How well do proxy species models inform conservation of surrogate species?