Negative effects of forest harvest on coastal tailed frog (Ascaphus truei) emerge after a decade-long experiment

Abstract

Perennial, low-order headwater streams provide critical ecosystem functions, including habitat for specialized aquatic species. Anthropogenic disturbances can degrade biological and physical conditions in headwater streams, with potentially corresponding negative effects on stream biodiversity. Headwater streams comprise most of the stream length in forested watersheds in the Pacific Northwest, the United States, including those that are globally important sources of wood biomass. To conserve stream biota, water quality, and related public resources, riparian buffers are retained as a management tool expected to ameliorate effects of anthropogenic disturbances, including timber harvests. From 2006 to 2015, we used a replicated Before-After Control-Impact (BACI) experiment to evaluate how coastal tailed frog (Ascaphus truei) density varied in three alternative riparian buffer configurations (continuous buffer, patchy buffer, and clearcut riparian area) and unharvested reference basins, in western Washington, the United States. We used count data to estimate tailed frog density for three years pre-harvest, the first two years post-harvest, and years seven and eight post-harvest. We compared relative change in density for each buffer configuration between pre- and post-harvest after controlling for temporal changes in the reference. We found no evidence of a decline in tailed frog density in the first two years post-harvest but saw evidence of substantial declines in larval and post-metamorphic coastal tailed frog densities in years seven and eight post-harvest. Specifically, we estimated a 65%, 93%, and 84% decline in average larval tailed frog density in basins with a continuous buffer, patchy buffer, and clearcut riparian area, respectively. We estimated 71% and 97% declines in post-metamorphic tailed frog densities in the continuous and patchy buffers. Because evidence of tailed frog declines became apparent only after eight years post-harvest, our results underscore the need for longer term studies to address lag effects in population responses to human land use and to understand how effectiveness of conservation practices can vary spatially and temporally.