Evaluation of groundwater quality for different uses, case study: Delfan ‎plain

Reduction in groundwater volume in one hand, and resources and human activities growth on the other hand has let to water quality depletion. Therefore study and assessing water quality for various applications including agriculture and irrigation, drinking and industry is essential for the proper and efficient management of water resources and economic growth that enable sustainable economic and social development. 

In this study, the quality of groundwater for different uses was classified in Delfan plain using calcium, magnesium, chlorine, sodium, sulfate, acidity, electrical conductivity, sodium absorption ratios, total hardness and total salts of 29 deep wells for 5 years (2011-2016), and vulnerable areas in terms of drip systems implementation was identified. Delfan Plain is located at 47ᵒ 27ʹ to 48ᵒ 18ʹE longitude and 33ᵒ 50ʹ to 34ᵒ 18ʹ N latitude. This study area is 1892 km2 in average elevation of 1856 m above sea level in northwest of Lorestan province. About 1729 km2 of this area belongs to the altitudes and the rest to the alluvial zones, indicating that the area is mountainous. The aquifer of this plain has been investigated in the present study with an area of 239 km2.
In this research, different interpolation methods were used for water quality zoning. The methods used include Kriging, Inverse Distance Weighting, Global Polynomial Interpolation, Local Polynomial Interpolation and Radial Basis Function. Kriging: An estimation method based on the weighted moving average logic and ‎known as the best unbiased linear estimator‏ .‏IDW: This method is based on this fact that points that are closer have greater values than those are farther. In this method, for each measurement point, a weight based on the ‎distance between that point and the position of the unknown point is ‎considered. These weights are then controlled by the weighting power so that ‎the larger weights reduce the effect of points farther from the estimated ‎point and the smaller weights are distributed evenly between adjacent ‎points. LPI: This method considers a short range of ‎variations in the input data and is sensitive to neighborhood intervals in ‎the shared window. As the window moves and the surface values at the center of ‎each window at each point are estimated by fitting a polygon. This method has higher ‎flexibility than the local estimator method. These two methods require no ‎assumptions for the data‏.‏ RBF: The radial function is a function as Φj ‎‎(X) = Φ (X-Xj) which is dependent on the distance between x = Rd and the fixed ‎point Rd ε Xj. In this function, Φ is a continuous and dependent function of ‎each RdεΩ subunit. R represents the Euclidean distance between any pair of ‎points in the set O. This method has 5 kernel functions: Completely ‎Regularized Spline, Spline with Tension, Multiquadratic, Inverse ‎Multiquadratic and Thin Plate Spline. It is necessary to evaluate the accuracy of each method or to select the ‎appropriate parameter. There are several methods in this regard, the most ‎important one is cross-validation. In this method, comparison between observed and estimated values is carried out using specific methods.In this research, to classify water quality, Scholer, Wilcox, and Langlier saturation indices were used.Scholler diagram is one of the most importantclassifications for drinking water ‎quality. Scholler diagram is a semi-logarithmic diagram showing the concentration of ‎the major ions in meq/liter. In this graph, water is divided by the chemical parameters ‎of sodium, chlorine, sulfate, calcium, magnesium and pH for drinking purposes‏.‏ To determine the quality of agricultural water, Wilcox classification is used which is one of ‎the most important classifications in this field. In this classification, ‎agricultural water quality according to EC and SAR is categorized into four groups of good ‎quality, acceptable, inappropriate and bad‏.‏ The Langlier saturation index is used to determine the quality of water for industrial use. It should be noted that this indicator is the most important indicator of water quality ‎assessment to the implementation of drip irrigation systems because it is used in low water ‎speed systems‏.‏ 

Results and discussion

‎According to RMSE index for different interpolation methods, ordinary kriging is a suitable method for zoning SAR, Na, Cl, SO4 and LSI index. RBF method is suitable for zoning. EC, TDS, and Mg, IDW1 method is for zoning TH and Ca, and IDW2 is suitable methods for pH zoning. The combination of qualitative layers of lowland aquifer for agricultural purposes ‎showed that 4.8% of the plain is good and 82.16% of plain is in poor, while bad quality points were not found. In terms of drinking, 48.14 percent of the aquifer ‎has good quality and 9.6 percent is has poor quality. The eastern part of the aquifer generally has ‎ the best quality for drinking and agriculture purposes. According to the zoning map, this index has the highest amount of LSI in the north and ‎east areas, so there are limitations for implementing drip systems in these areas.

Conclusion‎

Assessment of qualitative maps showed that there are some limitations in some parts of the ‎plain in terms of groundwater quality for different uses. These limitations include restrictions on ‎residential development, the use of water-sensitive crops, and the implementation of ‎drip irrigation systems during corrective measures such as using acid and temperature changes. ‎Hence, operating and maintaining of this system is faces to problem. In areas with poor ‎water quality and high LSI index, it is recommended to install treatment plants to improve the ‎quality of drinking water as well as to use a proportional crop pattern compatible to available water quality such as ‎the use of salinity tolerant crops, and not use the clog-sensitive dripper.