Wavelet analysis residual kriging vs. neural network residual kriging

This paper deals with the problem of spatial data mapping. A new method bas ed on wavelet interpolation and geostatistical prediction (kriging) is prop osed. The method - wavelet analysis residual kriging (WARK) - is developed in order to assess the problems rising for highly variable data in presence of spatial trends. In these cases stationary prediction models have very l imited application. Wavelet analysis is used to model large-scale structure s and kriging of the remaining residuals focuses on small-scale peculiariti es. WARK is able to model spatial pattern which features multiscale structu re. In the present work WARK is applied to the rainfall data and the result s of validation are compared with the ones obtained from neural network res idual kriging (NNRK). NNRK is also a residual-based method, which uses arti ficial neural network to model large-scale non-linear trends. The compariso n of the results demonstrates the high quality performance of WARK in predi cting hot spots, reproducing global statistical characteristics of the dist ribution and spatial correlation structure.