Kim et al. (2026) Multi-Model Comparison of Hydrologic Simulation Performance Using DWAT, PRMS, and TANK Models

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Water
• Year: 2026
• Date: 2026-01-06
• Authors: Deokhwan Kim, Wonjin Jang, Heechan Han, Hyoung-Sub Shin, H. Y. Kim, Cheolhee Jang
• DOI: 10.3390/w18020145

Research Groups

Not explicitly provided in the paper text.

Short Summary

This study compares the streamflow simulation performance of three hydrological models (DWAT, PRMS, TANK) across mountainous, mixed-use, and urbanized watersheds in the Republic of Korea, finding that DWAT generally exhibits the most stable and highest overall performance, but model efficacy is strongly dependent on watershed characteristics.

Objective

• To compare the streamflow simulation performance of a semi-distributed hydrological model (DWAT) and two conceptual models (PRMS, TANK) across diverse watershed types (mountainous, mixed-use, urbanized) in the Republic of Korea.

Study Configuration

• Spatial Scale: Three watersheds in the Republic of Korea: Okdong-gyo (mountainous), Wonbu-gyo (mixed-use), and Daegok-gyo (urbanized).
• Temporal Scale: Consistent calibration and validation periods for all models, using identical hydroclimatic datasets.

Methodology and Data

• Models used: DWAT (Dynamic Water Resources Assessment Tool, semi-distributed), PRMS (conceptual), TANK (conceptual).
• Data sources: Identical hydroclimatic datasets (specific type not detailed, but used for calibration and validation).

Main Results

• DWAT exhibited the most stable and highest overall streamflow simulation performance across all watersheds.
• DWAT showed particularly strong results in the urbanized Daegok-gyo basin (Nash-Sutcliffe Efficiency (NSE) = 0.85, coefficient of determination (R²) = 0.88).
• The TANK model performed best in the mixed-use Wonbu-gyo basin (NSE = 0.82, R² = 0.83).
• PRMS systematically underestimated streamflow, especially under high-flow and low-flow conditions.
• Statistical comparisons using Friedman and post hoc Dunn tests confirmed that performance differences among models were statistically significant (p < 0.001).
• Hydrological model performance strongly depends on watershed characteristics.

Contributions

• Provides a quantitative and statistically supported comparison of three distinct hydrological models across a range of watershed characteristics (mountainous, mixed-use, urbanized).
• Demonstrates the significant influence of watershed characteristics on the performance of hydrological models.
• Offers a robust basis for selecting appropriate runoff simulation models tailored to specific basin types.

Funding

Not explicitly provided in the paper text.

Citation

@article{Kim2026MultiModel,
  author = {Kim, Deokhwan and Jang, Wonjin and Han, Heechan and Shin, Hyoung-Sub and Kim, H. Y. and Jang, Cheolhee},
  title = {Multi-Model Comparison of Hydrologic Simulation Performance Using DWAT, PRMS, and TANK Models},
  journal = {Water},
  year = {2026},
  doi = {10.3390/w18020145},
  url = {https://doi.org/10.3390/w18020145}
}