5 The Geographic Distribution of Soil Lead Concentration : Description and Concerns

Pollution of inhabited geographic landscapes is often a public health concern. Ascertaining the degree and scope of such pollution is a difficult task and frequently involves analysis of soil samples. This article describes ways to statistically analyze the geographic distribution of soil samples to better understand polluted landscapes, in part to determine whether more samples are needed; special reference is made to lead. The soundness of this description rests on the identically distributed assumption of statistical analysis as well as attributes of soil samples supporting evaluations of constant variance; methods are outlined for assessing both. Empirical findings are summarized regarding the distribution of soil lead concentration in the environment extracted from three types of geographic landscapes: an urban area, superfund sites, and a flood plain. These regions are described in terms of the statistical frequency distribution of soil lead concentration and the nature and degree of spatial autocorrelation latent in the geographic distribution of soil lead concentration. The importance of knowing such results is demonstrated by examining three concerns: 1) why a spatial analysis of soil samples is worth undertaking; 2) the cost-effectiveness of spatial sampling; and 3) the ability to predict soil lead contamination at unsampled locations based on data from sampled locations.