On evaluating species distribution models with random background sites in place of absences when test presences disproportionately sample suitable habitat

On evaluating species distribution models with random background sites in place of absences when test presences disproportionately sample suitable habitat