Species Distribution Model of Shorea contorta S. Vidal in the Philippines Using Maximum Entropy

The existing forest policies pushed low supplies and high export of timber in the Philippines. This study projected the current and potential distribution of Shorea contorta (S. Vidal) in the Philippines using Maximum Entropy (MaxEnt) approach. The location of Shorea contorta in protected areas and future harvestable areas were also identified and its potential revenue for timber production was assumed. MaxEnt is a machine learning algorithm that estimates a selected species probability distribution of occurrence. The occurrence data (presence-absence) and environmental variables were used as inputs in running the model. Two models were generated, Full Model and Final Model. Principal component analysis (PCA) tool was used in decreasing the number and selecting the environmental variables. The full model had an ROC of 0.755 area under the curve (AUC), whereas the final model showed a value of 0.772 AUC value and land cover having the highest contribution on the distribution of S. contorta. Unlike the full model which utilized all variables, the final model only included suitable variables, excluding variables that were highly correlated to prevent overestimation of the results. Areas appropriate for the species were approximately 7.1 million ha, whereas unsuitable areas were 20 million ha. The highest potential harvesting area for the species was in Agusan del Sur, which covered 518,570.42 ha. S. contorta grows slowly when planted in stressful conditions such as damaged soils. This improves when maize is intercropped and fertilized. The species' financial performance was found to be poor, and it is the least likely to be financially viable due to its longer rotation duration when compared to traditionally cultivated exotics. This can reduce the internal rate of return and net present values, even if the price of timber rises.