Historical disturbances to rivers appear to constrain contemporary distribution of a river-dependent frog,

Abstract

Frogs dependent on lotic environments are sensitive to disturbances that alter the hydrology (e.g., water impoundments), substrate (e.g., debris torrents), and riparian vegetation (e.g., wildfires) of river ecosystems. Although rivers are often very dynamic, disturbances can push environmental baselines outside of narrowly defined ecological tolerances under which a species evolved. Short-lived lotic-dependent organisms, restricted to movements within the water or the riparian corridor, are at risk of local extirpations owing to such disturbances if they fragment and isolate affected populations from recolonizing source populations. In Oregon, USA, the foothill yellow-legged frog (Rana boylii) is at its northernmost range margin and has experienced an approximately 41% range contraction compared to their historical distribution. To inform conservation and management, we used species distribution models to identify environmentally suitable watersheds based on intrinsic baseline environmental variables, and then examined potential effects of human-caused alterations to rivers, including splash dams used to ferry timber downstream prior to 1957, large water impoundments, and adjacency to agricultural croplands. We used machine-learning in program Maxent and three different river layers that varied in extent and location of mapped rivers but contained distinct information to produce species distribution models which we then combined into a single ensemble model. Stream order, annual precipitation, and precipitation frequency were the highest ranked baseline environmental variables in most models. Watersheds with highly suitable baseline conditions in our ensemble model were negatively correlated with anthropogenic disturbances to rivers. Foothill yellow-legged frogs appeared to be sensitive to human-caused disturbances to rivers, perhaps indicative of their narrow ecological tolerance to in-river conditions. We do not anticipate variables in our model to change much through time. Rather, for conservation we identified potential legacy (spash dams) and ongoing human-caused disturbances that are more likely to change conditions for the species in the short- and long-term.