ABSTRACT
This chapter develops the concept of the random product process and discusses the distributions to which they give rise in greater detail. A lognormal process is one in which the random variable of interest results from the product of many independent random variables multiplied together. The lognormal probability model describes a distribution in which the logarithm of the random variable of interest is normally distributed. The chapter discusses the parameters of the lognormal distribution and the relationship among the parameters of the lognormal probability model. It develops a statistical theory of rollback for making predictions about the distribution of the concentrations of a pollutant after its source has been controlled by multiplying the source strength by a fixed proportion. To develop this theory, the chapter considers the general properties of two independent random variables multiplied together, examining how a multiplicative change in one random variable affects the overall product.
