Snowmelt runoff model (SRM) for regulated watersheds with regulation-correction

Abstract

We expanded the Snowmelt Runoff Model (SRM) to simulate streamflow in regulated watersheds, resulting in a modified framework termed the Expanded SRM (E-SRM) that integrates multi-year automated batch processing, nested iterators, and a seasonal divider algorithm for streamflow simulation. A parsimonious regulation-correction approach was developed that conceptually divides the watershed into a pristine upstream “daughter” subwatershed and a larger, regulated “mother” watershed. Hydrological parameter transferability was assumed between the daughter and mother watersheds. We applied the E-SRM to the Morony watershed in Montana, USA (~ 59,400 km²; elevation range: 860–3418 m). The area was subdivided into the Morony and Canyon Ferry watersheds, with the latter treated as a pristine basin for calibration. Following calibration at Canyon Ferry, regulation-correction was applied using streamflow from the Canyon Ferry, Hauser, and Holter dams. Validation was conducted over the entire Morony watershed. Three methodological scenarios were evaluated: (1) 21-year calibration and 21-year validation; (2) 11-year calibration and 21-year validation; and (3) calibration using odd years with validation on odd and even years. In all scenarios, comparison between observed and regulation-corrected streamflow showed improved performance across multiple model assessment metrics. These included both absolute (e.g. Nash–Sutcliffe Efficiency: from − 0.16 to 0.74, − 0.43 to 0.59, − 0.16 to 0.67; Kling–Gupta Efficiency) and relative (e.g. Root-Mean-Square Error, Normalized Root-Mean-Square Error, Mean Absolute Error, and Volume Difference) indicators, highlighting the significant impact of flow regulation on SRM performance. The E-SRM framework offers new opportunities for research and practical application in water resource management.