Search Results (1,117)

This paper presents a Systematic Mapping Study (SMS) on data-driven approaches in flood forecasting from 2019 to 2024, a period marked by transformative developments in Deep Learning (DL) technologies. Analysing 363 selected studies, this paper provides an overview of the technological evolution in this field, methodological approaches, evaluation practices and geographical distribution of studies. The study revealed that meteorological and hydrological factors constitute approximately 76% of input variables, with rainfall/precipitation and water level measurements forming the core predictive basis. Long Short-Term Memory (LSTM) networks emerged as the dominant algorithm (21% of implementations), whilst hybrid and ensemble approaches showed the most dramatic growth (from 2% in 2019 to 10% in 2024). The study also revealed a threefold increase in publications during this period, with significant geographical concentration in East and Southeast Asia (56% of studies), particularly China (36%). Several research gaps were identified, including limited exploration of graph-based approaches for modelling spatial relationships, underutilisation of transfer learning for data-scarce regions, and insufficient uncertainty quantification. This SMS provides researchers and practitioners with actionable insights into current trends, methodological practices, and future directions in data-driven flood forecasting, thereby advancing this critical field for disaster management. Full article

The need for systems that forecast and respond proactively to meteorological disasters is growing amid climate variability. Although the early warning system in South Korea includes countermeasure information, it remains limited in terms of data recency, granularity, and regional adaptability. Additionally, its closed architecture hinders interoperability with external systems. This study aims to redesign the countermeasure function as an independent cloud-based platform grounded in the common standard terminology framework in South Korea. A multi-dimensional data model was developed using attributes such as crop type, cultivation characteristics, growth stage, disaster type, and risk level. The platform incorporates user-specific customization features and history tracking capabilities, and it is structured using a microservices architecture to ensure modularity and scalability. The proposed system enables real-time management and dissemination of localized countermeasure suggestions tailored to various user types, including central and local governments and farmers. This study offers a practical model for enhancing the precision and applicability of agrometeorological response information. It is expected to serve as a scalable reference platform for future integration with external agricultural information systems. Full article

In Taiwan, hydro-meteorological data are fragmented across multiple agencies, limiting the effectiveness of coordinated flood response. To address this challenge and the increasing uncertainty associated with extreme rainfall, a real-time disaster prevention platform has been developed. This system integrates multi-source data and geospatial information through a cluster-based architecture to enhance pluvial flood management. Built on a Service-Oriented Architecture (SOA) and incorporating Internet of Things (IoT) technologies, AI-based convolutional neural networks (CNNs), and 3D drone mapping, the platform enables automated alerts by linking sensor thresholds with real-time environmental data, facilitating synchronized operational responses. Deployed in New Taipei City over the past three years, the system has demonstrably reduced flood risk during severe rainfall events. Region-specific action thresholds and adaptive strategies are continually refined through feedback mechanisms, while integrated spatial and hydrological trend analyses extend the lead time available for emergency response. Full article

Accurate short-term weather forecasting plays a vital role in disaster response, agriculture, and energy management, where timely and reliable predictions are essential for decision-making. Graph neural networks (GNNs), known for their ability to model complex spatial structures and relational data, have achieved remarkable success in meteorological forecasting by effectively capturing spatial dependencies among distributed weather stations. However, most existing GNN-based approaches rely on pairwise station connections, limiting their capacity to represent higher-order spatial interactions. Moreover, their dependence on supervised learning makes them vulnerable to spatial heterogeneity and temporal non-stationarity. This paper introduces a novel spatial–temporal pretraining framework, Hypergraph-enhanced Meteorological Pretraining (HyMePre), which combines hypergraph neural networks with self-supervised learning to model high-order spatial dependencies and improve generalization across diverse climate regimes. HyMePre employs a two-stage masking strategy, applying spatial and temporal masking separately, to learn disentangled representations from unlabeled meteorological time series. During forecasting, dynamic hypergraphs group stations based on meteorological similarity, explicitly capturing high-order dependencies. Extensive experiments on large-scale reanalysis datasets show that HyMePre outperforms conventional GNN models in predicting temperature, humidity, and wind speed. The integration of pretraining and hypergraph modeling enhances robustness to noisy data and improves generalization to unseen climate patterns, offering a scalable and effective solution for operational weather forecasting. Full article

Seed germination is a critical phase in the plant lifecycle of Camellia oleifera (oil tea), directly influencing seedling establishment and crop reproduction. In this study, we examined transcriptomic and physiological changes across five defined germination stages (G0–G4), from radicle dormancy to cotyledon emergence. Using RNA sequencing (RNA-seq), we assembled 169,652 unigenes and identified differentially expressed genes (DEGs) at each stage compared to G0, increasing from 1708 in G1 to 10,250 in G4. Functional enrichment analysis revealed upregulation of genes associated with cell wall organization, glucan metabolism, and Photosystem II assembly. Key genes involved in cell wall remodeling, including cellulose synthase (CESA), phenylalanine ammonia-lyase (PAL), 4-coumarate-CoA ligase (4CL), caffeoyl-CoA O-methyltransferase (COMT), and peroxidase (POD) showed progressive activation during germination. A Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis revealed dynamic regulation of phenylpropanoid and flavonoid biosynthesis, photosynthesis, carbohydrate metabolism, and hormone signaling pathways. Transcription factors such as indole-3-acetic acid (IAA), ABA-responsive element binding factor (ABF), and basic helix–loop–helix (bHLH) were upregulated, suggesting hormone-mediated regulation of dormancy release and seedling development. Physiologically, cytokinin (CTK) and IAA levels peaked in G4, antioxidant enzyme activities were highest in G2, and starch content increased toward later stages. These findings provide new insights into the molecular mechanisms underlying seed germination in C. oleifera and identify candidate genes relevant to rootstock breeding and nursery propagation. Full article

The inter-provincial border region in eastern China, encompassing the junction of Jiangsu, Anhui, Shandong, and Henan provinces, serves as a crucial zone that connects the important economic zones of Beijing–Tianjin–Hebei and the Yangtze River Delta. It is of great significance to study the temporal variation characteristics, spatial distribution patterns, and driving factors of PM_2.5 concentrations in this region. Based on the PM_2.5 concentration observation data, ground meteorological data, environmental data, and socio-economic data from 2022 to 2024, this study conducted in-depth and systematic research by using advanced methods, such as spatial autocorrelation analysis and geographical detectors. The research results show that the concentration of PM_2.5 rose from 2022 to 2023, but decreased from 2023 to 2024. From the perspective of seasonal variations, the concentration of PM_2.5 shows a distinct characteristic of being “high in winter and low in summer”. The monthly variation shows a “U”-shaped distribution pattern. In terms of spatial changes, the PM_2.5 concentration in the inter-provincial border region of eastern China (Jiangsu, Anhui, Shandong, Henan) forms a gradient difference of “higher in the west and lower in the east”. The high-concentration agglomeration areas are mainly concentrated in the Henan part of the study region, while the low-concentration agglomeration areas are distributed in the eastern coastal parts of the study region. The analysis of the driving factors of the PM_2.5 concentration based on geographical detectors reveals that the average temperature is the main factor affecting the PM_2.5 concentration. The interaction among the factors contributing to the spatial differentiation of the PM_2.5 concentration is very obvious. Temperature and population density (q = 0.92), temperature and precipitation (q = 0.95), slope and precipitation (q = 0.97), as well as DEM and population density (q = 0.96), are the main combinations of factors that have continuously affected the spatial differentiation of the PM_2.5 concentration for many years. The research results from this study provide a scientific basis and decision support for the prevention, control, and governance of PM_2.5 pollution. Full article

This study evaluated the wind speed simulation performance of the Weather Research and Forecasting (WRF) model at three resolutions in Hebei Province based on wind speed data from 2022. The results show that the simulation effectiveness of the WRF model for wind speeds at different heights varies significantly under different seasons and topographic conditions. In general, the model simulates the wind speed at the high level most accurately, followed by the mid level, and the simulation of low level wind speed shows the largest bias. Increasing the model resolution significantly improves the simulation of low-level wind speed, and the 5 km resolution performs best at most stations; while for the mid-level and high-level wind speeds, increasing the resolution does not significantly improve the simulation effect, and the high-resolution simulation has a greater bias at some stations. In terms of topographic features, wind speeds are generally better simulated in mountainous areas than in the plains during spring, summer, and autumn, while the opposite is true in winter. These findings provide scientific reference for WRF model optimal resolution selection and wind resource assessment. Full article

A reliable precipitation dataset with high spatial resolution is essential for climate research in the Tarim Basin. This study evaluated the performances of four models, namely a random forest (RF), a long short-term memory network (LSTM), a support vector machine (SVM), and a feedforward neural network (FNN). FNN, which was found to be superior to the other models, was used to integrate eight precipitation datasets spanning from 1990 to 2022 across the Tarim Basin, resulting in a new monthly high-resolution (0.1°) precipitation dataset named MoHiPr-TB. This dataset was subsequently bias-corrected by the China Land Data Assimilation System version 2.0 (CLDAS2.0). Validation results indicate that the corrected MoHiPr-TB not only accurately reflects the spatial distribution of precipitation but also effectively simulates its intensity and interannual and seasonal variations. Moreover, MoHiPr-TB is capable of detecting the precipitation–elevation relationship in the Pamir Plateau, where precipitation initially increases and then decreases with elevation, as well as the synchronous variation of precipitation and elevation in the Tianshan region. Collectively, this study delivers a high-accuracy precipitation dataset for the Tarim Basin, which is anticipated to have extensive applications in meteorological, hydrological, and ecological research. Full article

The spatiotemporal changes and driving mechanisms of rain-on-snow (ROS) events and their intensities are crucial for responding to disasters triggered by such events. However, there is currently a lack of detailed assessment of the seasonal variations and driving mechanisms of ROS events and their intensities in China. Therefore, this study utilized daily meteorological data and daily snow depth data from 513 stations in China during 1978–2020 to investigate spatiotemporal variations of ROS events and their intensities. Also, based on the detrend and partial correlation analysis model, the driving factors of ROS events and their intensity were explored. The results showed that ROS events primarily occurred in northern Xinjiang, the Qinghai–Tibet Plateau, Northeast China, and central and eastern China. ROS events frequently occurred in the middle and lower Yangtze River Plain in winter but were easily overlooked. The number and intensity of ROS events increased significantly (p < 0.05) in the Changbai Mountains in spring and the Altay Mountains and the southeast part of the Qinghai–Tibet Plateau in winter, leading to heightened ROS flood risks. However, the number and intensity of ROS events decreased significantly (p < 0.05) in the middle and lower Yangtze River Plain in winter. The driving mechanisms of the changes for ROS events and their intensities were different. Changes in the number of ROS events and their intensities in snow-rich regions were driven by rainfall days and quantity of rainfall, respectively. In regions with more rainfall, these changes were driven by snow cover days and snow water equivalent, respectively. Air temperature had no direct impact on ROS events and their intensities. These findings provide reliable evidence for responding to disasters and changes triggered by ROS events. Full article

This study presents AD-MRCNN, an advanced deep learning framework for automated atmospheric front detection that addresses two critical limitations in existing methods. First, current approaches directly input raw meteorological data without optimizing feature compatibility, potentially hindering model performance. Second, they typically only provide frontal category information without identifying individual frontal systems. Our solution integrates two key innovations: 1. An intelligent adapter module that performs adaptive feature fusion, automatically weighting and combining multi-source meteorological inputs (including temperature, wind fields, and humidity data) to maximize their synergistic effects while minimizing feature conflicts; the utilized network achieves an average improvement of over 4% across various metrics. 2. An enhanced instance segmentation network based on Mask R-CNN architecture that simultaneously achieves (1) precise frontal type classification (cold/warm/stationary/occluded), (2) accurate spatial localization, and (3) identification of distinct frontal systems. Comprehensive evaluation using ERA5 reanalysis data (2009–2018) demonstrates significant improvements, including an 85.1% F1-score, outperforming traditional methods (TFP: 63.1%) and deep learning approaches (Unet: 83.3%), and a 31% reduction in false alarms compared to semantic segmentation methods. The framework’s modular design allows for potential application to other meteorological feature detection tasks. Future work will focus on incorporating temporal dynamics for frontal evolution prediction. Full article

Meteorological observations are essential for climate modelling, prediction, early warning systems, decision-making processes, and disaster management. These observations are critical to societal development and the safeguarding of human activities and livelihoods. Spatial interpolation techniques play a pivotal role in addressing gaps between observation sites, enabling the generation of continuous meteorological datasets. However, due to the inherent complexity of atmosphere–surface interactions, no single interpolation technique has proven universally effective in achieving consistently accurate results for meteorological variables. This study proposes a novel interpolation model based on Fuzzy Adaptive Optimal Fusion (FAOF). The FAOF model integrates fuzzy theory by constructing station-specific fuzzy sets and sub-method element pools, employing a nonlinear membership function with error as the independent variable. An iterative accuracy index is used to identify the optimal parameter combination, facilitating adaptive data fusion and interpolation optimisation. The model’s performance is evaluated against 10 individual methods from the method pool. Experimental results demonstrate that FAOF effectively combines the strengths of multiple methods, achieving significantly enhanced interpolation accuracy. Additionally, the model consistently performs well across diverse regions and meteorological variables, underscoring its robustness and strong generalisation capability. Full article

Meiyu is a critical component of the summer rainy season over the Yangtze–Huaihe River Basin (YHRB) in China, and the Meiyu onset date (MOD), serving as a key indicator of Meiyu, has garnered substantial attention. This article demonstrates an in-phase relationship between MOD and the preceding spring Barents–Kara Seas ice concentration (BKSIC) during 1979–2023. Specifically, the loss of spring BKSIC promotes an earlier MOD. Further analysis indicates that decreased spring BKSIC reduces the reflection of shortwave radiation, thereby enhancing oceanic solar radiation absorption and warming sea surface temperature (SST) in spring. The warming SST persists into summer and induces significant deep warming in the BKS through enhanced upward longwave radiation. The BKS deep warming triggers a wave train propagating southeastward to the East Asia–Northwest Pacific region, leading to a strengthened East Asian Subtropical Jet and an intensified Western North Pacific Subtropical High in summer. Under these conditions, the transport of warm and humid airflows into the YHRB is enhanced, promoting convective instability through increased low-level warming and humidity, combined with enhanced wind shear, which jointly contribute to an earlier MOD. These results may advance the understanding of MOD variability and provide valuable information for disaster prevention and mitigation. Full article

Radar echo extrapolation is a critical forecasting tool in the field of meteorology, playing an especially vital role in nowcasting and weather modification operations. In recent years, spatiotemporal sequence prediction models based on deep learning have garnered significant attention and achieved notable progress in radar echo extrapolation. However, most of these extrapolation network architectures are built upon convolutional neural networks, using radar echo images as input. Typically, radar echo intensity values ranging from −5 to 70 dBZ with a resolution of 5 dBZ are converted into 0–255 grayscale images from pseudo-color representations, which inevitably results in the loss of important echo details. Furthermore, as the extrapolation time increases, the smoothing effect inherent to convolution operations leads to increasingly blurred predictions. To address the algorithmic limitations of deep learning-based echo extrapolation models, this study introduces three major improvements: (1) A Deep Convolutional Generative Adversarial Network (DCGAN) is integrated into the ConvLSTM-based extrapolation model to construct a DCGAN-enhanced architecture, significantly improving the quality of radar echo extrapolation; (2) Considering that the evolution of radar echoes is closely related to the surrounding meteorological environment, the study incorporates specific physical variable products from the initial zero-hour field of RMAPS-NOW (the Rapid-update Multiscale Analysis and Prediction System—NOWcasting subsystem), developed by the Institute of Urban Meteorology, China. These variables are encoded jointly with high-resolution (0.5 dB) radar mosaic data to form multiple radar cells as input. A multi-channel radar echo extrapolation network architecture (MR-DCGAN) is then designed based on the DCGAN framework; (3) Since radar echo decay becomes more prominent over longer extrapolation horizons, this study departs from previous approaches that use a single model to extrapolate 120 min. Instead, it customizes time-specific loss functions for spatiotemporal attenuation correction and independently trains 20 separate models to achieve the full 120 min extrapolation. The dataset consists of radar composite reflectivity mosaics over North China within the range of 116.10–117.50°E and 37.77–38.77°N, collected from June to September during 2018–2022. A total of 39,000 data samples were matched with the initial zero-hour fields from RMAPS-NOW, with 80% (31,200 samples) used for training and 20% (7800 samples) for testing. Based on the ConvLSTM and the proposed MR-DCGAN architecture, 20 extrapolation models were trained using four different input encoding strategies. The models were evaluated using the Critical Success Index (CSI), Probability of Detection (POD), and False Alarm Ratio (FAR). Compared to the baseline ConvLSTM-based extrapolation model without physical variables, the models trained with the MR-DCGAN architecture achieved, on average, 18.59%, 8.76%, and 11.28% higher CSI values, 19.46%, 19.21%, and 19.18% higher POD values, and 19.85%, 11.48%, and 9.88% lower FAR values under the 20 dBZ, 30 dBZ, and 35 dBZ reflectivity thresholds, respectively. Among all tested configurations, the model that incorporated three physical variables—relative humidity (rh), u-wind, and v-wind—demonstrated the best overall performance across various thresholds, with CSI and POD values improving by an average of 16.75% and 24.75%, respectively, and FAR reduced by 15.36%. Moreover, the SSIM of the MR-DCGAN models demonstrates a more gradual decline and maintains higher overall values, indicating superior capability in preserving echo structural features. Meanwhile, the comparative experiments demonstrate that the MR-DCGAN (u, v + rh) model outperforms the MR-ConvLSTM (u, v + rh) model in terms of evaluation metrics. In summary, the model trained with the MR-DCGAN architecture effectively enhances the accuracy of radar echo extrapolation. Full article

Hurricanes can cause massive power outages and pose significant disruptions to society. Accurately monitoring hurricane power outages will improve predictive models and guide disaster emergency management. However, many challenges exist in obtaining high-quality data on hurricane power outages. We systematically evaluated machine learning (ML) approaches to reconstruct historical hurricane power outages based on high-resolution (1 km) satellite night light observations from the Defense Meteorological Satellite Program (DMSP) and other ancillary information. We found that the two-step hybrid model significantly improved model prediction performance by capturing a substantial portion of the uncertainty in the zero-inflated data. In general, the classification and regression tree-based machine learning models (XGBoost and random forest) demonstrated better performance than the logistic and CNN models in both binary classification and regression models. For example, the xgb+xgb model has 14% less RMSE than the log+cnn model, and the R-squared value is 25 times larger. The Interpretable ML (SHAP value) identified geographic locations, population, and stable and hurricane night light values as important variables in the XGBoost power outage model. These variables also exhibit meaningful physical relationships with power outages. Our study lays the groundwork for monitoring power outages caused by natural disasters using satellite data and machine learning (ML) approaches. Future work should aim to improve the accuracy of power outage estimations and incorporate more hurricanes from the recently available Visible Infrared Imaging Radiometer Suite (VIIRS) Day/Night Band (DNB) data. Full article

Short-term precipitation forecasting is a core task in meteorological science, aiming to achieve accurate predictions by modeling the spatiotemporal evolution of radar echo sequences, thereby supporting meteorological services and disaster warning systems. However, existing spatiotemporal sequence prediction methods still struggle to disentangle complex spatiotemporal dependencies effectively and fail to capture the nonlinear chaotic characteristics of precipitation systems. This often results in ambiguous predictions, attenuation of echo intensity, and spatial localization errors. To address these challenges, this paper proposes a unified spatiotemporal sequence prediction framework based on spatiotemporal masking, which comprises two stages: self-supervised pre-training and task-oriented fine-tuning. During pre-training, the model learns global structural features of meteorological systems from sparse contexts by randomly masking local spatiotemporal regions of radar images. In the fine-tuning stage, considering the importance of the temporal dimension in short-term precipitation forecasting and the complex long-range dependencies in spatiotemporal evolution of precipitation systems, we design an RNN-based cyclic temporal mask self-encoder model (MAE-RNN) and a transformer-based spatiotemporal attention model (STMT). The former focuses on capturing short-term temporal dynamics, while the latter simultaneously models long-range dependencies across space and time via a self-attention mechanism, thereby avoiding the smoothing effects on high-frequency details that are typical of conventional convolutional or recurrent structures. The experimental results show that STMT improves 3.73% and 2.39% in CSI and HSS key indexes compared with the existing advanced models, and generates radar echo sequences that are closer to the real data in terms of air mass morphology evolution and reflection intensity grading. Full article