Detecting seabird responses to invasive species eradication

Maximising survey efficiency can help reduce the trade-off between spending limited conservation resources on evaluating performance of past interventions and directing those resources towards future interventions. Seabird responses to island eradications are often poorly evaluated owing to financial, logistical and methodological challenges associated with remote field work and species ecology. We surveyed an assemblage of threatened seabirds following the worlds largest island eradication of multiple invasive species, testing multiple survey designs and outputs. We compared the outcomes of two important choices made during survey design: 1) whether to use unbiased or targeted surveys; and 2) implementing design-based or model-based analyses. An unbiased whole-island stratified randomised survey design performed well in terms of confidence in the final population estimates for widespread species, but poorly for localised recolonising species. For widespread species, model-based analyses resulted in slightly lower population estimates with narrower confidence intervals than traditional design-based approaches but failed to capture the realised niches of recolonising species, resulting in population estimates three orders of magnitude higher than current best estimates. We conclude that a multi-method approach to survey design best captures the size and distribution of recovering populations when the study system is ecologically diverse--importantly our results suggest there is no single strategy for efficient surveys of diverse seabird communities following large island invasive species eradications.