We recently used existing occupancy survey data to develop a simulation assessing the power of 30 sample design scenarios to detect nine rates of true population change in the presence of sampling noise (occupancy surveys) and process noise (variability in availability for detection). We found that no sample design had > 50% success at detecting 1% annual increases or decreases in true occupancy rate over a 10 year period. For stronger rates of annual change in across-site occupancy rate (e.g., 2-4% annual change), increasing numbers of sample sites were clearly associated with stronger power to detect true change, whereas increasing numbers of site visits were more weakly associated with stronger power to detect change. Segmented regression found that the increase in power for each sampling site added was lower above 44 sites than the power gained when adding sampling sites up to 44 sites. We found multiple equivalencies-of-power for different sampling designs, providing guidance for developing a monitoring program to maximize power to detect change while minimizing cost.