Evaluation of Ground Water Level using Hybrid Models(Case Study:Khorramabad Plain)

Document Type : Research/Original/Regular Article

Authors

1 Assistant Professor, Department of Soil Conservation and Watershed Management, Lorestan Province Agriculture and Natural Resources Research and Education Center, Agricultural Research, Education and Extension Organization, Khorramabad, Iran

2 PhD in Water Sciences and Engineering, Department of Soil Conservation and Watershed Management, Lorestan Province Agriculture and Natural Resources Research and Education Center, Agricultural Research, Education and Extension Organization, Khorramabad, Iran

Abstract

Abstract

Introduction

Groundwater resources are susceptible to both climate changes through (A) direct interaction with surface water sources such as rivers and lakes and (B) indirect interaction via the feeding process. Climate change indirectly affects the discharge and storage of groundwater by changing the nutritional conditions induced by rainfall and runoff; therefore, identifying and analyzing effective parameters such as climatic parameters can greatly help predict serious hazards threatening groundwater resources such as subsidence and drought . Moreover, considering that the relationship between climatic parameters and groundwater resources is complex and non-linear, the application of artificial intelligence models including modern hybrid models is a good solution to solving these problems.

Given the non-parametric nature of these models, they are independent of the concept of prediction and the relationship between input variables and output data . As one classic feature of artificial intelligence models, they are capable of performing stochastic analysis of dynamics, patterns, and features associated with input variables used to simulate groundwater surface variables. Therefore, they are more feasible than other conceptual and statistical methods (such as experimental approaches and physics-based models). In general, AI-based models have many local applications. Therefore, these models have a great potential for various applications including hydrological and hydrogeological phenomena.

In general, according to research findings, it is necessary to provide a solution and make proper forecasting of groundwater resources in order to prevent subsidence and drought phenomena around the world and in Iran. Therefore, in Iran, Khorramabad plain located in Lorestan province, which is very important in terms of drinking and agriculture whose products in this plain feed on groundwater for growth and development, has been subjected to illegal harvesting and digging of illegal wells. The level of groundwater resources has declined sharply in recent years. Therefore, groundwater level changes are more than necessary for forecasting and management measures to improve it. Therefore, the objective of this study is to analyze and predict climatic parameters using groundwater level forecasting in Khorramabad plain using integrated vector regression model with the help of Wavelet Transform (WT) and modern optimization algorithms such as bat algorithm and Grey Wolf Optimizer algorithm. The basis of climatic parameters is the groundwater level and abstraction from the aquifer.



Materials and Methods

In this study, . In addition, the Artificial neural network model approach was used to predict the surface water level, assuming that the amount of groundwater abstraction from the Khorramabad plain was consistent with that in the previous statistical period. Since the ANN model is subject to errors according to recent findings, the strategy of optimizing the adjustment parameters using meta-heuristic algorithms was adopted to reduce the model error. In recent years, several studies have investigated the ANN hybrid model using meta-heuristic algorithms; however, this study used new algorithms that have not been studied in hydrologic or hydrogeologic processes to reduce the problems and challenges associated with this model and introduced a new algorithm to facilitate the simulation process. In addition, future groundwater level can be predicted using dependent parameters and its decline be prevented, thus causing irreparable damage to Iran's groundwater resources. Given that this is one of the most fundamental and important problems in Iran's water issues, this study employs a new creative algorithm and black widow spider algorithms. To facilitate the groundwater level simulation, new meta-heuristic algorithms with WANN support vector wave regression model, having acceptable performance according to several studies, were used. This approach can ensure taking an effective step in simulating and predicting groundwater levels.

Based on the structure of the SVR, the most basic step is to determine the tuning parameters. The coefficients of these parameters are usually determined through trial and error in ANN. Many factors affect the viability of trial and error and the accuracy of model prediction. Given the nature of trial and error, the predictive power may be generally reduced. Numerous solutions have been proposed by various researchers to address this fundamental weakness. One of these solutions adopted by researchers is to calculate the coefficients of parameter adjustment and optimize these coefficients by using meta-heuristic algorithms. Many decision-making problems can be expressed as finite optimization problems that have several decision variables which are constrained by a few constraints. Hybrid optimization problems are usually easily articulated, but are difficult to solve. Two sets of algorithms for solving hybrid problems include exact and approximate algorithms. Accurate algorithms guarantee finding the best solution, but the problem is that these algorithms are not applicable to difficult problems and the time to find solutions to difficult problems will increase exponentially. For most difficult problems, the exact algorithm is unsatisfactory. If the optimal answer is not achievable by using the exact algorithm in practice, we turn to the approximate algorithm. The approximate algorithm, commonly known as heuristic methods, seeks an appropriate and near-optimal solution. This method shortens the computation time compared to the previous method, but does not guarantee finding the best solution. A meta-initiative is the general framework of an algorithm that can provide solutions to the same problem with minor variations of various problems. There are many meta-heuristic algorithms such as Genetic Algorithm, Forbidden Search Simulation, Bat , gray wolves.

Results and Discussion

In this study, upon using climate change modeling, meteorological parameters (temperature and precipitation) for the years 2002-2022 were predicted and, then, by using ultra-exploratory hybrid models such as WSVR, Bat-ANN, and GWO-ANN, the groundwater level decline in Khorramabad plain located in Iran was predicted with the help of rainfall, temperature and harvest parameters associated with the four piezometric aquifers (Sarab Pardeh, Sali, Pol baba, and Naservand).

Conclusion

According to the statistical time periods of 2002-2022, WSVR, Bat-ANN, and GWO-ANN hybrid models in the combined structure including all input parameters had better performance due to higher memory, and WSVR model was more accurate with less error due to the separation of signals into two categories of high-pass and low-pass in WT wavelet transform.

Keywords

Main Subjects


منابع
عالم‌پور رجبی، فرناز، قربانی، محمدعلی، و اسدی، اسماعیل (1402). مدل‌سازی فرایند تبخیر با استفاده از الگوریتم هیبریدی پرنده کوت و شبکه عصبی مصنوعی. مدل‌سازی و مدیریت آب و خاک، 4(2)، 1-12. doi: 10.22098/MMWS.2023.12692.1266
زهساز، کیمیا، دربندی، صابره، و میرزانیا، احسان (1402). مدل‌سازی زمانی و مکانی بارش با استفاده از MLR، ANN و الگوریتم هیبریدی HBA-ANN. مدل‌سازی و مدیریت آب و خاک، 4(2)، 12-25. doi: 10.22098/MMWS.2023.12779.1273
میرزانیا، احسان، ملک احمدی، حسین، شاه‌محمدی، یادگار، و ابراهیم‌زاده، علی (1400). تأثیر موجک بر افزایش دقت مدل‌های تخمینی در مدل‌سازی بارش-رواناب (مطالعۀ موردی: حوضۀ صوفی‌چای). مدل­سازی و مدیریت آب و خاک، 1(3)، 69-79.
 doi: MMWS.2021.9335.1035/10.22098
 
References
Alampour Rajabi, F., Ghorbani, M.A., & Asadi, E. (2023). Modeling of the evaporation process using the hybrid algorithm of the Kot bird and the artificial neural network. Water and Soil Modeling and Management, 4(2), 1-12.
        doi:10.22098/MMWS.2023.12692.1266
         [In Persian]
Amuda, A., Brest, J., & Mezura-Montes, E. (2013). Structured Population Size Reduction Differential Evolution with Multiple Mutation Strategies on CEC 2013 real parameter optimization. In Proceedings of the IEEE Congress on Evolutionary Computation, Cancun, Mexico, Pp. 1925–1931.
         doi: 10.1109/CEC.2013.6557794
Bahmani, R., Taha, B.M., & Ouarda, J. (2021) Groundwater level modeling with hybrid artificial intelligence techniques. Journal of Hydrology, 595(4), 625-642.
       doi: 10.1016/j.jhydrol.2020.125659
Dehghani, R., & Babaali, H. (2022). Evaluation of Statistical Models and Modern Hybrid Artificial Intelligence in Simulation of Runoff Precipitation Process. Sustain. Water Resour. Manag, 154(8), 225-242.
        doi: 10.1007/s40899-022-00743-9
Dehghani, R., & Torabi Poudeh, H. (2020). Applying hybrid artificial algorithms to the estimation of river flow: a case study of Karkheh catchment area. Arabian Journal of Geosciences, 6(2), 426-448.
        doi: 10.1007/s12517-021-07079-2
Dehghani, R., & Torabi, H. (2021a). Application of novel hybrid artificial intelligence algorithms to groundwater simulation. International Journal of Environmental Science and Technology, 14(3), 242-258.
         doi: 10.1007/s13762-021-03596-5
Dehghani, R., & Torabi, H. (2021b). Dissolved oxygen concentration predictions for running waters with using hybrid machine learning techniques. Modeling Earth Systems and Environment, 18(4), 324-341.
Dehghani, R., & Torabi, H. (2022). The effect of climate change on groundwater level and its prediction using modern meta- heuristic model. Groundwater for Sustainable Development, 16(4), 224-238.
         doi: 10.1016/j.gsd.2021.100702
Dehghani, R., Torabi Poudeh, H., Younesi, H., & Shahinejad, B. (2020a). Daily Streamflow Prediction Using Support Vector Machine-Artificial Flora (SVM-AF) Hybrid Model. Acta Geophysica, 68(6), 51-66.
         doi: 10.1007/s11600-020-00472-7
Dehghani, R., Torabi Poudeh, H., Younesi, H., & Shahinejad, B. (2020b). Forecasting Daily River Flow Using an Artificial Flora–Support Vector Machine Hybrid Modeling Approach (Case Study: Karkheh Catchment, Iran). Air, Soil, and Water Research, 14(2), 22-35.
        doi: 10.1177/1178622120969659
Emamgholizadeh, S., Moslemi, K., & Karami, G. (2014). Prediction the Groundwater Level of Bastam Plain (Iran) by Artificial Neural Network (ANN) and Adaptive Neuro-Fuzzy Inference System (ANFIS). Water Resources Management, 28(3), 5433–5446.
         doi: 10.1007/s11269-014-0810-0
Feng, F., Ghorbani, H., & Radwan, A. (2023). Predicting groundwater level using traditional and deep machine learning algorithms. Frontiers in Environmental Science,12(4), 266-280. doi: 10.3389/fenvs.2024.1291327
Ghorbani, M.A., Deo, R.C., Yaseen, Z.M., Kashani, M.H., & Mohammadi, B. (2017a). Pan evaporation prediction using a hybrid multilayer perceptron-firefly algorithm (MLP-FFA) model: case study in North Iran. Theoretical and Applied Climatology, 133(3–4),1119–1131.
         doi: 10.1007/s00704-017-2244-0
Jeihouni, E., Mohammadi, M., Eslamian, S., & Zareian, M.J. (2019). Potential impacts of climate change on groundwater level through hybrid soft-computing methods: a case study—Shabestar plain, Iran. Environmental Monitoring and Assessment, 191(10), 620-632, doi: 10.1007/s10661-019-7784-6
Kaya, Y., Üneş, F., Demirci, M., Taşar, B., & Varçin, H. (2018). Groundwater level prediction using artificial neural network and M5 tree models. Aerul si Apa Componente ale Mediului, 6(4), 371-381. doi:10.1504/IJHST.2016.079356
Kisi, O., Karahan, M., & Sen, Z. (2006). River suspended sediment modeling using fuzzy logic approach, Hydrology of Process, 20(2), 4351-4362. doi:10.1002/hyp.6166
Malekzadeh, M., Kardar, S., Saeb, K., Shabanlou, S., & Taghavi, L. (2019). A novel approach for prediction of monthly ground water level using a hybrid wavelet and non-tuned self-adaptive machine learning model. Water Resources Management, 33(4),1609–1628.
        doi: 10.1007/s11269-019-2193-8
Mirzania, E., Malek Ahmadi, H., Shah Mohammadi, Y., & Ebrahimzadeh, A. (2021). The effect of wavelet on increasing the accuracy of estimation models in rainfall-runoff modeling (case study: Sofichai basin). Water and soil modeling and management, 1(3), 69-79.
 doi: 10.22098/MMWS.2021.9335.1035 [In Persian]
Mirzavand, M., & Ghazavi, R. (2015). A Stochastic Modelling technique for groundwater level forecasting in an arid environment using time series methods. Water Resources Management, 29, 1315–1328. doi: 10.1007/s11269-014-0875-9
Moravej, M., Amani, P., & Hosseini-Moghari, M. (2020). Groundwater level simulation and forecasting using interior search algorithm-least square support vector regression (ISA-LSANN). Groundwater for Sustainable Development, 11(1), 447-462.
        doi: 10.1016/j.gsd.2020.100447
Nagy, H., Watanabe, K., & Hirano, M. (2002). Prediction of sediment load concentration in rivers using artificial neural network model. Journal of Hydraulics Engineering, 128(6), 558-559.
         doi: 10.1061/(ASCE)0733-9429(2002)128:6(588)
Nourani, V., Alami, M.T., & Vousoughi, F.D. (2015). Wavelet-entropy data pre-processing approach for ANN-based groundwater level modeling. Journal of Hydrology, 524(3), 255–269. doi: 10.1016/j.jhydrol.2015.02.048
Ostu, N. (1979). A Threshold Selection Method from Gray-Level Histograms [J]. IEEE Transactions on Systems Man and Cybernetics, 9(1), 62-66. doi: 10.1109/TSMC.1979.4310076
Sakizadeh, M., Mohamed, M.M.A., & Klammler, H. (2019). Trend Analysis and Spatial Prediction of Groundwater Levels Using Time Series Forecasting and a Novel Spatio-Temporal Method. Water Resources Management, 33(2), 1425–1437.
         doi: 10.1007/s11269-019-02208-9
Sharafati, A., Haji, S.B., & Neshat, A. (2020). A new artificial intelligence strategy for predicting the groundwater level over the Rafsanjan aquifer in Iran. Journal of Hydrology, 591(4), 336-352. doi: 10.1016/j.jhydrol.2020.125468
Shin, S., Kyung, D., Lee, S., Taik & Kim, J., & Hyun, J. (2005). An application of support vector machines in bankruptcy prediction model. Expert Systems with Applications, 28(4), 127-135.
         doi: 10.1016/j.eswa.2004.08.009
Vapnik, V.N. (1998). Statistical learning theory. Wiley, New York
Wang, D., Safavi, A.A., & Romagnoli, J.A. (2000). Wavelet-based adaptive robust M-estimator for non-linear system identification, AIChE Journal, 46(4), 1607-1615.
        doi: 10.1002/aic.690460812
Wu, Z., Lu, C., Sun, Q., Lu, W., He, X., & Qin, T. (2023). Predicting groundwater level based on machine learning: a case study of the hebei plain. Water, 15 (4), 823.
        doi:10.3390/w15040823
Zehsaz, K., Darbandi, S., & Mirzania, E. (2023). Temporal and spatial modeling of precipitation using MLR, ANN and HBA-ANN hybrid algorithm. Water and Soil Modeling and Management, 4(2), 12-25.
         doi: 10.22098/MMWS.2023.12779.1273 [In Persian]