Hydrological forecasting plays an important role in water resource management, supporting socio-economic development and managing water-related risks in river basins. There are many flow forecasting techniques that ha...Hydrological forecasting plays an important role in water resource management, supporting socio-economic development and managing water-related risks in river basins. There are many flow forecasting techniques that have been developed several centuries ago, ranging from physical models, physics-based models, conceptual models, and data-driven models. Recently, Artificial Intelligence (AI) has become an advanced technique applied as an effective data-driven model in hydrological forecasting. The main advantage of these models is that they give results with compatible accuracy, and require short computation time, thus increasing forecasting time and reducing human and financial effort. This study evaluates the applicability of machine learning and deep learning in Hanoi water level forecasting where it is controlled for flood management and water supply in the Red River Delta, Vietnam. Accordingly, SANN (machine learning algorithm) and LSTM (deep learning algorithm) were tested and compared with a Physics-Based Model (PBM) for the Red River Delta. The results show that SANN and LSTM give high accuracy. The R-squared coefficient is greater than 0.8, the mean squared error (MSE) is less than 20 cm, the correlation coefficient of the forecast hydrology is greater than 0.9 and the level of assurance of the forecast plan ranges from 80% to 90% in both cases. In addition, the calculation time is much reduced compared to the requirement of PBM, which is its limitation in hydrological forecasting for large river basins such as the Red River in Vietnam. Therefore, SANN and LSTM are expected to help increase lead time, thereby supporting water resource management for sustainable development and management of water-related risks in the Red River Delta.展开更多
Hydrologic models generally represent the most dominant processes since they are mere simplifications of physical reality and thus are subject to many significant uncertainties. As such, a coupling strategy is propose...Hydrologic models generally represent the most dominant processes since they are mere simplifications of physical reality and thus are subject to many significant uncertainties. As such, a coupling strategy is proposed. To this end, the coupling of the artificial neural network (ANN) with the Xin'anjiang conceptual model with a view to enhance the quality of its flow forecast is presented. The approach uses the latest observations and residuals in runoff/discharge forecasts from the Xin'anjiang model. The two complementary models (Xin'anjiang & ANN) are used in such a way that residuals of the Xin'anjiang model are forecasted by a neural network model so that flow forecasts can be improved as new observations come in. For the complementary neural network, the input data were presented in a patterned format to conform to the calibration regime of the Xin'anjiang conceptual model, using differing variants of the neural network scheme. The results show that there is a substantial improvement in the accuracy of the forecasts when the complementary model was operated on top of the Xin'anjiang conceptual model as compared with the results of the Xin'anjiang model alone.展开更多
The currently used hydrological forecast system in China is mainly focused on flood,and the flood forecasting frameworks are typically based on point discharge measurements and predictions at discrete locations,hence ...The currently used hydrological forecast system in China is mainly focused on flood,and the flood forecasting frameworks are typically based on point discharge measurements and predictions at discrete locations,hence they can’t provide spatio-temporal information of various hydrological elements,such as surface runoff,soil moisture,ground water table,and flood inundation extents over large scales and at high spatial resolutions.The use of distributed hydrological model has recently appeared to be the most suitable option to bridge this gap.An open source GIS-based distributed hydrological forecast system was established recently,and the watershed delineation and hydrological modelling were integrated together seamlessly.The time and human consuming work of processing the spatial data in building distributed hydrological model could be reduced significantly,and the spatial distribution of hydrological information could be quickly simulated and predicted using this system.The system was applied successfully to forecast the flood caused by super strong typhoon"Mangkhut"which attacked the south China in2018.展开更多
Time series analysis plays an important role in hydrologic forecasting,while the key to this analysis is to establish a proper model.This paper presents a time series neural network model with back propagation proced...Time series analysis plays an important role in hydrologic forecasting,while the key to this analysis is to establish a proper model.This paper presents a time series neural network model with back propagation procedure for hydrologic forecasting.Free from the disadvantages of previous models,the model can be parallel to operate information flexibly and rapidly.It excels in the ability of nonlinear mapping and can learn and adjust by itself,which gives the model a possibility to describe the complex nonlinear hydrologic process.By using directly a training process based on a set of previous data, the model can forecast the time series of stream flow.Moreover,two practical examples were used to test the performance of the time series neural network model.Results confirm that the model is efficient and feasible.展开更多
The development and implementation of a real-time flood forecasting system with a hydro-meteorological operational alert procedure during the MAP-D-PHASE Project is described in this paper. This chain includes both pr...The development and implementation of a real-time flood forecasting system with a hydro-meteorological operational alert procedure during the MAP-D-PHASE Project is described in this paper. This chain includes both probabilistic and deterministic forecasts. The hydrological model used to generate the runoff simulations is the rainfall-runoff distributed FEST-WB model, developed at Politecnico di Milano. The observed data to run the control simulations were supplied by ARPA-Piemonte. The analysis is focused on Maggiore Lake basin, an Alpine basin between North-West of Italy and Southern Switzerland. Two hindcasts during the D-PHASE period are discussed in order to evaluate certain effects regarding discharge forecasts due to hydro-meteorological sources of uncertainties. In particular, in the June convective event it is analysed how the effect of meteorological model spatial resolution can influence the discharge forecasts over mountain basins, while in the November stratiform event how the effect of the initial conditions of soil moisture can modify meteorological warnings. The study shows how the introduction of alert codes appears to be useful for decision makers to give them a spread of forecasted QDFs with the probability of event occurrence, but also how alert warnings issued on the basis of forecasted precipitation only are not always reliable.展开更多
The mid-long term hydrology forecasting is one of most challenging problems in hydrological studies. This paper proposes an efficient dynamical system prediction model using evolutionary computation techniques. The ne...The mid-long term hydrology forecasting is one of most challenging problems in hydrological studies. This paper proposes an efficient dynamical system prediction model using evolutionary computation techniques. The new model overcomes some disadvantages of conventional hydrology forecasting ones. The observed data is divided into two parts; the slow 'smooth and steady' data, and the fast 'coarse and fluctuation' data. Under the divide and conquer strategy, the behavior of smooth data is modeled by ordinary differential equations based on evolutionary modeling, and that of the coarse data is modeled using gray correlative forecasting method. Our model is verified on the test data of the mid-long term hydrology forecast in the northeast region of China. The experimental results show that the model is superior to gray system prediction model (GSPM).展开更多
Hydrological monitoring and seasonal forecasting is an active research field because of its potential applications in hydrological risk assessment, preparedness and mitigation. In recent decades, developments in groun...Hydrological monitoring and seasonal forecasting is an active research field because of its potential applications in hydrological risk assessment, preparedness and mitigation. In recent decades, developments in ground and satellite measurements have made the hydrometeorological information readily available, and advances in information technology have facilitated the data analysis in a real-time manner. New progress in climate research and modeling has enabled the prediction of seasonal climate with reasonable accuracy and increased resolution. These emerging techniques and advances have enabled more timely acquisition of accurate hydrological fluxes and status, and earlier warning of extreme hydrological events such as droughts and floods. This paper gives current state-of-the-art understanding of the uncertainties in hydrological monitoring and forecasting, reviews the efforts and progress in operational hydrological monitoring system assisted by observations from various sources and experimental seasonal hydrological forecasting, and briefly introduces the current monitoring and forecasting practices in China. The grand challenges and perspectives for the near future are also discussed, including acquiring and extracting reliable information for monitoring and forecasting, predicting realistic hydrological fluxes and states in the river basin being significantly altered by human activity, and filling the gap between numerical models and the end user. We highlight the importance of understanding the needs of the operational water management and the priority to transfer research knowledge to decision-makers.展开更多
文摘Hydrological forecasting plays an important role in water resource management, supporting socio-economic development and managing water-related risks in river basins. There are many flow forecasting techniques that have been developed several centuries ago, ranging from physical models, physics-based models, conceptual models, and data-driven models. Recently, Artificial Intelligence (AI) has become an advanced technique applied as an effective data-driven model in hydrological forecasting. The main advantage of these models is that they give results with compatible accuracy, and require short computation time, thus increasing forecasting time and reducing human and financial effort. This study evaluates the applicability of machine learning and deep learning in Hanoi water level forecasting where it is controlled for flood management and water supply in the Red River Delta, Vietnam. Accordingly, SANN (machine learning algorithm) and LSTM (deep learning algorithm) were tested and compared with a Physics-Based Model (PBM) for the Red River Delta. The results show that SANN and LSTM give high accuracy. The R-squared coefficient is greater than 0.8, the mean squared error (MSE) is less than 20 cm, the correlation coefficient of the forecast hydrology is greater than 0.9 and the level of assurance of the forecast plan ranges from 80% to 90% in both cases. In addition, the calculation time is much reduced compared to the requirement of PBM, which is its limitation in hydrological forecasting for large river basins such as the Red River in Vietnam. Therefore, SANN and LSTM are expected to help increase lead time, thereby supporting water resource management for sustainable development and management of water-related risks in the Red River Delta.
文摘Hydrologic models generally represent the most dominant processes since they are mere simplifications of physical reality and thus are subject to many significant uncertainties. As such, a coupling strategy is proposed. To this end, the coupling of the artificial neural network (ANN) with the Xin'anjiang conceptual model with a view to enhance the quality of its flow forecast is presented. The approach uses the latest observations and residuals in runoff/discharge forecasts from the Xin'anjiang model. The two complementary models (Xin'anjiang & ANN) are used in such a way that residuals of the Xin'anjiang model are forecasted by a neural network model so that flow forecasts can be improved as new observations come in. For the complementary neural network, the input data were presented in a patterned format to conform to the calibration regime of the Xin'anjiang conceptual model, using differing variants of the neural network scheme. The results show that there is a substantial improvement in the accuracy of the forecasts when the complementary model was operated on top of the Xin'anjiang conceptual model as compared with the results of the Xin'anjiang model alone.
基金supported by the National Key Research and Development Program of China(Grant No.2018YFC1508100)
文摘The currently used hydrological forecast system in China is mainly focused on flood,and the flood forecasting frameworks are typically based on point discharge measurements and predictions at discrete locations,hence they can’t provide spatio-temporal information of various hydrological elements,such as surface runoff,soil moisture,ground water table,and flood inundation extents over large scales and at high spatial resolutions.The use of distributed hydrological model has recently appeared to be the most suitable option to bridge this gap.An open source GIS-based distributed hydrological forecast system was established recently,and the watershed delineation and hydrological modelling were integrated together seamlessly.The time and human consuming work of processing the spatial data in building distributed hydrological model could be reduced significantly,and the spatial distribution of hydrological information could be quickly simulated and predicted using this system.The system was applied successfully to forecast the flood caused by super strong typhoon"Mangkhut"which attacked the south China in2018.
文摘Time series analysis plays an important role in hydrologic forecasting,while the key to this analysis is to establish a proper model.This paper presents a time series neural network model with back propagation procedure for hydrologic forecasting.Free from the disadvantages of previous models,the model can be parallel to operate information flexibly and rapidly.It excels in the ability of nonlinear mapping and can learn and adjust by itself,which gives the model a possibility to describe the complex nonlinear hydrologic process.By using directly a training process based on a set of previous data, the model can forecast the time series of stream flow.Moreover,two practical examples were used to test the performance of the time series neural network model.Results confirm that the model is efficient and feasible.
文摘The development and implementation of a real-time flood forecasting system with a hydro-meteorological operational alert procedure during the MAP-D-PHASE Project is described in this paper. This chain includes both probabilistic and deterministic forecasts. The hydrological model used to generate the runoff simulations is the rainfall-runoff distributed FEST-WB model, developed at Politecnico di Milano. The observed data to run the control simulations were supplied by ARPA-Piemonte. The analysis is focused on Maggiore Lake basin, an Alpine basin between North-West of Italy and Southern Switzerland. Two hindcasts during the D-PHASE period are discussed in order to evaluate certain effects regarding discharge forecasts due to hydro-meteorological sources of uncertainties. In particular, in the June convective event it is analysed how the effect of meteorological model spatial resolution can influence the discharge forecasts over mountain basins, while in the November stratiform event how the effect of the initial conditions of soil moisture can modify meteorological warnings. The study shows how the introduction of alert codes appears to be useful for decision makers to give them a spread of forecasted QDFs with the probability of event occurrence, but also how alert warnings issued on the basis of forecasted precipitation only are not always reliable.
基金Supported by the National Natural Science Foundation of China(60133010,70071042,60073043)
文摘The mid-long term hydrology forecasting is one of most challenging problems in hydrological studies. This paper proposes an efficient dynamical system prediction model using evolutionary computation techniques. The new model overcomes some disadvantages of conventional hydrology forecasting ones. The observed data is divided into two parts; the slow 'smooth and steady' data, and the fast 'coarse and fluctuation' data. Under the divide and conquer strategy, the behavior of smooth data is modeled by ordinary differential equations based on evolutionary modeling, and that of the coarse data is modeled using gray correlative forecasting method. Our model is verified on the test data of the mid-long term hydrology forecast in the northeast region of China. The experimental results show that the model is superior to gray system prediction model (GSPM).
基金National Natural Science Foundation of China,No.41425002National Basic Research Program of China,No.2012CB955403+2 种基金National Youth Top-notch Talent Support Program in ChinaChina Special Fund for Meteorological Research in the Public Interest(Major projects),No.GYHY201506001-7The Beijing Science and Technology Plan Project,No.Z141100003614052
文摘Hydrological monitoring and seasonal forecasting is an active research field because of its potential applications in hydrological risk assessment, preparedness and mitigation. In recent decades, developments in ground and satellite measurements have made the hydrometeorological information readily available, and advances in information technology have facilitated the data analysis in a real-time manner. New progress in climate research and modeling has enabled the prediction of seasonal climate with reasonable accuracy and increased resolution. These emerging techniques and advances have enabled more timely acquisition of accurate hydrological fluxes and status, and earlier warning of extreme hydrological events such as droughts and floods. This paper gives current state-of-the-art understanding of the uncertainties in hydrological monitoring and forecasting, reviews the efforts and progress in operational hydrological monitoring system assisted by observations from various sources and experimental seasonal hydrological forecasting, and briefly introduces the current monitoring and forecasting practices in China. The grand challenges and perspectives for the near future are also discussed, including acquiring and extracting reliable information for monitoring and forecasting, predicting realistic hydrological fluxes and states in the river basin being significantly altered by human activity, and filling the gap between numerical models and the end user. We highlight the importance of understanding the needs of the operational water management and the priority to transfer research knowledge to decision-makers.