Seasonal and Spatial Variations of Cadmium Concentration in Groundwater of the Hamedan-Bahar Plain using Geostatistics and Deterministic Methods

In this research, using different deterministic and geostatistical methods including kriging, inverse distance, local polynomial and radial basis functions, the seasonal and spatial distributions of cadmium (Cd) amounts in groundwater were evaluated in the Hamedan-Bahar plain. To do this groundwater of 30 wells in the plain were sampled in four seasons i.e. summer and autumn 2002, winter and spring 2003. To analyze the seasonal and spatial elements of the quality of ground water, the geostatistical software ArcGIS version 10 was used. The performance criteria for evaluating the used methods were Mean Absolute Error (MAE), Root Mean Square Error (RMSE) and General Standard Deviation (GSD) using the cross validation method. The results showed that local polynomial, kriging (disjunctive and simple), kriging method (probability and indicator) and kriging (probability and indicator) were the best methods to estimate Cd in summer, autumn, winter and spring, respectively. Also, kriging (probability and indicator) for summer and autumn as well as Radial Basis Functions method for winter and spring were illustrated as unsuitable methods for estimating Cd. The interpreted results showed, the best approach adopted for a period of time could not be utilized for other seasons. Also cadmium concentration was high in the northern parts of the plain (summer season) and in the Northeast parts (spring season). This concentration was up to 2.0 micrograms per liter in autumn season, which was covering all the plain. The trend of increase in cadmium concentrations was observed from eastern parts to western parts of the study area in the winter season.