Niche modelling of endangered philippine birds using GARP and MAXENT

Abstract

Researches in the area of environmental niche modeling has been using climatic parameters in modeling niches of bird species. However, local experts believe that human activity is a great cont ributor to the birds' habitat status — a condition not often tested on niche model accuracy. Genetic Algorithm for Rule-set Production (GARP) and Maximum Entropy (MaxEnt) are two of the most commonly used and efficient methods in niche modeling using climatic data. In conjunction, this study aims to test the accuracy of the bird niche models produced by both GARP and MaxEnt when dealing with human-related parameters. Bird sightings of six endangered Philippine bird species found in Negros were used for the study. Niche models/prediction models from GARP and MaxEnt underwent partial-area ROC analysis for model evaluation. Results of the tests show that the prediction models of the two niche modeling algorithms are mostly good and positive predictions with GARP showing more accurate results than MaxEnt. In addition, GARP showed lower accuracy results when human-related parameters were introduced as compared to having no human-related parameters during the modeling phase. MaxEnt, on the other hand, showed accuracy improvements when the parameters were used. MaxEnt was also proven to be an ideal algorithm than GARP in dealing with species with very few occurrences.