Under the ongoing implementation of China's Domestic Emission Control Area (DECA) policy, shipping emissions and their impacts on air quality have drawn increasing attention. However, most studies have focused on the early and intermediate phases of DECA implementation, leaving the emission characteristics and multi-pollutant responses in the post-DECA era insufficiently quantified. This study investigates the Yangtze River Basin, the world's busiest inland shipping corridor, by constructing a high-resolution emission inventory for 18,478 vessels based on approximately 230 million Automatic Identification System (AIS) trajectories in 2023, and coupling it with the WRF-Chem model to quantify the contribution of ship emissions to regional air quality. Results show that SO2 and PM2.5 emissions from Yangtze shipping have been substantially controlled in the post-DECA era, whereas NOx emissions reached 156.41 kt yr-1, accounting for 82.0% of the total emissions of major pollutants, and formed pronounced hotspots along heavily trafficked reaches, including Yichang-Jingzhou, Wuhan, and Nanjing-Shanghai. Concentration simulations indicate a stable positive contribution of ship emissions to NO2 with marked near-source effects, yielding annual mean contributions of 0.16-3.41 μg m-3 across 11 mainstem cities. PM2.5 increments are generally low, but nitrate constitutes a substantial fraction. O3 responds differently, showing clear seasonal variation driven by photochemical processes: it decreases in January and October due to near-source NO titration but increases in July under strong photochemical conditions. These findings suggest that, beyond sustaining fuel-sulfur control and particulate mitigation, inland shipping emission management should advance toward coordinated strategies centered on source-level NOx reduction.
Social media imagery (SMI) provides timely and fine-grained ground perspectives that are valuable for situational awareness and emergency response. Unlike satellite or aerial imagery, SMI can capture disaster impacts and ground-level conditions in a timely manner. However, geographic references in SMI are often vague or ambiguous, making accurate geolocalization challenging. To address this issue, we propose Disaster TD, a disaster toponym disambiguation framework that integrates multimodal large language models (MLLMs)-based semantic reasoning with cross-view geolocalization. First, MLLMs extract toponyms and generate candidate geolocations from noisy textual inputs. Then, cross-view matching between SMI, remote sensing imagery (RSI), and optionally street-view imagery (SVI) is used to verify and refine these candidate results. A Vision Transformer (ViT)-based visual foundation model, DINOv2, is used to bridge the domain gap between overhead and ground-level imagery. We evaluate DisasterTD on the Hurricane Harvey dataset, where SMI is augmented with collected RSI and SVI to construct a cross-view benchmark for disaster geolocalization. The dataset is divided into four categories based on toponym clarity and ambiguity, allowing a fine-grained performance analysis across scenarios. Results show that DisasterTD consistently outperforms MLLM-only and cross-view-only baselines without disambiguation, achieving geolocalization accuracies of 71.62% within 1000 m, 62.36% within 500 m, 57.99% within 250 m, 52.09% within 100 m, and 47.01% within 50 m, while reducing the mean and median errors to 11.33 and 0.68 km, respectively. The largest improvements appear in ambiguous toponyms, where semantic reasoning with cross-view evidence reduces candidate dispersion and errors. These findings demonstrate the effectiveness of integrating MLLM-based candidate generation with cross-view verification for fine-grained disaster geolocalization.
Wenping Yin, Ziqi Liu, Naixia Mou et al.· IEEE Transactions on Geoscie...· 0 citations
