A four‐dimensional ensemble‐variational (4DEnVar) data assimilation (DA) system was developed based on the global forecast system of the Global/Regional Assimilation and Prediction System (GRAPES‐GFS). Instead of using the adjoint technique, this system utilizes a dimension‐reduced projection (DRP) technique to minimize the cost function of the standard four‐dimensional variational (4DVar) DA. It dynamically predicts ensemble background error covariance (BEC) and realizes the explicit flow‐dependence of BEC in the variational configuration. An inflation technique based on a linear combination of analysis increments and balanced random perturbations, is utilized to overcome the problem of underestimation of BEC matrix (B‐matrix) during the assimilation cycle. To mitigate the spurious correlations in the ensemble B‐matrix caused by the insufficient ensemble members, an ensemble‐sample‐based subspace localization method is utilized. In order to evaluate the new system, single‐point observation experiments (SOEs) and observing system simulation experiments (OSSEs) were conducted with sounding and cloud‐derived wind data based on GRAPES‐GFS. The explicit flow‐dependent characteristic of the 4DEnVar system using a localized ensemble covariance was verified in the SOEs. In the OSSEs, the ensemble mean analysis of 4DEnVar outperforms the analysis of 4DVar. The deterministic forecast initialized from the 4DEnVar ensemble mean analysis has better performance in the short‐range forecasts, better (worse) performance in the early (late) period of the medium‐range forecasts in the Northern Extratropics, and opposite performance in the Southern Extratropics, and exhibits slightly worse effects in the Tropics. Moreover, the ensemble mean forecast initialized by the 4DEnVar system has higher forecast skills.

Localization techniques are commonly used in ensemble-based data assimilation (e.g., the Ensemble Kalman Filter (EnKF) method) because of insufficient ensemble samples. They can effectively ameliorate the spurious long-range correlations between the background and observations. However, localization is very expensive when the problem to be solved is of high dimension (say 10⁶ or higher) for assimilating observations simultaneously. To reduce the cost of localization for high-dimension problems, an approach is proposed in this paper, which approximately expands the correlation function of the localization matrix using a limited number of principal eigenvectors so that the Schür product between the localization matrix and a high-dimension covariance matrix is reduced to the sum of a series of Schür products between two simple vectors. These eigenvectors are actually the sine functions with different periods and phases. Numerical experiments show that when the number of principal eigenvectors used reaches 20, the approximate expansion of the correlation function is very close to the exact one in the one-dimensional (1D) and two-dimensional (2D) cases. The new approach is then applied to localization in the EnKF method, and its performance is evaluated in assimilation-cycle experiments with the Lorenz-96 model and single assimilation experiments using a barotropic shallow water model. The results suggest that the approach is feasible in providing comparable assimilation analysis with far less cost.

The study of cloud types is critical for understanding atmospheric motions and climate predictions; for example, accurately classified cloud products help improve meteorological predicting accuracies. However, the current satellite cloud classification methods generally analyze the threshold change in a single pixel and do not consider the relationship between the surrounding pixels. The classification development relies heavily on human recourses and does not fully utilize the data-driven advantages of computer models. Here, a new intelligent cloud classification method based on the U-Net network (CLP-CNN) is developed to obtain more accurate, higher frequency, and larger coverage cloud classification products. The experimental results show that the CLP-CNN network can complete a cloud classification task of 800 × 800 pixels in 0.9 s. The classification area covers most of China, and the classification task only needs to use the original L1-level data, which can meet the requirements of a real-time operation. With the Himawari-8 CLTYPE product and the CloudSat 2B-CLDCLASS product as the test comparison target, the CLP-CNN network results matched the Himawari-8 product highly by 76.8%. The probability of detection (POD) was greater than 0.709 for clear skies, deep-convection, and Cirrus–Stratus-type clouds. The probability of detection (POD) and accuracy are improved compared with other deep learning methods.

Reanalysis data play a vital role in weather and climate study as well as meteorological resource development and application. In this work, the East Asia Reanalysis System (EARS) was developed using the Weather Research and Forecasting (WRF) model and the Gridpoint Statistical Interpolations (GSI) data assimilation system. The regional reanalysis system is forced by the European Centre for Medium-Range Weather Forecasts (ECMWF) global reanalysis ERA-Interim data at 6 h intervals. Hourly surface observations are assimilated by the Four-Dimension Data Assimilation (FDDA) scheme during the WRF model integration; upper observations are assimilated in three-dimensional variational data assimilation (3D-VAR) mode at the analysis moment. It should be highlighted that many of the assimilated observations have not been used in other reanalysis systems. The reanalysis runs from 1980 to 2018, producing a regional reanalysis dataset covering East Asia and surrounding areas at 12 km horizontal resolution, 74 sigma levels, and 3 h intervals. Finally, an evaluation of EARS has been performed with respect to the root mean square error (RMSE), based on the 10-year (2008–2017) observational data. Compared to the global reanalysis data of ERA-Interim, the regional reanalysis data of EARS are closer to the observations in terms of RMSE in both surface and upper-level fields. The present study provides evidence for substantial improvements seen in EARS compared to the ERA-Interim reanalysis fields over East Asia. The study also demonstrates the potential use of the EARS data for applications over East Asia and proposes further plans to provide the latest reanalysis in real-time operation mode. Simple data and updated information are available on Zenodo at https://doi.org/10.5281/zenodo.7404918 (Yin et al., 2022), and the full datasets are publicly accessible on the Data-as-a-Service platform of the China Meteorological Administration (CMA) at http://data.cma.cn (last access: 19 May 2023).

In this study, numerical experiments on the sensitivity of initial perturbations to the development of southwest vortices (SWVs), which are frequently observed in Southwest China, are performed with weather research and forecasting model. Regarding the development of SWVs as a strong nonlinear process, conditional nonlinear optimal perturbation (CNOP) with an adjoint‐free algorithm is used. The sensitivity denotes how and where the perturbations of model variables at the initial time can influence non‐moving SWV enhancement. The CNOPs leading to the strengthening of the SWV observed at 5UTC July 17, 2010 are obtained. Three sensitive areas are identified based on the distribution of the most significant CNOP norms. Two of them are located on the Eastern Tibetan Plateau (TP), and the third one is found in the eastern Sichuan Basin (SCB) near the SWV center. Sensitivity experiments indicate that the strengthening of the SWV is mainly associated with perturbations at the middle level at approximately 500–450 hPa in the eastern part of the TP and below 450–400 hPa in the eastern SCB near the center of the SWV. Furthermore, the potential vorticity (PV) budget, divergence, vorticity, vertical velocity, diabatic heating, and southerly wind are also analyzed in different simulations. The results show that at the initial stage of SWV formation, the diabatic heating term is not the main cause of PV budget differences between simulations; it is the dynamical effect that is the main contributing factor. However, when it comes to the intensification stage, the diabatic heating caused by the release of latent heat of condensation becomes important.