摘要:Monitoring seabird-derived nutrients on islands following invasive mammal eradications may provide a useful, cost- and time-efficient indication of the recovery of ecosystem function; however, the technique has only been investigated on environmentally similar islands. How seabird-derived nutrients recover on islands with different characteristics, and how differences in sampling regimes affect results is poorly understood. To determine how different island characteristics (size, geographic location and invasion history) and aspects of the sampling regime (sample collection year, season and intra-island location) influence seabird-derived nutrients we collated nitrogen stable isotope (δ15N) data from three ecosystem components (soil, plants and spiders), collected on 28 islands around New Zealand. We investigated which variables best predict δ15N using linear-mixed effects models. Accounting for these variables and using still-invaded and never-invaded islands as controls for recovery, we then investigated changes in δ15N on islands at different stages following invasive mammal eradication. Island size, invasion history and the presence of seabirds in the direct vicinity of a sampling location all influenced δ15N. After accounting for these variables, δ15N increased with time since eradication in soils, plants and spiders, though there was still some variation that our chosen variables could not explain. This study demonstrates the importance of considering island characteristics and sampling methods when assessing seabird-derived nutrient recovery and highlights the need for additional targeted sample collection on islands to help separate out the effects of time since eradication and other confounding variables affecting δ15N. Improved understanding of these factors will be prerequisite for furthering this technique as a useful addition to the post-eradication monitoring tool kit.