1. Introduction
Due to rapid urbanization, human activities have had a significant impact on the ecological environment [
1]. Owing to the huge discharge of municipal wastewater and urban drainage into river basins, the effect is more pronounced in the water quality in these areas. This in turn affects the aquatic lives and agriculture, and hence the humans who reside in these areas. The challenge for any country is to undertake sustainable development without causing damage to the natural environment, e.g., avoiding crucial issues leading to the rapid deterioration and degradation of the water quality in the water supply intake points [
2,
3]. Further, cost-effective methods are more convenient for developing countries to protect their natural resources. Rapid population growth as a result of urbanization is commonly observed in the flat regions of many countries, hence the flat beds of river basins are more susceptible to being affected.
The high average annual range in biochemical oxygen demand (BOD), total coliform (TC) and low dissolved oxygen (DO) levels reported in the rivers of some South Asian countries are mainly due to the huge discharge of municipal wastewater and urban drainage into river basins [
4]. Further, many studies have found untreated sewage to be the most critical issue with respect to water contamination, because 40% of the global population do not have adequate sanitary facilities [
5,
6]. Changes to land use, land cover or river basins in watershed areas, such as high scale agricultural activities, unplanned infrastructure developments, and sand mining, change the water quality and water balance due to human activities. Furthermore, the contents of untreated disposal or spatial disposal items such as organic materials, salinization, nutrients, suspended sediment, enteric virus, and oil and grease are directly affecting the water quality in river basins [
7,
8]. Furthermore, the aquatic chemical composition is possibly influenced by human activities [
6]. This may be caused by improperly planned urbanization or uncontrolled development [
9]. This has been made evident by Hongmei Bu et al.’s work conducted on the Jinshui River in China, where they reported that the population had a significant impact on most of the water quality parameters [
10]. Thereby, they identified a population capacity for the river basin, and estimated the possible increase of population. The possibility to control the impact of anthropogenic activities by analyzing the impact levels of urban, suburban and rural areas on the river basin was described in the research done on a river in Shanghai, China [
1]. A fecal coliform analysis of a long-term, stream water-quality monitoring system in the city of Atlanta shows the effect of urbanization, whereby contamination exceeded Georgia’s water quality standard for all usage levels [
11]. A study of the spatial correlation between urbanization and water quality parameters based on a regional perspective showed that human activities are positively correlated with the degradation of water quality in the river. Further, urban population density is used to evaluate the impact of urbanization [
12].
The Kelani River in Sri Lanka is rich with biodiversity and many natural resources, and plays an important role in the sustainable development of the country. More than 25% of the Sri Lankan population take benefits from the river [
13]. Unfortunately, it is considered to be one of the most polluted rivers in Sri Lanka [
14]. Further, unplanned anthropogenic activities (towns) together with, industrial and agricultural activities have been highlighted as major threats [
15]. The lower and middle regions of the Kelani River, which consist of flatbed areas, have faced a threat from human activities due to urbanization. In the Ma Oya tributary in the lower region of the Kelani River, the exceeded standards were BOD (60%) and DO (80%). Further, standard levels of COD, BOD and DO have been exceeded in the Raggahawatte sub-stream in the middle reach of the river basin, which includes flatbeds, submontane and dense forests [
13]. Sri Lanka, as a developing country with an increasing population, needs to show adequate concern for the management of existing water resources while achieving sustainable economic development; the availability of drinking water is a primary requirement for the citizens. Considering this situation, the responsible authorities in Sri Lanka have already proposed many sustainable utilization approaches for the Kelani River [
16]. In this study, we propose an approach that can be used for analyzing the impact of population growth on the water quality of natural water bodies. Further, under the assumption that the existing water quality management approaches and the human impact remains the same, with no implementation of water treatment plants or no changes in the human activities, the study aimed to obtain the optimum population range that can be accommodated within the carrying capacity of the watershed. The spatial context for the case study in this research comprised the lower and middle regions of the Kelani River.
Complexity and uncertainty are leading factors for concern when analyzing spreading sources of pollution such as those due to anthropogenic activities on natural water ecosystems [
17]. Classification models can be used to analyze various influential factors in natural environmental processes and in this case, we analyze the impact of human activities using the water quality classification standards (WQCS) for a river basin [
18]. Among various classification models, such as the Bayesian network (BN), artificial neural networks (ANNs), decision trees and support vector machines, we propose to use the BN model to find the optimum population ranges that can be carried by the natural environment in a watershed [
19]. The BN is a network with nodes representing probabilistic variables and links representing probabilistic dependencies. The conditional probability distribution given to each variable represents the influence of the parent nodes. The BN was employed to derive the sustainable population ranges for specific water uses such as drinking, bathing and fishing through analyzing the probabilistic influence between the water quality parameters and the population.
The following two steps were employed in this research to achieve the goal of our research: In the first step, the influence of human activities on the water quality in the five watersheds of the Kelani River was evaluated considering three urbanization levels categorized based on the population density of watersheds. Further, the population index (PI) was calculated for each urbanization category and comparisons were made among the PI values of each year to evaluate the human impact. Then we evaluated the correlation between the water quality parameters, such as Total Coliform (TC), NO3−, DO, chemical oxygen demand (COD), BOD, and the population in five watersheds. The results of the first step were used to find the most correlated parameters with the population. The classification model was developed as the second step. TC, DO, and BOD were used to qualitatively define the population ranges using the BN classification model. Upon developing the BN, the probability distribution table of the population node in the BN were analyzed to quantitatively derive the sustainable population ranges to maintain a certain level of water quality.