[wxqc] High Frequency Updates verses Noise (or when the seagull, butterfly, and flea have their day)

[Sam Drinkard]

Hi Dave,

    Well, I don't think you are taking it too far.  I've read the theory 
on the butterfly, and logicaly, it makes sense, altho in real practice, 
it's hard to imagine something as mundane as the flap of a butterfly 
wings actually causing anything such as a tornado or hurricane!  Not 
sure I follow you on the connection of report data tho, but I guess the 
same thoughts would apply if the data were ingested into a model 
initialization.  I was not aware the other services were reporting on 5 
min intervals, and if a one minute interval is forthcoming, the only 
thing left would be real-time, or near real-time, sort of like 
wunderground is doing  with their "rapid-fire" updates.

    I've seen what some of the gravity wave propagation will do to a 
model run, especially if it turns out to be more than a little wave.  It 
can and does cause the model to go unstable, and crash or bomb out 
during that run.  There are tweaks in WRF to help that kind of behavior, 
but sometimes it still happens.

    I'll check some of those links.. I believe I've visited a couple of 
them.. The ensemble forecasting looks familiar.

Sam