DOI: https://doi.org/10.1038/s41586-023-06825-8
PMID: https://pubmed.ncbi.nlm.nih.gov/38172362
تاريخ النشر: 2024-01-03
المؤلف: Fernando Paolo وآخرون
الموضوع الرئيسي: دراسات الشعاب المرجانية والأنظمة البيئية البحرية
نظرة عامة
يقدم هذا القسم من ورقة البحث تحليلًا شاملاً للأنشطة الصناعية في محيطات العالم، باستخدام صور الأقمار الصناعية، وبيانات GPS للسفن، ونماذج التعلم العميق لرسم خريطة لهذه الأنشطة من 2017 إلى 2021. تشير النتائج إلى أن جزءًا كبيرًا من سفن الصيد الصناعية – 72-76% – لا يتم تتبعه علنًا، خاصة في مناطق مثل جنوب آسيا وجنوب شرق آسيا وأفريقيا. بالإضافة إلى ذلك، فإن 21-30% من أنشطة سفن النقل والطاقة مفقودة أيضًا من أنظمة التتبع العامة. تسلط الدراسة الضوء على انخفاض بنسبة 12 ± 1% في النشاط العالمي للصيد بسبب جائحة COVID-19، التي لم تتعافى إلى مستويات ما قبل الجائحة بحلول عام 2021، بينما ظلت قطاعات النقل والطاقة مستقرة نسبيًا.
تؤكد الورقة على الأهمية المتزايدة للطاقة المتجددة البحرية، حيث تجاوزت تركيبات الرياح البحرية الهياكل النفطية بحلول عام 2021. وتبرز التحديات في مراقبة التصنيع البحري بسبب نقص البيانات العالمية الرصدية وقيود أنظمة التتبع الحالية، مثل نظام التعريف التلقائي (AIS). يقترح المؤلفون أن مجموعة البيانات الخاصة بهم وتقنية رسم الخرائط يمكن أن تحدد النقاط الساخنة للأنشطة غير القانونية وتوفر رؤى حول تأثير الصيد الصناعي على مصائد الأسماك الحرفية والمناطق الاقتصادية الخالصة (EEZs). بشكل عام، تكشف الدراسة عن تحول في التصنيع البحري، مع انخفاض نشاط الصيد واستعداد قطاعات النقل والطاقة للنمو، مما يتطلب تحسين المراقبة لإدارة التأثيرات البيئية لهذه التغييرات.
نقاش
تسلط قسم النقاش من ورقة البحث الضوء على نتائج مهمة تتعلق بتصنيف ورسم خرائط السفن والبنية التحتية البحرية باستخدام رادار الفتحة الاصطناعية ثنائية الاستقطاب (SAR) والصور البصرية. حققت الدراسة دقة تزيد عن 98% في تصنيف البنية التحتية البحرية إلى فئات مثل النفط والرياح، وأكثر من 90% دقة في التمييز بين سفن الصيد وغير الصيد. تم تحليل ما مجموعه 67 مليون قطعة صورة، مما يكشف أن جزءًا كبيرًا (حوالي 63,300 سفينة) تم اكتشافه في أي لحظة معينة، حيث تشكل سفن الصيد 42-49% من هذه الاكتشافات. من الجدير بالذكر أن حوالي 72-76% من نشاط الصيد الصناعي لم يتم التقاطه بواسطة أنظمة المراقبة العامة، مما يشير إلى وجود تقارير ناقصة كبيرة عن النشاط العالمي للصيد، خاصة في مناطق مثل آسيا، التي تمثل 70% من جميع اكتشافات سفن الصيد.
تناقش الورقة أيضًا آثار تطوير البنية التحتية البحرية على أنماط حركة السفن. وُجد أن قوارب الجر تميل إلى تجنب الصيد ضمن 1 كم من الهياكل النفطية لتجنب تشابك الشباك، بينما قد تنجذب طرق الصيد الأخرى إلى هذه الهياكل بسبب تجمع الأسماك. تشير التحليلات إلى أن عدد الهياكل النفطية البحرية قد زاد بنسبة 16% خلال السنوات الخمس الماضية، بينما زادت تركيبات توربينات الرياح البحرية بشكل كبير، خاصة في الصين، التي شهدت زيادة بنسبة 900% من 2017 إلى 2021. لا توفر هذه الطريقة الشاملة لرسم الخرائط رؤى حول أنشطة الصيد وتوزيع البنية التحتية فحسب، بل تبرز أيضًا النقاط الساخنة المحتملة للصيد غير القانوني، مما يساهم في فهم أكثر دقة للأنشطة البحرية وتأثيراتها البيئية.
القيود
تنشأ قيود هذه الدراسة بشكل أساسي من قيود قدرات قمر صناعي Sentinel-1 والتحديات الكامنة في اكتشاف السفن الصغيرة. من الجدير بالذكر أن Sentinel-1 لا يقوم بأخذ عينات فعالة من المحيط المفتوح، لكن التحليل يشير إلى أن معظم الأنشطة الصناعية تحدث بالقرب من السواحل، حيث تكون بيانات نظام التعريف التلقائي (AIS) أكثر شيوعًا. تستبعد الدراسة التصنيفات ضمن 1 كم من الساحل بسبب الميزات الساحلية الغامضة ولا تغطي أجزاء كبيرة من المناطق القطبية الشمالية والجنوبية، حيث تعقد الجليد البحري عملية الكشف. على الرغم من هذه القيود، تحدث الغالبية العظمى من الأنشطة الصناعية، مثل الصيد والشحن، على بعد عدة كيلومترات من الشاطئ، حيث يكون الكشف أكثر قابلية للتطبيق.
تحد دقة صور Sentinel-1 (حوالي 20 م) من اكتشاف السفن التي تقل عن 15 م، حيث تتأثر معدلات الكشف للسفن الصغيرة (<25 م) بعوامل بيئية مثل سرعة الرياح وحالة البحر. بالإضافة إلى ذلك، تتأثر دقة تقديرات طول السفن بجودة بيانات AIS، التي قد تحتوي على أخطاء. قد تنشأ مشاكل التصنيف الخاطئ في المناطق ذات الحركة العالية من السفن غير الصيد، خاصة في المناطق الغنية أو حيث لا يتم تتبع نشاط الصيد علنًا. وبالتالي، من المحتمل أن الدراسة تقلل من تقدير عدد سفن الصيد الصناعية، خاصة في المياه الآسيوية، حيث تعمل العديد من السفن الصغيرة بدون AIS. يمكن أن تعزز الأبحاث المستقبلية قدرات الكشف وتعالج هذه التحديات، خاصة مع الإضافة المتوقعة لمزيد من أقمار SAR الصناعية.
DOI: https://doi.org/10.1038/s41586-023-06825-8
PMID: https://pubmed.ncbi.nlm.nih.gov/38172362
Publication Date: 2024-01-03
Author(s): Fernando Paolo et al.
Primary Topic: Coral and Marine Ecosystems Studies
Overview
This research paper section presents a comprehensive analysis of industrial activities in the world’s oceans, utilizing satellite imagery, vessel GPS data, and deep-learning models to map these activities from 2017 to 2021. The findings indicate that a significant portion of industrial fishing vessels—72-76%—are not publicly tracked, particularly in regions such as South Asia, Southeast Asia, and Africa. Additionally, 21-30% of transport and energy vessel activities are also missing from public tracking systems. The study highlights a 12 ± 1% decrease in global fishing activity due to the COVID-19 pandemic, which had not recovered to pre-pandemic levels by 2021, while transport and energy sectors remained relatively stable.
The paper emphasizes the growing importance of offshore renewable energy, with offshore wind installations surpassing oil structures by 2021. It underscores the challenges in monitoring oceanic industrialization due to inadequate global observational data and the limitations of existing tracking systems, such as the Automatic Identification System (AIS). The authors propose that their dataset and mapping technology can identify hotspots of illegal activities and provide insights into the impact of industrial fishing on artisanal fisheries and exclusive economic zones (EEZs). Overall, the study reveals a shift in ocean industrialization, with fishing activity declining and transport and energy sectors poised for growth, necessitating improved monitoring to manage the ecological impacts of these changes.
Discussion
The discussion section of the research paper highlights significant findings regarding the classification and mapping of offshore vessels and infrastructure using dual-polarization synthetic-aperture radar (SAR) and optical imagery. The study achieved over 98% accuracy in classifying offshore infrastructure into categories such as oil and wind, and more than 90% accuracy in distinguishing between fishing and non-fishing vessels. A total of 67 million image tiles were analyzed, revealing that a substantial portion (approximately 63,300 vessels) were detected at any given moment, with fishing vessels constituting 42-49% of these detections. Notably, around 72-76% of industrial fishing activity was not captured by public monitoring systems, indicating a significant underreporting of global fishing activity, particularly in regions like Asia, which accounted for 70% of all fishing vessel detections.
The paper also discusses the implications of offshore infrastructure development on vessel traffic patterns. It was found that trawlers tend to avoid fishing within 1 km of oil structures to prevent net entanglement, while other fishing methods may be attracted to these structures due to fish aggregation. The analysis indicates that the number of offshore oil structures has increased by 16% over the past five years, while offshore wind turbine installations have surged, particularly in China, which saw a 900% increase from 2017 to 2021. This comprehensive mapping approach not only provides insights into fishing activities and infrastructure distribution but also highlights potential hotspots for illegal fishing, thereby contributing to a more accurate understanding of maritime activities and their environmental impacts.
Limitations
The limitations of this study primarily stem from the constraints of the Sentinel-1 satellite’s capabilities and the inherent challenges in detecting smaller vessels. Notably, Sentinel-1 does not effectively sample the open ocean, but the analysis indicates that most industrial activities occur near shorelines, where Automatic Identification System (AIS) data is more prevalent. The study excludes classifications within 1 km of the coast due to ambiguous coastal features and does not cover significant portions of the Arctic and Antarctic regions, where sea ice complicates detection. Despite these limitations, the majority of industrial activities, such as fishing and shipping, occur several kilometers offshore, where detection is more feasible.
The resolution of Sentinel-1 imagery (approximately 20 m) limits the detection of vessels smaller than 15 m, with detection rates for smaller vessels (<25 m) being influenced by environmental factors like wind speed and sea state. Additionally, the accuracy of vessel length estimates is affected by the quality of AIS data, which can contain errors. Misclassification issues may arise in regions with high traffic from non-fishing vessels, particularly in affluent areas or where fishing activity is not publicly tracked. Consequently, the study likely underestimates the number of industrial fishing vessels, especially in Asian waters, where many smaller vessels operate without AIS. Future research could enhance detection capabilities and address these challenges, particularly with the anticipated addition of more SAR satellites.
