Abstract: To investigate the influence of rain gauges network configuration on the accuracy of rainfall spatial interpolation and hydrological modelling,we selected different distributions of six rain gauges over the study basin and compared the results of four interpolation methods for 9-year(2006-2014) daily rainfall data. Then we calculated the mean areal daily rainfall series for different network configurations and used as the inputs of Xinanjiang model to obtain the simulated runoff. The results show that: 1) evenly distributed rain gauges lead to smaller rainfall interpolation error and higher hydrological modeling efficiency; 2)when the rain gauges are evenly distributed, no significant differences among the four rainfall interpolation results are found, while when the rain gauges are unevenly distributed, such differences are obvious especially for small rain gauge density networks; and 3) for point rainfall interpolation, Kriging interpolation method in consideration of spatial autocorrelation in residuals performs better than the other methods in calculating actual daily rainfall; for mean areal rainfall,the difference between different interpolation methods is small so we suggest to adopt a simple interpolation method, such as Thiessen Polygons or inverse distance weighted method.

Key words: rain gauges, spatial interpolation, daily rainfall, Xinanjiang Model, Kriging