Comparative interpretation of count, presence–absence and point methods for species distribution models