CONCLUSIONS
In this study we tested if sample counts allow the detection of population trends in a wild mountain ungulate, using simulated and real census data. We show that in most of the cases in which sampling the entire area would be able to detect population trends, sample counts in only half of the sectors have a sufficient statistical power to draw reliable conclusions about the population trend. Furthermore, a very small proportion of the total area is required to detect a strong trend even in a short period. Therefore, in taxa such as mountain ungulates, for which the cost of monitoring is high, sample counts could constitute a viable alternative to reduce the amount of resources needed for censuses, to monitor a higher number of populations of the species and to detect rapid changes of conservation interest. The results can be potentially extended to other species of mammals, especially those who exhibit a long generation time, and sample counts could be used successfully and extensively to reduce costs of monitoring. Nevertheless, the applicability of the method to species that exhibit high trend variability or with a short generation time should be carefully evaluated. Furthermore, as estimating population trends under high trend variability and weak trends on a short term is challenging, long-term monitoring is fundamental to correctly assess population dynamics of wild animals. At last, abundance extrapolated from sample counts could be biased compared to that estimated with complete counts in the entire target area, and for this reason sampling the entire area is probably still a preferable option (where there are adequate resources) to estimate abundance instead of the population trend.
Future developments of this study can include evaluating the effectiveness of sample counts with real census data from other ungulate and non-ungulate species, testing the applicability of the method to the wider possible number of species, or investigating in other systems the effect of detectability and of a biased sector selection, in order to identify cheaper census procedures also in other species.