Comparison of Precision Coefficients for Streamflow Data Simulation

Journal article

Authors/Editors

Strategic Research Themes

Publication Details

Author list: Noparat Techapunratanakul; Pongpan Kanjanakaroon; Chaiwat Ekkawatpanit

Publication year: 2024

Volume number: 33

Issue number: 2

Start page: 104

End page: 123

Number of pages: 20

URL: https://li01.tci-thaijo.org/index.php/science_kmitl/article/view/258682

Languages: Thai (TH)

Abstract

Streamflow data is crucial for water resource development projects. Acquiring long-term streamflow data poses difficulties due to maintenance or technological issues. The use of a mathematical model may reconstruct missing streamflow data. Nevertheless, it is necessary to calibrate and validate simulation data using precision coefficients. Different precision coefficients can be used to evaluate the results of hydrological simulations. This paper presents the suitability and constraints of precision coefficients in evaluating streamflow outcomes using a hydrological model under three conditions: water balance, flood, and drought. The simulated data was generated with both systematic and dynamic errors. Twenty-one coefficients were selected based on the principle of relative and absolute errors to evaluate the simulated data. Regardless of systematic errors, the findings indicate that all simulated outcome conditions could be evaluated using several coefficients such as NSE, NSEj=1, NSEj=3, NSErQ, ln NSE, NSEsqrtQ, NSEiQ, MSE, RMSE, MAE, RMSRE, and AAPE. These coefficients could analyze the simulation’s accuracy as well as determine its accuracy tendency. In addition, the precision coefficient RLFDoH can be used to evaluate the simulation results for flood scenarios, and the precision coefficients NSEuL, BuL, and RLFDuL can be used to assess the simulations for drought scenarios. For dynamic errors, the coefficients of NSEj=1, NSEiQ, and RMSRE were suitable for assessing the water balance and flood conditions, while drought is aligned with the coefficients of NSEuL and RLFDuL. The results can serve as a criterion for determining appropriate coefficients for streamflow simulation.

Keywords

Hydrological Modeling, water resource