Application of Geostatistics in Time Series: Mashhad Annual Rainfall

There are numerous methods for data fillingin hydrology. Most, however, are based oncorrelations with nearby stations in a generalscheme of regionalization. These methods,though robust, fail to function when andwhere all the near stations are missed-datatoo. The Mashhad synoptic station has annualrainfall data over a 50 year period from 1951to 2000 and historic rainfall data from 1893 to1940, just before World War II. This timeseries has about 15 years of missed datawhich can not be filled by usual methods. So,the techniques of geostatistics and krigingwere adopted to this long-term time series asan alternative. The data showed a poorcorrelation at every time lag, showing that,while all the semi-variogram modelsperformed nearly equal, there was a highcorrelation among each of the others. Theresults included with polynomial regressionfits to different moving average orders, nailedat some reasonable estimates for missedrainfall values.