Caution Is Needed When Using Niche Models to Infer Changes in Species Abundance: The Case of Two Sympatric Raptor Populations.

Abstract

Simple SummaryThis study focuses on sympatric populations of hen harrier (Circus cyaneus) and Montagu’s harrier (Circus pygargus) breeding in mountain heathlands in the NW Iberian Peninsula. These populations have been affected by habitat loss caused by land-use change. Despite the mounting evidence supporting positive relationships between species abundance and habitat suitability, the capacity of ecological niche models to capture variations in population abundance remains largely unexplored. This study shows that habitat suitability predicted from niche models was significantly correlated with the relative abundance for hen harrier and, to a lesser extent, for Montagu’s harrier. However, the temporal variation in local population abundance was not significantly explained by habitat suitability changes predicted by the niche models. These findings call for caution in the use of niche models to infer changes in population abundance. The positive relationship between species abundance and habitat suitability supports the use of niche models to estimate abundance but does not guarantee the ability of these models to predict temporal variations in species abundance. These findings highlight (1) the need to include other possible biotic or abiotic factors involved in population abundance dynamics into niche models and (2) the need to establish specific monitoring protocols for tracking population dynamics.Despite the mounting evidence supporting positive relationships between species abundance and habitat suitability, the capacity of ecological niche models (ENMs) to capture variations in population abundance remains largely unexplored. This study focuses on sympatric populations of hen harrier (Circus cyaneus) and Montagu’s harrier (Circus pygargus), surveyed in 1997 and 2017 in an upland moor area in northwestern Spain. The ENMs performed very well for both species (with area under the ROC curve and true skill statistic values of up to 0.9 and 0.75). The presence of both species was mainly correlated with heathlands, although the normalized difference water index derived from Landsat images was the most important for hen harrier, indicating a greater preference of this species for wet heaths and peat bogs. The findings showed that ENM-derived habitat suitability was significantly correlated with the species abundance, thus reinforcing the use of ENMs as a proxy for species abundance. However, the temporal variation in species abundance was not significantly explained by changes in habitat suitability predicted by the ENMs, indicating the need for caution when using these types of models to infer changes in population abundance.