Probabilistic modelling of water quality in the Ramganga River, India, informed by sparce observational data

Impaired water quality continues to be a serious problem in surface waters worldwide. Despite extensive regulatory water quality monitoring implemented by the Government of India over the past two decades, the spatial and temporal resolution of water quality observations, the range of monitored contaminants and data related to characterisation of point source effluents are still limited. In addition, discharge data for trans-boundary rivers is considered sensitive information and is not publicly available. Hence, quantifying, and mitigating pollutant loads and planning effective mitigation strategies are hindered by data paucity and there is an urgent need for the development of decision support tools (DST) that can account for these uncertainties.

In this study, we tested the application of a probabilistic DST based on Bayesian Belief Networks, to evaluate pollution risk from nutrients (phosphate, nitrate, ammonia), sediments and heavy metals (Cd, Cr, Cu, Pb, Zn) in the Ramganga river basin (30,839 km²), the first major tributary of the Ganga in the state of Uttar Pradesh, India, and is understood to be a significant source of pollution into the Ganga River, contributed from a range of industries, domestic sources and intensive farming practices. Bayesian belief networks are graphical causal models that enable to integrate observational data (both spatial and temporal) with data from literature and expert knowledge within a probabilistic framework, whilst accounting for uncertainty.

The objectives of this study were to 1) develop a parsimonious conceptual model of the system that allows harnessing diverse but limited data, 2) evaluate the important components of the system to inform further data collection and management strategies, and 3) simulate plausible management scenarios. We simulated the impacts of point source management interventions on pollution risk, including provision of sufficient municipal sewage treatment plant (STP) capacity, enhanced STP treatment levels and sufficient industrial wastewater effluent treatment capacity. We found a clear effect of enhanced STP interventions on improved regulatory standard compliance for nitrate (from 92% to 95%) and phosphate (from 33% to 41%). However, the effect of interventions on heavy metal pollution risk was not clear, due to considerable uncertainties related to the lack of reliable discharge data and the characterisation of industrial effluent quality. The parsimonious DST helped to collate the available understanding related to water quality impacts from multiple pollutants in the Ramganga river basin, while sensitivity analysis highlighted critical areas for further data collection.