| Abstract: | Identifying attributes associated with good breeding habitat is critical for understanding animal population dynamics. However, the association between environmental heterogeneity and breeding probability has been often overlooked in habitat analyses. We evaluated habitat quality in a metapopulation of the endangered Iberian lynx Lynx pardinus by analyzing spatiotemporal patterns in breeding records. Data summarizing successful production of litters after emergence from dens over four years within 13 lynx territories were examined. We designed a set of generalized linear mixed models representing different hypotheses regarding how patterns in breeding records relate to environmental heterogeneity. Environmental heterogeneity was described by two characteristics: 1) a landscape index measured in lynx territories indicative of time‐averaged prey availability and 2) yearly variability in prey abundance not captured with this index. By including the random effect of the lynx territory we also accounted for other territory‐specific effects on reproduction. We found significant differences in yearly prey density dynamics among lynx territories. However, temporal variation in prey density contributed poorly to explaining lynx breeding. The most parsimonious model included the landscape structure as the only effect explaining breeding patterns. A multinomial‐model‐representation of the landscape hypothesis explained nearly 50% of variability in breeding records. Results pointed to the existence of a habitat quality gradient associated with particular landscape structures influencing lynx habitat selection and breeding performance. Underlying this gradient was time‐averaged prey availability. Probably as a result of long‐term fitness strategies in long‐lived territorial species, the short‐term fluctuations in prey availability had a minor influence. Our results illustrate how habitat inferences can be enhanced by incorporating the link between spatiotemporal patterns in reproduction and environmental heterogeneity. |
