Monitoring burrowing petrel populations: A sampling scheme for the management of an island keystone species

Abstract

Burrow-nesting petrels (order Procellariiformes) are keystone species in island ecosystems, where they modify habitat through guano deposition and burrow digging. Burrowing petrels are among the most threatened groups of birds, yet robust long-term monitoring data remain scarce because of the financial and logistical constraints of working on offshore breeding islands, the variety of surveying strategies used, and the birds' below-ground breeding behavior. We examined the sampling requirements of monitoring programs to detect changes in the number of breeding pairs of gray-faced petrels (Pterodroma gouldi), a common species in northern New Zealand. We first examined the relationship between burrow entrance density and breeding pair density using 4 years of data from 3 large colonies. We then conducted a simulation-based power analysis to assess the ability of different burrow-occupancy sampling regimes to detect changes in breeding bird abundance. Power to detect change was influenced by population growth rates, initial bird density, interannual variation in abundance, plot size, number of plots, intervals between surveys, time of year surveys are undertaken, and duration of the monitoring program. Our analyses suggest that, under the most suboptimal monitoring conditions, at least 45 randomly assigned 5-m-radius plots surveyed annually during the incubation period for !20 years will be required to detect a 1% annual change in breeding bird abundance. Because power will vary depending on project specifications, local conditions, and potential change, we created an online application with over 50,000 combinations of starting parameters (https://landcare.shinyapps.io/ petrels). This allows managers to determine the power of different combinations of survey intensities while maintaining consistency and maximizing efficiency. Ó 2015 The Wildlife Society.

Uncertainty in population estimates: a meta-analysis for petrels

“…In a more realistic scenario of population stochasticity and uncertain variance, even steeper declines would need to occur before a change could be reliably detected. While our simulation was based upon a comparison between estimates at two time points, and it is well known that power could be enhanced by more frequent sampling (Buxton et al, 2016), this is representative of the reality on many islands where petrel surveys may be repeated decades apart using different methodologies.…”

“…Estimates of petrel populations are commonly made to inform trend and threat assessments, but they currently perform rather poorly in this role. Therefore it might be appropriate to focus research on parameters better able to inform management such as productivity and survival (Caravaggi et al, 2019;Dilley et al, 2018) and population trends derived from repeat sampling (Buxton et al, 2016).…”

Detecting seabird responses to invasive species eradication

“…Our study provides further evidence that whole-island surveys, regardless of method, yield wide confidence limits when used to provide population estimates, which in turn may preclude their use for reliably detecting trends. Instead, if the primary goal is to measure population change, approaches that focus on monitoring more frequently within fixed plots are likely to be more accurate (Buxton et al 2016).…”

Productivity changes in the Mediterranean Sea drive foraging movements of yelkouan shearwater Puffinus yelkouan from the core of its global breeding range

“…Known breeding sites are mainly distributed in the central Mediterranean basin, from Menorca island and the Southern French coasts (Bourgeois & Vidal, 2008; Derhé, 2012) to the Sicilian Channel and the Aegean Sea, with a global population size recently re‐assessed at 21,000–36,000 pairs (Gaudard, 2018). However, as for most other burrowing petrels breeding at hardly accessible locations such as cliffs and caves, reliable long‐term trends and population estimates are scarce (Buxton et al., 2016). Hence, most population estimates have been achieved by imprecise methods such as counting birds while rafting on the water surface in the proximity of colonies (Bourgeois & Vidal, 2008; Raine et al., 2010).…”