A comparative study of the methods for estimating streamflow at ungauged sites

Abstract

Streamflow records are most fundamental data for water resources planning, hydropower design and management, water supply, irrigation, flood control, and maintaining ecosystem health. However, gauged streamflow data usually are not available to conduct such applications at rural or remote areas. Although many methods have been developed to estimate streamflow at ungauged sites, such as hydrological drainage area ratio method and regional flow duration curves (FDC) method, accuracy of estimated streamflow is still poor. In this study, an improved FDC method was developed and applied to an ungauged basin in the central south of China. Results show that the improved FDC have better performances than other traditional methods in the study area. Meanwhile, the proposed method in this study is more reasonable and interpretable than other methods in predicting regional streamflow at ungauged sites. By considering three more smaller percentiles, i.e. 1%, 2% and 5%, further analysis shows that derived empirical FDC has a better performance in the ungauged sites in this study. It suggests that inclusion of smaller quantiles can improve the accuracy of estimated streamflow both in high or low flow regimes.