Integrating multiple datasets into spatially-explicit capture-recapture models to estimate the abundance of a locally scarce felid

The conservation of animal populations often requires the estimation of population size. Low density and secretive behaviour usually determine scarce data sources and hampers precise abundance estimations of carnivore populations. However, joint analysis of independent scarce data sources in a common modeling framework allows unbiased and precise estimates of population parameters. We aimed to estimate the density of the European wildcat (Felis silvestris) in a protected area of Spain, by combining independent datasets in a spatially-explicit capture-recapture (SCR) framework. Data from live-capture with individual identification, camera-trapping without individual identification and radio-tracking concurrently obtained were integrated in a joint SCR and count data model. Ten live captures of five wildcats were obtained with an effort of 2034 trap-days, whereas seven wildcat independent events were recorded in camera traps with 3628 camera-days. Two wildcats were radio-tagged and telemetry information on their movements was obtained. The integration of the different data sources improved the precision obtained by the standard SCR model. The mean (± SD) density estimated with the integrated model (0.038 ± 0.017 wildcats/km2, 95% highest posterior density 0.013–0.082) is among the lowest values ever reported for this species, despite corresponding to a highly protected area. Among the likely causes of such low density, low prey availability could have triggered an extinction vortex process. We postulate that the estimated low density could represent a common situation of wildcat populations in the southern Iberia, highlighting the need for further studies and urgent conservation actions in the furthermost southwestern range of this species in Europe.