Storks (Ciconiidae) are renowned for defecating on their legs when exposed to high temperatures, a phenomenon known as ‘urohidrosis’. Previous work suggested that this behaviour can cool down the blood supply to the legs and thus prevent hyperthermia in captive storks when overheated. However, no study has quantified the magnitude or duration of its evaporative cooling effect in free-ranging birds. Here, we combine urohidrosis data with thermal imaging and microclimate data to investigate the thermoregulatory significance of urohidrosis in White storks Ciconia ciconia during the breeding season. We show that urohidrosis can reduce leg surface temperature by up to 6.7 ºC (4.40 ± 1.04 ºC). Yet its cooling effect was of short duration (lasting no more than 2.5 min) and decreased with time since defecation. Thus, for urohidrosis to significantly contribute to heat dissipation, storks must perform it repeatedly when overheated. Indeed, individuals can perform up to 11 urohidrosis events per hour; our estimates indicate that repeated urohidrosis could represent a significant amount of heat loss during short-time spans — equivalent to 4% of daily field metabolic rate for an adult stork. Our results points to an absence of differences in the cooling efficiency of urohidrosis between adults and nestlings, probably because all nestlings were recorded during the last phase of the ontogeny of thermoregulation. Besides, during the hottest days adult storks delivered water to their nestlings, which might allow them to replenish body water reserves to sustain urohidrosis and other heat dissipation behaviours such as panting or gular fluttering. Our results indicate that urohidrosis might buffer the impact of high temperatures in breeding storks, preventing overheating and lethal hyperthermia. Gaining knowledge about behavioural thermoregulation in the heat is therefore crucial to better predict the future persistence and vulnerability of species under different climate warming scenarios.

During the spring of 2021 (from late March to mid-June), we conducted a study on the thermoregulation and heat exchange of White storks Ciconia ciconia at a nesting colony in Extremadura, southwestern Spain (38º52’ N; 6º19’W). This colony hosted 12 active nests that were visited twice a week to obtain thermal images to study the potential role of different body parts (particularly beaks and legs) as thermal windows along a range of different environmental temperatures. We collected thermal images of each nest throughout the study period, from early incubation to the end of the breeding season.

Thermal imaging sessions were performed from 1 h after sunrise until maximum ambient daily temperature was reached (which was confirmed by tracking temperature with a portable weather station; see below), measuring no more than 2 nests per day (switching the camera between focal nests during thermal imaging sessions). We employed a handheld thermal infrared camera (FLIR E95, resolution = 464 x 348 pixels, FLIR Systems, USA), equipped with a 14º lens. We automatically took a thermal image every 20 seconds and simultaneously recorded local weather data – environmental temperature, Ta (Ta_Kestrel, ºC); wind speed, (m s-1); and relative humidity (%) – every 30 s using a portable weather station (Kestrel 5400, Kestrel Instruments, USA) mounted on a vane and placed on the colony (at 1.5 m height).

Although thisurohidrosis behaviour appeared to be frequent in White storks during late spring, we only could obtain thermal images of 22 urohidrosis events (four in adults and 18 in nestlings), from five different nests and seven individuals (three adults and four nestlings) during the 31th of May, and the 7th, 9th and 10th of June. A total of 173 thermal images of individuals showing urohidrosis were taken, but we discarded those images in which legs were not completely visible or those in which storks’ position could compromise accuracy of thermal measures, as incidence angle influences apparent emissivity – with increased angle resulting on declines in apparent emissivity (Playà-Montmany & Tattersall 2021). Then, we selected for the analyses those pictures in which a lateral (or almost lateral) view of the legs could be taken. Finally, 81 images from these 22 urohidrosis events met this criterion and were analyzed using the program FLIR ResearchIR (FLIR Systems, USA). For each image, we assumed an emissivity of 0.95 and set ambient temperature and relative humidity from values obtained from the weather station. Average surface temperature of the leg area covered by excreta was measured using the ‘freehand roi’ function, while average surface temperature from the adjacent region free of excreta was measured by using the ‘line roi’ function. We employed the ‘freehand roi’ function to measure urohidrosis due to the irregular shape of excreta marks along the leg. Overall, temperature measures were calculated on areas that had, on average, 20.73 ± 7.97 pixels. We calculated urohidrosis gradient as the difference between the surface temperature of the bare skin leg region and that of the region covered by excreta. When possible, we repeatedly calculated urohidrosis gradient for the same event to estimate its duration and magnitude over time.