pseudoabsences

Pseudoabsences, also known as background points or background samples, are locations that are sampled to represent the absence of a species in species distribution modeling. Unlike true absences, which are confirmed locations where a species is known not to occur, pseudoabsences are randomly selected points within the study area that are considered to be unoccupied by the species. These points are used in conjunction with presence records to train and evaluate species distribution models.

The selection of pseudoabsences is a crucial step in species distribution modeling. If pseudoabsences are too close to known presence locations, the model may incorrectly infer that the species can occur in those areas. Conversely, if they are too far, they may not adequately represent the environmental conditions under which the species is absent. Various methods exist for selecting pseudoabsences, including random sampling within the study extent, random sampling within a defined buffer around presence points, or using algorithms that aim to select points that are environmentally distinct from the presence locations.

The quality and quantity of pseudoabsences can significantly impact the performance and accuracy of species distribution models. Researchers often conduct sensitivity analyses to assess how different pseudoabsence selection strategies affect model outcomes. The goal is to generate a representative sample of the environmental space where the species is not found, thereby improving the model's ability to predict suitable habitats.