Modeling the corrosivity and scaling of drinking water distributed in Morocco using multiple linear regression

The present study concentrates on the modeling of water corrosivity and calcocarbonic balance of the drinking water distribution in Fez, Morocco.Utilising a range of linear regression techniques by Statistica 12.The study analyses twelve physicochemical parameters to formulate predictive models for the indices of water corrosivity (Larson and Leroy).A total of 156 water samples were collected from six different points over the duration of six months, and the samples analysed show that the means of the parameters are CAT (4.890.39)meq/l, CATs (4.600.37)meq/l, TH (5.860.71)meq/l, dissolved oxygen (8.530.22)mg/l, chloride (1.130.57)meq/l, oxidability (1.560.71)mg(O 2 /l), residual chlorine (0.590.16) meq/l, LSI (0.020.15), conductivity (0.830.17) mS/ cm, temperature (17.251.02)C, turbidity (0.580.17)NTU and sulfate (0.690.35) meq/l.The means value of the water indices are Larson index (0.180.16) and Leroy index (0.850.12).The models achieved good predictive accuracy, with R values of 92.02% and 98.76%, and a low standard error 0.04 and 0.01 for the Larson and Leroy indices, respectively.The findings underscore the inverse correlation between corrosivity and calco-carbonic balance, emphasizing the significance of maintaining water quality within acceptable standards to prevent corrosion and scaling in water distribution systems.The results of the study provide valuable insights for the management of water quality, particularly in regions exhibiting similar environmental conditions.The aim of this study is to identify the parameters responsible for this phenomena.