Modeling of multiple regression and multiple linear regressions for prediction of groundwater quality (case study: north of Shiraz)

Abstract

The aim of the study area was investigation groundwater quality and determination of relationship between effective parameters in groundwater quality in north of Fars province, southeast Iran. For determination of groundwater quality, parameters of calcium (Ca), pH, potassium (k), chlorine (Cl), magnesium (Mg), sodium (Na), electrical conductivity, sulfate (So4), total dissolved solids (TDS) were used. Using inverse distance weighting spatial distribution of each parameters was determined. Also using multiple linear regressions (MLR) relationship between each of parameters was determined. It was found that in the study area, all of the parameters expect Hco3 was low value in the north of the study area (high groundwater quality). While the maximum value of parameters was located in south of the study area (low groundwater quality). So the north of the study area was better quality than the south of the study area. The relationship between parameters by MLR showed that Cl and TDS had the strong positive correlation (r = 0.97). Also calcium and magnesium showed strong positive correlation (r = 0.68) and there was strong positive significant correlation between Ca, Na and Mg with So4 (about r = 0.6).