Optimizing point count survey locations with viewshed analysis and integer programming: A geospatial approach to wildlife detectability

Point counts are a common method for measuring wildlife abundance, but their accuracy can be limited by detectability biases influenced by observer location and environmental conditions. We present a geospatial approach integrating viewshed analysis with spatial optimization methods to optimize survey point placement and improve detectability estimates. We quantified visibility based upon both terrain-only and LiDAR enhanced viewshed analysis within 34 wetlands in Tampa, Florida, with a diverse array of geometries and vegetation structures. Based on this analysis, we computed optimal locations for 1 to 5 observers at each wetland. We also conducted a field validation at one of the larger, more accessible wetlands. The results demonstrate that incorporating LiDAR substantially reduced overestimation of visible area by about half in comparison to a surface-only analysis, though the amount depended on wetland shape, amount of open water, and tree canopy coverage. Optimal survey locations yielded a mean increase of about 6 times more visible area than the worst locations and about 1.5 times the average. The field validation found the LiDAR-based visibility analysis 90 % accurate. This framework provides practical tools to enhance wildlife survey design by explicitly accounting for visibility and observer placement, potentially improving data quality and optimizing field effort in ecological monitoring programs. • Surface terrain and LiDAR are used to estimate visibility for wildlife surveys. • LiDAR accounts for vegetation and structures that can block views, improving detectability estimates. • Integer programming is used to select observer locations that maximize the proportion of the study area that is visible. • Strategic placement of observers can substantially increase survey coverage compared to random survey locations.