Spatio-temporal trends and change factors of groundwater quality in an arid area with peat rich aquifers: Emergence of water environmental problems in Tanta District, Egypt

An integrated spatio-temporal statistical, P-spline modeling, Mann-Kendall and Thiel-Sen trend detection, and factorial analyses techniques were performed on 438 chemical and 274 microbiological data collected from twenty drinking water supply wells over four years (2010-2013) in Tanta District (Egypt). The prime objective was to characterize the spatio-temporal quality trends using indicators of turbidity, pH, TDS, Cl−, SO42−, Na+, Total Alkalinity, hardness (Total, Mg, and Ca), Fe2+, Mn2+, Al3+, Cu2+, Zn2+, F−, NH4+, NO2−, NO3−, PO43−, SiO2, bacterial, and algal contents. Factorial analysis was applied to identify the significant factors loading the quality variation. Out of the 20 wells, notable upward trends were significant (>95% level) for the total hardness (30%), total alkalinity (20%), TDS (15%), Fe2+ (15%), Mn2+ (15%), NO2− (10%), and 10% for the NH4+, PO43−, and SiO2. Attenuation rates (mg/l/year) were higher in NH4+ (av. 0.023) than Mn2+ (av. 0.013) and Fe+2 (av. 0.010), and remarkable average rates were 6.77 (TDS), 3.27 (total alkalinity), 2.12 (total hardness), 0.79 (SiO2), 0.011 (PO43−), and 0.002 (NO2−) in decreasing order. High precision of the trend estimate was confirmed for the Mn2+, NH4+, and Fe2+ data. Five factors were found to explain 78.8% of the total variance of the quality variables and in particular, a significant load of hardness parameters, Total alkalinity, TDS, Mn2+, NH4+, and Fe2+ in decreasing order were identified. The spatio-temporal variation in pollutants originated from organic matter degradation, either naturally from the aquifer peaty sediments or anthropogenic due to improper well head protection in the urban centers or from the agricultural drains in low relief areas. Our findings have important societal implications regarding the management of the limited and valuable water resources in arid and semi-arid lands. The adopted methodologies could be readily applied to similar alluvial aquifers elsewhere.