Implications of Land Use/Land Cover Change on Land Surface Temperature and Hydrological Cycle Components of the Baro Akobo River Basin of Ethiopia
摘要
Land Use Land Cover (LULC) change is a major environmental concern that negatively affects natural systems, disrupts the water-energy balance, and impairs key components of the hydrological cycle. The Central Baro Akobo River Basin (BARB) of Ethiopia has been experiencing massive forest cover loss from the dynamics of LULC change. However, the implications of this on land surface temperature (LST) and the hydrological cycle have not been well studied in the basin. To achieve this, we utilized remotely sensed and reanalysis datasets that were processed and analyzed in Google Earth Engine (GEE) platform, including implementing the Classification and Regression Trees (CART) algorithm to classify LULC. The results revealed that, 53.7% of the study area was covered by forest in 1991, then it continued to decrease in inconsistent trends over the study periods (1991–2021). It was decreased by 11.2% in 2001, 27% in 2010, 21.1% in 2014, 24.3% in 2017 and 22.2% in 2021 from 1991’s (base year). On the other hand, artificial (developed) areas covered 2.5% of the total area in 1991, was continued in increasing trend over the study period except for 2014 (10.8%) and reached 20.2% in 2021, the highest percentage coverage over the study periods in comparison to 1991’s. Farm land also showed increasing trend over the study period in general except for 2021, for which it was decreased by 2.1% from 1991’s. Annual mean LST spatially varied 1–3 °C 2001, 5–13 °C in 2010, 1 °C in 2014, 2–3 °C in 2017 and 1–3 °C 2021 in comparison with 1991’s annual mean LST. When specific locations were considered, annual mean LST increased by 11 °C (31%) in Itang, followed by Lare 8 °C (22%) and Gambella Zuria 8 °C (25%) in 2010 from 1991. The correlation analysis among LST, normalized difference vegetation index (NDVI), actual evapotranspiration (ET), and soil moisture (SM) revealed that, LST and ET had a very strong significant correlation coefficient of 0.96 at p < 0.01 for the highland area but reduced to a strong significant correlation coefficient of 0.78 at p < 0.01 for lowland area. LST and NDVI had moderate negative correlation but significant correlation coefficient values of -0.41 at P < 0.05 and − 0.50 at P < 0.01 for the highland and lowland areas, respectively. The findings of this study can support efforts on water resources management implementation in the basin.
Graphical Abstract