A refined integrated forecasting procedure is considered based on integration of two alternative prediction schemes (the method for optimal extrapolation and the modified method of clustering arguments underline these schemes) and developed to solve the problems of spatial extrapolation of mesometeorological fields into the territory uncovered with aerological data. The quality of this forecasting procedure is numerically estimated based on the data of many-year observations at the radiosonde network being typical of the mesometeorological site.