River baseflow in supplying reservoirs inflows of Tehran metropolis: A machine learning modeling based on influencing factors

Abstract

Study regions

Five watersheds Taleqan, Karaj, Latian, Lar, and Mamlou, located in Salt Lake Basin, around the Tehran Province in Northern Iran.

Study focus

This study investigates the dynamics of Baseflow (BF) in five reservoirs critical to Tehran’s water supply, using an 18-year dataset (1999–2016). While three digital filter methods were used to estimate daily baseflow in the studied reservoirs, the results from the Chapman-Maxwell method were selected for further investigation. Accordingly, daily streamflow data were processed using this method to separate baseflow and were aggregated monthly. The Baseflow Index (BFI), calculated as the ratio of mean BF to total streamflow, revealed BF contributions ranging from 55 % to 89 %, with soil moisture and snowmelt identified as dominant drivers. The BFAST algorithm detected breakpoints in BF trends, linking shifts to climatic variability and human activities like dam operations. Cross-correlation analysis highlighted SM (0–290 cm depth) as the strongest predictor of BF (CCF: 0.80–0.89), with immediate response times, while Smelt exhibited a seasonal lag (2–3 months). Snow cover, temperature, and vegetation (NDVI) also influenced BF, with NDVI showing a negative correlation due to increased water uptake. A Random Forest model, validated with 70 % training and 30 % testing data, confirmed SM’s primacy (R² up to 0.90 for Karaj Dam), followed by Smelt and humidity index. Breakpoints in BF trends, underscored the impact of land-use changes and climate shifts.

New hydrological insights for the region

In populated urban areas like Tehran Metropolis where streamflow is critical for domestic water supply, analyzing the role of BF in streamflow of reservoirs supplying the water demands and identifying its driving factors within watersheds is crucial for sustainable water management. The findings advocate for watershed-specific strategies, including enhanced soil moisture retention and adaptive reservoir management, to mitigate water scarcity. This study provides a framework for sustainable water management in semi-arid regions, emphasizing the integration of remote sensing and hydrological modeling to address climate and anthropogenic pressures.