Evaluating the relationship between local food availability and wetland landscape structure in determining dabbling duck habitat use during spring migration

Wetlands in the Nebraska's Rainwater Basin (RWB) havedecreased by 90 percent over the past two centuries and are subject to on-going degradationof quality from urban and agricultural land-use practices. Losses in wetland habitatquantity and quality are important because the RWB serves as a critical spring staging areato [about]7 million dabbling ducks, including approximately 50 percent of North America'smid-continent mallard (Anas platyrhynchos) population, and 30 percent of North America'stotal Northern pintail (A. acuta) population. During spring, waterfowl depend on wetlandhabitat for aquatic invertebrates and plant materials to accumulate the energy and proteinneeded to complete migration and initiate egg production. If demands for quality foodresources are not met, waterfowl may arrive at breeding grounds in poorer body condition,and consequently be less likely to achieve reproductive success. This cross-seasonal effectis believed to be driven by excessive habitat loss at mid-latitudes, introduction ofinvasive plant species, and depletion of food resources by fall migrants. Given the importanceof food resource acquisition at mid-latitude stopover sites and subsequent effects onrecruitment, the goal of this study was to improve understanding of food resourceavailability in wetlands and the relationship to habitat use by spring-migrating waterfowl.I conducted weekly waterfowl surveys and quantified local habitat characteristics includingseed density (kg/ha), invertebrate density (kg/ha), energy derived from food resources(kcal/ha), water depth, wetland area, vegetative cover, and several water quality parametersat 32 wetlands in spring 2014 and 35 wetlands in spring 2015. Additionally, I quantifiedwetland habitat surrounding each study site by assessing wetland area and number of wetlands(greater than 1ha) within 2.5km and 5km of a study site. Study sites were located on publiclands managed by the Nebraska Game and Parks Commission and the U. S. Fish and WildlifeService, private conservation easement lands enrolled in the Wetlands Reserve Program (WRP),and on private lands managed for agriculture (cropped and non-cropped). A set of speciesdistribution models were developed to explain spring dabbling duck density and speciesrichness in the RWB. I hypothesized that a combination of local (food density, energy, waterdepth, wetland area, and vegetative cover) and landscape variables would explain thegreatest amount of variability in dabbling duck density. In 2014 (a dry year), energy, seeddensity, water depth, wetland area, and wetland density in the surrounding landscape werepositively associated with dabbling duck density; however, invertebrate density andvegetative cover had no influence on dabbling duck density. In 2015 (wet year), seed densityand energy were positively associated with dabbling duck density; however, water depth,wetland area, vegetative cover, invertebrate density, and wetland area in the surroundinglandscape had no influence on dabbling duck density. Wetland area and water depth were theonly useful explanatory variables for explaining species richness in 2014, whereas in 2015dabbling duck species richness was best explained by wetland area and vegetative cover. Iused non-parametric analyses to compare seed density, and true metabolizable energy (TME) atthree wetland types; public, WRP, and cropped wetlands. Seed density did not vary amongwetland types in 2014 or 2015. Median seed density estimates during both years at public,WRP, and cropped wetlands were 593kg/ha (x = 621kg/ha), 561kg/ha (x = 566kg/ha), and 419kg/ha (x = 608kg/ha) respectively. Seed density was consistent between years for public and WRP wetlands, but varied between years for cropped units (p less than 0.05). Variation in seed density between years at cropped wetlands was likely influenced by the presence/absence of agricultural waste grains. Cumulative TME varied among wetland type in 2014 and 2015, with greater TME at cropped wetlands (median = 2431kcal/kg) than public (median = 1740kcal/kg) and WRP wetlands (median = 1781kcal/kg), however TME did not differ between WRP and public wetlands. TME was consistent among wetland types between 2014 and 2015. Seed density estimates from this study were statistically greater than estimates currently used for management planning in the RWB, however, TME estimates were statistically less than estimates currently assumed for WRP and public wetlands in the region. My estimates for mean aquatic invertebrate density were approximately 40-fold less than estimates for mean seed density. Benthic communities accounted for 68 percent of the total invertebrate density, however invertebrate diversity was greater in nektonic communities. Neonicotinoid synthetic insecticides are believed to have a deleterious effect on aquatic invertebrate communities in agricultural areas, although their occurrence in RWB wetlands were previously unknown. I detected trace levels of neonicotinoids in 92 percent of water samples collected in wetlands sampled in the RWB during the spring of 2015. I predicted a relatively high detection rate given the intensity of row crop production in the region, though concentrations were lower than expected. Concentrations at 26 wetlands sampled fell below toxicity benchmarks proposed by the Canadian Environmental Quality Guidelines, and only 11 percent of wetlands sampled had concentrations exceeding the most conservative benchmark proposed by the Environmental Protection Agency. Neonicotinoids concentrations were minimal at wetlands with vegetative buffers strips greater than or equal to 50m between a wetland and a cropped field, relative to wetlands with vegetative buffers strips less than 50m. Although neonicotinoid levels were below lethal concentrations for all aquatic invertebrates identified in this study, I observed a negative association between neonicotinoid concentrations and aquatic invertebrate density (g/m2).