←back to thread

144 points scubakid | 5 comments | | HN request time: 0.662s | source
Show context
blargthorwars ◴[] No.44417463[source]
And it does so on seas that aren't even remotely close to China. They're looting the entire Pacific Ocean.
replies(4): >>44417496 #>>44417661 #>>44417681 #>>44417948 #
1. darth_avocado ◴[] No.44417948[source]
The article only talks about visible fishing activity. But China operates many “dark fleets” where many unregistered boats sail along registered boats. They are fishing way more than the 44% that is being reported. These fleets will no doubt destroy ecosystems beyond repair.

https://news.mongabay.com/2020/10/new-evidence-suggests-chin...

https://www.abc.net.au/news/2020-12-19/how-china-is-plunderi...

https://time.com/6328528/investigation-chinese-fishing-fleet...

replies(2): >>44418102 #>>44428323 #
2. mensetmanusman ◴[] No.44418102[source]
Thankfully with satellites and machine learning nothing is dark these days.
replies(2): >>44418689 #>>44428441 #
3. darth_avocado ◴[] No.44418689[source]
That is not true at all. A very small percentage of our oceans are comprehensively monitored. There is a reason why MH370 could not be located via satellite imaging. Tracking fleets of boats all over the planet is going to be very tricky.

https://www.weforum.org/stories/2017/08/how-data-can-heal-ou...

4. mapt ◴[] No.44428323[source]
The US has around 300 vessels for distant-water fishing.

Five years ago, China was assessed to have 17,000.

5. mapt ◴[] No.44428441[source]
It's still very patchy, and likely will be for another decade of growth. Earth has 500 million km^2 of surface area. A 24-bit, 10 meter resolution recording of Earth's surface (a shoebox sized satellite in LEO can do this) at 1 image per minute, plus a 1 meter resolution recording at 1 image per hour, would make tracking relatively easy (modulo weather), and necessitate 36 uncompressed petabytes per day of data handling, probably more like 360 uncompressed petabytes when you factor in overlaps of satellite planes and imaging areas.

We can do this for, say, all the airfields in Russia, today. But the world is a much bigger place.

As laser communications equipment becomes more standardized and the LEO comms and LEO remote sensing constellations continue to grow, this will slowly expand in scope. But tracking and deconfliction of moving objects is just an inherently difficult thing to do with any confidence; It wouldn't be feasible at all without various AI algorithms. There are far lower hanging fruit as yet unpicked.