Spatial Land Use and Land Cover (Lulc) Changes in Swat River Basin from 1992-2022

Abstract

Land Use and Land Cover (LULC) changes significantly influence environmental processes and watershed dynamics. This study examines the spatio-temporal LULC changes in the Swat River Basin, Khyber Pakhtunkhwa, Pakistan, over a period of three decades (1992–2022). Multi-temporal Landsat imagery acquired from Landsat-5 TM and Landsat-8 OLI was processed using Google Earth Engine (GEE). A Random Forest (RF) supervised classification algorithm was employed to classify six LULC classes: snow cover, forest, agriculture, barren land, water bodies, and settlements. Classification accuracy was evaluated using Overall Accuracy (OA), Kappa Coefficient, Producer Accuracy (PA), and Consumer Accuracy (CA), yielding satisfactory results with the highest OA of 87% and Kappa value of 0.84. The findings reveal substantial changes in the basin's landscape. Snow cover decreased dramatically from 30% to 3%, while forest cover declined from 36% to 17% between 1992 and 2022. Conversely, agriculture expanded from 4% to 19%, barren land increased from 28% to 54%, and built-up areas grew from 1% to 5%. Water bodies remained relatively stable throughout the study period. The observed changes are primarily attributed to climate change, population growth, urban expansion, agricultural development, and deforestation. The study highlights the effectiveness of GEE and RF algorithms for long-term LULC monitoring and provides valuable information for sustainable watershed management and environmental planning in the Swat River Basin.

Keywords: Land Use and Land Cover (LULC); Spatio-temporal Change Detection; Swat River Basin; Google Earth Engine (GEE); Random Forest Classification; Landsat Imagery; Watershed Management; Climate Change; Deforestation; Urban Expansion; Land Cover Dynamics; Remote Sensing.