Owning to the nature of flood events, near-real-time flood detection and mapping is essential for disaster prevention, relief, and mitigation. In recent years, the rapid advancement of deep learning has brought endless possibilities to the field of flood detection. However, deep learning relies heavily on training samples and the availability of high-quality flood datasets is rather limited. The present study collected 16 flood events in the Yangtze River Basin and divided them into three categories for different purpose: training, testing, and application. An efficient methodology of dataset-generation for training, testing, and application was proposed. Eight flood events were used to generate strong label datasets with 5296 tiles as flood training samples along with two testing datasets. The performances of several classic convolutional neural network models were evaluated with those obtained datasets, and the results suggested that the efficiencies and accuracies of convolutional neural network models were obviously higher than that of the threshold method. The effects of VH polarization, VV polarization, and the involvement of auxiliary DEM on flood detection were investigated, which indicated that VH polarization was more conducive to flood detection, while the involvement of DEM has a limited effect on flood detection in the Yangtze River Basin. Convolutional neural network trained by strong datasets were used in near-real-time flood detection and mapping for the remaining eight flood events, and weak label datasets were generated to expand the flood training samples to evaluate the possible effects on deep learning models in terms of flood detection and mapping. The experiments obtained conclusions consistent with those previously made on experiments with strong datasets.

Accurate landslide susceptibility mapping is vital for disaster forecasting and risk management. To address the problem of limited accuracy of individual classifiers and lack of model interpretability in machine learning-based models, a coupled multi-model framework for landslide susceptibility mapping is proposed. Using Jiuzhaigou County, Sichuan Province, as a case study, we developed an evaluation index system incorporating 14 factors. We employed three base models—logistic regression, support vector machine, and Gaussian Naive Bayes—assessed through four ensemble methods: Stacking, Voting, Bagging, and Boosting. The decision mechanisms of these models were explained via a SHAP (SHapley Additive exPlanations) analysis. Results demonstrate that integrating machine learning with ensemble learning and SHAP yields more reliable landslide susceptibility mapping and enhances model interpretability. This approach effectively addresses the challenges of unreliable landslide susceptibility mapping in complex environments.

Complex terrain, the sparse distribution of rain gauges, and the poor resolution and quality of satellite data in remote areas severely restrict the development of watershed hydrological modeling, meteorology, and ecological research. In this study, based on the relationship between cloud optical and physical properties and precipitation, a daily geographically weighted regression (GWR) precipitation downscaling model was constructed for the Three Rivers Source region, China, for the period from 2010 to 2014. The GWR precipitation downscaling model combined three different satellite precipitation datasets (CMORPH, IMERG, and ERA5) which were downscaled from a coarse resolution (0.25° and 0.1°) to a fine resolution (1 km). At the same time, the preliminary downscaling results were calibrated and verified by employing the geographic difference analysis (GDA) and geographic ratio analysis (GRA) methods combined with rainfall data. Finally, the analytical hierarchy process (AHP) and the entropy weight method (EW) were adopted to fuse the three downscaled and calibrated satellite precipitation datasets into the merged satellite precipitation dataset (MSP), which provides a higher quality of data (CC = 0.790, RMSE = 2.189 mm/day, and BIAS = 0.142 mm). In summary, the downscaling calibration and precipitation fusion scheme proposed in this study is suitable for obtaining high-resolution daily precipitation data in the Three Rivers Source region with a complex climate and topography.

Synthetic Aperture Radar (SAR) technology offers unparalleled advantages by delivering high-quality images under all-weather conditions, enabling effective flood monitoring. This capability provides massive remote sensing data for flood mapping, while recent rapid advances in deep learning (DL) offer methodologies for large-scale flood mapping. However, the full potential of deep learning in large-scale flood monitoring utilizing remote sensing data remains largely untapped, necessitating further exploration of both data and methodologies. This paper presents an innovative approach that harnesses convolutional neural networks (CNNs) with Sentinel-1 SAR images for large-scale inundation detection and dynamic flood monitoring in the Yangtze River Basin (YRB). An efficient CNN model entitled FloodsNet was constructed based on multi-scale feature extraction and reuse. The study compiled 16 flood events comprising 32 Sentinel-1 images for CNN training, validation, inundation detection, and flood mapping. A semi-automatic inundation detection approach was developed to generate representative flood samples with labels, resulting in a total of 5296 labeled flood samples. The proposed model FloodsNet achieves 1–2% higher F1-score than the other five DL models on this dataset. Experimental inundation detection in the YRB from 2016 to 2021 and dynamic flood monitoring in the Dongting and Poyang Lakes corroborated the scheme’s outstanding performance through various validation procedures. This study marks the first application of deep learning with SAR images for large-scale flood monitoring in the YRB, providing a valuable reference for future research in flood disaster studies. This study explores the potential of SAR imagery and deep learning in large-scale flood monitoring across the Yangtze River Basin, providing a valuable reference for future research in flood disaster studies.

Cyanobacterial harmful algal blooms (CyanoHABs) pose serious threats to inland water ecology and environmental security. Despite advances, accurately capturing their long-term dynamics for effective governance remains challenging due to limited datasets and segmentation model accuracy. This study addresses these gaps by developing a CyanoHABs dataset and a new model, Multi-scale Spatial Attention Network (MSA-Net), to automatically extract cyanobacterial blooms from Landsat imagery using deep learning. With enhanced multi-scale feature extraction and a Hybrid Attention Mechanism in an encoder–decoder framework, MSA-Net outperforms other models, providing robust methods and data support for remote sensing-based CyanoHAB monitoring in complex environments. The proposed MSA-Net achieved an F1-score of 88.34% and a Precision of 88.74%, outperforming all baseline methods. Ablation experiments further verified the effectiveness of the proposed Spatial Feature Enhancement Mechanism (SFEM) and Hybrid Attention Mechanism (HAM) in enhancing segmentation performance. The results highlight MSA-Net’s effectiveness in accurately identifying CyanoHABs, particularly in complex, multi-scale environments. This study provides a transferable deep learning framework and long-term remote sensing dataset for CyanoHAB monitoring, offering valuable insights for the large-scale assessment and management of eutrophic inland waters worldwide.