A comparison of geocomputational models for validating geospatial distribution of water quality index

Water resources management is a vital component of maintaining good water quality. The surveillance program is one of the essential tools for regularly monitoring water quality. But, it's a very expensive and long term process. To date, several tools and techniques have been developed for assessing water quality. The water quality index (WQI) is one of them. Recently, this technique has been widely used for assessing water quality. Its application has increased rapidly due to its allows converting a vast amount of water quality information into a unitless numerical expression using simple mathematical functions. For the purposes of predicting WQIs at each grid point, various geospatial techniques were used. The aim of this research was to identify the best geospatial predictive model for the spatial distribution of WQIs for coastal water quality. In this research, eight widely used interpolation techniques were utilized for the interpolation of WQIs: local polynomial interpolation (LPI), global polynomial interpolation (GPI), inverse distance weighted interpolation (IDW), radial basis function (RBF), simple kriging (SK), universal kringing (UK), disjunctive kriging (DK), and empirical Bayesian kriging (EBK). This study has been carried out in Cork Harbour, Ireland, as a case study for assessing coastal water quality using the weighted quadratic mean (WQM) WQI model. According to the cross-validation results, the UK (RMSE = 6.0, MSE = 0.0, MAE = 4.3, and R2 = 0.8) and EBK (RMSE = 6.2, MSE = 0.0, MAE = 4.6, and R2 = 0.78) methods performed excellently in predicting WQIs at each grid point in Cork Harbour, respectively. The findings of this study reveal that the EBK geospatial computational model could be effective in predicting WQIs in Harbour.