Area of habitat reveals biases in extent of occurrence across heterogeneous landscapes: Implications for conservation assessments and a new R package

The extent of occurrence (EOO) is a key metric used by the IUCN to assess the conservation status of species. However, EOO often overestimates species' ranges, particularly those with narrow ranges, potentially resulting in the underestimation of their extinction risk. In contrast, the area of habitat (AOH) has emerged recently, as a more accurate alternative for creating species range maps and refining conservation analyses. Here, we investigate the relationship between EOO and AOH across a gradient of environmental heterogeneity. We hypothesize that AOH-to-EOO ratio decreases as environmental heterogeneity increases, which may lead to the overestimation of species ranges in areas with diverse climates or varying elevations. Using a Random Forest classifier, we tested whether AOH could be used to estimate the extinction risk. For that, we used the most comprehensive dataset of georeferenced occurrences of angiosperms in the high-altitude grasslands of the Brazilian Atlantic Forest. We conducted the analyses using 'AOHr', a new R package proposed here, that enables the estimation of AOH, potential conservation assessment, visualization of ranges, and exporting shapefiles. Our results suggest that over 85% of the taxa analyzed in our study are potentially at risk. Species in heterogeneous environments have greater discrepancy between EOO and AOH, underscoring the tendency of EOO to overestimate ranges. Finally, we show that AOH alone was able to predict the IUCN extinction risk categories with an average accuracy of 75.5%. EOO may not reliably represent species ranges, especially in climatically diverse and heterogeneous areas. AOH, by incorporating habitat suitability and excluding unsuitable regions, offers a more realistic alternative. In this context, the 'AOHr' package facilitates AOH mapping and automated conservation assessments.