how about a pot sitting on a gas stove (whose flames can also be "over a thousand degrees F")?
I'm not an engineer, but this does not appear to me to accurately model a house fire. I think that there is going to be a difference between being engulfed in 1000+ degree heat vs being over a 1000+ degree heat source.
By way of example, let's say that you have a 22 quart canning pot filled with water and you were to suspend it over a Bunsen burner. That burner can reach a temperature of 2000+ degrees F at the tip of its inner cone, but how long do you think it will take that 2000+ degree burner to boil 10 quarts of water? Perhaps it will never boil?
I think that a house fire would transfer significantly more heat to the bucket of water than a gas stove would. I frankly have no idea how long it would take for water in a bucket to boil off in a house fire, but I am confident that it would be faster than sticking a pot of boiling water on a gas range.
Another issue with the "disk in a bucket" plan is that in dry climates, much care must be taken to maintain the water level of the bucket, because significant water loss would be expected via evaporation.
As always, the best way to keep data safe during a fire is for the data not to be in the fire.