A comparative study of soil moisture accounting and soil conservation service curve number methods for modelling water yield in an ungauged tropical forest catchment

ABSTRACT

Flowchart illustrating a catchment modeling workflow comparing continuous SMA and event-based SCS-CN methods, highlighting the superior performance of the SMA framework in simulating water yield.

Accurate estimation of continuous water yield from ungauged tropical catchments is vital for reservoir management. This study evaluates the physically based Soil Moisture Accounting (SMA) and the empirical Soil Conservation Service Curve Number (SCS-CN) in HEC-HMS for the Perlis State Park catchment in Malaysia. Models were validated against downstream discharge records using hydrograph analysis, Taylor diagrams, and volumetric assessments. Results indicate that the SMA method was substantially more suitable for this continuous simulation objective than the SCS-CN method, successfully capturing long-term streamflow, low-flow persistence, and seasonal variability (achieving monthly calibration correlations of r ≈ 0.85, n = 108, P < 0.001). In contrast, the event-driven SCS-CN method systematically overestimated wet-season peak discharges (>80 m3/s) and inflated monthly yields. Applying calibrated SMA parameters to ungauged forest nodes revealed that the central Wang Mu and Mata Ayer sectors dominate reservoir inflow, contributing peak monthly yields of 18 × 106 m3 and 11 × 106 m3, respectively. The SMA method captured the forest's vital hydrological buffering capacity, which sustains essential baseflow during the pronounced dry season. Ultimately, these findings validate that the continuous SMA method is a highly suitable tool for the specific continuous water-yield and baseflow-focused objectives of this data-scarce tropical region.