On Tue, May 22, 2018 at 5:18 AM, Rob O'Connor <xxxxxx@ozemail.com.au> wrote:

Richard Aiken wrote:
> Yes. But given enough data points and computing power,
> weather should also be fairly predictable.

Nope. You end up with a probability distribution for whatever parameters you can measure.
You can't say "next Thursday Bitburg will have a temperature min/max of x to y with/without rain".

But we don't need to predict next Thursday. We "just" need to know the current conditions in real time and whether or not these are likely to result in the formation of unwanted weather. If so, we use our "tornado cannon" to tweak conditions toward neutral.

And yes, we could get it wrong. But - as was previously pointed out - there is massive inertia in the full system. So we'll have plenty of time to see our "corrections" sliding sideways and thus apply further tweaks as needed.

Richard Aiken

"Never insult anyone by accident." Robert A. Heinlein

"I studied the Koran a great deal. I came away from that study with the conviction there have been few religions in the world as deadly to men as Muhammed." Alexis de Tocqueville

"We know a little about a lot of things; just enough to make us dangerous." Dean Winchester (fictional monster hunter portrayed by Jensen Ackles)

"It has been my experience that a gun doesn't care who pulls its trigger." Newton Knight (as portrayed by Matthew McConaughey), to a scoffing Confederate tax collector facing the weapons held by Knight's young children and wife.