Significance of spatial variability in precipitation for stream flow modeling in Maha Oya river basin, Sri Lanka

Abstract

Hydrological modeling and forecasting are being carried out all over the world to enable proper water resource management. Relationship between spatial variability of precipitation and runoff modeling remains an open scientific topic, which arises frequently in hydrologic modeling. The objective of this study is to find the effect of spatial variability of precipitation on stream flow modeling of Maha Oya river basin. A model was developed by using HEC-HMS 4.2.1 software and automatic calibration using Nelder-Mead method with Nash- Sutcliffe coefficient as objective function was applied. Pearson Correlation (PC) coefficient and Mean Relative Error (MRE) were also used to evaluate the model performance. Change in accuracy in simulated stream flow with the increase in number of rainfall stations used to produce input rainfall was compared. Overall accuracy of the modeling increased with number of rainfall stations. Minimum number of stations required to satisfy the objective function was three which gave Nash- Sutcliffe efficiency, Pearson correlation coefficient and mean relative error as 0.61, 0.78 and 0.13, respectively. The change in accuracy after three stations was not significant as it was before three stations. Based on these results, it could be concluded that there is a significant impact on stream flow modeling from spatial variability of rainfall in Maha Oya basin and this significance reduces after a certain limit of consideration given to the spatial nature of rainfall.