Occurrence vs abundance models: Differences between species with varying aggregation patterns

Abstract

Predicting distribution has become a fundamental component in conservation or wildlife management. Modelling is increasingly used to identify important areas (e.g., those areas more suitable for a species or more likely to hold high densities). Models often use presence/absence rather than abundance data, partly because measuring abundance is more difficult than measuring presence. We aimed to test if the relationship between occurrence models and predicted abundance varied for two sibling species that differ in the level of nest aggregation: the Montagu's harrier (a semi-colonial raptor species) and the hen harrier (more territorial). We modelled presence/absence distribution and the number of pairs of each species with GLM and large-scale environmental variables, and compared predicted results of both sets of models. In the case of the hen harrier, predictions of the presence/absence model reliably identified areas with highest densities for the species. In contrast, in the Montagu's harrier, there were large apparently favourable areas where predicted breeding density was low. Our results indicate that breeding system is likely to shape the relationship between presence/absence vs density models. In species that are randomly or evenly spaced, even if spatial variations in density occur, using results of presence/absence models is likely to be adequate for population monitoring. In contrast, in the case of semi-colonial species, it is necessary to take into account both occurrence and abundance models to identify areas of conservation importance or concern. There are a considerable number of birds which are semi-colonial or aggregated species, thus these results could have general implications.