Archives of Current Research International

Accurate estimation of direct runoff hydrographs from small ungauged micro-watersheds remains one of the central challenges in applied watershed hydrology, particularly in data-scarce, semi-arid agricultural regions. This study presents a comprehensive comparative evaluation of three unit hydrograph-based conceptual models — the Nash Instantaneous Unit Hydrograph (IUH), the Geomorphological Instantaneous Unit Hydrograph (GIUH), and the Representative Unit Hydrograph (RUH) — applied to the Kalmandargi Tanda-1 gauged micro-watershed (55.4 ha) in Kalburgi District, Karnataka, India. Rainfall-runoff data from 2020 to 2024, encompassing 15 observed storm events, were used for model calibration and validation. Nash IUH parameters (n and k) were estimated by three methods: moments, L-BFGS-B optimization, and Differential Evolution (DE) optimization. Excess rainfall was computed using the Intensity-Infiltration method with dynamic infiltration rates. The Nash IUH with DE-GM optimization consistently outperformed other models, yielding mean NSE of 0.91 and RMSE of 0.031 m³/s across all events. GIUH performed competitively (mean NSE = 0.87), while RUH (simple mean) showed acceptable but comparatively lower accuracy (mean NSE = 0.82). A geomorphological characterization of 14 surrounding ungauged micro-watersheds was carried out to assess morphometric similarity. A weighted Similarity Index (SI) based on Euclidean distance was developed using Analytic Hierarchy Process (AHP) weights for seven geomorphological parameters, enabling objective identification of hydrologically similar donor watersheds. Nash IUH and GIUH parameters were scaled and transferred to five selected ungauged micro-watersheds with SI ≥ 0.70. The study demonstrates that geomorphological similarity-based parameter transfer, coupled with the Nash IUH DE-GM optimization approach, provides a reliable framework for ungauged watershed runoff prediction in semi-arid micro-watershed settings.