Curves for Curlew: Identifying Curlew breeding status from GPS tracking data

Releasing a tagged Curlew. Rachel Taylor

Author(s): Bowgen, K.M., Dodd, S.G., Lindley, P., Burton, N.H.K. & Taylor, R.C.

Published: December 2022  

Journal: Ecology and Evolution Volume: 12

Article No.: e9509

Digital Identifier No. (DOI): 10.1002/ece3.9509

View journal article

The Curlew is of significant conservation concern in the UK, and poor breeding success is thought to be one driver of this species' decline. However, breeding behaviour can be particularly difficult to observe first-hand. Can GPS tracking data help address these knowledge gaps?

The Curlew is of significant conservation concern in the UK, but many questions still remain about their breeding behaviour. This is partially due to the species’ cryptic nature and sensitivity to human disturbance, which can make the birds difficult to observe in the field. Transitions between breeding stages, for example when incubation stops after hatching, or when a nest fails, can be difficult to observe first-hand. In this study, BTO scientists used data from GPS tracked Curlew in north Wales to uncover valuable information on parental breeding status and nest survival rates. 

GPS tagging offers an alternative to visual observation. Lightweight tags attached to the birds’ backs collect information on their location at predetermined time intervals. BTO scientists set out to use this technology and specialised statistical software to analyse the birds’ movement patterns and discover the locations of their nests. During this analysis, the researchers found that they could also detect transitions between breeding stages and identify cases of nest failure. This allowed them to calculate nest survival rates, which could provide clues as to why Curlew populations in north Wales are struggling. 

During the breeding seasons of 2016, 2018 and 2019, the research team attached GPS tags to adult Curlews at three locations in north Wales. These tags determined each bird’s location every 15 minutes and transmitted it to a stationary receiver, whenever the bird travelled within its range. Ultimately, 23 birds provided enough data to be used in the analysis. 

The scientists needed to ‘train’ the statistical programme to look for specific characteristics within the Curlews’ movement data. They did this using a dataset collected from a bird whose breeding status was confirmed through visual observation. This gave the researchers a baseline against which they could search for similar behavioural patterns in other birds’ GPS data. During incubation, this focal bird returned repeatedly to the same place (its nest) and remained there for long periods of time. After its young had hatched, it was recorded using more sites multiple times close together, suggesting it was accompanied by chicks which could not travel far. Crucially, breeding failure was characterised by the focal bird (mostly) deserting the breeding site and developing a willingness to travel much further. 

When the scientists applied these principles to the data collected from the other Curlew, they detected 25 breeding attempts, of which 17 were thought to be failures. A range of other behavioural patterns was found including transitions between stages and the identification of non-breeding birds (26% of those tagged). Overall nest survival rate was very low (nests had 15.1% chance of survival) suggesting that the Curlew studied had poor levels of breeding success. The scientists found that the computer programme needed a minimum of three consecutive days of data to detect a target event (e.g. a nest failure or the arrival of chicks) that occurred during that period. They were often able to estimate the timing of an event within a few hours. If the duration of the GPS data collection intervals was increased, for example to every 30 or 60 minutes, longer periods of consecutive data were needed (six and nine days respectively) to maintain this accuracy. Although this revealed potential difficulties in correctly assigning breeding status using data from other studies, the researchers aim to attempt it in future.  

This study demonstrates that GPS data can be used to detect changes in breeding stage in adult Curlew. It is believed to be the first time such a technique has been used to investigate behaviour of Curlew throughout the entirety of the breeding season (pre-breeding through to post-breeding). This means that the methods and results could be extremely valuable to other scientists asking similar questions. This study shows that data on breeding behaviour can be collected whilst limiting disturbance to the birds and relieving pressure on labour-intensive conservation initiatives. 

Abstract

Identifying the breeding status of cryptic bird species has proved problematic without intense or inherently expensive monitoring. Most, if not all, intensive bird monitoring comes with disturbance risks and many projects now rely on tagging of individuals to provide remote information on movements. Given the importance of breeding status when targeting conservation interventions novel methods are needed. This study aimed to identify breeding status in Eurasian Curlew (Numenius arquata) from GPS tag movement patterns using the “recurse” package in R. This package identifies foci of activity (using K-means clustering) based on revisitations. Using a training data set from an individual of known breeding status, we visually assessed the frequency of revisits to particular locations to identify prebreeding, incubation, chick guarding, and post-breeding stages to an accuracy of a within at most half a day and thus breeding outcomes. Limited validation was provided by additional field observations. Based on our results, we estimate a low daily nest survival rate of 0.935 during incubation, that only a small proportion of individuals successfully raised young, and that there was a high proportion (26%) of non-breeders in the population. The Eurasian Curlew is a species of high conservation concern across Europe, and our results concur with recent studies highlighting that population declines are likely to be driven by low levels of productivity. The acquisition of improved knowledge on the behaviors of individuals at each stage of breeding enables more targeted conservation efforts and reduces the need for additional monitoring visits that may cause increased disturbance and risk of nest failure. We hope that the approach outlined may be developed to provide practitioners who have detailed knowledge of the behavior of their study species with a practical means of assessing the breeding status and outcomes of their study populations.

Notes

This work was funded as part of a larger project funded by Natural Resources Wales, with the initial tagging in 2016 was carried out in collaboration with the RSPB (purchase of initial tags). The tracking of Curlew and production of this paper were additionally supported by major donors and the generous contributions to the BTO's Curlew Appeal.

Staff Author(s)


Related content