English 
Français 
Español 
2023-02 Effectiveness of acoustic monitoring for estimating population trends and recolonisation of burrow-nesting petrels
Project Leader: Richard Phillips, British Antarctic Survey
Co-investigators: Rachel Buxton, Carleton University, Institute of Environmental and Interdisciplinary Sciences, Ottawa Ontario Canada
Amount awarded: AUD 36,600
Knowledge of population trends of threatened species are essential to inform adaptive management, yet are lacking for many petrels because nesting sites are remote, most species are nocturnal at colonies, and they nest in burrows or crevices. As such, detection error is a major issue, and uncertainty in estimates of population numbers and trends are often very high, making it difficult or impossible to fully determine impacts of threats or the outcomes of management actions (Bird et al. 2021). A typical index of abundance in studies of petrels is the count of burrows; however, burrow occupancy can be low and variable across space and time (Sutherland and Dann 2012). Determining burrow occupancy (i.e., whether a bird is present or breeding) through grubbing, scoping, or playback, is labor intensive and involves its own set of assumptions.
Advances in passive acoustic recording technologies offers large-scale sampling of ecological conditions, animal distribution, and abundance (Gasc et al. 2015). Acoustic monitoring is low cost and scalable, causes minimal disturbance, eliminates observer bias and generates a permanent data record. However, the effectiveness of acoustic monitoring for estimating breeding density and burrow occupancy is unclear, and our study is designed to address this knowledge gap.
The White-chinned Petrel (Procellaria aequinoctialis) is considered to be decreasing globally and is listed as Vulnerable by IUCN. It is the most frequently killed species in Southern Ocean fisheries, and suffers heavy predation from invasive alien predators at several island groups (Phillips et al. 2016). The purpose of this project is to test the viability of acoustic recording for monitoring breeding populations of White-chinned Petrels, including recolonisation following habitat restoration.
To quantify vocal activity we will use predictive random forest models from acoustic indices to summarise variation in acoustic energy (Buxton et al. 2018). We will explore relationships between vocal activity, nesting density, and burrow occupancy across the breeding season and sites. We will also use the more frequent data from the intensive study plot to test relationships between daily vocal activity and burrow occupancy. Finally, we will take account of the uncertainty (predictive power) of the relationships between vocal activity, breeding density and burrow occupancy in a simulation-based power analysis (e.g. Buxton et al. 2016a) to assess the utility of deploying a network of acoustic loggers for monitoring changes in breeding bird abundance and distribution over large spatial scales. Thus, we will test the power to detect changes in White-chinned Petrel populations following habitat restoration, depending on initial bird density, annual variation in abundance, number of recorders, time of year recordings are collected, duration of recordings and spatial scale.
Our overall goal is to create a user-friendly and cost-effective tool to monitor population trends and conservation outcomes for White-chinned Petrel at any site, which we expect to be readily applicable to other Procellaria species and potentially other burrowing petrels.