Uncertainty in the Spatial Interpolation of Urban Rain Gauges Data

Rainfall data is useful in many fields such as urban management, agriculture, and so on. Spatial interpolation is widely used to interpolation continue rainfall data from discrete rainfall gauges. The uncertainty in spatial interpolation is change in different region. Paper focus on urban small area of Beijing city, Xicheng District and analyses uncertainty of spatial interpolation from four aspects: rainfall gauge number, density, position, spatial interpolation methods. RMSE and cross-validation is adopted to evaluate the accuracy of interpolation and the lowest RMSE is taken as optimal. The results suggest that more gauges can get a good performance with low error compared to little stations; and dense gauges network gets high accuracy than sparse station. Ordinary kriging is simple than other method and has a good estimation (except cokriging) in small area spatial interpolation. Co-kriging has a high accuracy in interpolation but complex in computation and must be considering in the other variables.