Quantifying whether different demographic models produce incongruent results on population dynamics of two long-term studied rodent species

1. Population density (ind/ha) of long-term (>15 years) series of CMR populations, using distinct demographic models designed for both open and closed populations, were analysed for two sympatric species of rodents (Myodes glareolus and Apodemus flavicollis) from a mountain area in central Italy, in order to test the relative performance of various employed demographic models. In particular, the hypothesis that enumeration models systematically underestimate the population size of a given population was tested. 2. Overall, we compared the performance of 7 distinct demographic models, including both closed and open models, for each study species. Although the two species revealed remarkable intrinsic differences in demography traits (for instance, a lower propensity for being recaptured in Apodemus flavicollis), the Robust Design appeared to be the best fitting model, showing that it is the most suitable model for long-term studies. 3. Among the various analysed demographic models, Jolly-Seber returned the lower estimates of population density for both species. Thus, this demographic model could not be suggested for being applied for long-term studies of small mammal populations because it tends to remarkably underestimate the effective population size. Nonetheless, yearly estimates of population density by Jolly-Seber correlated positively with yearly estimates of population density by closed population models, thus showing that interannual trends in population dynamics were uncovered by both types of demographic models, although with different values in terms of true population size