Nonbiological factors affecting track censuses: implications for sampling design and reliability

Track census is a widely used method for rapid faunal assessments, which assumes that differences in track count numbers mainly reflect differences in species abundance due to some biological factors. However, some methodological and climatic variables might affect results of track censuses. Here, we tested the effect of climatic variables, such as maximum temperature, humidity, wind speed or days since last rain, and methodological factors such as censusing day period, distance from transect to vegetation edge, substrate condition or observer, on the number of tracks of mammal carnivores and some of their potential prey detected in sandy substrates. We sampled 2?×?2 km² located within the scrubland area of Doñana National Park (southwestern Spain) for carnivore and several potential prey tracks. Our results showed differences in the number of tracks detected between observers and a significant interaction between observers and the day period when censuses were carried out. Moreover, the variables increasing the quality of the substrate (higher environmental humidity, lower wind speed and days since last rain) not only led to a greater detection of carnivore tracks but, depending on the size of the species sampled other variables such as distance from transects to the vegetation border, also affected results. We recommend restricting sampling to certain fixed weather conditions when planning to monitor relative animal abundance from track censuses. When not possible, climatic or methodological variables should be included as covariates in analyses that try to test for the biological factors affecting wildlife abundance, taking into account that these variables, which affect the number of tracks detected could vary between years.