A machine learning approach to mapping suitable areas for forest vegetation in the eThekwini municipality

Abstract

Driven by climate change, global forests are undergoing significant transformations in growth, ecology, and distribution, necessitating informed restoration and conservation strategies, particularly in the eThekwini Municipality where anthropogenic activities exacerbate these trends. Modelling current forest suitability (2023) utilized bioclimatic variables from the WorldClim dataset, alongside elevation and slope from the Shuttle Radar Topography Mission (SRTM) dataset, with remote sensing data acquired from Landsat 9 and Sentinel 2A. Future forest suitability (2021–2040) was projected also using bioclimatic variables from two Global Climate Models (GCMs) under four WorldClim Shared Socioeconomic Pathway (SSP)-based Representative Concentration Pathway (RCP) scenarios. Employing Random Forests (RF), Light Gradient Boosting (LightGBM), and Artificial Neural Networks (ANN), data processing was carried out using Google Earth Engine (GEE), QGIS and Python, with model accuracy primarily assessed using the Receiver Operating Characteristic (ROC) curves and the Area Under the ROC Curve (AUC). LightGBM demonstrated superior performance, achieving AUCs of 96.88% and 93.75% for current and future suitability mapping, respectively, with annual precipitation and vegetation changes identified as crucial variables. Currently, 30% of the municipality's land is deemed suitable, primarily concentrated in the central region. Future projections highlight the mountainous north-western region as most suitable, notably under the SSP370 scenario with a projected suitable area of 63%. Strategic recommendations include prioritizing reforestation efforts, engaging private landowners, exploring urban reforestation opportunities, and implementing continuous monitoring for adaptive management, thereby enhancing carbon sequestration, biodiversity conservation, and ecosystem resilience. This study provides valuable insights for informed decision-making in forest restoration and conservation, despite inherent uncertainties.