Predicting Cork Oak Suitability in Maâmora Forest Using Random Forest Algorithm

Abstract

Maamora is considered the most important cork-oak forest in the world with regard to surface. Therefore, anthropic pressure, including cork harvesting, grazing and soft acorn picking up by local communities, has harmful consequences on forest regeneration and the forest become older exceeding harvesting age. Thus, its sustainability depends on the managers’ ability to succeed cork oak plantations. This work presents an assessment approach to evaluate Quercus suber suitability to its plantation which is based on a random forest algorithm (RF). In fact, this suitability has been assessed through analyzing management data related to previous plantation success rates (SR). Then a relationship between SR and a set of environmental and social factors has been investigated using the RF. Application of the fitted model to continuous maps of all involved factors enabled establishment of suitability maps which would help managers to make more rational decisions in terms of cork oak regeneration, ensuring Maamora forest sustainability.