Estimation of Nitrate Pollution in Groundwater using Fuzzy Rule based System

Contamination of groundwater by nitrate is a serious problem in many areas. High concentration of nitrate in water is harmful for human beings. After nitrate is converted to nitrite within the gastrointestinal tract, nitrite reacts with hemoglobin in the blood and impairs oxygen transport to tissues. Accurate prediction of nitrate concentration is imperative for any strategy for remediation of nitrate concentration in groundwater. However, there is imprecision and uncertainty in parameters responsible for nitrate concentration in groundwater. Using the actual quantity of nitrogen-based fertilizers, extent of the applied area and recharge in groundwater are some of the parameters, which are not known with certainty. For the complex pollution scenario like nitrate pollution in groundwater where few numerical data exist and where only ambiguous or imprecise information may be available, fuzzy reasoning provides a way to understand the system behaviour by allowing one to interpolate approximately between observed input and output situations. The present study utilised average fertilizer applied and amount of groundwater recharge as inputs to predict nitrate pollution in groundwater. In this study, three fuzzy models are developed to evaluate the effect of the number of input variables and the number of partitions on the model performance. Results demonstrate that model efficiency can be improved by increasing the number of input parameters and fuzzy partitions of input and output domain. The output of the fuzzy system is also compared with the regression model. It is obvious from the comparison that a developed FIS (Fuzzy Inference System) model gave better performance than the model developed from regression analysis.