حسین نوروزی قوشبلاغ

Limitation of surface water resources and excessive water utilization from aquifers, as well as the pollutants intrusion through the agricultural, urban and industrial activities, cause irreparable damage to groundwater. Groundwater system doesn’t respond quickly to contaminants and contaminants reaching time to the groundwater and its release into the aquifer is usually long. Remediation of contaminated groundwater and re-use often takes a lot of time and money and sometimes finding an alternative source of water is not possible. Therefore, the best and most effective solution is preventing the contamination entry to this valuable source. Determining the groundwater contamination level is one of the most important hydrogeological studies, which in this regard, the identification of susceptible areas and aquifer vulnerability assessment has great importance. Determining groundwater contamination is one of the most important hydrogeological studies. In this regard, identification of the contamination risk areas the aquifer vulnerability assessment has great importance. The Miandoab study area as a most important plain of the Urmia lake basin is one of the agricultural areas in Iran, especially in grape production, therefore, because of overuse of various chemical and animal fertilizers, it can be a nitrate-contaminated plain. Therefore, considering the importance of groundwater in the Miandoab region, which is also used for drinking, the vulnerability assessment of this plain is necessary. In the present study, the study of contaminant risk areas using DRASTIC vulnerability method has been investigated and the vulnerability map has been optimized using a genetic algorithm. 2. Materials and Methods The Miandoab Plain with an area of approximately 1150 Km2 is located in the south of Urmia Lake and is a part of the Alborz-Azarbayjan structural zone from the geological viewpoint. The average annual rainfall, based on the thirty years (1989-2018) data from Malekan and Miandoab synoptic stations is about 284 mm per year. This region, based on empirical Emberger method (1952) and using the statistics meteorology data, has a cold and semi-arid climate. Figure 1 shows the location of the study area. Fig 1. Geographic location of the study area The DRASTIC model has been used to mapping the groundwater vulnerability to pollution in many areas. Since this method is used in different places without any changes, it cannot consider the effects of pollution type and characteristics. Therefore, the method needs to be calibrated and corrected for specific aquifer and pollution. The DRASTIC model was improved with several methods such as the artificial neural network. In this study, the genetic algorithm is proposed for groundwater vulnerability.

The DRASTIC map was created by applying the weights for each parameter and integrating and overlaying the layers. According to the results of the DRASTIC model for plain, 15, 10, 17, 33 and 25 percent of the plain, respectively located in areas with very low, moderate, high and very high vulnerability. In the genetic algorithm method, the optimal weights of the parameters were obtained by maximizing the objective function. Based on the genetic algorithm method, groundwater depth, hydraulic conductivity, and unsaturated medium have the most effect on the vulnerability of groundwater in the region, respectively. Fig 2 shows the vulnerability map of Miandoab plain based on DRASTIC and optimized DRASTIC. The optimized map using the genetic algorithm method shows that about 18, 11, 28, 26 and 17 percent of the plains are located in very low, moderate, high and very vulnerable areas. According to the results of the model, the central parts of the Miandoab plain have been introduced as more vulnerable areas. A) B) Fig 2. Vulnerability map using: A) General DRASTIC, B) optimized DRASTIC using genetic algorithm

In the present study, the study of contaminating risk areas using DRASTIC vulnerability method has been investigated and the vulnerability map has been optimized using a genetic algorithm. The optimized map using the genetic algorithm method shows that about 18, 11, 28, 26 and 17 percent of the plains are located in very low, moderate, high and very vulnerable areas. According to the results of the model, the central parts of the Miandoab plain have been introduced as more vulnerable areas. Based on the results of correlation index (CI), optimized DRASTIC using genetic algorithm has the highest priority in identifying areas at high contaminate risk due to having the highest correlation coefficient (CI) with nitrate. In general, identifying the susceptible areas to contamination using appropriate methods, sources and contributing factors can be used for proper management and monitoring of groundwater.

In the present study, the study of contaminating risk areas using DRASTIC vulnerability method has been investigated and the vulnerability map has been optimized using a genetic algorithm. The optimized map using the genetic algorithm method shows that about 18, 11, 28, 26 and 17 percent of the plains are located in very low, moderate, high and very vulnerable areas. According to the results of the model, the central parts of the Miandoab plain have been introduced as more vulnerable areas. Based on the results of correlation index (CI), optimized DRASTIC using genetic algorithm has the highest priority in identifying areas at high contaminate risk due to having the highest correlation coefficient (CI) with nitrate. In general, identifying the susceptible areas to contamination using appropriate methods, sources and contributing factors can be used for proper management and monitoring of groundwater.

In the present study, the study of contaminating risk areas using DRASTIC vulnerability method has been investigated and the vulnerability map has been optimized using a genetic algorithm. The optimized map using the genetic algorithm method shows that about 18, 11, 28, 26 and 17 percent of the plains are located in very low, moderate, high and very vulnerable areas. According to the results of the model, the central parts of the Miandoab plain have been introduced as more vulnerable areas. Based on the results of correlation index (CI), optimized DRASTIC using genetic algorithm has the highest priority in identifying areas at high contaminate risk due to having the highest correlation coefficient (CI) with nitrate. In general, identifying the susceptible areas to contamination using appropriate methods, sources and contributing factors can be used for proper management and monitoring of groundwater.

Groundwater system studies to understanding its behavior, requires the exploratory drilling wells, pumping test and geophysical experiments, which can carried out with most cost. For this reason, simulation of groundwater flows by mathematical and computer models, which is an indirect method to groundwater studies, is being spent a few costs. In this research, the efficiency of artificial neural network, fuzzy logic and random forest models has been investigated in groundwater level estimation of Boukan plain. Parameters of precipitation, temperature, flow rate and water level within time period of the previous month were used as input and the water table in each period were selected as output through monthly scale (2006-2017). To evaluating the performance of models, Correlation coefficient, root mean square error and coefficient of mean absolute error were used. The results showed that the Fuzzy Logic and Random Forest models are able to estimate water levels with acceptable accuracy. In terms of accuracy, fuzzy logic model with the highest correlation coefficient (0.96), lowest root mean square error (0.068 m0) and mean absolute error (0.056 m) was recognized as a best the model in the groundwater level prediction.

Groundwater system studies to understanding its behavior, requires the exploratory drilling wells, pumping test and geophysical experiments, which can carried out with most cost. For this reason, simulation of groundwater flows by mathematical and computer models, which is an indirect method to groundwater studies, is being spent a few costs. In this research, the efficiency of artificial neural network, fuzzy logic and random forest models has been investigated in groundwater level estimation of Boukan plain. Parameters of precipitation, temperature, flow rate and water level within time period of the previous month were used as input and the water table in each period were selected as output through monthly scale (2006-2017). To evaluating the performance of models, Correlation coefficient, root mean square error and coefficient of mean absolute error were used. The results showed that the Fuzzy Logic and Random Forest models are able to estimate water levels with acceptable accuracy. In terms of accuracy, fuzzy logic model with the highest correlation coefficient (0.96), lowest root mean square error (0.068 m0) and mean absolute error (0.056 m) was recognized as a best the model in the groundwater level prediction.

Estimating aquifer hydrogeological parameters is essential for the studies or management of groundwater resources. There are several different methods such as pumping test, simulation or modeling of groundwater, geophisical modeling to estimate these parameters. Although analysis and evaluation of pumping test data is the best way to achieve this purpose, it is costly, time consuming and the gained results are from limited points. Malekan plain aquifer is one of the marginal plains of Urmia Lake which suffered more ground water declination and Salinization in last decades and it needs qualitative and quantitative management. In this study, artificial neural networks, fuzzy logic and random forest models have been used to estimate the transmission of aquifers and the performance each of these models has been investigated. Inputs of presented models included related geophysical and hydrogeological variables to transmissivity such as transverse resistivity (Rt), electric conductivity (EC), alluvium thickness (B), and hydraulic conductivity (k). Based on the results of all models, random forest model has higher accuracy and ability to predict transmissivity parameter. According to this model, electrical conductivity (EC), aquifer environment (A) and hydraulic gradient (H) parameters were identified as the most important parameters to predict the transmissivity, respectively.

Lack of proper recognition about the susceptible area to contamination in aquifers, may cause the contamination in these resources, and it may happen that these resources cannot be used. Groundwater resources in the Malikan region strongly affected by nitrate fertilizers leachates, due to the presence of grape gardens and intensive agriculture. So in this area, identifying the vulnerable area with the proper method is very important. In this study, Improved DRASTIC methods using fuzzy logic and catastrophe theory were used to vulnerability assessment of Malikan plain aquifer. For optimization, DRASTIC parameters which including seven hydrogeological parameters and the value of the vulnerability index respectively are defined as the input and output of the models. Validation of the models was performed using nitrate concentration data and correlation coefficient with the vulnerability index in the region. The results of models showed that the improved DRASTIC model using catastrophe theory with higher correlation index (CI) with nitrate concentration has provided a better result than the fuzzy logic for the aquifer vulnerability assessments. Based on catastrophe theory, the 56, 23, 21% of aquifer respectively located in low, medium and high vulnerability area, and central parts of plain was identified as a high vulnerable zones.