Relative abundance and activity patterns explain method-related differences in mammalian species richness estimates

Species richness and species-specific occupancy are key variables for biodiversity monitoring. While there are many field techniques to assess presence of mammalian species, their relative performance and cost effectiveness can be affected by biological variables that are related to availability (relative abundance, diurnal activity) or by traits that are related to visibility (body size, coloration, habitat selection) of a species. We conducted four transect surveys, driven during daytime hours and continuous monitoring with camera traps over a 16-week period in Lake Manyara National Park, Tanzania. Transect and camera-trap surveys yielded estimates of 20 and 38 mammalian species, respectively. In absolute terms, diurnal vehicle transects were less costly than camera trapping, but in relative terms (cost/species), camera trapping was more cost effective. Logistic regression showed that mammal species with reduced availability (low relative abundance and low proportion of diurnal activity) were less likely to be detected during diurnal line-transect surveys. Species-specific detection probabilities and occupancy estimates were lower during transect surveys compared to the camera-trap survey. Similar to differences in naive species occupancy, method-related differences in modeled detection probabilities were positively correlated with relative abundance. Results suggest that factors influencing availability of a species explain method-related differences in detection probabilities and species richness estimates.