Typhoons are some of the most serious natural disasters, and the key to disaster prevention and mitigation is typhoon level classification. How to better use data of satellite cloud pictures to achieve accurate classification of typhoon levels has become one of classification the hot issues in current studies. A new framework of deep learning neural network, Graph Convolutional–Long Short-Term Memory Network (GC–LSTM), is proposed, which is based on the data of satellite cloud pictures of the Himawari-8 satellite in 2010–2019. The Graph Convolutional Network (GCN) is used to process the irregular spatial structure of satellite cloud pictures effectively, and the Long Short-Term Memory (LSTM) network is utilized to learn the characteristics of satellite cloud pictures over time. Moreover, to verify the effectiveness and accuracy of the model, the prediction effect and model stability are compared with other models. The results show that: the algorithm performance of this model is better than other prediction models; the prediction accuracy rate of typhoon level classification reaches 92.35%, and the prediction accuracy of typhoons and super typhoons reaches 95.12%. The model can accurately identify typhoon eye and spiral cloud belt, and the prediction results are always kept in the minimum range compared with the actual results, which proves that the GC–LSTM model has stronger stability. The model can accurately identify the levels of different typhoons according to the satellite cloud pictures. In summary, the results can provide a theoretical basis for the related research of typhoon level classification.

The lack of accurate estimation of intense precipitation is a universal limitation in precipitation retrieval. Therefore, a new rainfall retrieval technique based on the Random Forest (RF) algorithm is presented using the Advanced Himawari Imager-8 (Himawari-8/AHI) infrared spectrum data and the NCEP operational Global Forecast System (GFS) forecast information. And the gauge-calibrated rainfall estimates from the Global Precipitation Measurement (GPM) product served as the ground truth to train the model. The two-step RF classification model was established for (1) rain area delineation and (2) precipitation grades’ estimation to improve the accuracy of moderate rain and heavy rain. In view of the imbalance categories’ distribution in the datasets, the resampling technique including the Random Under-sampling algorithm and Synthetic Minority Over-sampling Technique (SMOTE) was implemented throughout the whole training process to fully learn the characteristics among the samples. Among the features used, the contributions of meteorological variables to the trained models were generally greater than those of infrared information; in particular, the contribution of precipitable water was the largest, indicating the sufficient necessity of water vapor conditions in rainfall forecasting. The simulation results by the RF model were compared with the GPM product pixel-by-pixel. To prove the universality of the model, we used independent validation sets which are not used for training and two independent testing sets with different periods from the training set. In addition, the algorithm was validated against independent rain gauge data and compared with GFS model rainfall. Consequently, the RF model identified rainfall areas with a Probability Of Detection (POD) of around 0.77 and a False-Alarm Ratio (FAR) of around 0.23 for validation, as well as a POD of 0.60–0.70 and a FAR of around 0.30 for testing. To estimate precipitation grades, the value of classification was 0.70 in validation and in testing the accuracy was 0.60 despite a certain overestimation. In summary, the performance on the validation and test data indicated the great adaptability and superiority of the RF algorithm in rainfall retrieval in East Asia. To a certain extent, our study provides a meaningful range division and powerful guidance for quantitative precipitation estimation.

In the context of global climate change and urban expansion, extreme urban weather events occur frequently and cause significant social problems and economic losses. To study the climate risks associated with rapid urbanization in the global context of climate change, the vulnerability degree of urban agglomeration is constructed by the Grey Model (GM (1, 1)). Based on the sixth phase of the Coupled Model Intercomparison Project (CMIP6) data sets SSP1-2.6, SSP2-4.5, and SSP5-8.5, drought, heat wave, and flood hazards under different emission scenarios are calculated. The vulnerability degree of the urban agglomeration and the climate change hazard were input into the climate change risk assessment model to evaluate future climate change risk. The analysis results show regional differences, with the Beijing–Tianjin–Hebei urban agglomeration having good urban resilience, the Yangtze River Delta urban agglomeration having slightly higher overall risk, and the Pearl River Delta urban agglomeration having the highest relative risk overall. On the whole, the higher the emission intensity is, the greater the risk of climate change to each urban agglomeration under different emission scenarios.

The boundary layer height (BLH) determines the interface between the lower and the free atmosphere, and it is a key variable in numerical simulations and aerosol and environmental pollution studies. This article proposes a novel method in conjunction with numerical regularization to analyze the climate characteristics of the marine boundary layer height (MBLH) using 2007–2011 GPS-RO data from the COSMIC mission. The MBLH corresponds to the smallest gradient, which is calculated using the numerical regulation method where the regulation parameters are determined by the double-parameter model function method. The results reveal the relationship between the MBLH and ocean currents for the first time. A low MBLH is associated with cold seasons and seas where cold ocean currents prevail whereas a high MBLH is related to warm seasons and seas where warm currents prevail. This correlation was validated by comparing the obtained results with different occultation data including atmprf and echprf, which also showed that atmprf is more sensitive to convective cloud top capture. To test the credibility of the results, the standard deviation was used to express the MBLH confidence level. The results show that the standard deviation of the MBLH was highest in low latitudes and lowest in the middle and high latitudes. Furthermore, we analyzed the trends in interannual MBLH variations, which display significant seasonal variations and spatial distributions that correspond with the current and subsolar point. Finally, we conducted a case study in the South China Sea, and identified a distinctive seasonal change and interannual decline in MBLH.

In this paper, we propose a new method of numerical differentiation to determine the height of the top layer of the atmospheric boundary layer. In this method, a regularization technique is used to convert the problem of calculating the differential of the curve of the corners into the problem of finding the minimum value of the objective function. The two-parameter model function method is used to select the regularization parameters. Finally, the maximum gradient method is used to determine the top height of the boundary layer. Firstly, the effectiveness of the new method is validated through two numerical experiments. The experimental results show that as the noise of the occultation data increases, the fluctuation of the height of the boundary layer top obtained by the difference method and the numerical differentiation method combined with the L curve scheme increases. And the height obtained by the two-parameter model function method is very stable, which shows that the new method can filter the noise well, thereby retaining the main information about the profile. Then, based on the COSMIC angle data in January, April, July and October 2007－2011, the new method is used to analyze the seasonal characteristics of the height of the global oceanic and atmospheric boundary layer, compared with the seasonal distribution obtained by “zbalmax” with the occultation data. The results show that the seasonal distribution characteristics of the two data are very consistent: the height of the boundary layer is higher in the area where the sea surface temperature is higher than that in the surrounding sea area; on the contrary, the height of the boundary layer top is lower. In the sea area where the warm current passes, the height of the boundary layer is higher; in the sea area where the cold current passes, the height of the boundary layer is lower.