Two indices, P1 and I, employing ranked data and following the beta distribution, are described. These indices, which are true statistical measures of water quality, can be used with any set of parameters. Further, these indices correlate highly with biological and subjective-engineering assessments of water quality.
Three indices, employing raw data and based on the chi-square distribution, are described. Two of these, B1 and B2, do not have sufficient discrimination to serve as general measures of water quality, although they have use in specific circumstances. The index C has been found to be a measure of excessive variability (at a station), and can be used to quantitatively compare the (data) variability among stations.
The five indices described here are evaluated against a set of operational criteria which a mathematically sound index should meet.