Detecting Silent Stressors: Trace Element Effects On Nutritional Status Of Declining Scoter Ducks Of Puget Sound, Usa

White-winged scoters (Melanitta fusca: WWSC) and surf scoters (M. perspicillata; SUSC) have declined by over 60% in recent decades. Identifying contributing factors from within a mosaic of sublethal, multiple stressors is challenging. In urbanized Puget Sound, Washington, USA where scoters winter, changes in prey availability explained only a portion of local declines, suggesting that other "silent stressors" such as sublethal contaminants might play a role. Past studies of pollutant effects on scoters used Fisherian statistics that often revealed few correlates; however, novel statistical approaches could detect and provide more insights about sublethal impacts. Our objectives were to (1) relate pollutant accumulation to health of the birds, and (2) compare permutational multivariate statistics with traditional approaches in identifying sublethal health effects. We collected scoters from three locations in Puget Sound in December 2005 and March 2006, and measured cadmium (Cd), mercury (Hg), and selenium (Se) levels in livers and kidneys. To assess impacts of low contaminants levels in tissues on nutritional status (whole-body mass, lipid, and protein; and triglycerides, beta-hydroxybutyrate, and uric acid in blood), we compared statistical methods. Permutational multivariate methods use Monte Carlo techniques to assess how an integrated matrix of physiological responses in each animal respond to contaminants. Univariate regressions revealed very few and inconsistent relationships. In contrast, multivariate models showed that liver Hg and Se explained 25% of the variance in nutritional status of white-winged scoters; and in surf scoters, Cd, Hg, and Se in tissues explained 14 to 27% of nutritional status depending on site. The influence of these factors equals other aspects of habitat such as foraging conditions. Our study indicates that permutational multivariate statistics can be a powerful tool for identifying sublethal contaminant associations that, with non-contaminant stressors, can influence nutritional status and thus, contribute to population dynamics. (C) 2020 Elsevier B.V. All rights reserved.