The Australian zebra finch is an arid-adapted passerine with high hygric demands, that is projected to be at risk from increasing temperature and aridity throughout its distribution by the end of the century. We examine here how individual zebra finches modify their drinking behaviour in response to the climatic conditions of ambient temperature, relative humidity, rainfall, wind speed, and day length over a two-month period during an Austral summer drought in arid central Australia. Visits to water by individual zebra finches increased with increasing ambient temperature and day length and decreased with increasing relative humidity, wind speed, and rainfall. These findings are the most comprehensive data for the drinking behaviour of individual wild birds globally and highlight the importance of regular drinking by individual finches to maintain water balance and thermoregulation. Our data highlight the importance of water availability for birds in the arid zone in a warming climate.

Over 500 wild zebra finches were implanted with PIT-tags (Minichip; Micro Products Australia) in the Gap Hills area of Fowlers Gap Arid Zone Research Station, far-west New South Wales (31°05′13.1″S 141°42′17.4″E) during the 50 months prior to December 2017 (Brandl et al. 2019; Funghi et al. 2019). From late November 2017 until after the conclusion of this study the only free water available within 4km of the centre of the study site was in two watering troughs located ~1km apart. The troughs were protected from large animals by a mesh fence and access to the water for small birds was restricted to an 11 cm diameter opening surrounded by the antenna of a custom-made battery-powered PITtag reader (RFIDRW-E-232; Priority 1 Design, Melbourne, Australia), that recorded the individual’s unique identification code, date and time. This study focuses on a total of 219 individually tagged adult zebra finches that were detected using the PITtag readers as they visited the water troughs between 16^th December 2017 and 18^th February 2018.  We obtained daily climatic data from the Australian Bureau of Meteorology for the Fowlers Gap automated weather station (AWS 046128 located 16 km from the study site) and considered mean daily wind speed, mean daily relative humidity, daily rainfall total (only a trace amount) and the daily maximum ambient temperature for the study period.

We collected data for at least one trough on 61 of the 65 days over which the study was conducted (not every trough recorded birds every day due to logistical reasons such as reader function, battery life, and memory card capacity). We used the interval between the birds’ recorded first and last visits to estimate the period of each day monitored, because visiting birds were not necessarily recorded for the entire day of these 61 days (e.g. if the readers’ batteries became flat). Recording hours (a similar concept to trap nights or trap hours to quantify effort for trapping studies; Thompson and Thompson 2007; Saska et al. 2021) at both water points varied from 6 h 15 min to a maximum of 26 h 55 min, with a mean daily duration of 19 h 55 min ± 5 h 38 min (i.e. the total summed duration of the drinking times across both troughs, e.g. 26 hours and 55 minutes is the sum of over 13 hours of monitoring each day at each of the two troughs).

A reading was considered an independent event when it occurred at least 15 minutes after a previous record for the same unique ID, avoiding double-counting birds as they entered and exited the trough and reducing the likelihood of multiple counts of individuals that repeatedly came and left the trough in a short period without drinking (Griffith, Cooper, pers. obs.). When a bird was recorded during a 15-minute interval, it was assigned a value of 1 for that period, even if multiple records were detected, and was assigned a value of 0 for the period if it was not detected within the 15-minute window.