Monitoring Hawaiian waterbirds: evaluation of sampling methods to produce reliable estimates

We conducted field trials to assess several different methods of estimating the abundance of four endangered Hawaiian waterbirds: the Hawaiian duck (Anas wyvilliana), Hawaiian coot (Fulica alai), Hawaiian common moorhen (Gallinula chloropus sandvicensis) and Hawaiian stilt (Himantopus mexicanus knudseni). At two sites on Oʽahu, James Campbell National Wildlife Refuge and Hamakua Marsh, we conducted field trials where both solitary and paired observers counted birds and recorded the distance to observed birds. We then compared the results of estimates using the existing simple count, distance estimates from both point- and line-transect surveys, paired observer count estimates, bounded count, and Overton estimators. Comparing covariate recorded values among simultaneous observations revealed inconsistency between observers. We showed that the variation among simple counts means the current direct count survey, even if interpreted as a proportional index of abundance, incorporates many sources of uncertainty that are not taken into account. Analysis revealed violation of model assumptions that allowed us to discount distance-based estimates as a viable estimation technique. Among the remaining methods, point counts by paired observers produced the most precise estimates while meeting model assumptions. We present an example sampling protocol using paired observer counts. Finally, we suggest further research that will improve abundance estimates of Hawaiian waterbirds.