest effects are likely to occur. To that end, we are working within the USGS Southeast Regional Assessment Project (SERAP) to develop a multi-scale framework for simulating climate change effects on stream fishes and mussels. We are piloting this work in the Apalachicola-Chattahoochee-Flint (ACF) river basin, which contains over 130 native species of fishes and mussels, including multiple species classified as imperiled. At a coarse level of resolution, our approach uses expert-knowledge, species distributions, and climate projections to estimate probabilistic effects of temperature and streamflow change on future population status of fish and mussel species in individual ACF sub-basins (n=246, defined as hydrologic response units in a basin-wide precipitation-runoff model). We have used input from regional aquatic ecologists and resource managers familiar with the ACF system to develop a conceptual model of expected species responses to climate change.
Results/Conclusions
In a facilitated workshop and follow-up meetings, regional experts identified changes in stream temperature, duration and frequency of low flows, peak flows, and flow variability prior to spawning and during spawning and rearing seasons as primary drivers of species persistence. However, effects of climate drivers are expected to depend on channel dynamics (e.g., degrading, aggrading or stable), which are modeled as conditional on geomorphic characteristics (land use, geology, stream size distribution, channel slopes) mapped and summarized for sub-basins. Species-specific vulnerability is conceptualized as a function of species traits (fluvial dependence, stream size preference, temperature and dissolved oxygen tolerance, and reproductive and dispersal characteristics). The models for fishes and mussels are constructed as Bayes networks that link probabilistic climate conditions to future population status via effects on habitat and reproductive potential. Although developed for the ACF, conditioning these models on geomorphic context and species traits will allow model application to other basins and species. We intend the coarse-resolution models as a first-step in developing conservation strategies. Moreover, analyzing model sensitivity and comparing predictions derived from expert-judgment to aggregated predictions from fine-resolution simulation models should be useful for identifying and prioritizing key model uncertainties.