Application of integrated back‐propagation network and self organizing map for flood forecasting

Much of the nonlinearity and uncertainty regarding the flood process is because hydrologic data required for estimation are often tremendously difficult to obtain. This study employed a back‐propagation network (BPN) as the main structure in flood forecasting to learn and to demonstrate the sophisticated nonlinear mapping relationship. However, a deterministic BPN model implies high uncertainty and poor consistency for verification work even when the learning performance is satisfactory for flood forecasting. Therefore, a novel procedure was proposed in this investigation which integrates linear transfer function (LTF) and self‐organizing map (SOM) to efficiently determine the intervals of weights and biases of a flood forecasting neural network to avoid the above problems. A SOM network with classification ability was applied to the solutions and parameters of the BPN model in the learning stage, to classify the network parameter rules and to obtain the winning parameters. The outcomes from the previous stage were then used as the ranges of the parameters in the recall stage. Finally, a case study was carried out in Wu‐Shi basin to demonstrate the effectiveness of the proposal. Copyright © 2009 John Wiley & Sons, Ltd.     

Water–air temperature relationships in a Devon river system and the role of flow

The nature of the water–air temperature relationship, and its moderation by discharge, were investigated for catchments ranging in size from 2·1 to 601 km² in the Exe basin, Devon, UK and for data relating to hourly, daily and weekly time bases. The sensitivity and explanatory power of simple water–air temperature regression models based on hourly data were improved by incorporation of a lag, which increased with catchment size, although relationships became more sensitive and less scattered as the time base of data increased from hourly to weekly mean values. Significant departures from linearity in water–air temperature relationships were evident for hourly, but not for daily mean or weekly mean, data. A clear tendency for relationships between water and air temperatures to be stronger and more sensitive for flows below median levels was apparent, and multiple regression analysis also revealed water temperature to be inversely related to discharge for all catchments and time‐scales. However, discharge had a greater impact in accounting for water temperature variation at shorter time‐scales and in larger catchments. Copyright © 2003 John Wiley & Sons, Ltd.     

Application of tank,NAM, ARMA and neural network models to flood forecasting

Two lumped conceptual hydrological models, namely tank and NAM and a neural network model are applied to flood forecasting in two river basins in Thailand, the Wichianburi on the Pasak River and the Tha Wang Pha on the Nan River using the flood forecasting procedure developed in this study. The tank and NAM models were calibrated and verified and found to give similar results. The results were found to improve significantly by coupling stochastic and deterministic models (tank and NAM) for updating forecast output. The neural network (NN) model was compared with the tank and NAM models. The NN model does not require knowledge of catchment characteristics and internal hydrological processes. The training process or calibration is relatively simple and less time consuming compared with the extensive calibration effort required by the tank and NAM models. The NN model gives good forecasts based on available rainfall, evaporation and runoff data. The black‐box nature of the NN model and the need for selecting parameters based on trial and error or rule‐of‐thumb, however, characterizes its inherent weakness. The performance of the three models was evaluated statistically. Copyright © 2000 John Wiley & Sons, Ltd.     

Prediction and modelling of rainfall–runoff during typhoon events using a physically-based and artificial neural network hybrid model

Abstract

The accurate prediction of hourly runoff discharge in a watershed during heavy rainfall events is of critical importance for flood control and management. This study predicts n-h-ahead runoff discharge in the Sandimen basin in southern Taiwan using a novel hybrid approach which combines a physically-based model (HEC-HMS) with an artificial neural network (ANN) model. Hourly runoff discharge data (1200 datasets) from seven heavy rainfall events were collected for the model calibration (training) and validation. Six statistical indicators (i.e. mean absolute error, root mean square error, coefficient of correlation, error of time to peak discharge, error of peak discharge and coefficient of efficiency) were employed to evaluate the performance. In comparison with the HEC-HMS model, the single ANN model, and the time series forecasting (ARMAX) model, the developed hybrid HEC-HMS–ANN model demonstrates improved accuracy in recursive n-h-ahead runoff discharge prediction, especially for peak flow discharge and time.     

Shared near neighbours neural network model: a debris flow warning system

The main purpose of this study is to develop a new type of artificial neural network based model for constructing a debris flow warning system. The Chen‐Eu‐Lan river basin, which is located in Central Taiwan, is assigned as the study area. The creek is one of the most well‐known debris flow areas where several damaging debris flows have been reported in the last two decades. The hydrological and geological data, which might have great influence on the occurrence of debris flows, are first collected and analysed, then, the shared near neighbours neural network (SNN + NN) is presented to construct the debris flow warning system for the watershed. SNN is an unsupervised learning method that has the advantage of dealing with non‐globular clusters, besides presenting computational efficiency. By using SNN, the compiled hydro‐geological data set can easily and meaningfully be clustered into several categories. These categories can then be identified as ‘occurrence’ or ‘no‐occurrence’ of debris flows. To improve the effectiveness of the debris flow warning system, a neural network framework is designed to connect all the clusters produced by the SNN method, whereas the connected weights of the network are adjusted through a supervised learning method. This framework is used and its applicability and practicability for debris flow warning are investigated. The results demonstrate that the proposed SNN + NN model is an efficient and accurate tool for the development of a debris flow warning system. Copyright © 2007 John Wiley & Sons, Ltd.     

Real‐time recurrent learning neural network for stream‐flow forecasting

Various types of neural networks have been proposed in previous papers for applications in hydrological events. However, most of these applied neural networks are classified as static neural networks, which are based on batch processes that update action only after the whole training data set has been presented. The time variate characteristics in hydrological processes have not been modelled well. In this paper, we present an alternative approach using an artificial neural network, termed real‐time recurrent learning (RTRL) for stream‐flow forecasting. To define the properties of the RTRL algorithm, we first compare the predictive ability of RTRL with least‐square estimated autoregressive integrated moving average models on several synthetic time‐series. Our results demonstrate that the RTRL network has a learning capacity with high efficiency and is an adequate model for time‐series prediction. We also investigated the RTRL network by using the rainfall–runoff data of the Da‐Chia River in Taiwan. The results show that RTRL can be applied with high accuracy to the study of real‐time stream‐flow forecasting networks. Copyright © 2002 John Wiley & Sons, Ltd.     

Estuary water-stage forecasting by using radial basis function neural network

The Radial basis function neural network (RBFNN) has been successfully applied to many tasks due to its powerful properties in classification and functional approximation. This paper presents a novel RBFNN for water-stage forecasting in an estuary under high flood and tidal effects. The RBFNN adopts a hybrid two-stage learning scheme, unsupervised and supervised learning. In the first scheme, fuzzy min–max clustering is proposed for choosing best patterns for cluster representation in an efficient and automatic way. The second scheme uses supervised learning, which is a multivariate linear regression method to produce a weighted sum of the output from the hidden layer. Since this network has only one layer using a supervised learning algorithm, its training process is much faster than the error back propagation based multilayer perceptrons. Moreover, only one parameter, θ, must be determined manually. The other parameters used in this model can be adjusted automatically by model training. The water-stage data of the Tanshui River under tidal effect are used to construct a water-stage forecasting model that can also be used during flood. The results show that the RBFNN can be applied successfully and provide high accuracy and reliability of water-stage forecasting in an estuary.     

Coupling the Xinanjiang model to a kinematic flow model based on digital drainage networks for flood forecasting

The Xinanjiang model, which is a conceptual rainfall‐runoff model and has been successfully and widely applied in humid and semi‐humid regions in China, is coupled by the physically based kinematic wave method based on a digital drainage network. The kinematic wave Xinanjiang model (KWXAJ) uses topography and land use data to simulate runoff and overland flow routing. For the modelling, the catchment is subdivided into numerous hillslopes and consists of a raster grid of flow vectors that define the water flow directions. The Xinanjiang model simulates the runoff yield in each grid cell, and the kinematic wave approach is then applied to a ranked raster network. The grid‐based rainfall‐runoff model was applied to simulate basin‐scale water discharge from an 805‐km² catchment of the Huaihe River, China. Rainfall and discharge records were available for the years 1984, 1985, 1987, 1998 and 1999. Eight flood events were used to calibrate the model's parameters and three other flood events were used to validate the grid‐based rainfall‐runoff model. A Manning's roughness via a linear flood depth relationship was suggested in this paper for improving flood forecasting. The calibration and validation results show that this model works well. A sensitivity analysis was further performed to evaluate the variation of topography (hillslopes) and land use parameters on catchment discharge. Copyright © 2009 John Wiley & Sons, Ltd.     

An analytical method of stage–fall–discharge rating

This paper presents an analytical method for establishing a stage–fall–discharge rating using hydraulic performance graphs (HPG). The rating curves derived from the HPG are used as the basis to establish the functional relation of stage, fall and discharge through regression analysis following the USGS procedure. In doing so, the conventional trial‐and‐error process can be avoided and the associated uncertainties involved may be reduced. For illustration, the proposed analytical method is applied to establish stage–fall–discharge relations for the Keelung River in northern Taiwan to examine its accuracy and applicability in an actual river. Based on the data extracted from the HPG for the Keelung River, one can establish a stage–fall–discharge relation that is more accurate than the one obtained by the conventionally used relation. Furthermore, the discharges obtained from the proposed rating method are verified through backwater analysis for measured high water level events. The results indicate that the analytical stage–fall–discharge rating method is capable of circumventing the shortcomings of those based on single‐station data and, consequently, enhancing the reliability of flood estimation and forecasting. Copyright © 2008 John Wiley & Sons, Ltd.     

Simulating and predicting river discharge time series using a wavelet‐neural network hybrid modelling approach

Accurate simulation and prediction of the dynamic behaviour of a river discharge over any time interval is essential for good watershed management. It is difficult to capture the high‐frequency characteristics of a river discharge using traditional time series linear and nonlinear model approaches. Therefore, this study developed a wavelet‐neural network (WNN) hybrid modelling approach for the predication of river discharge using monthly time series data. A discrete wavelet multiresolution method was employed to decompose the time series data of river discharge into sub‐series with low (approximation) and high (details) frequency, and these sub‐series were then used as input data for the artificial neural network (ANN). WNN models with different wavelet decomposition levels were employed to predict river discharge 48 months ahead of time. Comparison of results from the WNN models with those of the ANN models alone indicated that WNN models performed a more accurate prediction. Copyright © 2011 John Wiley & Sons, Ltd.     

Simultaneous hydrological prediction at multiple gauging stations using the NARX network for Kemaman catchment,Terengganu, Malaysia

ABSTRACT

This paper presents a neural network model capable of catchment-wide simultaneous prediction of river stages at multiple gauging stations. Thirteen meteorological parameters are considered in the input, which includes rainfall, temperature, mean relative humidity and evaporation. The NARX model is trained with a representative set of hourly data, with optimal time delay for both the input and output. The network trained using 120-day data is able to produce simulations that are in excellent agreement with field observations. We show that for application with one-step-ahead predictions, the loss in network performance is marginal. Inclusion of additional tidal observations does not improve predictions, suggesting that the river stage stations under consideration are not sensitive to tidal backwater effects despite the claim commonly made.

EDITOR D. Koutsoyiannis ASSOCIATE EDITOR F. Pappenberger     

Predicting and forecasting flow discharge at sites receiving significant lateral inflow

Two models, one linear and one non‐linear, were employed for the prediction of flow discharge hydrographs at sites receiving significant lateral inflow. The linear model is based on a rating curve and permits a quick estimation of flow at a downstream site. The non‐linear model is based on a multilayer feed‐forward back propagation (FFBP) artificial neural network (ANN) and uses flow‐stage data measured at the upstream and downstream stations. ANN predicted the real‐time storm hydrographs satisfactorily and better than did the linear model. The results of sensitivity analysis indicated that when the lateral inflow contribution to the channel reach was insignificant, ANN, using only the flow‐stage data at the upstream station, satisfactorily predicted the hydrograph at the downstream station. The prediction error of ANN increases exponentially with the difference between the peak discharge used in training and that used in testing. ANN was also employed for flood forecasting and was compared with the modified Muskingum model (MMM). For a 4‐h lead time, MMM forecasts the floods reliably but could not be applied to reaches for lead times greater than the wave travel time. Although ANN and MMM had comparable performances for an 8‐h lead time, ANN is capable of forecasting floods with lead times longer than the wave travel time. Copyright © 2007 John Wiley & Sons, Ltd.     

Improving the forecasts of extreme streamflow by support vector regression with the data extracted by self‐organizing map

During typhoons or storms, accurate forecasts of hourly streamflow are necessary for flood warning and mitigation. However, hourly streamflow is difficult to forecast because of the complex physical process and the high variability in time. Furthermore, under the global warming scenario, events with extreme streamflow may occur that leads to more difficulties in forecasting streamflows. Hence, to obtain more accurate hourly streamflow forecasts, an improved streamflow forecasting model is proposed in this paper. The computational kernel of the proposed model is developed on the basis of support vector machine (SVM). Additionally, self‐organizing map (SOM) is used to analyse observed data to extract data with specific properties, which are capable of providing valuable information for streamflow forecasting. After reprocessing, these extracted data and the observed data are used to construct the SVM‐based model. An application is conducted to clearly demonstrate the advantage of the proposed model. The comparison between the proposed model and the conventional SVM model, which is constructed without SOM, is performed. The results indicate that the proposed model is better performed than the conventional SVM model. Moreover, as regards the extreme events, the result shows that the proposed model reduces the forecasting error, especially the error of peak streamflow. It is confirmed that because of the use of data extracted by SOM, the improved forecasting performance is obtained. The proposed model, which can produce accurate forecasts, is expected to be useful to support flood warning systems. Copyright © 2012 John Wiley & Sons, Ltd.     

An RBF‐based model with an information processor for forecasting hourly reservoir inflow during typhoons

Inflow forecasting is essential for decision making on reservoir operation during typhoons. In this paper, a radial basis function (RBF)‐based model with an information processor is proposed for more accurate forecasts of hourly reservoir inflow. Firstly, based on the multilayer perceptron neural (MLP) network, an information processor is developed to pre‐process the typhoon information (namely, typhoon characteristics and rainfall) and to produce forecasts of rainfall. The forecasted rainfall and the observed inflow are then used as input to the RBF‐based model, which is a nonlinear function approximator, to produce forecasts of hourly inflow. For parameter estimation of the RBF‐based model, the fully‐supervised learning algorithm is used. Actual applications of the proposed model are performed to yield 1‐ to 6‐h ahead forecasts of inflow. To assess the improvement due to the use of the typhoon information processor, models without the typhoon information processor are constructed and compared with the proposed model. The results show that the proposed model performs the best and is capable of providing improved forecasts of hourly inflow, especially for long lead‐time. In conclusion, the proposed model with a typhoon information processor can extract useful information from typhoon characteristics and rainfall, and consequently improve the forecasting performance. Copyright © 2009 John Wiley & Sons, Ltd.     

Estimating river discharge from very high‐resolution satellite data: a case study in the Yangtze River,China

The measurement of river discharge is necessary for understanding many water‐related issues. Traditionally, river discharge is estimated by measuring water stage and converting the measurement to discharge by using a stage–discharge rating curve. Our proposed method for the first time couples the measurement of water‐surface width with river width–stage and stage–discharge rating curves by using very high‐resolution satellite data. We used it to estimate the discharge in the Yangtze (Changjiang) River as a case study. The discharges estimated at four stations from five QuickBird‐2 images matched the ground observation data very well, demonstrating that the proposed approach can be regarded as ancillary to traditional field measurement methods or other remote methods to estimate river discharge. Copyright © 2004 John Wiley & Sons, Ltd.     

A design‐variable‐based inelastic hysteretic model for beam–column connections

This paper presents a design‐variable‐based inelastic hysteretic model for beam–column connections. It has been well known that the load‐carrying capacity of connections heavily depends on the types and design variables even in the same connection type. Although many hysteretic connection models have been proposed, most of them are dependent on the specific connection type with presumed failure mechanisms. The proposed model can be responsive to variations both in design choices and in loading conditions. The proposed model consists of two modules: physical‐principle‐based module and neural network (NN)‐based module in which information flow from design space to response space is formulated in one complete model. Moreover, owing to robust learning capability of a new NN‐based module, the model can also learn complex dynamic evolutions in response space under earthquake loading conditions, such as yielding, post‐buckling and tearing, etc. Performance of the proposed model has been demonstrated with synthetic and experimental data of two connection types: extended‐end‐plate and top‐ and seat‐angle with double‐web‐angle connection. Furthermore, the design‐variable‐based model can be customized to any structural component beyond the application to beam–column connections. Copyright © 2007 John Wiley & Sons, Ltd.     

Impacts of uncertain river flow data on rainfall‐runoff model calibration and discharge predictions

In order to quantify total error affecting hydrological models and predictions, we must explicitly recognize errors in input data, model structure, model parameters and validation data. This paper tackles the last of these: errors in discharge measurements used to calibrate a rainfall‐runoff model, caused by stage–discharge rating‐curve uncertainty. This uncertainty may be due to several combined sources, including errors in stage and velocity measurements during individual gaugings, assumptions regarding a particular form of stage–discharge relationship, extrapolation of the stage–discharge relationship beyond the maximum gauging, and cross‐section change due to vegetation growth and/or bed movement. A methodology is presented to systematically assess and quantify the uncertainty in discharge measurements due to all of these sources. For a given stage measurement, a complete PDF of true discharge is estimated. Consequently, new model calibration techniques can be introduced to explicitly account for the discharge error distribution. The method is demonstrated for a gravel‐bed river in New Zealand, where all the above uncertainty sources can be identified, including significant uncertainty in cross‐section form due to scour and re‐deposition of sediment. Results show that rigorous consideration of uncertainty in flow data results in significant improvement of the model's ability to predict the observed flow. Copyright © 2010 John Wiley & Sons, Ltd.     

Real-time probabilistic sediment concentration forecasting using integrated dynamic network and error distribution heterogeneity

Sediment forecasting at a dam site is important for the operation and management of water and sediment in a reservoir. However, the forecast results generally have some uncertainties, which may hinder the operation of the dam. In this study, a real-time sediment concentration probabilistic forecasting model is proposed based on a dynamic network model. Under this framework, the Elman neural network (ENN) and nonlinear auto-regressive with exogenous inputs (NARX) neural network models were established for sediment concentration forecasting with different lead times. A hybrid algorithm, which combined the Levenberg–Marquardt algorithm and real-time recurrent learning, was used to train the model. Using the aforementioned method, the sediment concentration was forecast for at the Sanmenxia Dam, China, and, subsequently, the forecast results were evaluated. Among the selected lead time, the results at 5 h exhibited the highest accuracy and practical significance. Compared with the ENN model, the sediment concentration peak error using the NARX neural network was reduced by 4.5%, and the sediment yield error was reduced by 0.043%. Therefore, the NARX neural network was selected as the deterministic sediment forecasting model. Additionally, the probability density function of the sediment concentration was derived based on the heterogeneity of the error distribution, and the sediment concentration interval, with different confidence levels, expected values, and median values, was forecast. The Nash–Sutcliffe coefficient of efficiency for the sediment concentration, as forecasted based on the median value, was the highest (0.04 higher than that using a deterministic model), whereas the error of the sediment concentration peak and sediment yield remained unaltered. These results indicated the accuracy and superiority of the proposed real-time sediment probabilistic forecasting hybrid model.     

Multiple‐station neural network for modelling tidal currents across Shinnecock inlet,USA

This paper presents the development of a multiple‐station neural network for predicting tidal currents across a coastal inlet. Unlike traditional hydrodynamic models, the neural network model does not need inputs of coastal topography and bathymetry, grids, surface and bottom frictions, and turbulent eddy viscosity. Without solving hydrodynamic equations, the neural network model applies an interconnected neural network to correlate the inputs of boundary forcing of water levels at a remote station to the outputs of tidal currents at multiple stations across a local coastal inlet. Coefficients in the neural network model are trained using a continuous dataset consisting of inputs of water levels at a remote station and outputs of tidal currents at the inlet, and verified using another independent input and output dataset. Once the neural network model has been satisfactorily trained and verified, it can be used to predict tidal currents at a coastal inlet from the inputs of water levels at a remote station. For the case study at Shinnecock Inlet in the southern shore of New York, tidal currents at nine stations across the inlet were predicted by the neural network model using water level data located from a station about 70 km away from the inlet. A continuous dataset in May 2000 was used for the training, and another dataset in July 2000 was used for the verification of the neural network model. Comparing model predictions and observations indicates correlation coefficients range from 0·95 to 0·98, and the root‐mean‐square error ranges from 0·04 to 0·08 m s^?1 at the nine current locations across the inlet. Copyright © 2007 John Wiley & Sons, Ltd.     

Comparing the performance of empirical black‐box models for river flow forecasting in the Heihe River Basin,Northwestern China

For many practical reasons, the empirical black‐box models have become an increasingly popular modelling tool for river flow forecasting, especially in mountainous areas where very few meteorological observatories exist. In this article, precipitation data are used as the only input to estimate river flow. Using five empirical black‐box models—the simple linear model, the linear perturbation model, the linearly varying gain factor model, the constrained nonlinear system model and the nonlinear perturbation model–antecedent precipitation index—modelling results are compared with actual results in three catchments within the Heihe River Basin. The linearly varying gain factor model and the nonlinear perturbation model yielded excellent predictions. For better simulation accuracy, a commonly used multilayer feed‐forward neural network model (NNM) was applied to incorporate the outputs of the individual models. Comparing the performance of these models, it was found that the best results were obtained from the NNM model. The results also suggest that more reliable and precise predictions of river flow can be obtained by using the NNM model while also incorporating the combined outputs of different empirical black‐box models. Copyright © 2012 John Wiley & Sons, Ltd.