Thursday, August 9, 2007 - 4:40 PMOOS 44-10: Forecasting avian responses to Elwha River dam removals
John F. McLaughlin and Carly J. Gelarden. Western Washington University
Few dam removals have received comprehensive ecological study, and fewer still have considered effects on associated terrestrial components. We present four of six stages in a program to determine avian responses to pending Elwha dam removals: (1) pre-removal inventory, (2) modeling species densities and distributions, (3) forecasting post-removal habitat distributions and conditions, (4) forecasting species responses to changing post-removal conditions, (5) monitoring post-removal responses, and (6) evaluating consistencies and discrepancies between actual and forecasted responses. We completed a pre-removal avian inventory during the 2006 breeding season using the variable circular plot protocol implemented in several Pacific Northwest parks. We sampled 269 plots, and detected 1830 birds of 42 species. For 22 species, we determined point-specific and habitat-mean density estimates, after adjusting for species-specific differences in detectability. Total avian densities were lowest (8.4 birds/ha) in low elevation shrubs and greatest (10.0 birds/ha) in western hemlock forests. Total avian densities in all habitats sampled exceeded values obtained in the 2002-2003 avian inventory of Olympic National Park, suggesting that the Elwha riparian zone supports regionally high population densities. We developed three scenarios to forecast avian responses to post-removal habitat changes. These scenarios apply models containing constraints at local, patch, and landscape scales. The first assumes each species is limited by habitat only: restored habitats will be colonized rapidly at densities similar to existing habitats of the same type. The second adds effects of habitat area on species occupancy and density. The third includes landscape context by assuming colonization will be constrained by dispersal from existing habitats. This series of forecasts should bracket actual responses, which will be monitored as restoration proceeds.