Different environmental factors particularly precipitation and temperature influence the spatiotemporal variability of E. coli and water physicochemical parameters. Influence of rainfall on water bodies is a major concern particularly in a tropical country, where seasonal variations of surface water quality are primarily dominated by precipitation (Vermeulen and Hofstra 2013). A rainfall event can greatly influence the physicochemical characteristics of a surface water body by altering its hydrology and bringing contaminants into the water source via surface runoff (Ling et al. 2017). In the study area, rainy season and summer coincide (Islam et al. 2017), which is reflected by the observed significant positive correlation (R = 0.55) between temperature and precipitation.
The results showed elevated turbidity after rainfall events (correlation between turbidity and precipitation is 0.44), which was probably due to the resuspension of deposited sediment under high flow condition. This is supported by previous studies (e.g. Ling et al. 2017; Munoz-Nava et al. 2018). A significant positive correlation was found between COD and rainfall data, which shows the effects of rainfall on increasing the pollution load in the rivers. This is also in agreement with the study of Momou et al. (2017).
The weak correlation of E. coli concentration with pH, DO and BOD could be due to the relatively small spatiotemporal variation of these parameters, which allowed other parameters to have a stronger influence on E. coli concentrations. Weak influence of these parameters was also reported in other studies (e.g. Stocker et al. 2016; Momou et al. 2017).
Higher E. coli concentrations were observed during wet period compared to dry period, which agrees with previous studies (Walters et al. 2011; Abia et al. 2015; Aragonés et al. 2016). Precipitation was correlated positively with E. coli, because surface water can be contaminated with urban runoff, combined sewer overflow and resuspension from deposited sediments. Similar reasons regarding the positive correlations were also pointed out in previous studies (e.g. Ibekwe et al. 2011; Martinez et al. 2014; Aragonés et al. 2016; Islam et al. 2017). Water temperature and salinity showed negative correlation with E. coli concentrations. This is consistent with the results from previous studies (Adingra et al. 2012; Dastager 2015; Aragonés et al. 2016). Water temperature facilitates bacterial die-off, could be the reason of negative correlation. In southwest Bangladesh during monsoon (July–September), precipitation increases and water salinity reduce to below 1 ppt (Islam et al. 2017). The negative correlation with salinity was not due to the salinity induced die-off of E. coli, it might be due to the coincidence of low salinity with increased monsoon precipitation.
The regression model explains large part of the variation in E. coli concentration (R2 = 0.42) by taking different environmental parameters into account. The model results agree well with other studies that included similar parameters with this study, e.g. R2 of 0.46 in our previous study in the Betna river in Bangladesh (Islam et al. 2017); R2 of 0.20–0.41 in 23 sea beaches in Chicago (Whitman and Nevers 2008); R2 of 0.49–0.68 in Ribble drainage basin in the UK (Kay et al. 2005); and R2 of 0.49 in the Rhine, Meuse and Drentse Aa (Vermeulen and Hofstra 2013). The regression results indicate that water temperature and precipitation is the most important factors that can influence river faecal contamination greatly.
The measured mean E. coli concentrations (3.4–4.7 log cfu/100 mL) of this study are higher than our previous study (Islam et al., 2017) in Bangladesh (2.9–3.4) and the study of Liu et al. (2009) in China (1.8–3.4); but comparable with the study of Widmer et al. (2013) in Southeast Asia (2.8–4.3) and Adingra et al. (2012) in Côte d’Ivoire (2.55–3.47). All the E. coli samples violated the USEPA specified bathing water quality standards. The high E. coli concentrations and violation of standard are not strange for the study area, as a large volume of sewages from a densely populated urban area enter directly to the rivers without any treatment. These results indicate potential health risks with the use of the river water. A quantitative microbial health risk assessment is very essential, that can be a basis for reducing disease burden from microbial contamination.
The WQI in the river sampling sites after rain showed little improvement mainly due to dilution of E. coli concentrations and an increase in the DO level. The rainfall affected positively the river water quality and increased the index slightly compared to the dry period. However, rainfall events did not improve the rivers water quality substantially, which indicates that influence of seasonal changes to the rivers water quality is little. The overall WQI indicates poor quality of river water. When a water quality is poor, the water is unsuitable for use in domestic and recreational purposes, and has limited potential for aquaculture and irrigation (Munoz-Nava et al. 2018). The river water is currently being used for domestic purposes (e.g. washing cloths and utensils), bathing, fishing, discharging of untreated sewage and open defecation. All these practices are causing serious threat to the biodiversity of the river and public health by altering the physicochemical properties and increasing microbial concentrations of the river system.
The observed comparatively higher E. coli concentrations in the S2 and S3 sampling sites located near the populous area of the Khulna city and receive high sewage discharge through sewer drains; and lower concentrations at S6 located downstream and receive less sewer discharge from the city indicate that untreated wastewater from the sewer drains was the major contamination source in the rivers. Introduction of adequate level of wastewater treatment could result in considerable improvement of the river microbial water quality. This is also supported by other studies (Islam et al. 2018b; Vijay et al. 2016; Liu et al. 2015; Ouattara et al. 2013). Therefore, establishment of wastewater treatment plants to treat wastewater before it is discharged into the rivers are urgently needed. The current level of faecal contamination in the rivers is already very high. This could be further deteriorated by the increasing population growth, urbanization, and intensification of agriculture and aquaculture. As a result, uses of the highly contaminated river water will increase public health risks. Although establishment of effluent treatment plants has become mandatory for industries, there is no such strict regulation and initiative to establish domestic wastewater treatment plants in Bangladesh (Islam et al. 2018a). The implementation of wastewater treatment is a must to achieve the UN Sustainable Development Goal (SDG) 6 target 3 ‘to halve the proportion of untreated wastewater’ (UN-Water 2017; Islam et al. 2018a). At least secondary level of wastewater treatment should be established immediately in the highly populated city area (Islam et al. 2018a). Otherwise it will be impossible for Bangladesh to treat the increasing amount of waste and achieve SDG 6.3 by 2030.