三种地下水位动态预测模型在吉林西部的应用与对比

准确而可靠地预测地下水埋深对生态环境保护和水资源规划管理具有重要意义。针对吉林西部浅层地下水位动态变化的复杂性和非线性,提出了基于小波分析与人工神经网络相结合的预测方法小波神经网络(WA-ANN)模型。将研究区2002年1月2009年12月当月降水量、蒸发量、人工开采量和前月平均地下水埋深4个参数作为输入,当月平均地下水埋深作为输出,建立浅层地下水埋深预测模型,并与BP神经网络(BP-ANN)模型和自回归移动平均(ARIMA)模型进行比较,对比分析了三者的建模过程及其模拟精度。结果显示:相比两种ANN模型,ARIMA模型建模过程更为简单,计算效率更高;但WA-ANN模型的拟合精度高于BP-ANN和ARIMA模型,预测效果更好。总体来看,WA-ANN模型在浅层地下水埋深预测中具有一定的应用推广价值。     

A comparative study of shallow groundwater level simulation with WA–ANN and ITS model in a coastal island of south China

Accurate and reliable prediction of shallow groundwater level is a critical component in water resources management. Two nonlinear models, WA–ANN method based on discrete wavelet transform (WA) and artificial neural network (ANN) and integrated time series (ITS) model, were developed to predict groundwater level fluctuations of a shallow coastal aquifer (Fujian Province, China). The two models were testified with the monitored groundwater level from 2000 to 2011. Two representative wells are selected with different locations within the study area. The error criteria were estimated using the coefficient of determination (R ²), Nash–Sutcliffe model efficiency coefficient (E), and root-mean-square error (RMSE). The best model was determined based on the RMSE of prediction using independent test data set. The WA–ANN models were found to provide more accurate monthly average groundwater level forecasts compared to the ITS models. The results of the study indicate the potential of WA–ANN models in forecasting groundwater levels. It is recommended that additional studies explore this proposed method, which can be used in turn to facilitate the development and implementation of more effective and sustainable groundwater management strategies.     

Evaluation of dynamic regression and artificial neural networks models for real-time hydrological drought forecasting

In this study, application of a class of stochastic dynamic models and a class of artificial intelligence model is reported for the forecasting of real-time hydrological droughts in the Black River basin in the USA. For this purpose, the Standardized Hydrological Drought Index (SHDI) was adopted in different time scales to represent the hydrological drought index. Six probability distribution functions (PDF) were fitted to the discharge time series to obtain the best fit for SHDI calculation. Then, a dynamic linear spatio-temporal model (DLSTM) and artificial neural network (ANN) were used to forecast SHDI. Although results indicated that both models were able to forecast SHDI in different time scales, the DLSTM was far superior in longer lead times. The DLSTM could forecast SHDI up to 6 months ahead while ANN was only capable of forecasting SHDI up to 2 months ahead appropriately. For short lead times (1–6 months), the DLSTM has performed nearly perfect in test phase and CE oscillates between 0.97 and 0.86 while for ANN modeling, CE is between 0.72 and 0.07. However, the performance of DLSTM and ANN reduced considerably in medium lead times (7–12 months). Overall, the DLSTM is a powerful tool for appropriately forecasting SHDI at short time scales; a major advantage required for drought early warning systems.     

A comparative study of artificial neural networks,Bayesian neural networks and adaptive neuro-fuzzy inference system in groundwater level prediction

Predictive modeling of hydrological time series is essential for groundwater resource development and management. Here, we examined the comparative merits and demerits of three modern soft computing techniques, namely, artificial neural networks (ANN) optimized by scaled conjugate gradient (SCG) (ANN.SCG), Bayesian neural networks (BNN) optimized by SCG (BNN.SCG) with evidence approximation and adaptive neuro-fuzzy inference system (ANFIS) in the predictive modeling of groundwater level fluctuations. As a first step of our analysis, a sensitivity analysis was carried out using automatic relevance determination scheme to examine the relative influence of each of the hydro-meteorological attributes on groundwater level fluctuations. Secondly, the result of stability analysis was studied by perturbing the underlying data sets with different levels of correlated red noise. Finally, guided by the ensuing theoretical experiments, the above techniques were applied to model the groundwater level fluctuation time series of six wells from a hard rock area of Dindigul in Southern India. We used four standard quantitative statistical measures to compare the robustness of the different models. These measures are (1) root mean square error, (2) reduction of error, (3) index of agreement (IA), and (4) Pearson’s correlation coefficient (R). Based on the above analyses, it is found that the ANFIS model performed better in modeling noise-free data than the BNN.SCG and ANN.SCG models. However, modeling of hydrological time series correlated with significant amount of red noise, the BNN.SCG models performed better than both the ANFIS and ANN.SCG models. Hence, appropriate care should be taken for selecting suitable methodology for modeling the complex and noisy hydrological time series. These results may be used to constrain the model of groundwater level fluctuations, which would in turn, facilitate the development and implementation of more effective sustainable groundwater management and planning strategies in semi-arid hard rock area of Dindigul, Southern India and alike.     

Climate change and groundwater resources in Mekong Delta,Vietnam

Groundwater resources have considerable influences on the human population and socioeconomic development of Vietnam and the Mekong River Delta (MRD). This paper presents an overview of the relationship between climate change and groundwater in the MRD, including the challenges, strategies and technical measures. Our results showed that groundwater levels are related to other climate and hydrological variables (i.e., rainfall, river levels, etc.); therefore, the impacts of climate change on the groundwater resources of the Mekong delta are significant, especially on groundwater recharge. Based on the results of this study, it is recommended that groundwater development in the future should focus on reducing groundwater harvesting, enhancing groundwater quantity by establishing artificial works and exploiting surface water. This study suggests that the Artificial Neural Network (ANN) model is an effective tool for forecasting groundwater levels in periods of 1 month and 3 months for aquifers in the natural and tidal regime areas of the delta.     

Groundwater-level prediction using multiple linear regression and artificial neural network techniques: a comparative assessment

The potential of multiple linear regression (MLR) and artificial neural network (ANN) techniques in predicting transient water levels over a groundwater basin were compared. MLR and ANN modeling was carried out at 17 sites in Japan, considering all significant inputs: rainfall, ambient temperature, river stage, 11 seasonal dummy variables, and influential lags of rainfall, ambient temperature, river stage and groundwater level. Seventeen site-specific ANN models were developed, using multi-layer feed-forward neural networks trained with Levenberg-Marquardt backpropagation algorithms. The performance of the models was evaluated using statistical and graphical indicators. Comparison of the goodness-of-fit statistics of the MLR models with those of the ANN models indicated that there is better agreement between the ANN-predicted groundwater levels and the observed groundwater levels at all the sites, compared to the MLR. This finding was supported by the graphical indicators and the residual analysis. Thus, it is concluded that the ANN technique is superior to the MLR technique in predicting spatio-temporal distribution of groundwater levels in a basin. However, considering the practical advantages of the MLR technique, it is recommended as an alternative and cost-effective groundwater modeling tool.     

Forecasting of groundwater level in hard rock region using artificial neural network

In hardrock terrain where seasonal streams are not perennial source of freshwater, increase in ground water exploitation has already resulted here in declining ground water levels and deteriorating its’ quality. The aquifer system has shown signs of depletion and quality contamination. Thus, to secure water for the future, water resource estimation and management has urgently become the need of the hour. In order to manage groundwater resources, it is vital to have a tool to predict the aquifer response for a given stress (abstraction and recharge). Artificial neural network (ANN) has surfaced as a proven and potential methodology to forecast the groundwater levels. In this paper, Feed-Forward Network based ANN model is used as a method to predict the groundwater levels. The models are trained with the inputs collected from field and then used as prediction tool for various scenarios of stress on aquifer. Such predictions help in developing better strategies for sustainable development of groundwater resources.     

Rainfall-Runoff Modeling using Clustering and Regression Analysis for the River Brahmaputra Basin

In this research, k-means, agglomerative hierarchical clustering and regression analysis have been applied in hydrological real time series in the form of patterns and models, which gives the fruitful results of data analysis, pattern discovery and forecasting of hydrological runoff of the catchment. The present study compares with the actual field data, predicted value and validation of statistical yields obtained from cluster analysis, regression analysis with ARIMA model. The seasonal autoregressive integrated moving average (SARIMA) and autoregressive integrated moving average (ARIMA) models is investigated for monthly runoff forecasting. The different parameters have been analyzed for the validation of results with casual effects. The comparison of model results obtained by K-means & AHC have very close similarities. Result of models is compared with casual effects in the same scenario and it is found that the developed model is more suitable for the runoff forecasting. The average value of R² determined is 0.92 for eight ARIMA models. This shows more accuracy of developed ARIMA model under these processes. The developed rainfall runoff models are highly useful for water resources planning and development.     

Comparison of FFNN and ANFIS models for estimating groundwater level

Prediction of water level is an important task for groundwater planning and management when the water balance consistently tends toward negative values. In Maheshwaram watershed situated in the Ranga Reddy District of Andhra Pradesh, groundwater is overexploited, and groundwater resources management requires complete understanding of the dynamic nature of groundwater flow. Yet, the dynamic nature of groundwater flow is continually changing in response to human and climatic stresses, and the groundwater system is too intricate, involving many nonlinear and uncertain factors. Artificial neural network (ANN) models are introduced into groundwater science as a powerful, flexible, statistical modeling technique to address complex pattern recognition problems. This study presents the comparison of two methods, i.e., feed-forward neural network (FFNN) trained with Levenberg–Marquardt (LM) algorithm compared with a fuzzy logic adaptive network-based fuzzy inference system (ANFIS) model for better accuracy of the estimation of the groundwater levels of the Maheshwaram watershed. The statistical indices used in the analysis were the root mean square error (RMSE), regression coefficient (R ²) and error variation (EV).The results show that FFNN-LM and ANFIS models provide better accuracy (RMSE = 4.45 and 4.94, respectively, R ² is 93% for both models) for estimating groundwater levels well in advance for the above location.     

人工神经网络在水文水资源中的应用

人工神经网络理论被广泛地应用于水文水资源领域中各种问题的研究,依问题性质不同将其划分为4大类:(1)分类和识别问题;(2)预测预报问题;(3)优化计算问题;(4)基于神经网络的专家系统研制与开发问题,对人工神经网络在水文水资源中的应用现状作了较全面的介绍。还指出了目前应用中存在的主要问题以及今后的研究方向。     

人工神经网络(ANN)模型在地下水资源预测中的应用研究

分析了地下水系统影响因素的复杂性，提出对于研究程度不能满足分布参数模型计算要求的研究区域，更适于从系统的观点出发、建立适宜的集中参数模型，从整体上分析研究，以解决相关问题。结合沈阳市地下水资源评价与管理实例，尝试应用人工神经网络(ANN)技术在水资源系统模型研究中的新模式。构建了基于BP算法的ANN降水量和蒸发量的预测、地下水水位动态模拟、预测及开采量优化方面的应用模型，结果表明模型精度满足要求。     

基于人工神经网络的地下水状态方程—以山东莱州湾南岸地下水研究为例

山东莱州湾南岸地下水密度受多种因素影响,传统方法难以拟合,利用人工神经网络（简称ANN）高效的自组织学习能力和抗实验噪音能力及应用神经网络工具箱设计了三层反馈式神经网络模型,得出映射莱州湾南岸地下水密度与含盐量之间相关关系的基于ANN的地下水状态性关系的性质,基于ANN的地下水状态方程具有形式简单,易于使用等特点。     

Analysis and prediction of Caspian Sea level pattern anomalies observed by satellite altimetry using autoregressive integrated moving average models

In this study, we successfully present the analysis and forecasting of Caspian Sea level pattern anomalies based on about 15 years of Topex/Poseidon and Jason-1 altimetry data covering 1993–2008, which are originally developed and optimized for open oceans but have the considerable capability to monitor inland water level changes. Since these altimetric measurements comprise of a large datasets and then are complicated to be used for our purposes, principal component analysis is adopted to reduce the complexity of large time series data analysis. Furthermore, autoregressive integrated moving average (ARIMA) model is applied for further analyzing and forecasting the time series. The ARIMA model is herein applied to the 1993–2006 time series of first principal component scores (sPC1). Subsequently, the remaining data acquired from sPC1 is used for verification of the model prediction results. According to our analysis, ARIMA (1,1,0)(0,1,1) model has been found as optimal representative model capable of predicting pattern of Caspian Sea level anomalies reasonably. The analysis of the time series derived by sPC1 reveals the evolution of Caspian Sea level pattern can be subdivided into five different phases with dissimilar rates of rise and fall for a 15-year time span.     

Sensitivity analysis of groundwater level in Jinci Spring Basin (China) based on artificial neural network modeling

Jinci Spring in Shanxi, north China, is a major local water source. It dried up in April 1994 due to groundwater overexploitation. The groundwater system is complex, involving many nonlinear and uncertain factors. Artificial neural network (ANN) models are statistical techniques to study parameter nonlinear relationships of groundwater systems. However, ANN models offer little explanatory insight into the mechanisms of prediction models. Sensitivity analysis can overcome this shortcoming. In this study, a back-propagation neural network model was built based on the relationship between groundwater level and its sensitivity factors in Jinci Spring Basin; these sensitivity factors included precipitation, river seepage, mining drainage, groundwater withdrawals and lateral discharge to the associated Quaternary aquifer. All the sensitivity factors were analyzed with Garson’s algorithm based on the connection weights of the neural network model. The concept of “sensitivity range” was proposed to describe the value range of the input variables to which the output variables are most sensitive. The sensitivity ranges were analyzed by a local sensitivity approach. The results showed that coal mining drainage is the most sensitive anthropogenic factor, having a large effect on groundwater level of the Jinci Spring Basin.     

鲁能运河电厂长沟岩溶水源地水位模拟预测

长沟水源地是鲁能运河电厂供水水源地,通过预测水源地水位变化趋势,为电厂制定用水计划、及时调整开采布局提供依据。水位模拟预测是在五年长系列动态监测资料及地下水开采量调查基础上进行的,利用数值模型演练和预测了嘉祥单斜蓄水构造单元内在现状岩溶水开采状态下及增加20%开采量条件下的区域地下水流场变化特点,所模拟的曲线与实测曲线相吻合,所推算出的水文地质参数合理,为保证水源地合理、稳定、持续的开采提供了科学支持。     

Predicting groundwater level of wells in the Diyala River Basin in eastern Iraq using artificial neural network

Al-Mansourieh zone is a part of Al-Khalis City within the province of Diyala and located in the Diyala River Basin in eastern Iraq with a total area about 830 km2.Groundwater is the main water source for agriculture in this zone.Random well drilling without geological and hydraulic information has led the most of these wells to dry up quickly.Therefore,it is necessary to estimate the levels of groundwater in wells through observed data.In this study,Alyuda NeroIntelligance 2.1 software was applied to predict the groundwater levels in 244 wells using sets of measured data.These data included the coordinates of wells(x,y),elevations,well depth,discharge and groundwater levels.Three ANN structures(5-3-3-1,5-10-10-1 and 5-11-11-1)were used to predict the groundwater levels and to acquire the best matching between the measured and ANN predicted values.The coefficient of correlation,coefficient determination(R2)and sum-square error(SSE)were used to evaluate the performance of the ANN models.According to the ANN results,the model with the three structures has a good predictability and proves more effective for determining groundwater level in wells.The best predictor was achieved in the structure 5-3-3-1,with R2 about 0.92,0.89,0.84 and 0.91 in training,validation,testing and all processes respectively.The minimum average error in the best predictor is achieved in validation and testing processes at about 0.130 and 0.171 respectively.On the other hand,the results indicated that the model has the potential to determine the appropriate places for drilling the wells to obtain the highest level of groundwater.     

Feasibility of using artificial neural networks to forecast groundwater levels in real time

The purpose of this study is to develop the feed-forward back-propagation neural network (FFBPNN) to estimate the groundwater level (GL) of next hour according the current GL and past precipitation depth in the hillslope. The 72-h precipitation depth and the real-time groundwater levels are used as the model output layer determination variables. The output variables, are type 1, the GL, which has been used in many researches, and type 2, the groundwater level fluctuation (GLF), which is the difference between the current-time and the next-time groundwater level. The order of the water level fluctuation is less than that of the groundwater level by about one order of magnitude (ten times). The landslide area at the downstream of Wu-She Reservoir, Nantou County, Taiwan, is adopted as a field test area. Total 328 cases of Sinlaku typhoon were used to establish the prediction model of real-time GL. Another 327 cases of Jangmi typhoon were adopted to illustrate the model application. The result of model application shows that root-mean-square error of type 2 (=0.104 m) is smaller than that of type 1 (=0.408 m). In conclusion, the forecasting method used GLF gives a much better agreement with the measured values than that of GL.     

Forecasting of meteorological drought using Wavelet-ANFIS hybrid model for different time steps (case study: southeastern part of east Azerbaijan province, Iran)

Drought is accounted as one of the most natural hazards. Studying on drought is important for designing and managing of water resources systems. This research is carried out to evaluate the ability of Wavelet-ANN and adaptive neuro-fuzzy inference system (ANFIS) techniques for meteorological drought forecasting in southeastern part of East Azerbaijan province, Iran. The Wavelet-ANN and ANFIS models were first trained using the observed data recorded from 1952 to 1992 and then used to predict meteorological drought over the test period extending from 1992 to 2011. The performances of the different models were evaluated by comparing the corresponding values of root mean squared error coefficient of determination (R ²) and Nash–Sutcliffe model efficiency coefficient. In this study, more than 1,000 model structures including artificial neural network (ANN), adaptive neural-fuzzy inference system (ANFIS) and Wavelet-ANN models were tested in order to assess their ability to forecast the meteorological drought for one, two, and three time steps (6 months) ahead. It was demonstrated that wavelet transform can improve meteorological drought modeling. It was also shown that ANFIS models provided more accurate predictions than ANN models. This study confirmed that the optimum number of neurons in the hidden layer could not be always determined using specific formulas; hence, it should be determined using a trial-and-error method. Also, decomposition level in wavelet transform should be delineated according to the periodicity and seasonality of data series. The order of models with regard to their accuracy is as following: Wavelet-ANFIS, Wavelet-ANN, ANFIS, and ANN, respectively. To the best of our knowledge, no research has been published that explores coupling wavelet analysis with ANFIS for meteorological drought and no research has tested the efficiency of these models to forecast the meteorological drought in different time scales as of yet.     

Using statistical and artificial neural network models to forecast potentiometric levels at a deep well in South Texas

Reliable forecasts of monthly and quarterly fluctuations in groundwater levels are necessary for short- and medium-term planning and management of aquifers to ensure proper service of seasonal demands within a region. Development of physically based transient mathematical models at this time scale poses considerable challenges due to lack of suitable data and other uncertainties. Artificial neural networks (ANN) possess flexible mathematical structures and are capable of mapping highly nonlinear relationships. Feed-forward neural network models were constructed and trained using the back-percolation algorithm to forecast monthly and quarterly time-series water levels at a well that taps into the deeper Evangeline formation of the Gulf Coast aquifer in Victoria, TX. Unlike unconfined formations, no causal relationships exist between water levels and hydro-meteorological variables measured near the vicinity of the well. As such, an endogenous forecasting model using dummy variables to capture short-term seasonal fluctuations and longer-term (decadal) trends was constructed. The root mean square error, mean absolute deviation and correlation coefficient (R) were noted to be 1.40, 0.33 and 0.77 m, respectively, for an evaluation dataset of quarterly measurements and 1.17, 0.46, and 0.88 m for an evaluative monthly dataset not used to train or test the model. These statistics were better for the ANN model than those developed using statistical regression techniques.     

Environmental hazard analysis of a gypsum karst depression area with geophysical methods: a case study in Sivas (Turkey)

Because the mixture of seawater and freshwater in the Gyeongin-Ara Waterway in South Korea can lead to the intrusion of saline water into surrounding aquifers, systematic management through the establishment of a groundwater protection area is required. The analytic hierarchy process (AHP) model is used to delineate this protection area based on two primary factors and five secondary factors related to saline water intrusion. The study area is divided into 987 gridded cells with a unit size of 100 × 100 m, and the final evaluation score for each cell is calculated using the AHP model. Consequently, several artificial neural network models based on a multilayer perceptron are developed using the AHP’s secondary criteria and the evaluation score. Comparing the evaluation scores of ANN and AHP, more than 180 samples are required in the ANN model to insure high R2 between the original and estimated values. The ANN model is more consistent than the AHP model when determining groundwater protection area, because it can be re-constructed due to the changes in some secondary criteria and also changed due to a standardization process. The final evaluation score by the ANN model based on 300 samples, with the highest R2, is calculated and the regions with a score higher than 2.0 are selected as the groundwater protection area, accounting for 15% of the total cells. This area is similar to the range within approximately 200 m of the GA Waterway and also includes some changing sites in hydrogeochemistry and electric conductivity, which is produced by saline water intrusion. If the land-use type, groundwater levels, and some other criteria change at any cell, the ANN model can be re-executed to verify whether the cell belongs to a groundwater protection area. Considering that salinity of groundwater near the waterway can be affected by various factors including well depth, pumping conditions, and groundwater levels, the ANN model, which is a non-linear model, can be more effective for prediction than the AHP model.