Quantifying evaporation losses in the river-lake continuum in arid regions basins using isotope data

Understanding surface water evaporation losses is essential for quantifying water balance and optimizing resource management in arid inland basins. Arid inland regions face severe water scarcity driven by high evaporation and intensive human regulation, yet the magnitude and spatial variability of anthropogenic evaporation losses along the river-lake continuum remain insufficiently quantified. To capture the spatiotemporal variations of evaporation, a systematic observation network was established to continuously monitor surface water, precipitation, and meteorological parameters from 2019 to 2021. Based on these observations, we combined measurements of stable hydrogen and oxygen isotope values with the Craig-Gordon model, implemented in the Hydrocalculator, to quantitatively estimate evaporation losses across multiple hydrological units along the mountain-oasis-desert continuum. Results indicate that evaporation loss rates are 1.77 % for mountain rivers, 2.95 % for mountain reservoirs, 5.64 % for oasis rivers, 7.49 % for oasis reservoirs, and 33.17 % for terminal lakes. Cumulative losses associated with river-lake continuum account for 13.76 % of total surface water, including 10.44 % from artificial reservoirs and 33.17 % from ecological water transfers, highlighting the substantial role of human-made storage and diversion structures in enhancing evaporation losses. Reservoirs exhibit substantially higher evaporation losses than rivers in both mountainous and oasis regions, while terminal lakes experience the greatest overall loss. The combined isotope modeling approach strengthens quantitative understanding of surface water evaporation under intensive human regulation and demonstrates the amplified evaporation induced by human water management. These findings provide a scientific foundation for improving water balance assessments, optimizing reservoir operation, and guiding sustainable management in arid inland basins.