Forecasting of ocean currents is critical for both marine meteorological research and ocean engineering and construction.Timely and accurate forecasting of coastal current velocities offers a scientific foundation and...Forecasting of ocean currents is critical for both marine meteorological research and ocean engineering and construction.Timely and accurate forecasting of coastal current velocities offers a scientific foundation and decision support for multiple practices such as search and rescue,disaster avoidance and remediation,and offshore construction.This research established a framework to generate short-term surface current forecasts based on ensemble machine learning trained on high frequency radar observation.Results indicate that an ensemble algorithm that used random forests to filter forecasting features by weighting them,and then used the AdaBoost method to forecast can significantly reduce the model training time,while ensuring the model forecasting effectiveness,with great economic benefits.Model accuracy is a function of surface current variability and the forecasting horizon.In order to improve the forecasting capability and accuracy of the model,the model structure of the ensemble algorithm was optimized,and the random forest algorithm was used to dynamically select model features.The results show that the error variation of the optimized surface current forecasting model has a more regular error variation,and the importance of the features varies with the forecasting time-step.At ten-step ahead forecasting horizon the model reported root mean square error,mean absolute error,and correlation coefficient by 2.84 cm/s,2.02 cm/s,and 0.96,respectively.The model error is affected by factors such as topography,boundaries,and geometric accuracy of the observation system.This paper demonstrates the potential of ensemble-based machine learning algorithm to improve forecasting of ocean currents.展开更多
Currently,numerical models based on idealized assumptions,complex algorithms and high computational costs are unsatisfactory for ocean surface current prediction.Moreover,the complex temporal and spatial variability o...Currently,numerical models based on idealized assumptions,complex algorithms and high computational costs are unsatisfactory for ocean surface current prediction.Moreover,the complex temporal and spatial variability of ocean currents also makes the prediction methods based on time series data challenging.The deep network model can automatically learn and extract complex features hidden in large amount of complex data,so it is a promising method for high quality prediction of ocean currents.In this paper,we propose a spatiotemporal coupled attention deep network model STCANet that can extract abundant temporal and spatial coupling information on the behavior characteristics of ocean currents for improving the prediction accuracy.Firstly,Spatial Module is designed and implemented to extract the spatiotemporal coupling characteristics of ocean currents,and meanwhile the spatial correlations and dependencies among adjacent sea areas are obtained through Spatial Channel Attention Module(SCAM).Secondly,we use the GatedRecurrent-Unit(GRU)to extract temporal relationships of ocean currents,and design and implement the nearest neighbor time attention module to extract the interdependences of ocean currents between adjacent times,which can further improve the accuracy of ocean current prediction.Finally,a series of comparative experiments on the MediSea_Dataset and EastSea_Dataset showed that the prediction quality of our model greatly outperforms those of other benchmark models such as History Average(HA),Autoregressive Integrated Moving Average Model(ARIMA),Long Short-term Memory(LSTM),Gate Recurrent Unit(GRU)and CNN_GRU.展开更多
基金The fund from Southern Marine Science and Engineering Guangdong Laboratory(Zhuhai)under contract No.SML2020SP009the National Basic Research and Development Program of China under contract Nos 2022YFF0802000 and 2022YFF0802004+3 种基金the“Renowned Overseas Professors”Project of Guangdong Provincial Department of Science and Technology under contract No.76170-52910004the Belt and Road Special Foundation of the National Key Laboratory of Water Disaster Prevention under contract No.2022491711the National Natural Science Foundation of China under contract No.51909290the Key Research and Development Program of Guangdong Province under contract No.2020B1111020003.
文摘Forecasting of ocean currents is critical for both marine meteorological research and ocean engineering and construction.Timely and accurate forecasting of coastal current velocities offers a scientific foundation and decision support for multiple practices such as search and rescue,disaster avoidance and remediation,and offshore construction.This research established a framework to generate short-term surface current forecasts based on ensemble machine learning trained on high frequency radar observation.Results indicate that an ensemble algorithm that used random forests to filter forecasting features by weighting them,and then used the AdaBoost method to forecast can significantly reduce the model training time,while ensuring the model forecasting effectiveness,with great economic benefits.Model accuracy is a function of surface current variability and the forecasting horizon.In order to improve the forecasting capability and accuracy of the model,the model structure of the ensemble algorithm was optimized,and the random forest algorithm was used to dynamically select model features.The results show that the error variation of the optimized surface current forecasting model has a more regular error variation,and the importance of the features varies with the forecasting time-step.At ten-step ahead forecasting horizon the model reported root mean square error,mean absolute error,and correlation coefficient by 2.84 cm/s,2.02 cm/s,and 0.96,respectively.The model error is affected by factors such as topography,boundaries,and geometric accuracy of the observation system.This paper demonstrates the potential of ensemble-based machine learning algorithm to improve forecasting of ocean currents.
基金The authors would like to thank the financial support from the National Key Research and Development Program of China(Nos.2020YFE0201200,2019YFC1509100)the partial support by the Youth Program of Natural Science Foundation of China(No.41706010)the Fundamental Research Funds for the Central Universities(No.202264002).
文摘Currently,numerical models based on idealized assumptions,complex algorithms and high computational costs are unsatisfactory for ocean surface current prediction.Moreover,the complex temporal and spatial variability of ocean currents also makes the prediction methods based on time series data challenging.The deep network model can automatically learn and extract complex features hidden in large amount of complex data,so it is a promising method for high quality prediction of ocean currents.In this paper,we propose a spatiotemporal coupled attention deep network model STCANet that can extract abundant temporal and spatial coupling information on the behavior characteristics of ocean currents for improving the prediction accuracy.Firstly,Spatial Module is designed and implemented to extract the spatiotemporal coupling characteristics of ocean currents,and meanwhile the spatial correlations and dependencies among adjacent sea areas are obtained through Spatial Channel Attention Module(SCAM).Secondly,we use the GatedRecurrent-Unit(GRU)to extract temporal relationships of ocean currents,and design and implement the nearest neighbor time attention module to extract the interdependences of ocean currents between adjacent times,which can further improve the accuracy of ocean current prediction.Finally,a series of comparative experiments on the MediSea_Dataset and EastSea_Dataset showed that the prediction quality of our model greatly outperforms those of other benchmark models such as History Average(HA),Autoregressive Integrated Moving Average Model(ARIMA),Long Short-term Memory(LSTM),Gate Recurrent Unit(GRU)and CNN_GRU.
