Assessment of Probability Distribution of Dissolved Oxygen Deficit

Water quality modeling and prediction in a stream environment is complicated by the occurrence of a number of random processes. Due to the complexity of water quality model, the analytical derivation of the exact probability distribution of the dissolved oxygen (DO) deficit is difficult. This paper proposes an approximate but practical method by using first‐order analysis of uncertainty in estimating the statistical moments of the DO deficit. The statistical moments estimated were then used in an appropriate probability distribution for the DO deficit concentration. Practical issues involved are: (1) Which is the appropriate probability distribution for the DO deficit concentration? and (2) is this appropriate distribution function sensitive to the distribution of water quality parameters? This paper examines a number of commonly used probability distributions for their appropriateness in describing the random characteristics of the DO deficit concentration under various conditions. Of the distributions considered in the investigation, lognormal distribution proved to be the best model.