Cryptosporidium and Giardia Distribution in Water: Re-Evaluation

The data and the analysis applied to it in a 1996 paper on the statistical distribution of Cryptosporidium in a reservoir are re-examined with the objective of clarifying general understanding of the way in which Cryptosporidium oocysts and Giardia cysts are distributed in water. General objectives of monitoring for these protozoan parasites are summarized along with essential assumptions used in statistical analysis. The 1996 analysis is reviewed and an alternative analysis is proposed. The key distinction is that the Poisson model assumption of a homogeneous population was not appropriate as applied to the data set consisting of 52 consecutive weekly samples, leading to likely misinterpretation of the data. The critical issue is interpretation of negative (zero) results, whether as absence and hence intermittent presence or as continuous presence but below the limit of detection. The alternate analysis shows that the typically skewed annual data set can be effectively described as lognormal. The lognormal distribution of the 1996 data is compared to previously published data on Cryptosporidium and Giardia from reservoirs elsewhere with apparent similarity. The application of the Poisson model to understanding the relation between ambient concentrations of Cryptosporidium and Giardia, sample volumes, and recovery efficiency is described with its importance to effective planning of monitoring and data analysis.