Spatiotemporal forecasting of surface soil moisture(SSM)is recognized as a critical scientific issue in precision agricultural irrigation,regional drought monitoring,and early warning systems for extreme precipitation...Spatiotemporal forecasting of surface soil moisture(SSM)is recognized as a critical scientific issue in precision agricultural irrigation,regional drought monitoring,and early warning systems for extreme precipitation.However,long-term forecasting continues to pose formidable challenges because of the complexity observed across both the spatial and temporal scales.In this study,we used a daily SSM dataset at a 0.05°×0.05°spatial resolution over the Qilian Mountains,China and proposed a hybrid Convolutional Long Short-Term Memory(ConvLSTM)-Nudging model,which combined deep neural networks with data assimilation to increase the accuracy of long-term SSM forecasting.We trained and evaluated the SSM predictive performance of four models(Convolutional Neural Network(CNN),Long Short-Term Memory(LSTM),ConvLSTM,and ConvLSTM with Squeeze-and-Excitation(SE)attention mechanism(ConvLSTM-SE))in both short-term and long-term scenarios.The results showed that all the models perform well under short-term predictions,but the accuracy decrease substantially in long-term predictions.Therefore,we integrated Nudging technique during the long-term prediction phase to assimilate observational information and rectify model biases.Comprehensive evaluations demonstrate that Nudging significantly improves all the models,with ConvLSTM-Nudging achieving the best performance under the 200-d forecasting scenario.Relative to those of the best-performing ConvLSTM model for long-term forecasts,when observation noiseδ=0.00 and observation fraction obs=50.0%,the coefficient of determination(R2)of ConvLSTM-Nudging increases by approximately 82.1%,while its mean absolute error(MAE)and root mean squared error(RMSE)decrease by approximately 84.8%and 77.3%,respectively;the average Pearson correlation coefficient(r)improves by approximately 23.6%,and Bias is reduced by 98.1%.These results demonstrated that although pure deep learning models achieve high accuracy in the short-term predictions,they are prone to error accumulation and systematic drift in long-term autoregressive predictions.Integrating data assimilation with deep learning and continuously correcting the state through observation can effectively suppress long-term biases,thereby achieving robust long-term SSM forecasting.展开更多
This study investigated the simulations of three months of seasonal tropical cyclone （TC） activity over the western North Pacific using the Advanced Research WRF Model. In the control experiment （CTL）, the TC freq...This study investigated the simulations of three months of seasonal tropical cyclone （TC） activity over the western North Pacific using the Advanced Research WRF Model. In the control experiment （CTL）, the TC frequency was considerably overestimated. Additionally, the tracks of some TCs tended to have larger radii of curvature and were shifted eastward. The large-scale environments of westerly monsoon flows and subtropical Pacific highs were unreasonably simulated. The overestimated frequency of TC formation was attributed to a strengthened westerly wind field in the southern quadrants of the TC center. In comparison with the experiment with the spectral nudging method, the strengthened wind speed was mainly modulated by large-scale flow that was greater than approximately 1000 km in the model domain. The spurious formation and undesirable tracks of TCs in the CTL were considerably improved by reproducing realistic large-scale atmospheric monsoon circulation with substantial adjustment between large-scale flow in the model domain and large-scale boundary forcing modified by the spectral nudging method. The realistic monsoon circulation took a vital role in simulating realistic TCs. It revealed that, in the downscaling from large-scale fields for regional climate simulations, scale interaction between model-generated regional features and forced large-scale fields should be considered, and spectral nudging is a desirable method in the downscaling method.展开更多
With the Weather Research and Forecasting model （WRFV3.2.1）, the application of specmun nudging tech- niques in numerical simulation of the genesis and development of typhoon Longwang （2005） is evaluated in this w...With the Weather Research and Forecasting model （WRFV3.2.1）, the application of specmun nudging tech- niques in numerical simulation of the genesis and development of typhoon Longwang （2005） is evaluated in this work via four numerical experiments with different nudging techniques. It is found that, due to the ability to capture the large-scale fields and to keep the meso-to small-scale features derived from the model dynamics, the experiment with spectrum nudging technique can simulate the formation, intensification and motion of Longwang properly. The improve- ment on the numerical simulation of Longwang induced by the spectrum nudging depends on the nudging coefficients. A weak spectrum nudging does not make significant improvement on the simulation of Longwang. Although the experi- ment with four-dimensional data assimilation, i.e., FDDA, also derives the genesis and movement of Longwang appro- priately, it fails to simulate the intensifying process of Longwang properly. The reason is that, as the large-scale features derived from the model are nudged to the observational data, the meso- to small-processes produced by the model dy- namics important to the intensification of typhoon are nearly smoothed by FDDA.展开更多
The performance of spectral nudging in an investigation of the 2010 East Asia summer monsoon was assessed using the Weather Research and Forecasting (WRF) model, forced by 1-degree NCEP Global Final Analysis (FNL). Tw...The performance of spectral nudging in an investigation of the 2010 East Asia summer monsoon was assessed using the Weather Research and Forecasting (WRF) model, forced by 1-degree NCEP Global Final Analysis (FNL). Two pairs of experiments were made, spectral nudging (SP) and non-spectral nudging (NOSP), with five members in each group. The members were distinguished by different initial times, and the analysis was based on the ensemble mean of the two simulation pairs. The SP was able to constrain error growth in large-scale circulation in upper-level, during simulation, and generate realistic regional scale patterns. The main focus was the model ability to simulate precipitation. The Tropical Rainfall Measuring Mission (TRMM) 3B42 product was used for precipitation verification. Mean precipitation magnitude was generally overestimated by WRF. Nevertheless, SP simulations suppressed overestimation relative to the NOSP experiments. Compared to TRMM, SP also improved model simulation of precipitation in spatial and temporal distributions, with the ability to reproduce movement of rainbands. However, extreme precipitation events were suppressed in the SP simulations.展开更多
An ensemble-based assimilation method is proposed for correcting the subsurface temperature field when nudging the sea surface temperature(SST) observations into the Max Planck Institute(MPI) climate model,ECHAM5/MPI-...An ensemble-based assimilation method is proposed for correcting the subsurface temperature field when nudging the sea surface temperature(SST) observations into the Max Planck Institute(MPI) climate model,ECHAM5/MPI-OM. This method can project SST directly to subsurface according to model ensemble-based correlations between SST and subsurface temperature. Results from a 50 year(1960–2009) assimilation experiment show the method can improve the subsurface temperature field up to 300 m compared to the qualitycontrolled subsurface ocean temperature objective analyses(EN4), through reducing the biases of the thermal states, improving the thermocline structure, and reducing the root mean square(RMS) errors. Moreover, as most of the improvements concentrate over the upper 100 m, the ocean heat content in the upper 100 m(OHT100 m)is further adopted as a property to validate the performance of the ensemble-based correction method. The results show that RMS errors of the global OHT100 m convergent to one value after several times iteration,indicating this method can represent the relationship between SST and subsurface temperature fields well, and then improve the accuracy of the simulation in the subsurface temperature of the climate model.展开更多
With the aid of Meteorological Information Composite and Processing System (MICAPS), satellite wind vectors derived from the Geostationary Meteorological Statellite-5 (GMS-5) and retrieved by National Satellite Meteor...With the aid of Meteorological Information Composite and Processing System (MICAPS), satellite wind vectors derived from the Geostationary Meteorological Statellite-5 (GMS-5) and retrieved by National Satellite Meteorology Center of China (NSMC) can be obtained. Based on the nudging method built in the fifth-generation Mesoscale Model (MM5) of Pennsylvania State University and Na- tional Center for Atmospheric Research, a data preprocessor is developed to convert these satellite wind vectors to those with specified format required in MM5. To examine the data preprocessor and evaluate the impact of satellite winds from GMS-5 on MM5 simulations, a series of numerical experimental fore- casts consisting of four typhoon cases in 2002 are designed and implemented. The results show that the preprocessor can process satellite winds smoothly and MM5 model runs successfully with a little extra computational load during ingesting these winds, and that assimilation of satellite winds by MM5 nudging method can obviously improve typhoon track forecast but contributes a little to typhoon intensity forecast. The impact of the satellite winds depends heavily upon whether the typhoon bogussing scheme in MM5 was turned on or not. The data preprocessor developed in this paper not only can treat GMS-5 satellite winds but also has capability with little modification to process derived winds from other geostationary satellites.展开更多
基于全球海洋模式OGCTM(Ocean General Circulation and Tide Model),利用19年(1992—2011年)的卫星高度计资料调和分析得到全球分潮调和常数回报逐时正压潮水位;采用Nudging(牛顿松弛逼近)方法进行潮汐同化,针对Nudging松弛项的差分方...基于全球海洋模式OGCTM(Ocean General Circulation and Tide Model),利用19年(1992—2011年)的卫星高度计资料调和分析得到全球分潮调和常数回报逐时正压潮水位;采用Nudging(牛顿松弛逼近)方法进行潮汐同化,针对Nudging松弛项的差分方案以及松弛系数进行了数值试验研究。一系列试验结果表明,不论采用显式还是隐式Nudging松弛方案,模式结果的偏差会随着松弛系数的增加先减小后增大,松弛系数最优值为1×10^(-5);当松弛系数太大时,会造成模式的溢出。Nudging松弛项采用隐式差分方案可以显著提高松弛系数的阈值;松弛系数在适当的取值范围内时,加入Nudging松弛项的数值试验描述的潮汐特征要明显优于未加Nudging松弛项的数值试验描述的潮汐特征;在最优Nudging松弛方案下,Nudging松弛方法对全日分潮的振幅模拟准确度可提高50%,对半日分潮的振幅模拟准确度可提高56%。试验结果也表明了Nudging方法作为一种简单的同化方法的可行性和有效性。展开更多
基金funded by the National Natural Science Foundation of China(42461053)the Department of Education of Gansu Province:Higher Education Innovation Fund Project(2023B-064)+1 种基金the Youth Doctoral Fund Project(2024QB-014)the Natural Science Foundation of Gansu Province(25JRRA012).
文摘Spatiotemporal forecasting of surface soil moisture(SSM)is recognized as a critical scientific issue in precision agricultural irrigation,regional drought monitoring,and early warning systems for extreme precipitation.However,long-term forecasting continues to pose formidable challenges because of the complexity observed across both the spatial and temporal scales.In this study,we used a daily SSM dataset at a 0.05°×0.05°spatial resolution over the Qilian Mountains,China and proposed a hybrid Convolutional Long Short-Term Memory(ConvLSTM)-Nudging model,which combined deep neural networks with data assimilation to increase the accuracy of long-term SSM forecasting.We trained and evaluated the SSM predictive performance of four models(Convolutional Neural Network(CNN),Long Short-Term Memory(LSTM),ConvLSTM,and ConvLSTM with Squeeze-and-Excitation(SE)attention mechanism(ConvLSTM-SE))in both short-term and long-term scenarios.The results showed that all the models perform well under short-term predictions,but the accuracy decrease substantially in long-term predictions.Therefore,we integrated Nudging technique during the long-term prediction phase to assimilate observational information and rectify model biases.Comprehensive evaluations demonstrate that Nudging significantly improves all the models,with ConvLSTM-Nudging achieving the best performance under the 200-d forecasting scenario.Relative to those of the best-performing ConvLSTM model for long-term forecasts,when observation noiseδ=0.00 and observation fraction obs=50.0%,the coefficient of determination(R2)of ConvLSTM-Nudging increases by approximately 82.1%,while its mean absolute error(MAE)and root mean squared error(RMSE)decrease by approximately 84.8%and 77.3%,respectively;the average Pearson correlation coefficient(r)improves by approximately 23.6%,and Bias is reduced by 98.1%.These results demonstrated that although pure deep learning models achieve high accuracy in the short-term predictions,they are prone to error accumulation and systematic drift in long-term autoregressive predictions.Integrating data assimilation with deep learning and continuously correcting the state through observation can effectively suppress long-term biases,thereby achieving robust long-term SSM forecasting.
基金funded by the Korea Meteorological Administration Research and Development Program under grant KMIPA 2015–2083
文摘This study investigated the simulations of three months of seasonal tropical cyclone （TC） activity over the western North Pacific using the Advanced Research WRF Model. In the control experiment （CTL）, the TC frequency was considerably overestimated. Additionally, the tracks of some TCs tended to have larger radii of curvature and were shifted eastward. The large-scale environments of westerly monsoon flows and subtropical Pacific highs were unreasonably simulated. The overestimated frequency of TC formation was attributed to a strengthened westerly wind field in the southern quadrants of the TC center. In comparison with the experiment with the spectral nudging method, the strengthened wind speed was mainly modulated by large-scale flow that was greater than approximately 1000 km in the model domain. The spurious formation and undesirable tracks of TCs in the CTL were considerably improved by reproducing realistic large-scale atmospheric monsoon circulation with substantial adjustment between large-scale flow in the model domain and large-scale boundary forcing modified by the spectral nudging method. The realistic monsoon circulation took a vital role in simulating realistic TCs. It revealed that, in the downscaling from large-scale fields for regional climate simulations, scale interaction between model-generated regional features and forced large-scale fields should be considered, and spectral nudging is a desirable method in the downscaling method.
基金Nature Science Foundation of China(41475046,41130964)State Key Program of China(2012CB417201)
文摘With the Weather Research and Forecasting model （WRFV3.2.1）, the application of specmun nudging tech- niques in numerical simulation of the genesis and development of typhoon Longwang （2005） is evaluated in this work via four numerical experiments with different nudging techniques. It is found that, due to the ability to capture the large-scale fields and to keep the meso-to small-scale features derived from the model dynamics, the experiment with spectrum nudging technique can simulate the formation, intensification and motion of Longwang properly. The improve- ment on the numerical simulation of Longwang induced by the spectrum nudging depends on the nudging coefficients. A weak spectrum nudging does not make significant improvement on the simulation of Longwang. Although the experi- ment with four-dimensional data assimilation, i.e., FDDA, also derives the genesis and movement of Longwang appro- priately, it fails to simulate the intensifying process of Longwang properly. The reason is that, as the large-scale features derived from the model are nudged to the observational data, the meso- to small-processes produced by the model dy- namics important to the intensification of typhoon are nearly smoothed by FDDA.
基金Supported by the Knowledge Innovation Program of Chinese Academy of Sciences (No. KZCX2-YW-Q11-02)
文摘The performance of spectral nudging in an investigation of the 2010 East Asia summer monsoon was assessed using the Weather Research and Forecasting (WRF) model, forced by 1-degree NCEP Global Final Analysis (FNL). Two pairs of experiments were made, spectral nudging (SP) and non-spectral nudging (NOSP), with five members in each group. The members were distinguished by different initial times, and the analysis was based on the ensemble mean of the two simulation pairs. The SP was able to constrain error growth in large-scale circulation in upper-level, during simulation, and generate realistic regional scale patterns. The main focus was the model ability to simulate precipitation. The Tropical Rainfall Measuring Mission (TRMM) 3B42 product was used for precipitation verification. Mean precipitation magnitude was generally overestimated by WRF. Nevertheless, SP simulations suppressed overestimation relative to the NOSP experiments. Compared to TRMM, SP also improved model simulation of precipitation in spatial and temporal distributions, with the ability to reproduce movement of rainbands. However, extreme precipitation events were suppressed in the SP simulations.
基金The National Key R&D Program of China under contract No. 2017YFA0604201the National Natural Science Foundation of China under contract Nos 41876012 and 41861144015.
文摘An ensemble-based assimilation method is proposed for correcting the subsurface temperature field when nudging the sea surface temperature(SST) observations into the Max Planck Institute(MPI) climate model,ECHAM5/MPI-OM. This method can project SST directly to subsurface according to model ensemble-based correlations between SST and subsurface temperature. Results from a 50 year(1960–2009) assimilation experiment show the method can improve the subsurface temperature field up to 300 m compared to the qualitycontrolled subsurface ocean temperature objective analyses(EN4), through reducing the biases of the thermal states, improving the thermocline structure, and reducing the root mean square(RMS) errors. Moreover, as most of the improvements concentrate over the upper 100 m, the ocean heat content in the upper 100 m(OHT100 m)is further adopted as a property to validate the performance of the ensemble-based correction method. The results show that RMS errors of the global OHT100 m convergent to one value after several times iteration,indicating this method can represent the relationship between SST and subsurface temperature fields well, and then improve the accuracy of the simulation in the subsurface temperature of the climate model.
基金Supported by Major State Basic Research Development Program of China (973 Program, No.2005CB4223-01) and Key Technologies R & D Program of China (No.2001BA603B-01).
文摘With the aid of Meteorological Information Composite and Processing System (MICAPS), satellite wind vectors derived from the Geostationary Meteorological Statellite-5 (GMS-5) and retrieved by National Satellite Meteorology Center of China (NSMC) can be obtained. Based on the nudging method built in the fifth-generation Mesoscale Model (MM5) of Pennsylvania State University and Na- tional Center for Atmospheric Research, a data preprocessor is developed to convert these satellite wind vectors to those with specified format required in MM5. To examine the data preprocessor and evaluate the impact of satellite winds from GMS-5 on MM5 simulations, a series of numerical experimental fore- casts consisting of four typhoon cases in 2002 are designed and implemented. The results show that the preprocessor can process satellite winds smoothly and MM5 model runs successfully with a little extra computational load during ingesting these winds, and that assimilation of satellite winds by MM5 nudging method can obviously improve typhoon track forecast but contributes a little to typhoon intensity forecast. The impact of the satellite winds depends heavily upon whether the typhoon bogussing scheme in MM5 was turned on or not. The data preprocessor developed in this paper not only can treat GMS-5 satellite winds but also has capability with little modification to process derived winds from other geostationary satellites.
文摘基于全球海洋模式OGCTM(Ocean General Circulation and Tide Model),利用19年(1992—2011年)的卫星高度计资料调和分析得到全球分潮调和常数回报逐时正压潮水位;采用Nudging(牛顿松弛逼近)方法进行潮汐同化,针对Nudging松弛项的差分方案以及松弛系数进行了数值试验研究。一系列试验结果表明,不论采用显式还是隐式Nudging松弛方案,模式结果的偏差会随着松弛系数的增加先减小后增大,松弛系数最优值为1×10^(-5);当松弛系数太大时,会造成模式的溢出。Nudging松弛项采用隐式差分方案可以显著提高松弛系数的阈值;松弛系数在适当的取值范围内时,加入Nudging松弛项的数值试验描述的潮汐特征要明显优于未加Nudging松弛项的数值试验描述的潮汐特征;在最优Nudging松弛方案下,Nudging松弛方法对全日分潮的振幅模拟准确度可提高50%,对半日分潮的振幅模拟准确度可提高56%。试验结果也表明了Nudging方法作为一种简单的同化方法的可行性和有效性。
