Mapping of oil spills in China Seas using optical satellite data and deep learning,Journal of Hazardous Materials

当前位置： X-MOL 学术 › J. Hazard. Mater. › 论文详情

Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)

Mapping of oil spills in China Seas using optical satellite data and deep learning
Journal of Hazardous Materials ( IF 12.2 ) Pub Date : 2024-09-10 , DOI: 10.1016/j.jhazmat.2024.135809
Lifeng Wang , Yingcheng Lu , Mingxiu Wang , Wei Zhao , Hang Lv , Shuxian Song , Yuntao Wang , Yanlong Chen , Wenfeng Zhan , Weimin Ju

Oils spilled into the ocean can form various weathered oils (non-emulsified oil slicks (NEOS), oil emulsions (OE)) which threaten the oceanic and coastal environments and ecosystems. Optical remote sensing has the unique ability to discriminate oil types and quantify oil volumes as their spectral contrasts with oil-free seawater. Here, a deep learning-based model is developed for identification, classification, and quantification of various oil types. Based on the oil-contained datasets collected from 7 satellite sensors from April 2019 to August 2023, the origin, quantity, and spatial distribution of oils spilled from ships and rigs in the China Seas are mapped in detail. We found that oil spill incidents are primarily from ship discharges (85.8 %), while platform leaks lead to more oil emulsions (58.6 % compared to 13.1 % from ships), which illuminates that the drilling oils are the main source of oil spill pollution in China Seas. The spilled oils correlate with major port locations, including offshore Qingdao and Rongcheng, Bohai Bay, the adjacent areas of Beihai, and Hue and Danang in Vietnam. This study provides new insights into the assessment and management of offshore and marine oil spills.

更新日期：2024-09-10

点击分享查看原文

点击收藏

阅读更多本刊新发论文本刊介绍/投稿指南