Evaluation of Visibility Correction Factors for Waterfowl Surveys in Wyoming

The waterfowl breeding population survey for Wyoming was redesigned and evaluated in 1993 and 1994. Ten regional strata were systematically sampled with aerial strip transects 400 m wide. Double sampling was employed to correct aerial counts for visibility bias, estimated from a small number of transects. Nine transects in 1993 and 13 in 1994 were surveyed both from the air and from the ground (truthing transects). Visibility correction factors (VCFs) were calculated as ratios of ground counts to air counts. Precision of VCFs was inadequate (CVs > 50%) in 1993. To increase precision of VCFs in 1994, truthing transects were subjectively placed in areas with more potential waterfowl habitat, and their number and average length were increased. These modifications improved the precision of VCFs (CV = 27.2% for mallard, Anas platyrhynchos), but waterfowl distributions and counts appear to be more variable in Wyoming than in the prairie potholes of northcentral North America. Violating assumptions of this double sampling scheme affects accuracy and precision of VCFs. Important assumptions are (1) air and ground crews survey the same waterfowl populations, and (2) truthing transects are representative of the other transects. Violations of the first assumption include deviations from intended flight paths and variability of flight altitudes on truthing transects. Flight path error was more than 30% in both years. Average altitudes on truthing transects had a range of 19 m in 1993 and 12 m in 1994. In 1994, the second assumption was violated because truthing transects were placed subjectively. Average flight altitudes were lower (P = 0.009) and density of all waterfowl greater (P = 0.002) on truthing transects than on the other transects ; and species compositions were not closely correlated (r 2 = 0.765) between truthing and other transects. Due to the variability of waterfowl distributions and the above-mentioned procedural problems, adequately precise and accurate VCFs might not be achievable or cost-effective in Wyoming with traditional aerial survey techniques.