Monitoring Forest Changes in Yunnan Province Using Landsat 8 and Random Forest on GEE

Protecting and sustain ably utilizing forest resources is of paramount importance for systematically monitoring and analyzing changes in forest cover across Yunnan Province. In this study, we employed the random forest classifier on the Google Earth Engine (GEE) platform, utilizing Landsat 8 images to monitor and analyze forest changes in Yunnan Province from 2013 to 2023. The primary objectives of this research are to assess alterations in forest cover by employing Random Forest (RF) algorithms that incorporate multispectral features, topographic features, and textural characteristics. The findings of the study reveal the following: (1) The Random Forest (RF) algorithm outperforms the Support Vector Machine (SVM) algorithm with respect to forest mapping accuracy. The overall average accuracy of the forest cover classification stands at 87.94%, accompanied by a Kappa coefficient of 0.76.(2) Between 2013 and 2023, the total area covered by forests in Yunnan Province expanded from 228,876.74 km<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> to 233,238.57 km ”, indicating an approximate average growth rate of about 1.91 % over this ten-year period. The maj ority of these forests are located in southwestern and northwestern regions of Yunnan, characterized by robust vegetation coverage.(3) Alterations in forest area have significant implications for the water cycle as they influence transpiration rates which subsequently affect precipitation patterns. A reduction in forested areas results in decreased precipitation levels. Furthermore, variations in surface temperature impact evapotranspiration processes along with precipitation outcomes—changes that invariably influence both forestration growth and spatial distribution. This study provides a scientific foundation for effective management practices regarding forestry resources within Yunnan Province while also serving as a vital reference point for leveraging remote sensing technology aimed at ecological monitoring initiatives.