Determinants of the acoustic behaviour of red deer during breeding in a wild alpine population, and implications for species survey

A necessary step in the application of animal communication studies to population ecology involves quantitative descriptions of acoustic variation within individuals. Stereotypy is in fact fundamental to allow individual recognition based on call structure and, in turn, to establish the number of different calling individuals in a population (complementing censuses), or to track individuals through years by the acoustic characteristics of their signals (complementing monitoring). Here we centred on the acoustic behaviour of red deer stags during reproduction to analyze its ecological determinants and to test whether calls are individual specific. Scanty information is available on how day time and weather may affect red deer acoustic activity, and whether spectrotemporal analyses of calls could permit to discriminate among different individuals. We centred on roars, the loud repeated calls given by stags throughout the rut, in a mountainous forest habitat, over two rutting seasons. Significant inter-annual differences were found in the duration of the roaring period and in the magnitude of roaring rates. Roaring rate dropped in the wettest days, suggesting that counts by ear may be biased by changeable environmental conditions within and among breeding seasons. By analysing the spectrotemporal features of roars, we found that several spectro-temporal variables varied more among than within individuals. Discriminant analysis classified a posteriori the 72% of the calls to the individuals that uttered them, and the fundamental frequency was the acoustic parameter that best discriminated individuals. Therefore, in areas in which visual information is limited by habitat characteristics, surveys by ear should be performed in dry weather and over several nights to well cover the roaring period. This count method could be complemented by spectrografic analysis of stag call bouts to assist in censusing populations and reduce the probability of double counts when stags are moving.