FengYun-4A (FY-4A)’s Geostationary Interferometric Infrared Sounder (GIIRS) is the first hyperspectral infrared sounder on board a geostationary satellite, enabling the collection of infrared detection data with high temporal and spectral resolution. As clouds have complex spectral characteristics, and the retrieval of atmospheric profiles incorporating clouds is a significant problem, it is often necessary to undertake cloud detection before further processing procedures for cloud pixels when infrared hyperspectral data is entered into assimilation system. In this study, we proposed machine-learning-based cloud detection models using two kinds of GIIRS channel observation sets (689 channels and 38 channels) as features. Due to differences in surface cover and meteorological elements between land and sea, we chose logistic regression (lr) model for the land and extremely randomized tree (et) model for the sea respectively. Six hundred and eighty-nine channels models produced slightly higher performance (Heidke skill score (HSS) of 0.780 and false alarm rate (FAR) of 16.6% on land, HSS of 0.945 and FAR of 4.7% at sea) than 38 channels models (HSSof 0.741 and FAR of 17.7% on land, HSS of 0.912 and FAR of 7.1% at sea). By comparing visualized cloud detection results with the Himawari-8 Advanced Himawari Imager (AHI) cloud images, the proposed method has a good ability to identify clouds under circumstances such as typhoons, snow covered land, and bright broken clouds. In addition, compared with the collocated Advanced Geosynchronous Radiation Imager (AGRI)-GIIRS cloud detection method, the machine learning cloud detection method has a significant advantage in time cost. This method is not effective for the detection of partially cloudy GIIRS’s field of views, and there are limitations in the scope of spatial application.

Observations of sea surface wind field are critical for typhoon prediction. The scatterometer observation is one of the most important sources of sea surface winds, which provides both wind speed and wind direction information. However, the spatial resolution of scatterometer wind is low. Synthetic Aperture Radar (SAR) can provide a more detailed wind structure of the tropical cyclone. In addition, the cross-polarization observation of SAR can provide more detailed information of high speed wind (>25 m·s ${}^{-1}$ ) than the scatterometer. Nevertheless, due to the narrow swath of SAR, the number of retrieved sea surface wind data used in the data assimilation is limited, and another limitation of SAR wind observation is that it does not provide true wind direction information. In this paper, the joint assimilation of the Advanced Scatterometer (ASCAT) wind and Sentinel-1 SAR wind was investigated. Another limitation in the current operational typhoon prediction is the inefficient quality control (QC) method used in the data assimilation since a large number of high speed wind observations was rejected by the traditional Gaussian distribution QC. We introduce the Huber norm distribution quality control (QC) into the data assimilation successfully. A numerical simulation experiment of typhoon by Lionrock (2016) is conducted to test the proposed method. The experimental results showed that the new quality control scheme not only greatly increases the availability of wind data in the area of the typhoon center, but also improves the typhoon track prediction, as well as the intensity prediction. The joint assimilation of scatterometer and SAR winds does have a positive impact on the typhoon prediction.

High-resolution synthetic aperture radar (SAR) wind observations provide fine structural information for tropical cycles and could be assimilated into numerical weather prediction (NWP) models. However, in the conventional method assimilating the u and v components for SAR wind observations (SAR_uv), the wind direction is not a state vector and its observational error is not considered during the assimilation calculation. In this paper, an improved method for wind observation directly assimilates the SAR wind observations in the form of speed and direction (SAR_sd). This method was implemented to assimilate the sea surface wind retrieved from Sentinel-1 synthetic aperture radar (SAR) in the basic three-dimensional variational system for the Weather Research and Forecasting Model (WRF 3DVAR). Furthermore, a new quality control scheme for wind observations is also presented. Typhoon Lionrock in August 2016 is chosen as a case study to investigate and compare both assimilation methods. The experimental results show that the SAR wind observations can increase the number of the effective observations in the area of a typhoon and have a positive impact on the assimilation analysis. The numerical forecast results for this case show better results for the SAR_sd method than for the SAR_uv method. The SAR_sd method looks very promising for winds assimilation under typhoon conditions, but more cases need to be considered to draw final conclusions.

Polar sea ice is undergoing rapid change, with recent record-low extents in both hemispheres, raising the demand for skillful predictions from days to seasons for navigation, ecosystem management, and climate risk assessment. Accurate sea ice prediction is essential for understanding coupled climate processes, supporting safe polar operations, and informing adaptation strategies. Physics-based numerical models remain the backbone of operational forecasting, but their skill is limited by uncertainties in coupled ocean–ice–atmosphere processes, parameterizations, and sparse observations, especially in the marginal ice zone and during melt seasons. Statistical and empirical models can provide useful baselines for low-dimensional indices or short lead times, yet they often struggle to represent high-dimensional, nonlinear interactions and regime shifts. This review synthesizes recent progress of DL for key sea ice prediction targets, including sea ice concentration/extent, thickness, and motion, and organizes methods into (i) sequential architectures (e.g., LSTM/GRU and temporal Transformers) for temporal dependencies, (ii) image-to-image and vision models (e.g., CNN/U-Net, vision Transformers, and diffusion or GAN-based generators) for spatial structures and downscaling, and (iii) spatiotemporal fusion frameworks that jointly model space–time dynamics. We further summarize hybrid strategies that integrate DL with numerical models through post-processing, emulation, and data assimilation, as well as physics-informed learning that embeds conservation laws or dynamical constraints. Despite rapid advances, challenges remain in generalization under non-stationary climate conditions, dataset shift, and physical consistency (e.g., mass/energy conservation), interpretability, and fair evaluation across regions and lead times. We conclude with practical recommendations for future research, including standardized benchmarks, uncertainty-aware probabilistic forecasting, physics-guided training and neural operators for long-range dynamics, and foundation models that leverage self-supervised pretraining on large-scale Earth observation archives.