Leiming Ma

Urban severe convective rainfall, a major urban hazard modulated by microscale cloud processes, mesoscale storm dynamics, and synoptic-scale forcing, severely threatens the safety of coastal megacities and urban infrastructure. As a core megacity in the Yangtze River Delta (YRD) region, Shanghai is significantly affected by the combined effects of the East Asian monsoon, complex urban-topography interactions, and sea-land breeze circulations, which not only amplify the intensity and frequency of extreme rainfall events in the city but also across the YRD region. These events further trigger severe urban waterlogging, traffic congestion, and damage to urban infrastructure. To address these challenges and advance integrated urban hazard forecasting, we develop a Physics-Informed Cross-Scale Synergistic Framework (CSSF) for urban severe convective rainfall hazards, which integrates multi-band radar observations and ordinal regression to update Numerical Weather Prediction (NWP) model forecasts. The integrated system features three core components aligned with urban hazard monitoring and forecasting needs: (1) Multi-Band Radar Observation Fusion Module, which fuses S/X-band radar data from a high-density network with machine learning-based synthesis loss functions, capturing fine-scale convective features to provide observational constraints for model forecast updates; (2) Gated Vertical Information Propagation, an improved ConvLSTM architecture with skip connections and bidirectional vertical information flow, which links radar-derived convective features to the temporal evolution of NWP forecasts, ensuring physical consistency in update processes; (3) Ordinal Regression-Based Forecast Update Module, which employs ordinal regression to integrate radar observations into NWP model outputs, updating forecast results to better reflect real-time convective development and extreme rainfall characteristics in urban areas. The integrated system is validated using 2019–2023 multi-band radar data from Shanghai’s high-density monitoring network, ERA5 reanalysis data, and NWP forecasts. Validation results demonstrate that the CSSF outperforms traditional ConvLSTM models and operational NWP systems in capturing convective initiation, extreme rainfall intensity, and urban waterlogging-prone areas, primarily by using radar observations and ordinal regression to dynamically update model forecasts, bridging the gap between observation and NWP forecast. This study presents a holistic integrated system for urban severe convective rainfall hazards, highlighting the value of radar-ordinal regression integration for NWP forecast updates, and provides a transferable approach to enhance urban hazard resilience and support evidence-based disaster risk management for climate-vulnerable cities.

Current deep learning methods for predicting tropical cyclone intensity are restricted by inadequate utilization of climate factors, insufficient feature fusion, and the inability to automatically adapt input data to changes in the scale of cyclone structures. To address these challenges, this article proposes a multimodal data fusion model guided by sequential infrared satellite images (IRGMF). The IRGMF model integrates a broader range of climate factors for a more comprehensive, enriched feature representation of tropical cyclone intensity. In addition, we designed a temporal attention module that allows the model to focus on critical moments in the sequential data for a more precise capture of the dynamic features of climate factors as they evolve. To enhance prediction accuracy, we also developed an attention-cropping module that adaptively crops irrelevant information outside the scale of a tropical cyclone from reanalysis data. To validate the effectiveness of the proposed model, we conducted experiments on multiple datasets, including CMA-BST, ERA5, and GridSat-B1. Experimental results show that, compared to existing methods, the proposed IRGMF model demonstrates significant advantages in predicting tropical cyclone intensity and provides superior accurate information for meteorological forecasting and decision support.

Rainfall forecasting holds significant importance across a wide range of sectors, including disaster prevention, energy planning, and agriculture. In the past decade, artificial intelligence (AI) has emerged as a revolutionary approach, aiming to overcome the long-standing limitations of traditional numerical weather prediction (NWP) models and statistical downscaling models (SDMs) for rainfall forecasting. This chapter briefly introduces the remarkable progress made in AI-based rainfall forecasting. It mainly focuses on three major aspects: physical-constrained machine learning (ML), multi-modal data fusion, and extreme event prediction. AI-based models can be used to resolve the subgrid-scale parameterization problems (e.g., convective parameterization) that troubled NWP models for a long time. For instance, DeepMind’s GraphCast employs dynamic graph neural networks to generate a high-resolution global forecast. Making 10-day forecasts with GraphCast takes less than a minute on a single Google TPU v4 machine. Regarding multi-modal data fusion, systems such as National Oceanic and Atmospheric Administration (NOAA) Multi-Radar Multi-Sensor (MRMS) combine various data sources and significantly improve the accuracy of forecasts. For the extreme rainfall prediction, the application of adversarial training and attention mechanisms has also led to improvements. The review finally suggests the future research directions. It emphasizes how AI is updating rainfall forecasting technology, enabling it to better meet the challenges posed by a changing climate.

Severe convective weather systems, characterized by their small spatial scale and rapid, violent development, frequently give rise to disasters such as rainstorms, lightning, gales, and hail. Accurate forecast of such systems has long been a challenging issue of weather forecasting and a dilemma for disaster prevention and mitigation in Shanghai. This research introduces an intelligent forecasting technology for severe convective rainfall systems in Shanghai, encompassing adaptive radar observation of strong convection targets, identification and prediction of typical convective features, machine learning-based correction of numerical prediction errors, and system integration.In this technology, to address the problem of unbalanced samples with a scarcity of heavy rainfall cases, an autoencoder for noise reduction and ordinal boosting regression module is designed. A FocalLoss method is employed to weight the Loss function, thereby transforming the regression task of precipitation values into multiple classification tasks. An adaptive scale selection method is constructed to better represent the heavy rainfall system. In this method, the spatio-temporal scales of convective systems are adaptively selected, enabling targeted correction of heavy rainfall prediction.Finally, an intelligent monitoring and early warning system capable of predicting the three-dimensional structure and evolution of strong convective systems has been established and put into operation. This system was evaluated during the flood seasons of 2022-2023. The results indicate that the TS score for 24-hour heavy rainfall (50mm) was significantly enhanced compared to the operational numerical weather prediction system of Shanghai Meteorological Service (SMS). This technology has been extended to application in urban disaster prevention and mitigation.

Tropical cyclone (TC) is a highly catastrophic weather event, and accurate estimation of intensity is of great significance. The current proposed TC intensity estimation model focuses on training using satellite images from a single or two channels, and the model cannot fully capture features related to TC intensity, resulting in low accuracy. To this end, we propose a double-layer encoder-decoder model for estimating the intensity of TC, which is trained using images from three channels: infrared, water vapor, and passive microwave. The model mainly consists of three modules: wavelet transform enhancement module, multi-channel satellite image fusion module, and TC intensity estimation module, which are used to extract high-frequency information from the source image, generate a three channel fused image, and perform TC intensity estimation. To validate the performance of our model, we conducted extensive experiments on the TCIR dataset. The experimental results show that the proposed model has MAE and RMSE of 3.76m/s and 4.62m/s for TC intensity estimation, which are 15.70% and 20.07% lower than advanced Dvorak technology (ADT), respectively. Therefore, the model proposed in this article has great potential in accurately estimating TC intensity.

Currently, the extraction of the size characteristics of a tropical cyclone (TC) from satellite data is mainly based on subjective human experience and archived satellite data. Few objective methods have been developed to estimate the size of a TC. To overcome this problem, this study introduces a new deep learning model, called a Tropical Cyclone Multi-task Estimation (TCMSE) model. The goal of using the model is to accurately extract the spatio-temporal characteristics of the size of a TC from infrared satellite images acquired during the current and previous three moments, combined with a multi-task learning approach and physically assisted task learning, to produce more reliable estimates of the wind radii of a TC. We used the TCMSE model to estimate the TC wind radius on the infrared satellite image of the Northwest Pacific Ocean to verify the performance of the model. Experimental results show that the proposed deep spatiotemporal feature extraction network, which combines multi-task learning and physically assisted tasks, is superior to existing methods for estimating the wind radii of a TC. More specifically, the TCMSE model reduces the MAE error in estimating the TC wind radii by 6%, compared to the multi-task learning approach and 14%, compared to separately learning a single wind radii task.

Weather radar echo extrapolation is one of the essential means for weather nowcasting. It has been considerably inspired over the last decade by deep learning. However, the internal similarity of the echo evolution process has little been exploited. To investigate this merit, a deep contrastive model with an encoder–projector structure is proposed in this letter, which projects the subsequences sampled from the same evolution process into the neighborhood of latent space by contrastive learning. Thus, the internal evolution similarity of the input echo sequence itself can be discovered and exploited for promoting prediction. To make the training smoother, we also adopt a cumulative sampling strategy that follows a simple-to-hard manner. Experimental results on two real-world radar datasets demonstrate the superiority of our model in comparison to state-of-the-art. The effectiveness of the sampling strategy and extrapolation ability on limited input is also analyzed and verified. Training code and pretrained models are available at https://github.com/tolearnmuch/ESCL .

The properties of aerosols are highly uncertain owing to the complex changes in their composition in different regions. The radiative properties of different aerosol types differ considerably and are vital for studying aerosol regional and/or global climate effects. Traditional aerosol-type identification algorithms, generally based on cluster or empirical analysis methods, are often inaccurate and time-consuming. In response, our study aimed to develop a new aerosol-type classification model using an innovative hybrid algorithm to improve the precision and efficiency of aerosol-type identification. This novel algorithm incorporates an optical database, constructed using the Mie scattering model, and employs a random forest algorithm to classify different aerosol types based on the optical data from the database. The complex refractive index was used as a baseline to assess the performance of our hybrid algorithm against the traditional Gaussian kernel density clustering method for aerosol-type identification. The hybrid algorithm demonstrated impressive consistency rates of 90 %, 85 %, 84 %, 84 %, and 100 % for dust, mixed-coarse (mixed, course-mode aerosol), mixed-fine (mixed, fine-mode aerosol), urban/industrial, and biomass burning aerosols, respectively. Moreover, it achieved remarkable precision, with evaluation metric indexes for micro-precision, micro-recall, micro-F1-score, and accuracy of 95 %, 89 %, 91 %, and 89 %, respectively. Lastly, a global map of aerosol types was generated using the new hybrid algorithm to characterize aerosol types across the five continents. This study, utilizing a novel approach for the classification of aerosol, will help improve the accuracy of aerosol inversion and determine the sources of aerosol pollution.

Aerosols have great uncertainty owing to the complex changes in their composition in different regions. The radiation properties of different aerosol types differ considerably and are vital in studying aerosol regional and/or global climate effects. Traditional aerosol-type identification algorithms, generally based on cluster or empirical analysis methods, are often inaccurate and time-consuming. Hence, we aimed to develop a new aerosol-type classification model using an innovative hybrid algorithm to improve the precision and efficiency of aerosol-type identification. An optical database was built using Mie scattering and a complex refractive index was used as a baseline to identify different aerosol types by applying a random forest algorithm to train the aerosol optical parameters obtained from the Aerosol Robotic Network sites. The consistency rates of the new model with the traditional Gaussian density cluster method were 90 %, 85 %, 84 %, 84 %, and 100 % for dust, mixed-coarse, mixed-fine, urban/industrial, and biomass burning aerosols, respectively. The corresponding precision of the new hybrid algorithm (F-score and accuracy scores) was 95 %, 89 %, 91 %, and 89 %. Lastly, a global map of aerosol types was generated using the new model to characterize aerosol types across the five continents. This study utilizing a novel approach for the classification of aerosol will help improve the accuracy of aerosol inversion and determine the sources of aerosol pollution.