Search

This study aimed to predict evaporation from dam reservoirs using artificial intelligence considering climate change. Mahabad Dam, near Lake Urmia, in northwestern Iran, is used to investigate the proposed approach. There are three parts to the study presented herein. In the first part, two machine learning models, namely group method of data handling (GMDH) and least squares support vector regression (LS-SVR), were used to model the inflow to the dam reservoir. Temperature, precipitation, and inflow during the previous month from 1990 to 2017 were used as input data. In the second part, the evaporation from the dam reservoir was modeled using the adaptive neuro-fuzzy inference system (ANFIS) and optimized ANFIS using Harris hawks optimization (HHO) and the arithmetic optimization algorithm (AOA) optimization algorithms. The input parameters in this part were temperature, precipitation, inflow to the dam reservoir, along with evaporation from the dam reservoir in the previous month. In the third part, precipitation and temperature were predicted using the fifth report of IPCC based on RCP2.6, RCP4.5, and RCP8.5 scenarios for the period 2020-2040. Out of 28 models presented in the fifth report, EC-ERATH and FIO-ESM had the greatest similarity with observational data of temperature and precipitation, respectively. The results of scatter plots and Taylor's diagram showed the higher performance of LS-SVR (root mean square error (RMSE), mean absolute percentage error (MAPE), and Nash-Sutcliffe efficiency (NSE) of 8.65, 4.69, and 0.96) compared to GMDH (RMSE, MAPE, and NSE of 11.65, 7.81, and 0.93) in modeling the inflow. Moreover, both hybrid modes (AOA-ANFIS and HHO-ANFIS) improved the performance of ANFIS in modeling the evaporation from the dam reservoir. The RMSE, MAPE, and NSE values for ANFIS were 0.56, 0.52, and 0.89, respectively, while these values for the AOA-ANFIS (RMSE, MAPE, and NSE of 0.31, 0.24, and 0.93) and HHO-ANFIS (RMSE, MAPE, and NSE of 0.20, 0.30, and 0.96) were improved remarkably. The impact of climate change reduced the inflow to the dam reservoir by about 0.45, 0.80, and 1.7 MCM in RCP2.6, RCP4.5, and RCP8.5 scenarios, respectively. Also, the effect of climate change caused the evaporation from the dam reservoir to increase by about 0.2, 0.9, and 1 MCM in RCP2.6, RCP4.5, and RCP8.5 scenarios, respectively. The findings of this study show that the correct management of dam reservoirs needs to consider the potential effects of climate change in the future. Moreover, the hybrid machine learning models used in this study can be used to predict the amount of evaporation in other reservoirs.

Evaporation is one of the key factors in the hydrological cycle, and it is one of the most critical parameters in hydrological, agricultural, and meteorological studies, especially in arid and semi-arid regions. By estimating evaporation, it is possible to make a significant contribution to studies related to water balance, management and design of irrigation systems, estimation of crop amount, and management of water resources. In this regard, in the present study, using artificial neural network (ANN) and its hybrid algorithms including Harris Hawks Optimization (HHO), Whale Optimization Algorithm (WOA), and Particle Swarm Optimization (PSO), the daily evaporation value from the reservoir of Qaleh Chay Ajab Shir dam was estimated. For this purpose, the effective parameters in the evaporation process were introduced to each of the models in the form of different input patterns, and the evaporation value from the reservoir was also considered as output parameter. The results showed that the best selected model for ANN is the P3 model including three parameters of minimum air temperature, and daily evaporation data with NASH of 0.89, RMSE of 1.5 mm/day, and MAE of 1.1 mm/day, which was optimized by applying hybrid algorithms to train the neural network. The results disclosed that all three models had a good performance in estimating the daily evaporation value, so that the value of the correlation coefficient for all three models is in the range of 0.95–0.99. Based on evaluation criteria, ANN–HHO has better performance than the two other algorithms in estimating daily evaporation value. The values of NASH, RMSE and MAE for the selected pattern of the test data are 0.943, 0.908 and 0.736 mm/day, respectively. For better analysis, Taylor diagram is used (RMSD = 0.98, CC = 0.97, STD = 4 for ANN–HHO). The results of this diagram also showed that the ANN–HHO model provides acceptable performance when compared with other models. Considering the promising results of the models in predicting the daily evaporation from dam, it is suggested to use the existing approach for landscaping the groundwater balance and design of irrigation systems.

In this study, a new 3-stage approach that consists of clustering, simulation, and optimization stages is proposed for the simulation of groundwater level (GWL) in an arid region of eastern Iran. In the first stage, K-means clustering was used to divide the study aquifer into five different clusters based on precipitation, water recharge, water discharge, transmissivity, earth level, and water table. In the second stage, to simulate GWL in each cluster, several input variables, such as water level at the previous month, aquifer discharge, aquifer recharge, evaporation, temperature, and precipitation, were used in the form of various input patterns that were fed to an artificial neural network (ANN). Finally, in the third stage, two advanced optimization methods, i.e., particle swarm optimization (PSO) and whale optimization algorithm (WOA), were utilized to optimize the ANN results. Various patterns were identified as suitable clusters based on the studied models. A pattern including water level at the previous month, aquifer discharge, aquifer recharge, and precipitation was identified as the best model for four clusters, except for cluster 3. The validation with root mean squared error (RMSE), mean absolute percentage error (MAPE), and Nash Sutcliffe index (NSE) revealed RMSE = 0.01, NSE = 0.97, and MAPE = 0.13 for the first cluster, RMSE = 0.011, NSE = 0.99, and MAPE = 0.22 for the second cluster, RMSE = 0.003, NSE = 0.99, and MAPE = 0.30 for the fourth cluster, and RMSE = 0.001, NSE = 0.98, and MAPE = 0.05 for the fifth cluster. For the third cluster, a pattern including water level at the previous month, aquifer discharge, and aquifer recharge was identified as the best model resulting in RMSE = 0.006, NSE = 0.99, and MAPE = 0.05. Finally, according to the results, the ANN–PSO model was applied to three clusters, while the ANN–WOA model was applied to the remaining clusters. In general, this study showed that optimization algorithms can improve the simulation accuracy of ANN, and the efficient use of each method depends on the clustering type. The application of the approach proposed here can be extended to other aquifers that have a relatively large area and limited data availability.

Groundwater is a crucial source of water supply in drought conditions, and an auxiliary water source in wet seasons. Due to its increasing importance in view of climate change, predicting groundwater level (GWL) needs to be improved to enhance management. We used adaptive neuro-fuzzy inference systems (ANFIS) to predict the GWL of the Urmia aquifer in northwestern Iran under various input scenarios using precipitation, temperature, groundwater withdrawal, GWL during the previous month, and river flow. In total, 11 input patterns from various combinations of variables were developed. About 70% of the data were used to train the models, while the rest were used for validation. In a second step, several metaheuristic algorithms, such as genetic algorithm (GA), particle swarm optimization (PSO), ant colony optimization for continuous domains (ACOR), and differential evolution (DE) were used to improve the model and, consequently, prediction performance. The results showed that (i) RMSE, MAPE, and NSE of 0.51 m, 0.00037 m, and 0.86, respectively, were obtained for the ANFIS model using all input variables, indicating a rather poor performance, (ii) metaheuristic algorithms were able to optimize the parameters of the ANFIS model in predicting GWL, (iii) the input pattern that included all input variables resulted in the most appropriate performance with RMSE, MAPE, and NSE of 0.28 m, 0.00019 m, and 0.97, respectively, using the ANIFS-ACOR hybrid model, (iv) results of Taylor’s diagram (CC = 0.98, STD = 0.2, and RMSD = 0.30), as well as the scatterplot (R2 = 0.97), showed that best prediction was achieved by ANFIS-ACOR, and (v) temperature and evaporation exerted stronger influence on GWL prediction than groundwater withdrawal and precipitation. The findings of this study reveal that metaheuristic algorithms can significantly improve the performance of the ANFIS model in predicting GWL.

Due to limited groundwater resources in arid and semi-arid areas, conjunctive use of surface water and groundwater is becoming increasingly important. In view of this, there are needs to improve the methods for conjunctive use of surface and groundwater. Using numerical models, optimization algorithms, and machine learning, we created a new comprehensive methodological structure for optimal allocation of surface and groundwater resources and optimal extraction of groundwater. The surface and groundwater system was simulated by MODFLOW to reflect groundwater transport and aquifer conditions. The important Marvdasht aquifer in the south of Iran was used as an experimental study area to test the methodology. In this context, we developed an optimal conjunctive exploitation model for dry and wet years using two new evolutionary algorithms, i.e., whale optimization algorithm (WOA) and firefly algorithm (FA). These were used in combination with the group method of data handling (GMDH) and least squares support vector machine (LS-SVM) to estimate sustainable groundwater withdrawal. The results show that the FA is more efficient in calculating optimal conjunctive water supply so that about 61% of water needs were met in the worst scenario for surface water resources, while it was 52% using the WOA. By applying the optimal conjunctive model during the simulation period, the groundwater level increased by about 0.4 and 0.55 m using the WOA and FA, respectively. The results of Taylor's diagram, box plot diagram, and rock diagram with error evaluation criteria, i.e., root mean square error (RMSE), mean absolute error (MAE), and Nash-Sutcliffe efficiency (NSE), showed that the GMDH (RMSE = 6.04 MCM, MAE = 3.89 MCM, and NSE = 0.99) was slightly better than LS-SVM (RMSE = 6.36 MCM, MAE = 4.50 MCM, and NSE = 0.98) to estimate optimal groundwater use. The results show that machine learning models are cost- and time-effective solutions to estimate optimal exploitation of groundwater resources in complex combined surface and groundwater supply problems. The methodology can be used to better estimate sustainable exploitation of groundwater resources by water resources managers.

Evaporation is a crucial factor in hydrological studies; its precise measurement has always been challenging due to the costly recording tolls. Therefore, machine learning models that can give reliable predictive results with the least information available have been recommended for evaporation prediction. This study was conducted in the central of Iran using the data related to the Doroudzan dam. Several hydrological and meteorological variables, including inflow and outflow of the reservoir, lake area behind the dam, temperature, overflow from the reservoir, precipitation, and evaporation at the previous month, were considered input data to predict the evaporation at the current month. Monthly data from October 1999 to September 2020 were used during the modeling. First, the single adaptive neuro-fuzzy inference system (ANFIS) and least-squares support vector regression (LS-SVR) models were evaluated for predicting the amount of evaporation using different scenarios defined based on the different combinations of input variables. The results showed that LS-SVR with RMSE = 2.77, MAPE = 2.48, and NSE = 0.93 provided a better prediction than ANFIS. Second, the Harris hawks optimization (HHO) algorithm was used to optimize the parameters of ANFIS to check for the possibility of performance improvement. The hybrid ANFIS-HHO model predicted the evaporation with RMSE = 2.35, MAPE = 1.55, and NSE = 0.95, respectively. The Taylor's diagram also demonstrated the superior performance of the hybrid ANFIS-HHO model than the LS-SVR and ANFIS models. The best scenario for all three models included all input variables but the area behind the dam into the models. The methodology proposed in this study is useful for predicting the evaporation from dam reservoirs under the influence of various dam variables.

Water resources management requires a proper understanding of the status of available and exploitable water. One of the useful management tools is the use of simulation models that are highly efficient in spite of the complex problems in the groundwater sector. In the present study, three data-based models, namely, group method of data handling (GMDH), Bayesian network (BN), and artificial neural network (ANN), have been investigated to simulate the groundwater levels and assess the quantitative status of aquifers. Five observation wells were selected in Birjand aquifer using spatial clustering to analyze and evaluate the aquifer. To determine the effective variables in predicting groundwater level, 10 scenarios were developed by combining several variables, including groundwater level in the previous month, aquifer exploitation, surface recharge, precipitation, temperature, and evaporation. Results showed that the GMDH model with three input variables, i.e., the groundwater level in the previous month, aquifer exploitation, and precipitation, had the highest prediction performance, RMSE, NASH, MAPE, and R2 of which were obtained equal to 0.074, 0.97, 0.0037, and 0.97, respectively. Furthermore, Taylor’s diagram showed that the predicted values using the GMDH model had the highest correlation with the observational data. Hydrograph simulation was performed for 6 years to analyze the condition of the aquifer. The results showed that the groundwater level is in critical condition in this aquifer, and a 1.2-m groundwater loss was predicted for this aquifer. The findings of this study show that the management of the studied aquifer is necessary to improve its current situation.

Comprehensive national estimates of groundwater storage loss (GSL) are needed for better management of natural resources. This is especially important for data scarce regions with high pressure on groundwater resources. In Iran, almost all major groundwater aquifers are in a critical state. For this purpose, we introduce a novel approach using Artificial Intelligence (AI) and machine learning (ML). The methodology involves water budget variables that are easily accessible such as aquifer area, storage coefficient, groundwater use, return flow, discharge, and recharge. The GSL was calculated for 178 major aquifers of Iran using different combinations of input data. Out of 11 investigated variables, agricultural water consumption, aquifer area, river infiltration, and artificial drainage were highly associated to GSL with a correlation of 0.84, 0.79, 0.70, and 0.69, respectively. For the final model, 9 out of the totally 11 investigated variables were chosen for prediction of GSL. Results showed that ML methods are efficient in discriminating between different input variables for reliable GSL estimation. The Harris Hawks Optimization Adaptive Neuro-Fuzzy Inference System (HHO-ANFIS) and the Least-Squares Support Vector Machine (LS-SVM) gave best results. Overall, however, the HHO-ANFIS was most efficient to predict GSL. AI and ML methods can thus, save time and costs for these complex calculations and point at the most efficient data inputs. The suggested methodology is especially suited for data-scarce regions with a great deal of uncertainty and a lack of reliable observations of groundwater levels and pumping.

Accurate calculation of the longitudinal dispersion coefficient (Kx) of pollution is essential in modeling river pollution status. Various equations are presented to calculate the Kx using experimental, analytical, and mathematical methods . Although machine learning models are more reliable than experimental equations in the presence of uncertainties missing data, they have not been widely used in predicting Kx. In this study, the Kx of the river was predicted using machine learning methods, including least square-support vector machine (LS-SVM), adaptive neuro-fuzzy inference system (ANFIS), and ANFIS optimized by Harris hawk optimization (ANFIS-HHO), and the results were compared with that of the experimental methods. Several scenarios were designed by different combinations of input variables, such as the average depth of the flow (H), average flow velocity (U), and shear velocity (u⁎). The results showed that machine learning models had a more efficient performance to predict Kx than experimental equations. The ANFIS-HHO, with a scenario containing all the input variables, performed better than the other two models, with root mean square error, mean absolute percentage error, and coefficient of determination of 17.0, 0.22, and 0.97, respectively. Furthermore, the HHO algorithm slightly increased the prediction performance of the ANFIS. The discrepancy ratio (DR) evaluation criteria showed that experimental equations overestimated the values of Kx, while the machine learning models resulted in higher precision. Also, the results of Taylor's diagram showed the acceptable performance of the ANFIS-HHO model compared to other models. Given the promising results of the present study, it is expected that the proposed approach can be efficiently used for similar environmental modeling problems.

Agricultural months are the critical period for the allocation of surface water and groundwater resources due to the increased demands on water supplies and decreased recharge rate. This situation urges the necessity of using conjunctive water management to fulfill the entire water demand. Here, we proposed an approach for aquifer stabilization and meeting the maximum water demand based on the available surface and groundwater resources and their limitations. In this approach, we first used the MODFLOW model to simulate the groundwater level to control the optimal withdrawal and the resulting drop. We next used a whale optimization algorithm (WOA) to develop an optimized model for the planning of conjunctive use to minimize the monthly water shortage. In the final step, we incorporated the results of the optimized conjunctive model and the available field data into the least squares-support vector machine (LS-SVM) model to predict the amounts of water shortage for each month, particularly for the agricultural months. The results showed that during the period from 2005 to 2020, the most water shortage belonged to 2018, in which only about 52% of water demand was met with the contribution of groundwater (67%) and surface water (33%). However, the groundwater level could have increased by about 0.7 m during the study period by implementing the optimized model. The results of the third part revealed that LS-SVM could predict the water shortage with better performance with a root-mean-square error (RMSE), mean absolute percentage error (MAPE), and Nash–Sutcliffe Index of 5.70 m, 3.43%, and 0.89 m, respectively. The findings of this study will enable managers to predict the water shortage in future periods to make more informed decisions for water resource allocation.

The present study deals with the optimal extraction of groundwater resources. This approach has been developed for optimal integrated operation in an aquifer in Iran. The results of simulation / optimization models have been used to develop a predictive model based on machine learning. In the first stage, adaptive neuro-fuzzy inference system (ANFIS) was used to predict the Optimal Groundwater Exploitation (OGE) amount in each month having the amounts of several inputs, including the drop and amount of surface water at the end of the previous month and two months earlier, and the water demand of the current month. The results showed that the model’s performance in predicting the test data was undesirable. Therefore, to improve the prediction results, in the second stage several evolutionary optimization algorithms, i.e., particle swarm optimization (PSO), gray wolf optimization (GWO), and Harris hawk optimization (HHO), were used to train ANFIS model. The results indicated the appropriate performance of HHO in ANFIS training, which significantly improved the prediction accuracy of this model. The best scenario for the ANFIS-HHO model included all the input parameters, which resulted in RMSE = 1.45, MAE = 1.15, and R2 = 0.99 respectively, for the test data. In addition, the Taylor diagram (RMSD = 1.40, STD = 15.5 and CC = 0.99) showed ANFIS-HHO accuracy in estimating the OGE value. ANFIS-HHO was also able to improve the accuracy of anfis by RMSE = 4 and MAE = 2 MCM. In general, ANFIS-POS, ANFIS-GWO and ANFIS-HHO had good predictive accuracy compared to ANFIS. The results assure the authors to suggest the developed approach to experts for timely and cost-effective prediction of OGE in similar study areas.