PS 60-18
Flow anomalies determine long-term trajectories of native, but not non-native fish, in a desert stream

Friday, August 15, 2014
Exhibit Hall, Sacramento Convention Center
Albert Ruhí, Global Institute of Sustainability, Arizona State University, Tempe, AZ
Elizabeth E. Holmes, Northwest Fisheries Science Center, NOAA Fisheries Service, Seattle, WA
John N. Rinne, Southwest Forest Science Complex, Rocky Mountain Research Station, Flagstaff, AZ
John L. Sabo, School of Life Sciences, Arizona State University, Tempe, AZ
Background/Question/Methods

Environmental variation is central to ecological theory, and influences processes across all levels of organization. However, our ability to anticipate the impacts of stochastic events at the community level is rudimentary. We selected the upper Verde River, a relatively undisturbed and highly variable desert river in Arizona, to study how environmental stochasticity determines variation in temporal and spatial community patterns, contrasting native to non-native fishes’ responses. We quantified stochastic variation in discharge using Fourier (FFT) analyses on >15,000 daily observations. We subsequently coupled yearly streamflow anomalies (unexpected floods and droughts) with long-term time series of fish abundances (1994-2008) using Multivariate Autoregressive State-Space (MARSS) models. These time series models can disentangle biotic from abiotic interactions and quantify the effects of stochastic drivers in a precise and species-specific way. In our case, MARSS-FFT allowed us to test whether the widely described negative covariation between flow alteration and native fish abundance is a result of changes in streamflow, or a result of flow-mediated changes in non-native fish abundance.

Results/Conclusions

Abiotic (stochastic) drivers were paramount in determining long-term fish abundances, whereas biotic interactions played only a secondary role. As expected, anomalous droughts reduced the abundances of native species while floods increased them. However, in contrast to previous studies, we observed that the non-native assemblage did not respond inversely to streamflow anomalies, but was surprisingly unresponsive. Biological trait analyses showed that functional uniqueness was higher in native than in non-native fishes, with the native assemblage being typified by traits that conferred resilience from, or resistance against, flood disturbance. We also found that streamflow anomalies influenced the temporal variation in spatial composition patterns. In particular, global beta diversity was unaffected by streamflow anomalies, but droughts increased the relative importance of the nestedness component, making post-drought local communities relatively more predictable in composition. Overall, our results advance the notion that environmental variation is key for determining long-term fish abundances. We suggest this variation, rather than biotic interactions, may commonly underlie covariation between native and non-native faunas, especially in variable environments. If droughts continue to increase in frequency and magnitude due to climate change, and floods become increasingly muted due to regulation, fish assemblages in desert rivers may become taxonomically and functionally impoverished and dominated by non-native taxa.