Recognizing the importance and challenges inherent in the remote sensing of precipitation in typhoon monitoring, a study of the Advanced Geosynchronous Radiation Imager (AGRI) data from Feng-Yun 4A on typhoon precipitation was conducted. First, Typhoon Maria was selected to statistically analyze the AGRI infrared brightness temperature in the “precipitation” and “nonprecipitation” channels of the field of view. When there was precipitation, the brightness temperature of the AGRI channel changed significantly. Second, the shrunken locally linear embedding algorithm (SLLE) was adopted to carry out the retrieval of precipitation based on the brightness temperatures of AGRI infrared channels 8–14. The contribution rate of the brightness temperature at different channels to the objective function of precipitation retrieval model was obtained by the Bayesian model averaging (BMA). Based on the preliminary experimental “quantification” evaluation index, we concluded that the method adopted in this paper can be used to retrieve precipitation in infrared data and to retrieve the spiral cloud rain bands of a typhoon. Finally, based on the AGRI channel brightness temperature of a 10.8-micron window channel, we applied the membership degree information of a typhoon’s dominant cloud system from the fuzzy c-means (FCM) clustering method to modify precipitation retrieval results. The results were used to obtain the main morphological structure of typhoon precipitation. By further analyzing the temporal variation of dominant cloud system development using the FCM method, we concluded that the brightness temperature gradient can assist in the analysis of the variation of a typhoon’s intensity. This method can be applied to the continuous retrieval of large-scale precipitation. Precipitation retrieval via the AGRI can yield indicators for typhoon precipitation warnings and forecasts, thus providing a reliable reference tool for disaster prevention and mitigation.

Atmospheric water vapor is an important factor in the formation and evolution of extreme weather events, such as heavy rainfall, typhoons, and major droughts and floods. We analyzed the applicability of precipitable water vapor (PWV) values from the Global Navigation Satellite System Meteorology (GNSS/MET) stations in northeastern China. We examined the potential of PWV for precipitation forecasting using an enhanced bidirectional long short‐term memory (BiLSTM) network. Using radiosonde data, the accuracy of GNSS/MET PWV was evaluated, and the diurnal and seasonal differences in retrieval errors were analyzed. Results show that diurnal differences in retrieval errors are insignificant, while seasonal differences are pronounced, which can be attributed to the seasonal distribution of precipitation. Through case analyses of rainstorms and severe convective events, this study concludes that GNSS PWV varies several hours ahead of precipitation. Building on the earlier analyses of applicability assessment and precipitation warning signal identification, the improved BiLSTM framework is employed to investigate the application of GNSS PWV in hourly precipitation forecasting. Feature extraction and data resampling were utilized to enhance the physical interpretability of the binary nature of precipitation prediction and improve the model's generalization capability. Validation with 873 randomly split testing samples revealed a classification accuracy of 86.3% for precipitation prediction, with a regression RMSE of 2.73 mm. The intelligent precipitation forecasting methodology developed in this research can be applied to public‐sector precipitation monitoring and early warning services.

This study adopted an improved fuzzy clustering and optical flow method for the multiscale identification and forecasting of a cloud system based on the cloud images from a 10.8-micron infrared channel of the Advanced Geosynchronous Radiation Imager. First, we used the locally constrained fuzzy c-means (FCM) clustering method to identify typhoon-dominant cloud systems. Second, we coupled the background field-constrained optical flow method with the semi-Lagrangian scheme to forecast typhoon-dominant cloud systems. The experimental results for Typhoon Maria showed that the improved FCM method was able to effectively identify changes in the cloud system while retaining its edge information through the effective removal of the offset field. The identified dominant cloud system was consistent with the precipitation field of the Global Precipitation Measurement mission. We optimized the semi-Lagrangian nonlinear extrapolation of the optical flow field by introducing background field information, thus improving the forecast accuracy of the optical flow field. Based on the assessment indicators of structural similarity, normalized mutual information, peak signal-to-noise ratio, relative standard deviation, and root mean square error, the forecast results demonstrated that the forecast effect of the background field-constrained optical flow method was better than that of the standard optical flow method.

Satellite infrared detectors cannot penetrate clouds, especially precipitating clouds. Improving precipitation estimation accuracy based on infrared brightness temperature has always been important but challenging. In this paper, based on the infrared brightness temperature of the Advanced Geosynchronous Radiation Imager (AGRI) onboard China's Feng‐Yun 4A satellite, we develop and evaluate a new precipitation estimation method. First, using static data, physical characteristics of clouds, cloud image texture features, temporal motion features, and AGRI infrared channel brightness temperature, we construct features for a machine learning model. Then, we develop precipitation estimation methods. Precipitation is estimated in two steps: classification and regression. We employ a random forest classification model to identify whether there is precipitation in a given field of view. If there is precipitation, a multi‐model ensemble regression learning method is used to estimate the areas with this precipitation. The ensemble learning method uses convex optimization to integrate prediction results based on the optimization of hyperparameters of five basic models (i.e., those of random forest, XGBoost, LightGBM, decision tree, and extra tree models). Furthermore, two regression stacking ensemble models—the Least Absolute Shrinkage and Selection Operator (herein referred to as Stacking1‐LASSO) and K‐nearest neighbor (herein referred to as Stacking2‐KNN)—are used to predict the results of the aforementioned basic models. The results of basic models are used as inputs of these two stacking models. Finally, based on the Integrated Multi‐satellitE Retrievals for GPM (IMERG) precipitation product and rain gauge precipitation data, we conduct precipitation estimation experiments and evaluate our methods. The results show that ensemble learning models have greater accuracy in estimating precipitation than the basic models. When using IMERG precipitation as the target precipitation, ensemble learning models can estimate the central area of heavy precipitation during typhoons Ampil and Maria. The ensemble learning estimation effect is better than that of Stacking2‐KNN. Moreover, when rain gauge data is used as the target precipitation, ensemble learning can also estimate the center of heavy precipitation and with good consistency with recorded satellite brightness temperature data.

This study examines precipitation retrieval by the L1‐norm regularization using the infrared brightness temperatures from the Himawari‐8/Advanced Himawari Imager (AHI) of the typhoon Hagibis and analyses the advantages of high temporal resolution data. The AHI data covering the entire life cycle of typhoon Hagibis are implemented and we label “precipitation” and “non‐precipitation” fields‐of‐view (FOVs) based on the variation of brightness temperature gradient in AHI channels 8–16. In the precipitation FOVs, we implement the L1‐norm regularization where contribution rate of the “input variable” in the objective functional is obtained by approximating the degree of freedom for signal. The proposed approach is an effective method to obtain sparse solutions, which can express most or all atomic information by a small amount of atomic information. Several experiments reveal that the retrieved precipitation of typhoon Hagibis exhibits good structural similarity in consistency with the global precipitation measurement (GPM) precipitation values. In particular, the proposed approach provides better structural similarity (SSIM) and peak signal‐to‐noise ratio (PSNR) value compared with the random forest method, which is beneficial in extreme precipitation.