Attributing impacts of LULCC on the boundary layer climate of South Africa’s eastern escarpment

Land surface characteristics may influence the planetary boundary layer atmosphere and climate through exchanges of moisture, energy, and momentum near the surface. We attributed the impact of multitemporal landsat-derived land use/land cover change (LULCC) on temperature and precipitation variability in eastern South Africa using reanalysis data and satellite-derived estimates from 1979 to 2020. Landsat images were classified into different land cover classes using a machine learning random forest pixel-based supervised algorithm within the cloud-based Google Earth Engine. Time series analysis was employed to analyze cycles and trends in LULCC and hydrometeorological variables, whilst the variable importance model determined the most sensitive variable. The impacts of LULCC on the boundary layer climate were attributed via multiple linear regression. An uninterrupted rapid expansion of urban areas was observed, resulting in the transformation of grasslands, water bodies, forests, and croplands. Statistically significant changes in moisture and energy fluxes, and hydrometeorological variables were observed across the study period. Latent heat flux (LHF), as well as rainfall decreased, while maximum temperature, sensible heat, and potential evapotranspiration (PET) increased significantly. We found that LULCC is significantly impacting the boundary layer climate, with urban and bare land, grasslands, forests, and croplands influencing temperature positively while negatively influencing rainfall. Rainfall was most sensitive to changes in LHF, whilst the key driver of temperature variability was PET. Our results reinforce the significance of LULCC and associated feedbacks to understanding boundary layer processes, climate variability, and change.