U. Andawayanti, E. Suhartanto, R.D. Lufira, S. Ramadhania
Reliable discharge information is vital for sustainable water allocation, especially in data-scarce basins affected by climate variability. The Pancarglagas Watershed in East Java, Indonesia, has recently experienced drought-induced water shortages, while gaps in Automatic Water Level Recorder (AWLR) data have limited hydrological planning efforts. This study integrates the National Rural Electric Cooperative Association (NRECA) rainfall runoff model with a Genetic Algorithm (GA) to optimize its key parameters, Percent Subsurface (PSUB) and Groundwater Flow (GWF), for improved discharge estimation. Quality-controlled monthly rainfall, climate, and discharge data from 2004-2022 were used, with areal rainfall computed via Thiessen polygons and potential evapotranspiration derived from the FAO-56 Penman-Monteith method. Model calibration validation was conducted across five split-sample schemes using the Nash Sutcliffe Efficiency (NSE) and Pearson correlation coefficient (R). The optimal scheme (17 years calibration, 2 years validation) achieved NSE = 0.9941 and R = 0.9976 in calibration, and NSE = 0.5027 and R = 0.7924 in validation. Derived dependable flows Q80 = 0.52 m3/s, Q90 = 0.38 m3/s, and Q95 = 0.24 m3/s closely matched observed values, supporting their use in irrigation, industrial, and hydropower planning. This work demonstrates that NRECA, coupled with GA, can deliver planning-grade monthly discharge estimates in ungauged or partially gauged catchments, offering a transferable framework for water-resource management under data limitations and climate uncertainty. © 2026 Published under licence by IOP Publishing Ltd.
Water Resources Engineering Department, Faculty of Engineering, Universitas Brawijaya, Malang, Indonesia
