Human activities pose a significant threat to the water quality of rivers when pollution exceeds the threshold limit. Urban activities in particular are highlighted as one of the major causes of contamination in surface water bodies in Asian countries. Evaluation of sustainable human population capacities in river watersheds is necessary to maintain better freshwater ecosystems in a country while achieving its development goals as a nation. We evaluated the correlation between the growth rate of the population in a watershed area and water quality parameters of a river ecosystem. The Kelani River in Sri Lanka was selected for the study. The highest correlation coefficients of 0.7, 0.69, 0.69 (p < 0.01) corresponding to biochemical oxygen demand (BOD), dissolved oxygen (DO) and total coliform (TC) were obtained with the population in watersheds of the Kelani river in Sri Lanka. Thus, we propose a quantitative approach to estimating the population capacity of watersheds based on water quality classification standards (WQCS), employing the Bayesian network (BN) classification model. The optimum population ranges were obtained from the probability distribution table of the population node in the BN. The results showed that the population density should be approximately less than 2375 to keep the water quality in the watershed for bathing and drinking purposes and approximately less than 2672 for fish and other aquatic organisms. This research will offer a means that can used to understand the impact of population on water quality in river basins and confer direct influence on natural water bodies.

Abstract:Design of water quality monitoring network in a river basin is most necessary when the cost concerns of the process is considered. The most of the methods have followed a set of objectives in the last few decades to find optimized selection of sampling locations. One of the main contributions of this paper is to design an optimized selection of sampling sites network considering the newly identified monitoring sites rather than the points which are available in the current network that exists. This study used four criteria including one urbanization factor which is in the development pressure index (DPI) rather than the factors environmental pressure index (EPI) to evaluate the objectives. Multi objective analysis method and genetic algorithm were applied for find optimal network and three constraints are used to obtain a practical solution. The other main purpose of this study is to compare the efficiency between method of genetic algorithm and the brute-force approach by considering the computation time of fitness functions. Further, the fitness function was defined using the linear combination of the criteria. Both proposed that the optimal water quality monitoring networks are reasonably sufficient to enhance the existing network in Kelani River. The genetic algorithm has a high performance to find the fitness functions even though all possible combinations of monitoring sites are identified by brute-force approach.

No abstract