Bainian Liu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Bainian Liu has a total of 36 co-authors. Here are the top co-authors by number of shared publications:
- Xiaoqun Cao
College of Meteorology and Oceanography National University of Defense Technology Changsha China
12 shared publications - Yanan Guo
College of Meteorology and Oceanography, National University of Defense Technology, Changsha, China
5 shared publications - Xiang Xing
College of Meteorology and Oceanography, National University of Defense Technology, Changsha 410073, China
4 shared publications - Qunbo Huang
93110 Troops of PLA, Beijing 100843, China
3 shared publications - Jianping Wu
College of Meteorology and Oceanography, National University of Defense Technology, Changsha, China
2 shared publications - Jingzhe Sun
Beijing Institute of Applied Meteorology, Beijing 100029, China
2 shared publications
Latest Publications
- Open Access
- 1 March 2026
- Published by IOP Publishing in Journal of Physics: Conference Series
- Vol. 3178 (1)
Impact of sea surface temperature uncertainty on the numerical prediction of Typhoon Doksuri
Sea Surface Temperature (SST) serves as a critical boundary condition in numerical weather prediction systems, and its uncertainty significantly influences forecast accuracy. Based on the independently developed Yin-He Global Spectral Model (YHGSM), this study conducts a comparative analysis of the error characteristics of three widely used SST products—OSTIA, REMSS, and OISSTv2—over global and key regional domains such as the Western Pacific and North Pacific, as well as their impacts on typhoon track, intensity, and precipitation forecasts. Experimental results indicate notable differences in the statistical error features among the three SST products at both global and regional scales. Among them, OISSTv2 demonstrates the best performance globally, while OSTIA exhibits smaller errors in the Western Pacific region. The results indicate that SST uncertainties lead to an average typhoon track error ranging from 172 to 176 km, and systematic biases are observed in intensity prediction during the typhoon decay stage. Moreover, the choice of SST product significantly affects the forecasted spatial extent and intensity of heavy precipitation. This study provides important references for improving numerical weather prediction models.
- 1 March 2026
- Published by IOP Publishing in Journal of Physics: Conference Series
- Vol. 3178 (1)
Open AccessImpact of sea surface temperature uncertainty on the numerical prediction of Typhoon Doksuri
Sea Surface Temperature (SST) serves as a critical boundary condition in numerical weather prediction systems, and its uncertainty significantly influences forecast accuracy. Based on the independently developed Yin-He Global Spectral Model (YHGSM), this study conducts a comparative analysis of the error characteristics of three widely used SST products—OSTIA, REMSS, and OISSTv2—over global and key regional domains such as the Western Pacific and North Pacific, as well as their impacts on typhoon track, intensity, and precipitation forecasts. Experimental results indicate notable differences in the statistical error features among the three SST products at both global and regional scales. Among them, OISSTv2 demonstrates the best performance globally, while OSTIA exhibits smaller errors in the Western Pacific region. The results indicate that SST uncertainties lead to an average typhoon track error ranging from 172 to 176 km, and systematic biases are observed in intensity prediction during the typhoon decay stage. Moreover, the choice of SST product significantly affects the forecasted spatial extent and intensity of heavy precipitation. This study provides important references for improving numerical weather prediction models.
- Open Access
- 4 November 2022
- Published by MDPI AG in Journal of Marine Science and Engineering
- Vol. 10 (11) , 1653
Effect of Flow-Dependent Unbalanced Background Error Variances on Tropical Cyclone Forecasting
The background error variance in variational data assimilation can significantly affect a model’s initial field. Around extreme weather events, the variance of the unbalanced control variables have contributed highly to the total variance. This study investigates the effect of flow-dependent unbalanced variance on tropical cyclone (TC) forecasts using the ensemble of data assimilation (EDA) method. The analysis of TC Saudel (October 2020) shows that flow-dependent unbalanced variances can better represent the uncertainty in the background error, which is investigated in terms of magnitude and distribution. The vertical distribution of the temperature-explained variance ratio also shows that the contribution of the vorticity-balanced variance around Saudel is lower than the global average (in the troposphere). Single-observation experiments reveal that the structured flow-dependent errors of unbalanced control variables can also introduce corresponding structural information in analysis increments. As expected, the experiments in which the variances of all variables are flow-dependent in the one-month TC forecast performed better overall. Compared with the reference, these forecasts reduce the average absolute track and intensity errors by approximately 31% and 9%, respectively. The results demonstrate that EDA-based unbalanced variances can indeed improve the mean forecast skills of TC tracks and intensities despite instability at some lead times by improving the forecast of the circulation situation and providing a more appropriate balance relationship between variables.
- 4 November 2022
- Published by MDPI AG in Journal of Marine Science and Engineering
- Vol. 10 (11) , 1653
Open AccessEffect of Flow-Dependent Unbalanced Background Error Variances on Tropical Cyclone Forecasting
The background error variance in variational data assimilation can significantly affect a model’s initial field. Around extreme weather events, the variance of the unbalanced control variables have contributed highly to the total variance. This study investigates the effect of flow-dependent unbalanced variance on tropical cyclone (TC) forecasts using the ensemble of data assimilation (EDA) method. The analysis of TC Saudel (October 2020) shows that flow-dependent unbalanced variances can better represent the uncertainty in the background error, which is investigated in terms of magnitude and distribution. The vertical distribution of the temperature-explained variance ratio also shows that the contribution of the vorticity-balanced variance around Saudel is lower than the global average (in the troposphere). Single-observation experiments reveal that the structured flow-dependent errors of unbalanced control variables can also introduce corresponding structural information in analysis increments. As expected, the experiments in which the variances of all variables are flow-dependent in the one-month TC forecast performed better overall. Compared with the reference, these forecasts reduce the average absolute track and intensity errors by approximately 31% and 9%, respectively. The results demonstrate that EDA-based unbalanced variances can indeed improve the mean forecast skills of TC tracks and intensities despite instability at some lead times by improving the forecast of the circulation situation and providing a more appropriate balance relationship between variables.
- Open Access
- 21 July 2022
- Published by MDPI AG in Journal of Marine Science and Engineering
- Vol. 10 (7) , 998
Correction: Xing et al. An Investigation of Adaptive Radius for the Covariance Localization in Ensemble Data Assimilation. J. Mar. Sci. Eng. 2021, 9, 1156
The authors wish to make the following corrections to this paper
- 21 July 2022
- Published by MDPI AG in Journal of Marine Science and Engineering
- Vol. 10 (7) , 998
Open AccessCorrection: Xing et al. An Investigation of Adaptive Radius for the Covariance Localization in Ensemble Data Assimilation. J. Mar. Sci. Eng. 2021, 9, 1156
The authors wish to make the following corrections to this paper
- Open Access
- 21 October 2021
- Published by MDPI AG in Journal of Marine Science and Engineering
- Vol. 9 (11) , 1156
An Investigation of Adaptive Radius for the Covariance Localization in Ensemble Data Assimilation
The covariance matrix estimated from the ensemble data assimilation always suffers from filter collapse because of the spurious correlations induced by the finite ensemble size. The localization technique is applied to ameliorate this issue, which has been suggested to be effective. In this paper, an adaptive scheme for Schur product covariance localization is proposed, which is easy and efficient to implement in the ensemble data assimilation frameworks. A Gaussian-shaped taper function is selected as the localization taper function for the Schur product in the adaptive localization scheme, and the localization radius is obtained adaptively through a certain criterion of correlations with the background ensembles. An idealized Lorenz96 model with an ensemble Kalman filter is firstly examined, showing that the adaptive localization scheme helps to significantly reduce the spurious correlations in the small ensemble with low computational cost and provides accurate covariances that are similar to those derived from a much larger ensemble. The investigations of adaptive localization radius reveal that the optimal radius is model-parameter-dependent, vertical-level-dependent and nearly flow-dependent with weather scenarios in a realistic model; for example, the radius of model parameter zonal wind is generally larger than that of temperature. The adaptivity of the localization scheme is also illustrated in the ensemble framework and shows that the adaptive scheme has a positive effect on the assimilated analysis as the well-tuned localization.
- 21 October 2021
- Published by MDPI AG in Journal of Marine Science and Engineering
- Vol. 9 (11) , 1156
Open AccessAn Investigation of Adaptive Radius for the Covariance Localization in Ensemble Data Assimilation
The covariance matrix estimated from the ensemble data assimilation always suffers from filter collapse because of the spurious correlations induced by the finite ensemble size. The localization technique is applied to ameliorate this issue, which has been suggested to be effective. In this paper, an adaptive scheme for Schur product covariance localization is proposed, which is easy and efficient to implement in the ensemble data assimilation frameworks. A Gaussian-shaped taper function is selected as the localization taper function for the Schur product in the adaptive localization scheme, and the localization radius is obtained adaptively through a certain criterion of correlations with the background ensembles. An idealized Lorenz96 model with an ensemble Kalman filter is firstly examined, showing that the adaptive localization scheme helps to significantly reduce the spurious correlations in the small ensemble with low computational cost and provides accurate covariances that are similar to those derived from a much larger ensemble. The investigations of adaptive localization radius reveal that the optimal radius is model-parameter-dependent, vertical-level-dependent and nearly flow-dependent with weather scenarios in a realistic model; for example, the radius of model parameter zonal wind is generally larger than that of temperature. The adaptivity of the localization scheme is also illustrated in the ensemble framework and shows that the adaptive scheme has a positive effect on the assimilated analysis as the well-tuned localization.
- Open Access
Polar Vortex Multi-Day Intensity Prediction Relying on New Deep Learning Model: A Combined Convolution Neural Network with Long Short-Term Memory Based on Gaussian Smoothing Method
The variation of polar vortex intensity is a significant factor affecting the atmospheric conditions and weather in the Northern Hemisphere (NH) and even the world. However, previous studies on the prediction of polar vortex intensity are insufficient. This paper establishes a deep learning (DL) model for multi-day and long-time intensity prediction of the polar vortex. Focusing on the winter period with the strongest polar vortex intensity, geopotential height (GPH) data of NCEP from 1948 to 2020 at 50 hPa are used to construct the dataset of polar vortex anomaly distribution images and polar vortex intensity time series. Then, we propose a new convolution neural network with long short-term memory based on Gaussian smoothing (GSCNN-LSTM) model which can not only accurately predict the variation characteristics of polar vortex intensity from day to day, but also can produce a skillful forecast for lead times of up to 20 days. Moreover, the innovative GSCNN-LSTM model has better stability and skillful correlation prediction than the traditional and some advanced spatiotemporal sequence prediction models. The accuracy of the model suggests important implications that DL methods have good applicability in forecasting the nonlinear system and vortex spatial–temporal characteristics variation in the atmosphere.
Open AccessPolar Vortex Multi-Day Intensity Prediction Relying on New Deep Learning Model: A Combined Convolution Neural Network with Long Short-Term Memory Based on Gaussian Smoothing Method
The variation of polar vortex intensity is a significant factor affecting the atmospheric conditions and weather in the Northern Hemisphere (NH) and even the world. However, previous studies on the prediction of polar vortex intensity are insufficient. This paper establishes a deep learning (DL) model for multi-day and long-time intensity prediction of the polar vortex. Focusing on the winter period with the strongest polar vortex intensity, geopotential height (GPH) data of NCEP from 1948 to 2020 at 50 hPa are used to construct the dataset of polar vortex anomaly distribution images and polar vortex intensity time series. Then, we propose a new convolution neural network with long short-term memory based on Gaussian smoothing (GSCNN-LSTM) model which can not only accurately predict the variation characteristics of polar vortex intensity from day to day, but also can produce a skillful forecast for lead times of up to 20 days. Moreover, the innovative GSCNN-LSTM model has better stability and skillful correlation prediction than the traditional and some advanced spatiotemporal sequence prediction models. The accuracy of the model suggests important implications that DL methods have good applicability in forecasting the nonlinear system and vortex spatial–temporal characteristics variation in the atmosphere.
- Open Access
Investigation of Local Weighting Filtering on Randomization Technique Estimates in a Data Assimilation System
Mainstream numerical weather prediction (NWP) centers usually estimate the standard deviations of background error by using a randomization technique to calibrate specific parameters of the background error covariance model in variational data assimilation (VAR) systems. However, the sampling size of the randomization technique is typically several orders of magnitude smaller than that of model state variables, and using finite-sized estimates as a proxy for the truth can lead to sampling noise, which may contaminate the estimation of the standard deviation. The sampling noise is firstly investigated in an atmospheric model to show that the sampling noise has a symmetrical structure oscillating around the truth on a small scale. To alleviate the sampling noise, a heterogeneous local weighting filtering is proposed based on distance-weighted correlation and similarity-weighted correlation. Local weighting filtering is easy to implement in the VAR operational systems and has a low computational cost in the post-processing of reducing the sampling noise. The validity and performance of local weighting filtering method are examined in a realistic model framework to show that the proposed filtering is able to eliminate most of the sampling noise dramatically, the details of the filtered results are more visible, and the accuracy of the filtered results is almost the same as that estimated from the larger sample. The signal-to-noise ratio of the optimal filtered field is improved by nearly 20%. A comparison with the widely used spectral filtering approach in the operational system is considered, showing that the proposed filtering method is more efficient to implement in the filtering procedure and exhibits very good performance in terms of preserving the local anisotropic features of the estimates. These attractive results show the potential efficiency of the local weighting filtering method for solving the noise issue in the randomization technique.
Open AccessInvestigation of Local Weighting Filtering on Randomization Technique Estimates in a Data Assimilation System
Mainstream numerical weather prediction (NWP) centers usually estimate the standard deviations of background error by using a randomization technique to calibrate specific parameters of the background error covariance model in variational data assimilation (VAR) systems. However, the sampling size of the randomization technique is typically several orders of magnitude smaller than that of model state variables, and using finite-sized estimates as a proxy for the truth can lead to sampling noise, which may contaminate the estimation of the standard deviation. The sampling noise is firstly investigated in an atmospheric model to show that the sampling noise has a symmetrical structure oscillating around the truth on a small scale. To alleviate the sampling noise, a heterogeneous local weighting filtering is proposed based on distance-weighted correlation and similarity-weighted correlation. Local weighting filtering is easy to implement in the VAR operational systems and has a low computational cost in the post-processing of reducing the sampling noise. The validity and performance of local weighting filtering method are examined in a realistic model framework to show that the proposed filtering is able to eliminate most of the sampling noise dramatically, the details of the filtered results are more visible, and the accuracy of the filtered results is almost the same as that estimated from the larger sample. The signal-to-noise ratio of the optimal filtered field is improved by nearly 20%. A comparison with the widely used spectral filtering approach in the operational system is considered, showing that the proposed filtering method is more efficient to implement in the filtering procedure and exhibits very good performance in terms of preserving the local anisotropic features of the estimates. These attractive results show the potential efficiency of the local weighting filtering method for solving the noise issue in the randomization technique.
- Open Access
- 1 June 2021
- Vol. 7 (3) , 1639-1648
Variational Principles and Solitary Wave Solutions of Generalized Nonlinear Schrodinger Equation in the Ocean
- 1 June 2021
- Vol. 7 (3) , 1639-1648
Open AccessVariational Principles and Solitary Wave Solutions of Generalized Nonlinear Schrodinger Equation in the Ocean
- Open Access
Ensemble Empirical Mode Decomposition with Adaptive Noise with Convolution Based Gated Recurrent Neural Network: A New Deep Learning Model for South Asian High Intensity Forecasting
The intensity variation of the South Asian high (SAH) plays an important role in the formation and extinction of many kinds of mesoscale systems, including tropical cyclones, southwest vortices in the Asian summer monsoon (ASM) region, and the precipitation in the whole Asia Europe region, and the SAH has a vortex symmetrical structure; its dynamic field also has the symmetry form. Not enough previous studies focus on the variation of SAH daily intensity. The purpose of this study is to establish a day-to-day prediction model of the SAH intensity, which can accurately predict not only the interannual variation but also the day-to-day variation of the SAH. Focusing on the summer period when the SAH is the strongest, this paper selects the geopotential height data between 1948 and 2020 from NCEP to construct the SAH intensity datasets. Compared with the classical deep learning methods of various kinds of efficient time series prediction model, we ultimately combine the Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) method, which has the ability to deal with the nonlinear and unstable single system, with the Permutation Entropy (PE) method, which can extract the SAH intensity feature of IMF decomposed by CEEMDAN, and the Convolution-based Gated Recurrent Neural Network (ConvGRU) model is used to train, test, and predict the intensity of the SAH. The prediction results show that the combination of CEEMDAN and ConvGRU can have a higher accuracy and more stable prediction ability than the traditional deep learning model. After removing the redundant features in the time series, the prediction accuracy of the SAH intensity is higher than that of the classical model, which proves that the method has good applicability for the prediction of nonlinear systems in the atmosphere.
Open AccessEnsemble Empirical Mode Decomposition with Adaptive Noise with Convolution Based Gated Recurrent Neural Network: A New Deep Learning Model for South Asian High Intensity Forecasting
The intensity variation of the South Asian high (SAH) plays an important role in the formation and extinction of many kinds of mesoscale systems, including tropical cyclones, southwest vortices in the Asian summer monsoon (ASM) region, and the precipitation in the whole Asia Europe region, and the SAH has a vortex symmetrical structure; its dynamic field also has the symmetry form. Not enough previous studies focus on the variation of SAH daily intensity. The purpose of this study is to establish a day-to-day prediction model of the SAH intensity, which can accurately predict not only the interannual variation but also the day-to-day variation of the SAH. Focusing on the summer period when the SAH is the strongest, this paper selects the geopotential height data between 1948 and 2020 from NCEP to construct the SAH intensity datasets. Compared with the classical deep learning methods of various kinds of efficient time series prediction model, we ultimately combine the Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) method, which has the ability to deal with the nonlinear and unstable single system, with the Permutation Entropy (PE) method, which can extract the SAH intensity feature of IMF decomposed by CEEMDAN, and the Convolution-based Gated Recurrent Neural Network (ConvGRU) model is used to train, test, and predict the intensity of the SAH. The prediction results show that the combination of CEEMDAN and ConvGRU can have a higher accuracy and more stable prediction ability than the traditional deep learning model. After removing the redundant features in the time series, the prediction accuracy of the SAH intensity is higher than that of the classical model, which proves that the method has good applicability for the prediction of nonlinear systems in the atmosphere.
- Open Access
- 17 September 2020
- Published by Wiley in Advances in Meteorology
- Vol. 2020, 1-15
Comparison of the Atmospheric 200 hPa Jet’s Analyses between Proper Orthogonal Decomposition and Advanced Dynamic Mode Decomposition Method
In this paper, a frequently employed technique named the sparsity-promoting dynamic mode decomposition (SPDMD) is proposed to analyze the velocity fields of atmospheric motion. The dynamic mode decomposition method (DMD) is an effective technique to extract dynamic information from flow fields that is generated from direct experiment measurements or numerical simulation and has been broadly employed to study the dynamics of the flow, to achieve a reduced-order model (ROM) of the complex high dimensional flow field, and even to predict the evolution of the flow in a short time in the future. However, for standard DMD, it is hard to determine which modes are the most physically relevant, unlike the proper orthogonal decomposition (POD) method which ranks the decomposed modes according to their energy content. The advanced modal decomposition method SPDMD is a variant of the standard DMD, which is capable of determining the modes that can be used to achieve a high-quality approximation of the given field. It is novel to introduce the SPDMD to analyze the atmospheric flow field. In this study, SPDMD is applied to extract essential dynamic information from the 200 hPa jet flow, and the decomposed results are compared with the POD method. To further demonstrate the extraction effect of POD/SPDMD methods on the 200 hPa jet flow characteristics, the POD/SPDMD reduced-order models are constructed, respectively. The results show that both modal decomposition methods successfully extract the underlying coherent structures from the 200 hPa jet flow. And the DMD method provides additional information on the modal properties, such as temporal frequency and growth rate of each mode which can be used to identify the stability of the modes. It is also found that a fewer order of modes determined by the SPDMD method can capture nearly the same dynamic information of the jet flow as the POD method. Furthermore, from the quantitative comparisons between the POD and SPDMD reduced-order models, the latter provides a higher precision than the former, especially when the number of modes is small.
- 17 September 2020
- Published by Wiley in Advances in Meteorology
- Vol. 2020, 1-15
Open AccessComparison of the Atmospheric 200 hPa Jet’s Analyses between Proper Orthogonal Decomposition and Advanced Dynamic Mode Decomposition Method
In this paper, a frequently employed technique named the sparsity-promoting dynamic mode decomposition (SPDMD) is proposed to analyze the velocity fields of atmospheric motion. The dynamic mode decomposition method (DMD) is an effective technique to extract dynamic information from flow fields that is generated from direct experiment measurements or numerical simulation and has been broadly employed to study the dynamics of the flow, to achieve a reduced-order model (ROM) of the complex high dimensional flow field, and even to predict the evolution of the flow in a short time in the future. However, for standard DMD, it is hard to determine which modes are the most physically relevant, unlike the proper orthogonal decomposition (POD) method which ranks the decomposed modes according to their energy content. The advanced modal decomposition method SPDMD is a variant of the standard DMD, which is capable of determining the modes that can be used to achieve a high-quality approximation of the given field. It is novel to introduce the SPDMD to analyze the atmospheric flow field. In this study, SPDMD is applied to extract essential dynamic information from the 200 hPa jet flow, and the decomposed results are compared with the POD method. To further demonstrate the extraction effect of POD/SPDMD methods on the 200 hPa jet flow characteristics, the POD/SPDMD reduced-order models are constructed, respectively. The results show that both modal decomposition methods successfully extract the underlying coherent structures from the 200 hPa jet flow. And the DMD method provides additional information on the modal properties, such as temporal frequency and growth rate of each mode which can be used to identify the stability of the modes. It is also found that a fewer order of modes determined by the SPDMD method can capture nearly the same dynamic information of the jet flow as the POD method. Furthermore, from the quantitative comparisons between the POD and SPDMD reduced-order models, the latter provides a higher precision than the former, especially when the number of modes is small.
- Top Cited PapersOpen Access
- 26 August 2020
- Published by MDPI AG in Applied Sciences
- Vol. 10 (17) , 5917
Solving Partial Differential Equations Using Deep Learning and Physical Constraints
The various studies of partial differential equations (PDEs) are hot topics of mathematical research. Among them, solving PDEs is a very important and difficult task. Since many partial differential equations do not have analytical solutions, numerical methods are widely used to solve PDEs. Although numerical methods have been widely used with good performance, researchers are still searching for new methods for solving partial differential equations. In recent years, deep learning has achieved great success in many fields, such as image classification and natural language processing. Studies have shown that deep neural networks have powerful function-fitting capabilities and have great potential in the study of partial differential equations. In this paper, we introduce an improved Physics Informed Neural Network (PINN) for solving partial differential equations. PINN takes the physical information that is contained in partial differential equations as a regularization term, which improves the performance of neural networks. In this study, we use the method to study the wave equation, the KdV–Burgers equation, and the KdV equation. The experimental results show that PINN is effective in solving partial differential equations and deserves further research.
- 26 August 2020
- Published by MDPI AG in Applied Sciences
- Vol. 10 (17) , 5917
Top Cited PapersOpen AccessSolving Partial Differential Equations Using Deep Learning and Physical Constraints
The various studies of partial differential equations (PDEs) are hot topics of mathematical research. Among them, solving PDEs is a very important and difficult task. Since many partial differential equations do not have analytical solutions, numerical methods are widely used to solve PDEs. Although numerical methods have been widely used with good performance, researchers are still searching for new methods for solving partial differential equations. In recent years, deep learning has achieved great success in many fields, such as image classification and natural language processing. Studies have shown that deep neural networks have powerful function-fitting capabilities and have great potential in the study of partial differential equations. In this paper, we introduce an improved Physics Informed Neural Network (PINN) for solving partial differential equations. PINN takes the physical information that is contained in partial differential equations as a regularization term, which improves the performance of neural networks. In this study, we use the method to study the wave equation, the KdV–Burgers equation, and the KdV equation. The experimental results show that PINN is effective in solving partial differential equations and deserves further research.