Understanding the ecological value of novel ecosystems for native species persistence is vital for biodiversity conservation. Because of difficulties in obtaining long-term demographic data, and the ambiguity of indicators used to quantify ecological processes such as habitat selection, we still rely heavily on species lists to make important conservation decisions. High species richness has been documented in shade coffee plantations, but few studies go beyond species presence to assess population persistence. As part of a larger project evaluating the role of habitat selection in species adaptation to transformed landscapes, we assessed if landscape-level estimates of abundance corresponded with apparent survival, with the goal of using them as indicators of habitat preference. We present results of a capture-recapture study for twelve species of resident birds that are relatively common in shade coffee plots and mountain forest remnants in the Sierra Nevada de Santa Marta, Colombia. We combined data from two independent projects carried out in 17 sites and 12 sampling occasions from 2009 to 2015. To overcome sampling heterogeneity between habitats, we adapted a Bayesian formulation of the Jolly-Seber population model by making encounter probabilities conditional on sampling effort and using it to estimate demographic variables separately by habitat and occasion.
Results/Conclusions
Data for each species included 117-874 records (22-49% recaptures) for 89-548 individuals. Seven species were generally more abundant in forest, but nonetheless equally or more abundant in coffee during certain occasions. Five species were more generally abundant in coffee, but because effort was lower in plantations, patterns were less distinct. Species showed differences in apparent survival in up to 5 occasions. For abundance, 46% of 144 comparisons were higher in forest, 20% were higher in coffee, and 34% were unresolved. For apparent survival, 16% were higher in forest, 6% were higher in coffee, and 78% were unresolved. More conclusive inference from these patterns will require additional demographic and ecological information; however we highlight two important results: 1) that abundance (expected number of present individuals) and apparent survival (compound measure of site fidelity and true survival) of a species are not necessarily correlated and 2) that longer-term, habitat-specific demography can reveal very different patterns than those described with counts or presence/absence inventories. We recommend that care should be taken when assessing the conservation potential of novel ecosystems, and that demographic information should be used to improve conservation tools for birds in shade coffee specifically, and biodiversity in transformed landscapes in general.