Integrating Landsat and Sentinel-2 for the assessment of Rangelands vegetation dynamics

Abstract

Rangelands are vast ecosystems that support biodiversity and livestock grazing. Monitoring species composition in rangelands is vital for assessing ecosystem health and informing management decisions. This study utilized Landsat and Sentinel-2 data to map vegetation composition, monitor vegetation dynamics over time, as well as analyze the relationship between Normalized Difference Vegetation Index, and environmental variables (temperature and rainfall) in Gweru district, Zimbabwe. Land use and land cover maps identified water, forest, bareland, grassland, and built-up. The accuracy of Landsat data was 94.19% in (1985), 93.06% (2015) and 97.06% in (2023). Sentinel’s accuracy was lower at 92% in (2016) and 97.99% in (2023). From 1985-2023, bareland increased by 15.01% and 8% increase in built-up areas, while a significant decline of forest -25%. NDVI showed a strong relationship with rainfall (R² = 0.9748) and a weaker relationship with temperature (R² = 0.3333), highlighting the importance of integrating remote sensing in rangeland management.