Surface water area dynamics of the major lakes of Ethiopia (1985–2023): A spatio-temporal analysis

Abstract

Ethiopian lakes face multiple ecosystem threats from human and natural factors, including climate change. The long-term dynamics of water surface area in many lakes remain unknown. This study presents the first comprehensive account of major Ethiopian lakes’ long-term water surface area dynamics, providing essential information for prioritizing hydrological conservation strategies. We extracted the time-series water surface area of the lakes from Landsat images in the Google Earth Engine (GEE), using the Modified Normalized Water Index (MNDWI) and machine learning (Random Forest) methods. Random Forest (RF) outperformed MNDWI across many lakes, especially when water body spectral characteristics became complex. A comparison of the results with the Joint Research Center (JRC) Global Surface Water Mapping Layers showed a good agreement. Long-term water surface area dynamics showed both common and distinct trends. Common trends included: (a) similar recession and expansion periodicities despite lakes being in different climatic zones, (b) an increase in the inter-annual variability of water area after 2000, and (c) a common abrupt point in time (2000/2001). Distinct trends included: (i) long-term decline in five lakes (Abiyata, Langano, Shalla, Chamo, and Hayq), (ii) expansion in three lakes (Ziway, Hawassa, and Abaya), and (iii) non-homogeneous long-term trends in two lakes (Tana and Hardibo). These common and distinct trends show the interplay of climate phenomena and lake-specific factors. Climate factors govern inter-annual lake area change (decreasing/increasing) and recession and expansion cycle periods. The magnitudes of changes and recovery rates after each change cycle vary based on lake-specific factors. We recommend prioritizing conservation efforts for lakes with high water storage degradation risks, as indicated by significant water loss over the historical period and persistent long-term declining trends. This includes Abiyata, Chamo, Shalla, Langano, Hayq, and Hardibo, in order of priority.