Prediction of global marginal land resources for Pistacia chinensis Bunge by a machine learning method

Abstract

AbstractBiofuel has attracted worldwide attention due to its potential to combat climate change and meet emission reduction targets. Pistacia chinensis Bunge (P. chinensis) is a prospective plant for producing biodiesel. Estimating the global potential marginal land resources for cultivating this species would be conducive to exploiting bioenergy yielded from it. In this study, we applied a machine learning method, boosted regression tree, to estimate the suitable marginal land for growing P. chinensis worldwide. The result indicated that most of the qualified marginal land is found in Southern Africa, the southern part of North America, the western part of South America, Southeast Asia, Southern Europe, and eastern and southwest coasts of Oceania, for a grand total of 1311.85 million hectares. Besides, we evaluated the relative importance of the environmental variables, revealing the major environmental factors that determine the suitability for growing P. chinensis, which include mean annual water vapor pressure, mean annual temperature, mean solar radiation, and annual cumulative precipitation. The potential global distribution of P. chinensis could provide a valuable basis to guide the formulation of P. chinensis-based biodiesel policies.