Advances in Streamflow Forecasting From Traditional to Modern Approaches
Advances in Streamflow Forecasting: From Traditional to Modern Approaches covers the three major data-driven approaches of streamflow forecasting including traditional approach of statistical and stochastic time-series modelling with their recent developments, stand-alone data-driven approach such a...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | eBook |
| Language: | English |
| Published: |
Chantilly
Elsevier
2021
|
| Edition: | 1 |
| Subjects: | |
| ISBN: | 9780128206737, 012820673X |
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Table of Contents:
- Front Cover -- Advances in Streamflow Forecasting -- Advances in Streamflow Forecasting: From Traditional to Modern Approaches -- Copyright -- Dedication -- Contents -- Contributors -- About the editors -- Foreword -- Preface -- Acknowledgment -- 1 - Streamflow forecasting: overview of advances in data-driven techniques -- 1.1 Introduction -- 1.2 Measurement of streamflow and its forecasting -- 1.3 Classification of techniques/models used for streamflow forecasting -- 1.4 Growth of data-driven methods and their applications in streamflow forecasting -- 1.4.1 Time series modeling -- 1.4.2 Artificial neural network -- 1.4.3 Other AI techniques -- 1.4.4 Hybrid data-driven techniques -- 1.5 Comparison of different data-driven techniques -- 1.6 Current trends in streamflow forecasting -- 1.7 Key challenges in forecasting of streamflows -- 1.8 Concluding remarks -- References -- 2 - Streamflow forecasting at large time scales using statistical models -- 2.1 Introduction -- 2.2 Overview of statistical models used in forecasting -- 2.2.1 Forecasting in general -- 2.2.1.1 ARIMA models -- 2.2.1.2 Exponential smoothing models -- 2.2.1.3 General literature -- 2.2.1.4 Literature in hydrology -- 2.3 Theory -- 2.3.1 ARIMA models -- 2.3.1.1 Definition -- 2.3.1.2 Forecasting with ARIMA models -- 2.3.2 Exponential smoothing models -- 2.4 Large-scale applications at two time scales -- 2.4.1 Application 1: multi-step ahead forecasting of 270 time series of annual streamflow -- 2.4.2 Application 2: multi-step ahead forecasting of 270 time series of monthly streamflow -- 2.5 Conclusions -- Conflicts of interest -- Acknowledgment -- References -- 3 - Introduction of multiple/multivariate linear and nonlinear time series models in forecasting streamflow process -- 3.1 Introduction -- 3.1.1 Review of MLN time series models -- 3.2 Methodology -- 3.2.1 VAR/VARX model
- 8.5.3 Model calibration and validation -- 8.5.4 Sensitivity analysis of model configurations towards model performance -- 8.5.4.1 Influence of input variable combinations -- 8.5.4.2 Influence of model tree variants -- 8.5.4.3 Influence of data proportioning -- 8.5.5 Selection of best-fit model for streamflow forecasting -- 8.6 Summary and conclusions -- Acknowledgments -- References -- 9 - Averaging multiclimate model prediction of streamflow in the machine learning paradigm -- 9.1 Introduction -- 9.2 Salient review on ANN and SVR modeling for streamflow forecasting -- 9.3 Averaging streamflow predicted from multiclimate models in the neural network framework -- 9.4 Averaging streamflow predicted by multiclimate models in the framework of support vector regression -- 9.5 Machine learning-averaged streamflow from multiple climate models: two case studies -- 9.6 Conclusions -- References -- 10 - Short-term flood forecasting using artificial neural networks, extreme learning machines, and M5 model tree -- 10.1 Introduction -- 10.2 Theoretical background -- 10.2.1 Artificial neural networks -- 10.2.2 Extreme learning machines -- 10.2.3 M5 model tree -- 10.3 Application of ANN, ELM, and M5 model tree techniques in hourly flood forecasting: a case study -- 10.3.1 Study area and data -- 10.3.2 Methodology -- 10.4 Results and discussion -- 10.5 Conclusions -- References -- 11 - A new heuristic model for monthly streamflow forecasting: outlier-robust extreme learning machine -- 11.1 Introduction -- 11.2 Overview of extreme learning machine and multiple linear regression -- 11.2.1 Extreme learning machine model and its extensions -- 11.2.2 Multiple linear regression -- 11.3 A case study of forecasting streamflows using extreme machine learning models -- 11.3.1 Study area -- 11.4 Applications and results -- 11.5 Conclusions -- References
- 12 - Hybrid artificial intelligence models for predicting daily runoff -- 12.1 Introduction -- 12.2 Theoretical background of MLP and SVR models -- 12.2.1 Support vector regression model -- 12.2.2 Multilayer perceptron neural network model -- 12.2.3 Grey wolf optimizer algorithm -- 12.2.4 Whale optimization algorithm -- 12.2.5 Hybrid MLP neural network model -- 12.2.6 Hybrid SVR model -- 12.3 Application of hybrid MLP and SVR models in runoff prediction: a case study -- 12.3.1 Study area and data acquisition -- 12.3.2 Gamma test for evaluating the sensitivity of input variables -- 12.3.3 Multiple linear regression -- 12.3.4 Performance evaluation indicators -- 12.4 Results and discussion -- 12.4.1 Identification of appropriate input variables using gamma test -- 12.4.2 Predicting daily runoff using hybrid AI models -- 12.5 Conclusions -- References -- 13 - Flood forecasting and error simulation using copula entropy method -- 13.1 Introduction -- 13.2 Background -- 13.2.1 Artificial neural networks -- 13.2.2 Entropy theory -- 13.2.3 Copula function -- 13.3 Determination of ANN model inputs based on copula entropy -- 13.3.1 Methodology -- 13.3.1.1 Copula entropy theory -- 13.3.1.2 Partial mutual information -- 13.3.1.3 Input selection based on copula entropy method -- 13.3.2 Application of copula entropy theory in flood forecasting-a case study -- 13.3.2.1 Study area and data description -- 13.3.2.2 Flood forecasts at Three Gorges Reservoir -- 13.3.2.3 Flood forecasting at the outlet of Jinsha River -- 13.3.2.4 Performance evaluation -- 13.3.2.5 Results of selected model inputs -- 13.4 Flood forecast uncertainties -- 13.4.1 Distributions for fitting flood forecasting errors -- 13.4.2 Determination of the distributions of flood forecasting uncertainties at TGR -- 13.5 Flood forecast uncertainty simulation
- 13.5.1 Flood forecasting uncertainties simulation based on copulas
- 5.5 ANN application software and programming language -- 5.6 Conclusions -- 5.7 Supplementary information -- References -- 6 - Application of artificial neural network and adaptive neuro-fuzzy inference system in streamflow forecasting -- 6.1 Introduction -- 6.2 Theoretical description of models -- 6.2.1 Artificial neural network -- 6.2.2 Adaptive neuro-fuzzy inference system -- 6.3 Application of ANN and ANFIS for prediction of peak discharge and runoff: a case study -- 6.3.1 Study area description -- 6.3.2 Methodology -- 6.3.2.1 Principal component analysis -- 6.3.2.2 Artificial neural network -- 6.3.2.3 Adaptive neuro-fuzzy inference system -- 6.3.2.4 Assessment of model performance by statistical indices -- 6.3.2.5 Sensitivity analysis -- 6.4 Results and discussion -- 6.4.1 Results of ANN modeling -- 6.4.2 Results of ANFIS modeling -- 6.5 Conclusions -- References -- 7 - Genetic programming for streamflow forecasting: a concise review of univariate models with a case study -- 7.1 Introduction -- 7.2 Overview of genetic programming and its variants -- 7.2.1 Classical genetic programming -- 7.2.2 Multigene genetic programming -- 7.2.3 Linear genetic programming -- 7.2.4 Gene expression programming -- 7.3 A brief review of the recent studies -- 7.4 A case study -- 7.4.1 Study area and data -- 7.4.2 Criteria for evaluating performance of models -- 7.5 Results and discussion -- 7.6 Conclusions -- References -- 8 - Model tree technique for streamflow forecasting: a case study in sub-catchment of Tapi River Basin, India -- 8.1 Introduction -- 8.2 Model tree -- 8.3 Model tree applications in streamflow forecasting -- 8.4 Application of model tree in streamflow forecasting: a case study -- 8.4.1 Study area -- 8.4.2 Methodology -- 8.5 Results and analysis -- 8.5.1 Selection of input variables -- 8.5.2 Model configuration
- 3.2.2 Model building procedure -- 3.2.3 MGARCH model -- 3.2.3.1 Diagonal VECH model -- 3.2.3.2 Testing conditional heteroscedasticity -- 3.2.4 Case study -- 3.3 Application of VAR/VARX approach -- 3.3.1 The VAR model -- 3.3.2 The VARX model -- 3.4 Application of MGARCH approach -- 3.5 Comparative evaluation of models' performances -- 3.6 Conclusions -- References -- 4 - Concepts, procedures, and applications of artificial neural network models in streamflow forecasting -- 4.1 Introduction -- 4.2 Procedure for development of artificial neural network models -- 4.2.1 Structure of artificial neural network models -- 4.2.1.1 Neurons and connection formula -- 4.2.1.2 Transfer function -- 4.2.1.3 Architecture of neurons -- 4.2.2 Network training processes -- 4.2.2.1 Unsupervised training method -- 4.2.2.2 Supervised training method -- 4.2.3 Artificial neural network to approximate a function -- 4.2.3.1 Step 1: preprocessing of data -- 4.2.3.1.1 Data normalization techniques -- 4.2.3.1.2 Principal component analysis -- 4.2.3.2 Step 2: choosing the best network architecture -- 4.2.3.3 Step 3: postprocessing of data -- 4.3 Types of artificial neural networks -- 4.3.1 Multilayer perceptron neural network -- 4.3.2 Static and dynamic neural network -- 4.3.3 Statistical neural networks -- 4.4 An overview of application of artificial neural network modeling in streamflow forecasting -- References -- 5 - Application of different artificial neural network for streamflow forecasting -- 5.1 Introduction -- 5.2 Development of neural network technique -- 5.2.1 Multilayer perceptron -- 5.2.2 Recurrent neural network -- 5.2.3 Long short-term memory network -- 5.2.4 Gated recurrent unit -- 5.2.5 Convolutional neural network -- 5.2.6 WaveNet -- 5.3 Artificial neural network in streamflow forecasting -- 5.4 Application of ANN: a case study of the Ganges River