The Spatial Analysis of Acid Precipitation Data

Abstract

Kriging, an interpolation procedure that minimizes interpolation error and gives an accurate estimate of that error, is shown to be an appropriate objective analysis procedure for the study of spatial variability and structure in acid precipitation data. Variograms for H+, SO4, NO3, and NH4 are presented. They are shown to be clearly distance dependent in all cases, increasing with increasing distance between stations. The functional form of the increase, however, was not consistent. Typical isopleths with their corresponding one sigma confidence limits are computed. The spatial extent of these confidence limits is considerable, illustrating the difficulty of fine structure analysis of acid precipitation data with the existing network of sampling sites. Kriging is also a useful procedure for sampling network design and improvement. Illustrations are presented which show how increases in network density quantitatively improve spatial analysis by decreasing interpolation error.