The use of suitability models and remote sensing to map forest suitability: a systematic review

Abstract

The rapid decline of vegetated landscapes jeopardises vital ecosystem services underpinning climate change mitigation efforts. Restoring and repurposing these landscapes presents a powerful opportunity to recover lost services and strengthen the fight against climate change. This study undertook a comprehensive review of the scientific literature, focusing on suitability models and remote sensors used to identify areas suitable for forest vegetation. An exhaustive search and analysis of publications across the globe made over a 15-year period from 2008 to 2022 across three major databases (Scopus, ScienceDirect, and Web of Science) yielded 80 relevant publications. This analysis revealed a significant upward trend in research output, particularly since 2020. This surge reflects the increasing urgency of global landscape restoration initiatives. Additionally, the analysis of the reviewed articles revealed a rising preference for medium- to high-resolution remote sensing data, with Landsat emerging as the dominant sensor for forest suitability assessments. Notably, Maximum Entropy (MaxEnt) emerged as the most widely used model, followed by the increasingly popular Random Forest (RF). However, a concerning geographical disparity in research was identified. Publications were heavily concentrated in the Americas and Asia, while developing nations showed a significant research gap. This discrepancy highlights the critical need for increased research efforts in developing countries to equip them with the robust suitability models and advanced sensor technologies necessary for effective and targeted forest rehabilitation and restoration initiatives. Investing in research capacity building within developing nations holds immense potential to accelerate global landscape restoration efforts.