Midwest Fish & Wildlife Conference 2016

AUTHORS: Travis J Schepker*, University of Missouri; Elisabeth Webb, Missouri Cooperative Fish and Wildlife Research Unit; Ted LaGrange, Nebraska Game and Parks Commission

ABSTRACT: Wetland biologists rely on species distribution models (SDM) to generate an ecological explanation and prediction of how wetland-dependent species distribute and interact within a system. Typically, SDM are developed by identifying relationships between species distributions (response variable), and varying attributes within the physical environment they inhabit. Waterfowl are among the more frequently studied wetland dependent groups, however even waterfowl SDMs are far from complete. Previous studies have identified forage, vegetative dispersion, depth, and wetland area as factors influencing waterfowl distribution, however these variables are generally only applicable for local scale (within wetland) assessments. Waterfowl are highly mobile and capable of exploiting wetlands in the surrounding landscape to acquire food resources, form pair bonds, and avoid predation. Therefore it is important that SDMs incorporate relevant variables at multiple scales to accurately predict how waterfowl distribute themselves across an ecological complex. We conducted weekly avian surveys at 20-27 playas in Nebraska’s Rainwater Basin, to determine waterfowl density, and species richness during springs of 2014 and 2015. At the wetland scale, we assessed spring food resource phenology (seed and invertebrate biomass), vegetative cover, depth, and wetland area. At the landscape scale we used recurring aerial imagery to quantify change in total wetland area within a 4.6 km radius of our individual study sites. Local and landscape attributes were designated as independent variables, and waterfowl density and species richness were designated as dependent variables. From our independent variables we developed a priori candidate models and used multiple generalized linear mixed models and Akaike information criterion to evaluate models based on their ability to explain variation in waterfowl density and species richness. Preliminary analysis from 2014 data indicated invertebrate biomass as the only covariate in the top model predicting waterfowl density, however, wetland area within 4.6 km of a study site were covariates in competing models.