Prioritizing habitat conservation for the mottled duck (Anas fulvigula) along the western Gulf of Mexico coast

The Western Gulf of Mexico Coast provides important habitat for migrating and resident waterfowl. The mottled duck (Anas fulvigula) relies on this region for all its life-cycle events. Its steady population decline has earned it a "Red" status on the Audubon WatchList and a species of concern among state and federal agencies. The Western Gulf Coast (WGC) mottled duck population decline is caused by the destruction and degradation of large tracts of wetlands and native coastal prairies due to human activities in the western Gulf of Mexico Coast. There is general agreement among experts that nesting and brood-rearing habitat loss and degradation is the most important threat to WGC population and increasing recruitment is essential to the recovery of the population. Many gaps in knowledge of the species ecological requirements still exist. Decision Support Systems can be used as tools that aide stakeholders in decision making processes by consolidating available biological and ecological knowledge and taking into account temporal and spatial variation at the landscape level. The first goal of this dissertation was to use available biological knowledge of mottled duck nesting and brood-rearing requirements to develop a model that will aid managers in targeting areas for conservation and management. I used primarily ArcGIS software, but also other spatial software (e.g. ERDAS Imagine, FragSTATS), to develop four spatially explicit models that: 1) identify and prioritize currently suitable mottled duck nesting habitat for protection, 2) identify and prioritize currently suitable mottled duck brood-rearing habitat for protection, 3) identify and prioritize areas for grassland establishment, and 4) identify and prioritize wetland basins for freshwater enhancement. Results from the spatial models show that only 605 ha of currently suitable nesting habitat and 946 ha of currently suitable brood-rearing habitat fall in the highest priority (top 10%) for conservation range. Wetlands are the limiting resource in the Texas Mid Coast and the Laguna Madre initiative areas, while grassland habitat is the limiting resource in coastal Chenier Plain and Mississippi River Coastal Wetlands initiative areas. There is a high density of wetland basins of high priority for freshwater enhancement in coastal Louisiana and the upper Texas coast. There is a high density of areas of high priority for grassland establishment in inland areas in Texas and Louisiana. These results indicate the need for regional management and conservation goals, while also aiding stakeholders in targeting areas that will yield the largest biological return per conservation dollar. The decision support tool can serve as a valuable tool in the management decisionmaking process; however, managers should be aware that it is built on spatial datasets that are not inclusive of all required information and contain unknown amounts of inherent errors. Thus, a critical component was the assessment of performance and the validation of the model’s ability to predict currently suitable nesting and brood-rearing habitat. The second goal of this dissertation was to conduct several analyses to assess the performance of the Mottled Duck Decision Support Tool as well as provide recommendations to improve its performance in its ability to identify currently suitable nesting and brood-rearing habitat patches and landscapes. I used three separate measures to accomplish this goal. I 1) assessed the accuracy of patch suitability as defined by the model using randomly generated points surveyed from a helicopter, 2) analyzed the ability of the Mottled Duck Decision Support Tool to identify high priority habitat by investigating use of these landscapes by breeding mottled ducks from the cooperative Breeding Mottled Duck Survey during 2009 to 2014, and 3) used backwards elimination regression to determine which landscape variables explain the greatest variability in breeding mottled duck density. Testing model performance is vital to informing future Mottled Duck Decision Support Tool revisions and forming recommendations for improving the model. I found that the Mottled Duck Decision Support Tool adequately identified, with ≥79% accuracy, currently unsuitable mottled duck nesting or brood-rearing habitat, currently suitable nesting, and currently suitable brood-rearing habitats by using the currently available spatial datasets as input parameters. I also found that breeding mottled ducks were using landscapes with optimal spatial arrangement of nesting and brood-rearing habitat, which is reflected by higher mean priority rankings of nesting and brood-rearing habitat in the landscape. Finally, I found that edge density and edge evenness of brood-rearing habitat patches were landscape variables that were significant for explaining variation in breeding mottled duck densities in the coastal and in the inland zones of the Western Gulf Coast