February 01, 2005
Atmospheric turbulence model disruption.
Lots of things can mess up our atmospheric turbulence model and remove the simplicity used in advection (horizontal transport) forecasting, outlined in the hyperlink above.If that is an oversimplification, what is the other extreme?
Chaos...
...the thing that still makes weather unpredictable.Also known in meteorology as the "sensitive dependence on initial conditions", or more commonly, the "Butterfly Effect", the principle works like this: A small but finite change in a set of circumstances (initial conditions) can grow to a great degree over time. So much so that it can result in a world completely different.
Think of the old man who delivers the magazine to his younger self 60 years earlier in the second Back to the Future movie, and how radically this tiny event changed the world. Well in the physical world, this type of thing supposedly can happen, given enough time.
(I had one of these in my life recently, read about it.)
The Butterfly Effect title comes from the hypothesis of an early theorist.
Lorentz Who?
A physicist known as Edward Lorentz. This concept came into being:Predictability: Does the Flap of a Butterfly's Wings in Brazil set off a Tornado in Texas?
The title of one of his lectures.
If this is possible, how could you predict anything in a complicated system such as weather. Well, the important thing is that it takes some time for the ripple in the atmospheric turbulence model to have its effects. That's why you can still have a decent forecast for tomorrow or maybe the next couple days.
It is one of our limitations we encounter in modelling and numerical weather prediction. How can we ever achieve better longer term forecasts? The answer – resolution!
As we develop a more dense network of observations plus more data from all levels of the atmosphere, with the help of satellites, we can finally have an increasingly solid set of initial conditions for our models. As the speed and power of the super computers conducting the modeling continues to increase, we can handle models with ever increasing resolution, both spatial and temporal.
Maybe we can have effective long range forecasts in our lifetimes, but there exist some doubts.
Here's another good'er from Lorentz:
To the oft-heard question, 'Why can't we make better weather forecasts?' I have been tempted to reply, 'well, why should we be able to make any forecasts at all?'
In light of this, it has been postulated that daily forecasts for more that about two weeks will never be possible. So there, Accuweather!
Longer ranges are currently handled by using statistics to compare and average the results of multiple forecasting methods. Forecasters create Ensemble Forecasts or Spaghetti Plots as Canadian meteorologists like to call them. This is what we use to arrive at things like probability of precipitation , or above normal/below normal seasonal forecasts.
Anyway, the future is exciting.
Hope yours is too.