New fuzzy neural network–Markov model and application in mid- to long-term runoff forecast

ABSTRACT

In this paper, a mid- to long-term runoff forecast model is developed using an ideal point fuzzy neural network–Markov (NFNN-MKV) hybrid algorithm to improve the forecasting precision. Combining the advantages of the new fuzzy neural network and the Markov prediction model, this model can solve the problem of stationary or volatile strong random processes. Defined error statistics algorithms are used to evaluate the performance of models. A runoff prediction for the Si Quan Reservoir is made by utilizing the modelling method and the historical runoff data, with a comprehensive consideration of various runoff-impacting factors such as rainfall. Compared with the traditional fuzzy neural networks and Markov prediction models, the results show that the NFNN-MKV hybrid algorithm has good performance in faster convergence, better forecasting accuracy and significant improvement of neural network generalization. The absolute percentage error of the NFNN-MKV hybrid algorithm is less than 7.0%, MSE is less than 3.9, and qualification rate reaches 100%. For further comparison of the proposed model, the NFNN-MKV model is employed to estimate (training and testing for 120-month-ahead prediction) and predict river discharge for 156 months at Weijiabao on the Weihe River in China. Comparisons among the results of the NFNN-MKV model, the WNN model and the SVR model indicate that the NFNN-MKV model is able to significantly increase prediction accuracy.

Editor D. Koutsoyiannis; Associate editor Y. Gyasi-Agyei     

Long‐lead seasonal rainfall forecasting using time‐delay recurrent neural networks: a case study

A temporal artificial neural network‐based model is developed and applied for long‐lead rainfall forecasting. Tapped delay lines and recurrent connections are two different components that are used along with a static multilayer perceptron network to design a time‐delay recurrent neural network. The proposed model is, in fact, a combination of time‐delay and recurrent neural networks. The model is applied in three case studies of the Northwest, West, and Southwest basins of Iran. In addition, an autoregressive moving average with exogenous inputs (ARMAX) model is used as a baseline in order to be compared with the time‐delay recurrent neural networks developed in this study. Large‐scale climate signals, such as sea‐level pressure, that affect the rainfall of the study area are used as the predictors in the models, as well as the persistence between rainfall data. The results of winter‐spring rainfall forecasts are discussed thoroughly. It is demonstrated that in all cases the proposed neural network results in better forecasts in comparison with the statistical ARMAX model. Moreover, it is found that in two of three case studies the time‐delay recurrent neural networks perform better than either recurrent or time‐delay neural networks. The results demonstrate that the proposed method can significantly improve the long‐lead forecast by utilizing a non‐linear relationship between climatic predictors and rainfall in a region. Copyright © 2007 John Wiley & Sons, Ltd.     

RESEARCH ON FORECASTING STORM DEBRIS FLOWS

１ＩＮＴＲＯＤＵＣＴＩＯＮＢａｓｅｄｏｎｔｈｅｍｏｕｎｔａｉｎｓｔｒｅａｍｃｌａｓｉｆｉｃａｔｉｏｎａｎｄｈａｚａｒｄｚｏｎｅｍａｐｐｉｎｇ（Ｗａｎｇｅｔａｌ,１９９６;ａｎｄＷａｎｇｅｔａｌ,１９９８）,ａｓｗｅｌａｓｔｈｅｉｎｖｅｓｔｉｇａｔｉｏ...     

Performance of an artificial neural network on forecasting the daily occurrence and annual depth of rainfall at a tropical site

Performance of a feed‐forward back‐propagation artificial neural network on forecasting the daily occurrence and annual depth of rainfall at a single meteorological station is presented. Both short‐term and long‐term forecasting was attempted, with ground level data collected by the meteorological station in Colombo, Sri Lanka (79° 52′E, 6° 54′N) during two time periods, 1994–2003 and 1869–2003. Two neural network models were developed; a one‐day‐ahead model for predicting the rainfall occurrence of the next day, which was able to make predictions with a 74·3% accuracy, and one‐year‐ahead model for yearly rainfall depth predictions with an 80·0% accuracy within a ± 5% error bound. Each of these models was extended to make predictions several time steps into the future, where accuracies were found to decrease rapidly with the number of time steps. The success rates and rainfall variability within the north‐east and south‐west monsoon seasons are also discussed. Copyright © 2007 John Wiley & Sons, Ltd.     

Design and evaluation of SVR,MARS and M5Tree models for 1, 2 and 3-day lead time forecasting of river flow data in a semiarid mountainous catchment

Accurate and reliable river flow forecasts attained with data-intelligent models can provide significant information about future water resources management. In this study we employed a 50-model ensemble of three data-driven predictive models, namely the support vector regression (SVR), multivariate adaptive regression spline (MARS) and M5 model tree (M5Tree) to forecast river flow data in a semiarid and ecologically significant mountainous region of Pailugou catchment in northwestern China. To attain stable and accurate forecast results, 50 different models were trained by randomly sampling the entire river flow data into 80% for training and 20% for testing subsets. To attain a complete evaluation of the ensemble-model based results, the global mean of six quantitative statistical performance evaluation measures: the coefficient of correlation (R), mean absolute relative error (MAE), root mean squared error (RMSE), Nash–Sutcliffe efficiency coefficient (NS), relative RMSE, and the Willmott’s Index (WI), and Taylor diagrams, including skill scores relative to a persistence model, were selected to assess the performances of the developed predictive models. The results indicated that all of the averaged R value attained was higher than 0.900 and all of the averaged NS values were higher than 0.800, representing good performance of the SVR, MARS and M5Tree models applied in the 1-, 2- and 3-day ahead modeling horizon, and this also accorded with the deductions made through an assessment of the Willmott’s Index. However, the M5Tree model outperformed both the SVR and MARS models (with NS?=?0.917 vs. 0.904 and 0.901 for 1-day, 0.893 vs. 0.854 and 0.845 for 2-day, and 0.850 vs. 0.828 and 0.810 for 3-day forecasting horizons, respectively), which was in concurrence with the high value of WI. Therefore, based on the ensemble of 50 models, the performance of the M5Tree can be considered as superior to the SVR and MARS models when applied in a problem of river flow forecasting at multiple forecast horizon. A detailed comparison of the overall performance of all three models evaluated through Taylor diagrams and boxplots indicated that the 1-day ahead forecasting results were more accurate for all of the predictive models compared to the 2- and 3-day ahead forecasting horizons. Data-intelligent models designed in this study indicate that the M5Tree method could successfully be explored for short-term river flow forecasting in semiarid mountainous regions, which may have useful implications in water resources management, ecological sustainability and assessment of river systems.     

Improving flood forecasting using an input correction method in urban models in poorly gauged areas

ABSTRACT

Poorly monitored catchments could pose a challenge in the provision of accurate flood predictions by hydrological models, especially in urbanized areas subject to heavy rainfall events. Data assimilation techniques have been widely used in hydraulic and hydrological models for model updating (typically updating model states) to provide a more reliable prediction. However, in the case of nonlinear systems, such procedures are quite complex and time-consuming, making them unsuitable for real-time forecasting. In this study, we present a data assimilation procedure, which corrects the uncertain inputs (rainfall), rather than states, of an urban catchment model by assimilating water-level data. Five rainfall correction methods are proposed and their effectiveness is explored under different scenarios for assimilating data from one or multiple sensors. The methodology is adopted in the city of São Carlos, Brazil. The results show a significant improvement in the simulation accuracy.     

Uncertainty based modeling of rainfall-runoff: Combined differential evolution adaptive Metropolis (DREAM) and K-means clustering

Simulation of rainfall-runoff process in urban areas is of great importance considering the consequences and damages of extreme runoff events and floods. The first issue in flood hazard analysis is rainfall simulation. Large scale climate signals have been proved to be effective in rainfall simulation and prediction. In this study, an integrated scheme is developed for rainfall-runoff modeling considering different sources of uncertainty. This scheme includes three main steps of rainfall forecasting, rainfall-runoff simulation and future runoff prediction. In the first step, data driven models are developed and used to forecast rainfall using large scale climate signals as rainfall predictors. Due to high effect of different sources of uncertainty on the output of hydrologic models, in the second step uncertainty associated with input data, model parameters and model structure is incorporated in rainfall-runoff modeling and simulation. Three rainfall-runoff simulation models are developed for consideration of model conceptual (structural) uncertainty in real time runoff forecasting. To analyze the uncertainty of the model structure, streamflows generated by alternative rainfall-runoff models are combined, through developing a weighting method based on K-means clustering. Model parameters and input uncertainty are investigated using an adaptive Markov Chain Monte Carlo method. Finally, calibrated rainfall-runoff models are driven using the forecasted rainfall to predict future runoff for the watershed. The proposed scheme is employed in the case study of the Bronx River watershed, New York City. Results of uncertainty analysis of rainfall-runoff modeling reveal that simultaneous estimation of model parameters and input uncertainty significantly changes the probability distribution of the model parameters. It is also observed that by combining the outputs of the hydrological models using the proposed clustering scheme, the accuracy of runoff simulation in the watershed is remarkably improved up to 50% in comparison to the simulations by the individual models. Results indicate that the developed methodology not only provides reliable tools for rainfall and runoff modeling, but also adequate time for incorporating required mitigation measures in dealing with potentially extreme runoff events and flood hazard. Results of this study can be used in identification of the main factors affecting flood hazard analysis.     

A review of advances in flash flood forecasting

Flash flooding is one of the most hazardous natural events, and it is frequently responsible for loss of life and severe damage to infrastructure and the environment. Research into the use of new modelling techniques and data types in flash flood forecasting has increased over the past decade, and this paper presents a review of recent advances that have emerged from this research. In particular, we focus on the use of quantitative precipitation estimates and forecasts, the use of remotely sensed data in hydrological modelling, developments in forecasting models and techniques, and uncertainty estimates. Over the past decade flash flood forecast lead‐time has expanded up to six hours due to improved rainfall forecasts. However the largest source of uncertainty of flash flood forecasts remains unknown future precipitation. An increased number of physically based hydrological models have been developed and used for flash flood forecasting and they have been found to give more plausible results when compared with the results of conceptual, statistical, and neural network models. Among the three methods for deciding flash flood occurrence discussed in this review, the rainfall comparison method (flash flood guidance) is most commonly used for flash flood forecasting as it is easily understood by the general public. Unfortunately, no existing model is capable of making reliable flash flood forecasts in urban watersheds even though the incidence of urban flash flooding is increasing due to increasing urbanisation. Copyright © 2011 John Wiley & Sons, Ltd.     

Adaptive Neuro-Fuzzy Inference System for drought forecasting

Drought causes huge losses in agriculture and has many negative influences on natural ecosystems. In this study, the applicability of Adaptive Neuro-Fuzzy Inference System (ANFIS) for drought forecasting and quantitative value of drought indices, the Standardized Precipitation Index (SPI), is investigated. For this aim, 10 rainfall gauging stations located in Central Anatolia, Turkey are selected as study area. Monthly mean rainfall and SPI values are used for constructing the ANFIS forecasting models. For all stations, data sets include a total of 516 data records measured between in 1964 and 2006 years and data sets are divided into two subsets, training and testing. Different ANFIS forecasting models for SPI at time scales 1–12 months were trained and tested. The results of ANFIS forecasting models and observed values are compared and performances of models were evaluated. Moreover, the best fit models have been also trained and tested by Feed Forward Neural Networks (FFNN). The results demonstrate that ANFIS can be successfully applied and provide high accuracy and reliability for drought forecasting.     

Real-time flood forecasting with the use of inadequate data

Abstract

A modelling scheme is developed for real-time flood forecasting. It is composed of (a) a rainfall forecasting model, (b) a conceptual rainfall-runoff model, and (c) a stochastic error model of the ARMA family for forecast error correction. Initialization of the rainfall-runoff model is based on running this model on a daily basis for a certain period prior to the flood onset while parameters of the error model are updated through the Recursive Least Squares algorithm. The scheme is suitable for the early stages of operation of flood forecasting systems in the presence of inadequate historical data. A validation framework is set up which simulates real-time flood forecasting conditions. Thus, the effects of the procedures for rainfall-runoff model initialization, forecast error correction and rainfall forecasting are assessed. Two well-known conceptual rainfall-runoff models (the Soil Moisture Accounting model of the US National Weather Service River Forecast Service—SMA-NWSRFS and TANK) together with data from a Greek basin are used for illustration purposes.     

Hydroclimate variability and long-lead forecasting of rainfall over Thailand by large-scale atmospheric variables

Abstract

The development of statistical relationships between local hydroclimates and large-scale atmospheric variables enhances the understanding of hydroclimate variability. The rainfall in the study basin (the Upper Chao Phraya River Basin, Thailand) is influenced by the Indian Ocean and tropical Pacific Ocean atmospheric circulation. Using correlation analysis and cross-validated multiple regression, the large-scale atmospheric variables, such as temperature, pressure and wind, over given regions are identified. The forecasting models using atmospheric predictors show the capability of long-lead forecasting. The modified k-nearest neighbour (k-nn) model, which is developed using the identified predictors to forecast rainfall, and evaluated by likelihood function, shows a long-lead forecast of monsoon rainfall at 7–9 months. The decreasing performance in forecasting dry-season rainfall is found for both short and long lead times. The developed model also presents better performance in forecasting pre-monsoon season rainfall in dry years compared to wet years, and vice versa for monsoon season rainfall.

Editor Z.W. Kundzewicz

Citation Singhrattna, N., Babel, M.S. and Perret, S.R., 2012. Hydroclimate variability and long-lead forecasting of rainfall over Thailand by large-scale atmospheric variables. Hydrological Sciences Journal, 57 (1), 26–41.     

BIBLIOGRAPHIE—BIBLIOGRAPHY

ABSTRACT

With the increasing use of telemetry in the control of water resource systems, a considerable amount of effort is being devoted to the development of models and parameter estimation techniques for on-line use. A variety of models and parameter estimation algorithms have been considered, ranging from complex conceptual models of the soil moisture accounting type, which are traditionally calibrated off-line, to state-space/Kalman filter models which, perhaps, have enjoyed undue popularity in the recent literature due to their mathematical elegance. The fundamental assumptions underlying the various approaches are reviewed, and the validity of these assumptions in the hydrological forecasting context is assessed. The paper draws on some results obtained during a recent workshop at the Institute of Hydrology in making assessments of the relative merits of different models and parameter estimation algorithms; these results have been derived from an intercomparison of a number of real time forecasting models.     

Quantitative assessment of short‐term rainfall forecasts from radar nowcasts and MM5 forecasts

Short‐term Quantitative Precipitation Forecasts (QPFs) can be achieved from numerical weather prediction (NWP) models or radar nowcasting, that is the extrapolation of the precipitation at a future time from consecutive radar scans. Hybrid forecasts obtained by merging rainfall forecasts from radar nowcasting and NWP models are potentially more skilful than either radar nowcasts or NWP rainfall forecasts alone. This paper provides an assessment of deterministic and probabilistic high‐resolution QPFs achieved by implementing the Short‐term Ensemble Prediction System developed by the UK Met Office. Both radar nowcasts and hybrid forecasts have been performed. The results show that the performance of both deterministic nowcasts and deterministic hybrid forecasts decreases with increasing rainfall intensity and spatial resolution. The results also show that the blending with the NWP forecasts improves the performance of the forecasting system. Probabilistic hybrid forecasts have been obtained through the modelling of a stochastic noise component to produce a number of equally likely ensemble members, and the comparative assessment of deterministic and probabilistic hybrid forecasts shows that the probabilistic forecasting system is characterised by a higher discrimination accuracy than the deterministic one. Copyright © 2011 John Wiley & Sons, Ltd.     

Comparison of daily streamflow forecasts using extreme learning machines and the random forest method

ABSTRACT

Daily streamflow forecasting is a challenging and essential task for water resource management. The main goal of this study was to compare the accuracy of five data-driven models: extreme learning machine (basic ELM), extreme learning machine with kernels (ELM-kernel), random forest (RF), back-propagation neural network (BPNN) and support vector machine (SVR). The results show that the ELM-kernel model provided a superior alternative to the other models, and the basic ELM model had the poorest performance. To further evaluate the predictive capacities of the five models, the estimations of low flow and high flow in the testing dataset were compared. The RF model was slightly superior to the other models in predicting the peak flows, and the ELM-kernel model showed the highest prediction precision of low flows. There was no single model that showed obvious advantages over the other models in this study. Therefore, further exploration is required for the hydrological forecasting problems.     

Multiple criteria rainfall–runoff model calibration using a parallel genetic algorithm in a cluster of computers / Calage multi-critères en modélisation pluie–débit par un algorithme génétique parallèle mis en œuvre par une grappe d'ordinateurs

Abstract

Genetic algorithms are among of the global optimization schemes that have gained popularity as a means to calibrate rainfall–runoff models. However, a conceptual rainfall–runoff model usually includes 10 or more parameters and these are interdependent, which makes the optimization procedure very time-consuming. This may result in the premature termination of the optimization process which will prejudice the quality of the results. Therefore, the speed of optimization procedure is crucial in order to improve the calibration quality and efficiency. A hybrid method that combines a parallel genetic algorithm with a fuzzy optimal model in a cluster of computers is proposed. The method uses the fuzzy optimal model to evaluate multiple alternatives with multiple criteria where chromosomes are the alternatives, whilst the criteria are flood performance measures. In order to easily distinguish the performance of different alternatives and to address the problem of non-uniqueness of optimum, two fuzzy ratios are defined. The new approach has been tested and compared with results obtained by using a two-stage calibration procedure. The current single procedure produces similar results, but is simpler and automatic. Comparison of results between the serial and parallel genetic algorithms showed that the current methodology can significantly reduce the overall optimization time and simultaneously improve the solution quality.     

Ensemble rainfall forecasting with numerical weather prediction and radar‐based nowcasting models

The overall objective of this study is to improve the forecasting accuracy of the precipitation in the Singapore region by means of both rainfall forecasting and nowcasting. Numerical Weather Predication (NWP) and radar‐based rainfall nowcasting are two important sources for quantitative precipitation forecast. In this paper, an attempt to combine rainfall prediction from a high‐resolution mesoscale weather model and a radar‐based rainfall model was performed. Two rainfall forecasting methods were selected and examined: (i) the weather research and forecasting model (WRF); and (ii) a translation model (TM). The WRF model, at a high spatial resolution, was run over the domain of interest using the Global Forecast System data as initializing fields. Some heavy rainfall events were selected from data record and used to test the forecast capability of WRF and TM. Results obtained from TM and WRF were then combined together to form an ensemble rainfall forecasting model, by assigning weights of 0.7 and 0.3 weights to TM and WRF, respectively. This paper presented results from WRF and TM, and the resulting ensemble rainfall forecasting; comparisons with station data were conducted as well. It was shown that results from WRF are very useful as advisory of anticipated heavy rainfall events, whereas those from TM, which used information of rain cells already appearing on the radar screen, were more accurate for rainfall nowcasting as expected. The ensemble rainfall forecasting compares reasonably well with the station observation data. Copyright © 2012 John Wiley & Sons, Ltd.     

Effective typhoon characteristics and their effects on hourly reservoir inflow forecasting

This paper describes the identification of effective typhoon characteristics and the development of a new type of hourly reservoir inflow forecasting model with the effective typhoon characteristics. Firstly, a comparison of support vector machines (SVMs), which is a novel kind of neural networks (NNs), and back-propagation networks (BPNs) is made to select an appropriate NN-based model. The results show that SVM-based models are more appropriate than BPN-based models because of their higher accuracy and much higher efficiency. In addition, effective typhoon characteristics for improving forecasting performance are identified from all the collected typhoon information. Then the effective typhoon characteristics (the position of the typhoon and the distance between the typhoon center and the reservoir) are added to the proposed SVM-based models. Next, a performance comparison of models with and without effective typhoon characteristics is conducted to clearly highlight the effects of effective typhoon characteristics on hourly reservoir inflow forecasting. To reach a just conclusion, the performance is evaluated by cross validation, and the improvement in performance due to the addition of effective typhoon characteristics is tested by paired comparison t-tests at the 5% significance level. The results confirm that effective typhoon characteristics do improve the forecasting performance and the improvement increases with increasing lead-time, especially when the rainfall data are not available. For four- to six-hour ahead forecasts, the improvement due to the addition of effective typhoon characteristics increases from 3% to 18% and from 10% to 113% for Categories I (rainfall data are available) and II (rainfall data are not available), respectively. In conclusion, effective typhoon characteristics are recommended as key inputs for reservoir inflow forecasting during typhoons. The proposed SVM-based models with effective typhoon characteristics are expected to provide more accurate forecasts than BPN-based models. The proposed modeling technique is also expected to be useful to support reservoir operation systems and other disaster warning systems.     

降雨输入对实时洪水预报精度与预见期的影响

针对降雨输入不确定性对实时洪水预报影响的问题,本文采用不考虑未来预报降雨、考虑未来预报降雨、考虑预报降雨的降雨量误差和降雨时间误差4种方法,以陕西省两个半湿润流域（陈河流域和大河坝流域）为研究区域,分析不同预见期和不同降雨输入情况下洪水预报的精度.研究表明：相对于不考虑未来降雨情况,考虑未来降雨后在预报预见期较长时对预报结果精度提升较大,在预见期较短时对预报结果精度提升不显著;暴雨中心位置不同对预报精度影响也不同,当暴雨中心位于流域下游时降雨量误差对流量预报误差影响更大;降雨量误差主要影响洪量相对误差和洪峰相对误差,且这种影响是线性的,对确定性系数的影响是非线性的二次函数,降雨时间误差主要影响峰现时间误差.     

Multi-scale support vector algorithms for hot spot detection and modelling

The algorithmic approach to data modelling has developed rapidly these last years, in particular methods based on data mining and machine learning have been used in a growing number of applications. These methods follow a data-driven methodology, aiming at providing the best possible generalization and predictive abilities instead of concentrating on the properties of the data model. One of the most successful groups of such methods is known as Support Vector algorithms. Following the fruitful developments in applying Support Vector algorithms to spatial data, this paper introduces a new extension of the traditional support vector regression (SVR) algorithm. This extension allows for the simultaneous modelling of environmental data at several spatial scales. The joint influence of environmental processes presenting different patterns at different scales is here learned automatically from data, providing the optimum mixture of short and large-scale models. The method is adaptive to the spatial scale of the data. With this advantage, it can provide efficient means to model local anomalies that may typically arise in situations at an early phase of an environmental emergency. However, the proposed approach still requires some prior knowledge on the possible existence of such short-scale patterns. This is a possible limitation of the method for its implementation in early warning systems. The purpose of this paper is to present the multi-scale SVR model and to illustrate its use with an application to the mapping of Cs¹³⁷ activity given the measurements taken in the region of Briansk following the Chernobyl accident.     

Merging multiple precipitation sources for flash flood forecasting

We investigated the effectiveness of combining gauge observations and satellite-derived precipitation on flood forecasting. Two data merging processes were proposed: the first one assumes that the individual precipitation measurement is non-bias, while the second process assumes that each precipitation source is biased and both weighting factor and bias parameters are to be calculated. Best weighting factors as well as the bias parameters were calculated by minimizing the error of hourly runoff prediction over Wu-Tu watershed in Taiwan. To simulate the hydrologic response from various sources of rainfall sequences, in our experiment, a recurrent neural network (RNN) model was used.

The results demonstrate that the merged method used in this study can efficiently combine the information from both rainfall sources to improve the accuracy of flood forecasting during typhoon periods. The contribution of satellite-based rainfall, being represented by the weighting factor, to the merging product, however, is highly related to the effectiveness of ground-based rainfall observation provided gauged. As the number of gauge observations in the basin is increased, the effectiveness of satellite-based observation to the merged rainfall is reduced. This is because the gauge measurements provide sufficient information for flood forecasting; as a result the improvements added on satellite-based rainfall are limited. This study provides a potential advantage for extending satellite-derived precipitation to those watersheds where gauge observations are limited.