Improved Predictions of Water Quality Values and Design of Sampling Strategies Based on Entropy Theory

Two recent developments are presented, which can assist in: (i) the prediction of water quality values at certain types of discontinued water quality monitoring stations; and (ii) determination of an optimum sampling frequency which is the most effective balance between the costs of monitoring and the information obtained from the data so collected. The information entropy concepts of information loss and transferred information are refined as measures for evaluation of the “informativeness” and “representativeness” of temporal sampling frequencies. The method produces more accurate predictions of water quality more frequently than traditional regression techniques.