Pragmatic Approach for Establishing Predictive Computational Framework Incorporating Public Health and Water Quality

Despite major adverse health impacts, the use, and therefore exposure to the general public, of the heavy metals persevere attributed to their widespread application domains, especially in developing countries. Among all other exposure modes and sites, air and water intakes prove to be the most vulnerable (USEPA 2017). To assess this exposure-response relationship, a numerical model framework is developed, calibrated and evaluated to quantify and predict the pertinent health risks. An ensemble of ground data for the sub-region scale of North Indian regime with center as Sonipat city (mean spatial coordinates: 28.9931° N, 77.0151° E) is used to derive the hybrid model based on the downscaled scenario-based outputs including the effects of anthropogenic factors. The water quality data from the Central Ground Water Board (CGWB) are coupled with the data of 2870 patients from the Medical Council of India (MCI) for 10-year time span. The model is evaluated against the local prevalence study conducted on 25 hospitals comprising of government-, government-aided- and private hospitals of the city.

The patients affected with diseases caused by heavy metals are about 44 % of the total patients recorded and constituted 8% of the total population, thus making it a concern of paramount importance. Multi-objective multi-parameter MCMC methods using Metropolis-Hastings Algorithm with Gaussian proposals are used for initial sampling and optimization of the observations (no. of patients). The samples converged with an RMSE 0.18 (±0.069), NSE 0.78 (±0.152), and correlation coefficient 0.906 (±0.224). Prediction accuracy (89 ±2.5%) of the model for the constituent parameters fortifies the rational applicability of the model. The sensitivity analysis revealed that the water intake is the most sensitive amongst all other exposure modes. Regression analysis with R²=0.841 construed strong positive linear correlation between two variables viz. the number of patients and polluted water (with heavy metals). By real-time monitoring and forecasting, this research will serve and benefit local policymakers and decision-making process, particularly regarding water resources vulnerability and future availability with significant economic consequences. Tailoring the model form sub-regional to regional scale is well underway.