You get real differences of scale which makes smaller models inconsistent with what happens in a train derailment. Lac-Mégantic rail disaster for example had some serious explosions from oil.

According to the Wikipedia article there was much destruction, but... that happened when the oil flowed where it shouldn't and then burned in places where one really doesn't want it to burn.

Lighter-than-air gas is crucially different. If you let that leak out it won't flow downhill through the city, it'll go straight up. Of course you can argue that gas is dangerous, but not by example of an accident where most of the flammable material flowed away (and caused harm elsewhere) instead of either burning or exploding at the site.

LNG isn’t a gas, it’s a cryogenic liquid like liquid nitrogen except flammable. If we are talking say 100+ DOT-111 tank car‘s that’s 3,450,000+ gallons which isn’t going to simply fly into the air instantly. The phase change from liquid to has takes significant energy at the scale which takes time.

Again it’s a question of scale, you don’t think of say molasses as fast moving liquid but increase the scale to 2.3 million gallons and you can be talking about a 25 foot tall 35 mph wave. https://en.wikipedia.org/wiki/Great_Molasses_Flood

AIUI it's not flammable. It is flammable when mixed with air in the right proportion, but if it's mixed with air it is a gas and rises straight up, rather than flow as a wave across the landscape.

You can't have your cake and eat it.

Again small scale vs large scale. If your talking a square foot of the stuff then it’s going to rise like your suggesting. Cover a few acres in LNG you get complex boundary layer interactions and giant swirls of natural gas mixed with air.

Read up on how you can get temperature inversion layers in the atmosphere. In your model that kind of effect shouldn’t happen, but actual fluid dynamics gets complicated.

Explain how you get spontaneous combustion in the spots where the mixture is right.

Or were you thinking about transporting LNG together with something that that would help start a fire?

Train crashes have plentiful ignition sources as metal scrapes across metal. Oddly enough it’s surprisingly common for train crashes to result in fires ex: https://youtu.be/vww5V-xmtzs or https://baltimore.cbslocal.com/2016/07/06/recent-derailments... or the original disaster I mentioned.

At which point those secondary fires can easily ignite the literal pools on LNG being described.

I've thought about this a few times now.

Would you describe this as a noticeable risk compared to driving around in a car, in urban traffic, with gasoline? Or are you comparing it rather to an alternative that more or less avoids transport of LNG?