Predicting animal abundance through local ecological knowledge : An internal validation using consensus analysis