[quote author=3D Resident]
Actually, models don't use retrospective analysis to make forecasts. They take recent observations (surface and upper air) and insert them into complex prognostic mathematical equations (such as conservation of mass, momentum, vorticity, and of course thermodynamical equations). These equations are based on the physics of the atmosphere and enable a prediction of future atmospheric quantities. Thus, even if the weather is "chaotic" at the present time, there is no reason that the models cannot handle the observations and physics with an adequate level of skill. Thunderstorms and hurricanes are an example of very unstable weather, and models can handle these situations with reasonable accuracy most of the time, at least in my experience.
But modelling is but one tool in weather forecasting. There are many other aspects to forecasting the weather also. Forecasters use their own local knowledge, experience of past events and synoptic reasoning to adjust any output from computational models. Good forecasters always use these tools in addition to models, and they *always* ground truth models with what is currently happening to make sure the models can be trusted with any reasonable level of confidence. And there are some programs which climatologists and meteorologists have set up which look at past synoptic situations, and based on what happened during such times, make predictions as to what is likely to happen when such situations arise again. However this is not a modelling approach to forecasting, but rather a climatological approach.
As for inaccurate forecasts, well meteorology being an inexact science, it's bound to happen from time to time. But the way the official forecast is conveyed by various agencies such as the media, can play a large role in how accurate the forecast is perceived by the end user. As a forecaster myself, I know it can be very hard to describe what the weather will do the next day for a large region. Local effects can alter the "mean forecast" for that area and if you have only a couple of lines to describe the weather for the whole region, then it is common for some areas to get less-than-ideal forecasts.
[/quote]
Yes nicely said 3D; no professional offense taken I hope for the cheap shot on "darts" (if you had a dollar for every time someone said that...), anyway, these seem complex things subject to many localized influences and a good forecaster can project relatively well. Complex terrain and sudden high low mixing can add much variability that must be hard to accurately account for sometimes. Living in the western mountains, northerly and primarily westerly, sometimes southerly but rarely easterly flows with radiant heating and rapid cooling can produce some interesting fast changing events not previously prescribed.
Used to get spot forecasts for fire weather indicies and prediction and the forecaster did a darn good job considering the pressure he was under to get things right without much time to make the call - lots of good interpretative skill.
Would say that the biggest local changes atmospherically in the last 3 years here have been the intensity of electrical storms, also, cool springs and early deep freezes (-25c in November); had a frost in late August last year that hit the gardens and seemed to come out of nowhere - definitely not the norm.
Thanks for your professional account :)
