The concept of habitat and spatial extent are key features in landscape ecology. A non-precise definition of habitat and the wrong choice of the scale can affect model outcomes and our understanding about population conservation status. We proposed a framework and applied to five species representing different ecological profiles (1) to model species occurrences and (2) to evaluate habitat structure at nine different scale extents from local landscapes to entire species range. Then, we (3) evaluated the scale sensitivity of each metric and (4) assessed if the scale sensitivity of each metric changed according to species. Our model was succesfull in predicting species occurrence for all species. When we applied deductive suitability models, the total area of remaining habitat varied from 83% to 12% of the original extension of occurrence. On average, the proportion of habitat amount, fragmentation, and carrying capacity decreased and functional increased as scale extent increased. Habitat amount and fragmentation assessed locally would show the same pattern across species’ range, but carrying capacity and functional connectivity – which consider biological features – were affected by the choice of scale.
Also, the inclusion of species preferences on habitat modeling diminished commission errors arising from landscape-scale underestimation of species’ occurrences. Local landscapes samples were not able to represent species’ entire range feature and the way that individuals reach the remaining habitat depends on species’ features. Species conservation status should be assessed preferably at the range scale and include species biological features as an additional factor determining species occupancy inside geographic ranges.
