Design and Analysis for Assessment of Water Quality

Abstract

Quality of surface waters is a rlative concept expressed in terms of the variables (physical, chemical and biological) that are measured to meeet objectives of the study designed to assess it. The comparisons that are made between variables are important rather than their absolute values. Studies on water quality usually yield diverse data types and the study design should allow for the association of these data. Any variables that are measured to assess water quality (physical, chemical and biological) will have some degree of uncertainty associated with them. To make comparisons of water quality valid, the degree of uncertainty (precision) must be estimated and environmental variability must be accounted for in sampling. If an insufficient number of replicated collections is made to account for environmental variability, a null hypothesis of no effect may be accepted when it could have been refuted that more replicateds been collected. The more powerful parametric statistical methods used to associate diverse data types will require that the data have similar levels of precision and statistical distributions. The magnitude of the differences in the factores being assessed will determine the levels of precision required, which will, in turn, determine the number of replicate collections to be made, and the number of replicate collections will largely determine the cost of the studey. It is argued that more emphasis could be placed on chemical measures of water quality when the magnitude of the difference is larege and more on biological measures when differences are small.