A Hybrid Double Feedforward Neural Network for Suspended Sediment Load Estimation

Estimation of suspended sediment loads (SSL) in rivers is an important issue in water resources management and planning. This study proposes a hybrid double feedforward neural network (HDFNN) model for daily SSL estimation, by combining fuzzy pattern-recognition and continuity equation into a structure of double neural networks. A comparison is performed between HDFNN, multi-layer feedforward neural network (MFNN), double parallel feedforward neural network (DPFNN) and hybrid feedforward neural network (HFNN) models. Based on a case study on the Muddy Creek in Montana of USA, it is found that the HDFNN model is strongly superior to the other three benchmarking models in terms of root mean squared error (RMSE) and Nash-Sutcliffe efficiency coefficient (NSEC). HDFNN model demonstrates the best generalization and estimation ability due to its configuration and capability of physically dealing with different inputs. The peak value of SSL is closely estimated by the HDFNN model as well. The performances of HDFNN model in low and medium loads are satisfactory when investigated by partitioning analysis. Thus, the HDFNN is appropriate for modeling the sediment transport process with nonlinear, fuzzy and time-varying characteristics. It explores a practical alternative for use and can be recommended as an efficient estimation model for SSL.     

Rainfall Pattern Forecasting Using Novel Hybrid Intelligent Model Based ANFIS-FFA

In this study, a new hybrid model integrated adaptive neuro fuzzy inference system with Firefly Optimization algorithm (ANFIS-FFA), is proposed for forecasting monthly rainfall with one-month lead time. The proposed ANFIS-FFA model is compared with standard ANFIS model, achieved using predictor-predictand data from the Pahang river catchment located in the Malaysian Peninsular. To develop the predictive models, a total of fifteen years of data were selected, split into nine years for training and six years for testing the accuracy of the proposed ANFIS-FFA model. To attain optimal models, several input combinations of antecedents’ rainfall data were used as predictor variables with sixteen different model combination considered for rainfall prediction. The performances of ANFIS-FFA models were evaluated using five statistical indices: the coefficient of determination (R ² ), Nash-Sutcliffe efficiency (NSE), Willmott’s Index (WI), root mean square error (RMSE) and mean absolute error (MAE). The results attained show that, the ANFIS-FFA model performed better than the standard ANFIS model, with high values of R ² , NSE and WI and low values of RMSE and MAE. In test phase, the monthly rainfall predictions using ANFIS-FFA yielded R ² , NSE and WI of about 0.999, 0.998 and 0.999, respectively, while the RMSE and MAE values were found to be about 0.272 mm and 0.133 mm, respectively. It was also evident that the performances of the ANFIS-FFA and ANFIS models were very much governed by the input data size where the ANFIS-FFA model resulted in an increase in the value of R ² , NSE and WI from 0.463, 0.207 and 0.548, using only one antecedent month of data as an input (t-1), to almost 0.999, 0.998 and 0.999, respectively, using five antecedent months of predictor data (t-1, t-2, t-3, t-6, t-12, t-24). We ascertain that the ANFIS-FFA is a prudent modelling approach that could be adopted for the simulation of monthly rainfall in the present study region.     

Modeling River Mixing Mechanism Using Data Driven Model

Modeling river mixing mechanism in terms of pollution transmission in rivers is an important subject in environmental studies. Dispersion coefficient is an important parameter in river mixing problem. In this study, to model and predict the longitudinal dispersion coefficient (D _L ) in natural streams, two soft computing techniques including multivariate adaptive regression splines (MARS) as a new approach to study hydrologic phenomena and multi-layer perceptron neural network as a common type of neural network model were prepared. To this end, related dataset were collected from literature and used for developing them. Performance of MARS model was compared with MLP and the empirical formula was proposed to calculate D _L . To define the most effective parameters on D _L structure of obtained formula from MARS model and more accurate formula was evaluated. Calculation of error indices including coefficient of determination (R²) and root mean square error (RMSE) for the results of MARS model showed that MARS model with R²?=?0.98 and RMSE?=?0.89 in testing stage has suitable performance for modeling D _L . Comparing the performance of empirical formulas, ANN and MARS showed that MARS model is more accurate compared to others. Attention to the structure of developed MARS and the most accurate empirical formulas model showed that flow velocity, depth of flow (H) and shear velocity are the most influential parameters on D _L .     

Excess Stormwater Quantification in Ungauged Watersheds Using an Event-Based Modified NRCS Model

Quantifying runoff from a storm event is a crucial part of rainfall-runoff model development. The objective of this study is to illustrate inconsistencies in the initial abstraction (I _a) and curve number (CN) in the Natural Resources Conservation Service (NRCS) model for ungauged steep slope watersheds. Five alternatives to the NRCS model were employed to estimate stormwater runoff in 39 forest-dominated mountainous watersheds. The change to the parameterization (slope-adjusted CN and I _a) leads to more efficient modified NRCS models. The model evaluations based on root mean square error (RMSE), Nash-Sutcliffe coefficient E, coefficient of determination (R ² ), and percent bias (PB) indicated that our proposed model with modified I _a, consistently performed better than the other four models and the original NRCS model, in reproducing the runoff. In addition to the quantitative statistical accuracy measures, the proposed I _a modification in the NRCS model showed very encouraging results in the scatter plots of the combined 1799 storm events, compared to other alternatives. This study’s findings support modifications to the CN and the I _a in the NRCS model for steep slope ungauged watersheds and suggest additional changes for more accurate runoff estimations.     

Improving the Pressurized Flushing Efficiency in Reservoirs: an Experimental Study

Sediment flushing in many reservoirs of the world is accomplished with low efficiency. In this study, a new configuration was proposed for reservoir bottom outlet to increase the pressurized flushing efficiency. In the new configuration, a projecting semi-circular structure was connected to the upstream edge of bottom outlet. It was observed that by employing the projecting bottom outlet, the sediment removal efficiency increased significantly compared to the flushing via typical bottom outlet. In the case of new-configuration bottom outlet with L _sc /D _outlet  = 5.26 and D _sc /D _outlet  = 1.32, the dimensionless length, width and depth of flushing cone increased 280%, 45% and 14%, respectively, compared to the reference test. The proposed structure can ensure the sustainable use of reservoirs.     

A Comparison of Nonlinear Stochastic Self-Exciting Threshold Autoregressive and Chaotic <Emphasis Type="Italic">k</Emphasis>-Nearest Neighbour Models in Daily Streamflow Forecasting

A nonlinear stochastic self-exciting threshold autoregressive (SETAR) model and a chaotic k-nearest neighbour (k-nn) model, for the first time, were compared in one and multi-step ahead daily flow forecasting for nine rivers with low, medium, and high flows in the western United States. The embedding dimension and the number of nearest neighbours of the k-nn model and the parameters of the SETAR model were identified by a trial-and-error process and a least mean square error estimation method, respectively. Employing the recursive forecasting strategy for the first time in multi-step forecasting of SETAR and k-nn, the results indicated that SETAR is superior to k-nn by means of performance indices. SETAR models were found to be more efficient in forecasting flows in one and multi-step forecasting. SETAR is less sensitive to the propagated error variances than the k-nn model, particularly for larger lead times (i.e., 5 days). The k-nn model should carefully be used in multi-step ahead forecasting where peak flow forecasting is important by considering the risk of error propagation.     

A Probabilistic Nonlinear Model for Forecasting Daily Water Level in Reservoir

Accurate prediction and monitoring of water level in reservoirs is an important task for the planning, designing, and construction of river-shore structures, and in taking decisions regarding irrigation management and domestic water supply. In this work, a novel probabilistic nonlinear approach based on a hybrid Bayesian network model with exponential residual correction has been proposed for prediction of reservoir water level on daily basis. The proposed approach has been implemented for forecasting daily water levels of Mayurakshi reservoir (Jharkhand, India), using a historic data set of 22 years. A comparative study has also been carried out with linear model (ARIMA) and nonlinear approaches (ANN, standard Bayesian network (BN)) in terms of various performance measures. The proposed approach is comparable with the observed values on every aspect of prediction, and can be applied in case of scarce data, particularly when forcing parameters such as precipitation and other meteorological data are not available.     

An Analytical Solution of Groundwater Flow in Wedge-shaped Aquifers with Estuarine Boundary Conditions

The issue of the groundwater fluctuation due to tidal effect in a two-dimensional coastal leaky aquifer system has attracted much attention in recent years. The predictions of head fluctuation play an important role in dealing with groundwater managements and contaminant remediation problems in costal aquifers. This article presents a two-dimensional analytical model describing the groundwater flow in a coastal leaky aquifer of wedge shape affected by the tides and bounded by two estuarine rivers with an arbitrary included angle. The solution of the model is derived in the Polar coordinates by the Hankel transform and finite sine transform. The head fluctuation predicted by this new solution is compared with that by an existing solution for groundwater flow in a non-L shaped tidal aquifer. The groundwater fluctuation due to the joint effect of estuarine tides is explored based on the present solution. Moreover, the influences of the parameters such as diffusion (D_i), included angle (Ф), and tidal river coefficients (K₁, K₂) on the head fluctuation in the aquifer are also assessed and discussed. The results demonstrate that those parameters have significant effects on the head fluctuation in the leaky confined aquifer system. Moreover, the effect of D_i increases with Ф, and the effects of K₁ and K₂ on the normalized amplitude and phase lag of the groundwater fluctuation are significant when both parameter values are larger than 10^?5.     

Suitability of ANN-Based Daily Streamflow Extension Models: a Case Study of Gaoping River Basin,Taiwan

It is well known that sufficiently long and continuous streamflow data are required for accurate estimations and informed decisions in water-resources planning, design, and management. Although streamflow data are measured and available at most river basins, streamflow records often suffer from insufficient length or missing data. In this work, artificial neural networks (ANNs) are applied to extend daily streamflow records at Lilin station located in Gaoping River basin, southern Taiwan. Two ANNs, including feed forward back propagation (FFBP) and radial basis function (RBF) networks, associated with various time-lagged streamflow and rainfall inputs of nearby long-record stations are employed to extend short daily streamflow records. Performances of ANNs are evaluated by root-mean-square error (RMSE), coefficient of efficiency (CE), and histogram-matching dissimilarity (HMD). Inconsistency among these evaluation measures is solved by the technique for order performance by similarity to ideal solution (TOPSIS), a widely used multi-criteria decision-making approach, to find an optimal model. The results indicate that RBF-E1 (entire-year data training with Q _t and Q _t?1 inputs) has the minimum RMSE of 104.4 m³/s, second highest CE of 0.956, and third lowest HMD of 0.0096, which outperforms other ANNs and provide the most accurate reconstruction of daily streamflow records at Lilin station.     

Multiple Random Forests Modelling for Urban Water Consumption Forecasting

The precise forecasting of water consumption is the basis in water resources planning and management. However, predicting water consumption fluctuations is complicated, given their non-stationary and non-linear characteristics. In this paper, a multiple random forests model, integrated wavelet transform and random forests regression (W-RFR), is proposed for the prediction of daily urban water consumption in southwest of China. Raw time series were first decomposed into low- and high-frequency parts with discrete wavelet transformation (DWT). The random forests regression (RFR) method was then used for prediction using each subseries. In the process, the input and output constructions of the RFR model were proposed for each subseries on the basis of the delay times and the embedding dimension of the attractor reconstruction computed by the C-C method, respectively. The forecasting values of each subseries were summarized as the final results. Four performance criteria, i.e., correlation coefficient (R), mean absolute percentage error (MAPE), normalized root mean square error (NRMSE) and threshold static (TS), were used to evaluate the forecasting capacity of the W-RFR. The results indicated that the W-RFR can capture the basic dynamics of the daily urban water consumption. The forecasted performance of the proposed approach was also compared with those of models, i.e., the RFR and forward feed neural network (FFNN) models. The results indicated that among the models, the precision of the predictions of the proposed model was greater, which is attributed to good feature extractions from the multi-scale perspective and favorable feature learning performance using the decision trees.     

A Novel Approach to Modeling Wastewater Evaporation Based on Dimensional Analysis

Wastewater from municipal and industrial sources is becoming increasingly important in being reused, for example, for irrigation purposes. Wastewater is commonly stored in treatment lagoons in which evaporation is the main cause of water loss. Nonetheless, modeling wastewater evaporation (WWE) has received little attention. Driven by this knowledge gap, this study was performed to explore extent to which impurities affect water evaporation. A dimensional analysis was used to formulate WWE as a function of clear water evaporation (CWE), wastewater properties and climatic variables. We based our modeling approach on experimental data collected from the Neishaboor municipal wastewater treatment plant, Iran. As a result of this analysis, a multiplicative model to formulate WWE as a function of the influencing variables is proposed which indicated a reasonably well accuracy (RMSE?=?1.09 mm) for the WWE estimation. Clear water evaporation indicated to be the most correlated variable in the model such that a constant coefficient can also be used to estimate WWE from CWE at the cost of losing accuracy only by 4.6 %.     

Regionalization of Tank Model Using Landscape Metrics of Catchments

To address the challenges inherent in accessing spatiotemporal hydrological data, water resources professionals have developed various regionalization tools. The present study examines the possibility that changes in landscape metrics including mean shape index, mean perimeter-area ratio, mean patch size and patch density of land use/ land cover could result in variations in the optimized parameters of the conceptual rainfall-runoff Tank model. Data from 30 catchments that are geographically distributed in Germany was used to develop the procedure. Regression analysis-based modeling indicated that four out of twelve model parameters (r²?≥?0.40) can be explained by changes in catchment geometrics along with a set of landscape metrics of land use/land cover. They include: coefficient of infiltration flow (r²?=?0.48, p?<?0.03), intermediate flow (r²?=?0.77, p?<?0.02), water storage level for sub-surface flow (r²?=?0.57, p?<?0.05) and water storage level for intermediate flow (r²?=?0.85, p?<?0.01). Despite developing fairly reliable regression models, uncertainty analysis also revealed that uncertainty induced unreliability of the regionalized models is of significant importance.     

A Modified SCS-CN Method Incorporating Storm Duration and Antecedent Soil Moisture Estimation for Runoff Prediction

In one of the widely used methods to estimate surface runoff - Soil Conservation Service Curve Number (SCS-CN), the antecedent moisture condition (AMC) is categorized into three AMC levels causing irrational abrupt jumps in estimated runoff. A few improved SCS-CN methods have been developed to overcome several in-built inconsistencies in the soil moisture accounting (SMA) procedure that lies behind the SCS-CN method. However, these methods still inherit the structural inconsistency in the SMA procedure. In this study, a modified SCS-CN method was proposed based on the revised SMA procedure incorporating storm duration and a physical formulation for estimating antecedent soil moisture (V ₀ ). The proposed formulation for V ₀ estimation has shown a high degree of applicability in simulating the temporal pattern of soil moisture in the experimental plot. The modified method was calibrated and validated using a dataset of 189 storm-runoff events from two experimental watersheds in the Chinese Loess Plateau. The results indicated that the proposed method, which boosted the model efficiencies to 88% in both calibration and validation cases, performed better than the original SCS-CN and the Singh et al. (2015) method, a modified SCS-CN method based on SMA. The proposed method was then applied to a third watershed using the tabulated CN value and the parameters of the minimum infiltration rate (f _c ) and coefficient (β) derived for the first two watersheds. The root mean square error between the measured and predicted runoff values was improved from 6 mm to 1 mm. Moreover, the parameter sensitivity analysis indicated that the potential maximum retention (S) parameter is the most sensitive, followed by f _c . It can be concluded that the modified SCS-CN method, may predict surface runoff more accurately in the Chinese Loess Plateau.     

Modified Affinity Laws in Hydraulic Machines towards the Best Efficiency Line

The development of hydraulic and optimization models in water networks analyses to improve the sustainability and efficiency through the installation of micro or pico hydropower is swelling. Hydraulic machines involved in these models have to operate with different rotational speed, in order that in each instant to maximize the recovered energy. When the changes of rotational speed are determined using affinity laws, the errors can be significant. Detailed analyses are developed in this research through experimental tests to validate and propose new affinity laws in different reaction turbomachines. Once the errors have been analyzed, a methodology to modify the affinity laws is applied to radial and axial turbines. An empirical method to obtain the Best Efficiency Line (BEL) in proposed (i.e., based on all the Best Efficiency Points (BEPs) for different flows). When the experimental measurements and the calculated values by the empirical method are compared, the mean errors are reduced 81.81%, 50%, and 86.67% for flow, head, and efficiency parameters, respectively. The knowledge of BEL allows managers to define the operation rules to reach the BEP for each flow, improving the energy efficiency in the optimization strategies to be adopted.     

Soft Computing Techniques for Rainfall-Runoff Simulation: Local Non–Parametric Paradigm vs. Model Classification Methods

Accurate simulation of rainfall-runoff process is of great importance in hydrology and water resources management. Rainfall–runoff modeling is a non-linear process and highly affected by the inputs to the simulation model. In this study, three kinds of soft computing methods, namely artificial neural networks (ANNs), model tree (MT) and multivariate adaptive regression splines (MARS), have been employed and compared for rainfall-runoff process simulation. Moreover, this study investigates the effect of input size, including number of input variables and number of data time series on runoff simulation by the developed models. Inputs to the simulation models for calibration and validation purposes consist two parts: I1: five variables, including daily rainfall and runoff time series (30 years) with lag times, and I2: twelve variables, including daily rainfall and runoff time series (10 years). To increase the model performances, optimal number and type for input variables are identified. The efficiency of the training and testing performances using the ANNs, MT and MARS models is then evaluated using several evaluation criteria. To implement the methodology, Tajan catchment in the northern part of Iran is selected. Based on the results, it was found that using I1 as input to the developed models results in higher simulation performance. The results also provided evidence that MT (R = 0.897, RMSE = 6.70, RSE = 0.33) with set I2 is capable of reliable model for rainfall-runoff process compared with MARS (R = 0.892, RMSE = 7.47, RSE = 0.83) and ANNs (R = 0.884, RMSE = 7.40, RSE = 0.43) models. Therefore, size (length of data time series) and type of input variables have significant effects on the modeling results.     

Seasonality of Roughness - the Indicator of Annual River Flow Resistance Condition in a Lowland Catchment

Accurate estimation of flow resistance restricts the quality of the hydraulic model performance. In this study, we try to investigate the seasonal dynamic of the Manning’s roughness coefficient (n) based on the one-dimensional hydraulic model HEC-RAS in a German lowland area. We set up four river section models based on the 1 m digital elevation model and field measurements, in which the seasonal roughness factors were calibrated and validated with the gauge record. The results revealed that: 1) the Manning’s n varied from 46% to 135% from the base value in autumn; 2) adopting the seasonal roughness factor improved the quality of the model output; 3) the vegetation condition and water elevation dominated the Manning’s n in summer (April–September) and winter (October–March) half year respectively. Water temperature increased the flow resistence in winter half year; 4) the peak value of Manning’s n appeared in late summer due to the highest biomass, while the minimum roughness occurred in early-spring because of the combined influence of low biomass, high water level and relatively higher temperature. The involvement of seasonal roughness factor improved the model performance and the results are comparable to the previous research of the same area.     

Development of a GIS Interface for Estimation of Runoff from Watersheds

Development of accurate surface runoff estimation techniques from ungauged watersheds is relevant in Indian condition due to the non-availability of hydrologic gauging stations in majority of watersheds. Besides this, the high budgetary requirements for installation of gauging stations are another limiting factor in India, which leads to the use of surface runoff estimation techniques for ungauged watersheds. Natural Resources Conservation Services Curve Number (NRCS-CN) method is one of the most widely used methods for quick and accurate estimation of surface runoff from ungauged watershed. Also, the coupling of NRCS-CN techniques with the advanced Geographic Information System (GIS) capabilities automates the process of runoff prediction in timely and efficient manner. Keeping view of this, a GIS interface was developed using the in-built macro programming language, Visual Basic for Applications (VBA) of ArcGIS® tool to estimate the surface runoff by adopting NRCS-CN technique and its three modifications. The developed interface named as Interface for Surface Runoff Estimation using Curve Number techniques (ISRE-CN), was validated using the recorded data for the periods from 1993 to 2001 of a gauged watershed, Banha in the Upper Damodar Valley in Jharkhand, India. The observed runoff depths for different rainfall events in this study watershed was compared with the predicted values of NRCS-CN methods and its three modifications using statistical significance tests. It was revealed that using all the rainfall data for different AMC conditions, the modified CN I performed the best [R ² (coefficient of determination)?=?0.92; E (model efficiency)?=?0.89) followed by modified CN III method (R ²?=?0.88; E?=?0.87), while the modified CN II (R ²?=?0.42; E?=?0.36) failed to predict accurately the surface runoff from Banha watershed. Moreover, under AMC based estimations, the modified CN I method also performed best ( R ²?=?0.95; E?=?0.95) for AMC II condition, while the modified CN II performed the worst in all the AMC conditions. However, the developed Interface in ArcGIS® needs to be tested in other watershed systems for wider applicability of the modified CN methods.     

Simulating Microbial Water Quality in Data-Scarce Catchments: an Update of the WQSAM Model to Simulate the Fate of <Emphasis Type="BoldItalic">Escherichia coli</Emphasis>

Faecal-derived microbial pollution of fresh surface waters is a global problem. Water quality models can play an important role in the management of microbial pollution; however, most existing models are too complex and require a large amount of observed data for calibration, thereby excluding their use in data-scarce catchments. The Water Quality Systems Assessment Model (WQSAM) is a water quality water model structured on the concept of requisite simplicity, thereby limiting the complexity and data requirements of the model. Here, microbial water quality simulation functionality was added to WQSAM, with the aim of assessing whether a simplified representation of processes affecting microbial water quality is sufficiently accurate for purposes of water resource management. Simulations of microbial water quality were based on the inputs and fate of an indicator organism, Escherichia coli. Non-point source inputs were modelled by assigning microbial water quality ‘signatures’ to incremental flow components, whereas a similar signature was assigned to point source inputs. The instream fate of E. coli was based on a first-order rate equation, moderated by salinity and water temperature. The model was validated by application to the upper to middle Crocodile River Catchment, Mpumalanga, South Africa, for historical conditions. Model simulations were obtained that were representative of the variability of observed temperature, salinity and microbial water quality data. The simulations of E. coli were found to be most sensitive to the decay rate k ₀. It is argued here that the uncertainty in model results due to the use of a relatively simple model structure would be no more, or even less that that due to the application of a complex model to a catchment with insufficient observed data for adequate model calibration.     

A Comparative Assessment of Artificial Neural Network,Generalized Regression Neural Network,Least-Square Support Vector Regression,and K-Nearest Neighbor Regression for Monthly Streamflow Forecasting in Linear and Nonlinear Conditions

Monthly forecasting of streamflow is of particular importance in water resources management especially in the provision of rule curves for dams. In this paper, the performance of four data-driven models with different structures including Artificial Neural Network (ANN), Generalized Regression Neural Network (GRNN), Least Square-Support Vector Regression (LS-SVR), and K-Nearest Neighbor Regression (KNN) are evaluated in order to forecast monthly inflow to Karkheh dam, Iran, in linear and non-linear conditions while the optimized values of the model parameters are determined in the same condition via the Leave-One-Out Cross Validation (LOOCV) method. Results show that the performance of the models is different in linear and nonlinear conditions; the cumulative ranking of the models according to the three assessment criteria including NSE, RMSE and R² indicates that ANN performs best in linear conditions while LS-SVR, GRNN and KNN are in the next ranks, respectively. But in nonlinear conditions, the best performance belongs to LS-SVR, followed by KNN, ANN, and GRNN models.