Spatio-temporal changes in density and distribution of burrow-nesting seabird colonies after rat eradication

Abstract

Abstract:The size and distribution of colonies of burrow-nesting petrels is thought to be limited partly by the availability of suitable breeding habitat and partly by predation. Historically, the availability of safe nesting habitat was restricted in New Zealand, due to the introduction of rats by humans. More recently, however, habitat has been restored by rat eradication. Petrel colony growth is mediated by both positive and negative density dependence, although it is unclear if, or how, density dependence will affect patterns in post-eradication colony recovery. Here, using burrow density as a proxy for relative abundance, we tested whether petrel colonies increase in density or area after rat eradication by sampling along a chronosequence of (1) five islands from which rats were eliminated 1 to 26 years ago, (2) two islands that never had rats, and (3) an island with rats still present, while controlling for habitat availability. We also measured a time series of burrow densities in plots on each island to compare temporal changes after rat eradication. Using Bayesian hierarchical modelling, after controlling for nesting habitat, we found that mean burrow density increased with time since rat eradication. Burrows remained clustered (i.e. spatially structured), but became more randomly distributed on islands with more time since eradication. Point density mapping indicated that colony extent increased with time since rat eradication, with colonies filling over 70% of surveyed areas on islands by 25 years after eradication. Increases in burrow density and colony area, but maintenance of clustered distribution, suggest both positive and negative density dependence may operate during colony expansion. Understanding patterns in petrel colony recovery is important, not only due to the indispensable role of petrels as island ecosystem engineers, reflecting the recovery of ecosystem functioning, but also to help guide post-eradication monitoring strategies.

Mercury Islands and their role in understanding seabird island restoration

“…These findings were particularly relevant to Korapuki Island, which at the longest time since eradication of mammals (then 22 years) had the highest soil and plant δ 15 N and lowest C:N of the islands sampled (Jones 2010). In contrast to the scattered low density colonies recorded 10 years before mammal eradication (Hicks et al 1975), seabird colonies now extend over 70% of surveyed areas on the island (Buxton et al 2016).…”

“…Given that the seabird-driven ecosystem on Korapuki appears to be recovering naturally (Jones 2010;Buxton et al 2016), is restoration based on pre-determined assemblage composition justified? For example, the ecosystem on Korapuki Figure 5.…”

“…Four sources of information can help us predict the response of native species and ecosystems to the removal of invasive species from islands: 1) studies of the responses of plants and animals after eradication; 2) the use of reference sites that lack invasive species as defined above; 3) using chronosequences of islands from which invasive species have been removed (e.g. Buxton et al 2016); and 4) paleoecological studies of plant and animal remains in middens, caves, sand-dunes and wetlands (e.g. Towns & Ballantine 1993;Wilmshurst et al 2014).…”

Prioritizing eradication actions on islands: it's not all or nothing

“…These insights might come from detailed ecological studies on species recovery (Ringler, Russell & Le Corre ; Buxton et al . ), predictive modelling techniques (Raymond et al . ), or expert elicitation (Sutherland & Burgman ).…”

“…This requires insight into not only the impact of each invasive species on each species of concern, but also how the absence of an invasive species might affect other invasive species populations. These insights might come from detailed ecological studies on species recovery (Ringler, Russell & Le Corre 2015;Buxton et al 2016), predictive modelling techniques (Raymond et al 2011), or expert elicitation (Sutherland & Burgman 2015).…”