The YouTube algorithm works similar to how the Streisand effect works. First, people prefer blockbuster, Hollywood kind of scenes, of current extreme weather, hence incredible scenes will likely result in more users sharing and viewing a video. Secondly, how many views and how fast a video gets above certain view counts, determines how long and how often a video is recommended on YouTube.
A) A video with faked storm footage, which contains footage from movies, or footage from different storms, mixed with real footage, but posted under the title for a specific event: https://www.youtube.com/embed/rTQjpnUxp_I
B) Similar video covering the same event, but with the attempt to make it reliable: https://www.youtube.com/watch?v=Ki5GRBLFi8k Additional, this video also includes a segment on carbon emissions, how it effects storms. Thus, covers educational aspect too.
C) Example of tornado from 2015, making it by mistake into the Typhoon coverage of Mangkhut, original clip https://www.youtube.com/watch?v=wFABQ6jb9kw BBC in 2018 https://www.bbc.com/news/av/world-asia-45540282/typhoon-mangkhut-hong-kong-latest-to-face-storm What appears to be the original announcement https://twitter.com/daveystan1/status/1041289214615203840
Difference how these videos perform on YouTube
Video A “The fake”, was posted closer to the event happening, and is kind of trending with over 100,000 views. The more accurate video B, was posted about 24 hrs after the event begun, and got so far 1,300 views. Generally, due to the timing, the people following the event may already be satisfied with their interests of this event, interest exhaustion. However, even though video B was posted late in the event, it features distinct unique content with some of the related science, and has a good overall production quality, and length. Something not acknowledged by the YouTube algorithm, so it appears.
Summary The videos covering an event first, usually get most attention.
To get the best results on YouTube for extreme weather events, timing is of essence. Video view counts encourage users to post content as soon as possible, which naturally affects content reliability, content length, and overall production quality. Additional, it can be concluded that users would benefit from mixing in footage from different events, or even movies, because more dramatic footage will likely increase attention, again leading to more recommendations. There are many more examples, including examples of quality content trending, but those usually are more like breaking news announcements.
Content creators can prepare a video layout in advance, including music, breaks, intro and outro to improve on quality. Keeping it brief, possibly posting several videos, instead of just one, also helps with the timing, with video performance. In some instances it is possible to use a related movie scene, eg Antarctic iceberg break-up from the movie The Day After Tomorrow, but this may lead to copyright issues, and if added should always be pointed out.
.After an event has passed, it is likely more successful to post event compilations, or even compilations for different events, eg. The Top 10 Largest Hurricanes Ranked. False content attribution, eg. adding a movie clip without attribution, or adding content from a different event, may become a problem for YouTube channels, since this is generally considered Fake News.