Forecasting 'rules'

I meant to ask this a week or two ago when it happened, but I forgot at the time. This means that I can’t describe the exact circumstances, but the question I’m going to ask is still valid.

One day my VP2 console was ‘forecasting’ snow. I was rather surprised at the time, because (from memory) it was a reasonably nice day with the temperature up around 9-10degC. I know the ‘forecast’ covers the next 12-24 hours, that current conditions aren’t always an indication of what will happen in the near future, and the forecast is only a best guess. It didn’t snow as it happened and a few hours later the snow forecast disappeared. I can’t remember the other conditions at the time, so I can’t say whether there was some condition that might have triggered the snow forecast, but it set me wondering.

Based on the limited set of information available to the VP2 console, i.e. local pressure, temperature, humidity, rainfall and wind speed/direction, and the rates of change/direction of change for those parameters, what rules does the console use to create its forecast? I’ve done a bit of googling around and I can’t find anything that seems to given any simple ‘rules of thumb’ relating to such forecasts/predictions based on those parameters. Does anyone know of a (simple) set of rules that can be used to do this. They could useful if all you have is access to current/recent data without having the console nearby to apply the rules for you!

Note…I’m not suggesting that the VP console is the worlds best forecaster. It’s just my ever enquiring mind being intrigued at how it’s done!

Can’t help with the how but I’m amazed how close the VP forecast is. I haven’t looked a the forecast much in the past year but since I started using VirtualVP and VPLive I’ve been able to compare. Although not detailed it is usually close. I hope someone has the answer.

–Dave

you can get WD to scroll the VP forecast under the main screen icon dave…

Yes, I know that but I’ve always had the metar forcast there. Just my preference as the metar is more detailed. But VPLive can’t run the metar forecast so now have both. WD does it all and that’s why it’s my program of choice. But having VirtualVP alows me to use others so I take advantage of that. I like VPLive for the simplicity and I like that other one for… well there must be something.

–Dave

using WD’s custom screen you can make a display in WD as simple as you like
(just saying that as i often get people saying wd has too much info…well, then, customise it and create a custom screen, or customise the main screen, the weather dials screen, etc…just some info for people reading this thread and not aimed at you dave :wink:)

I like WD the way it is. Could make it simple but then I’d miss all the good stuff.

Didn’t take it as being aimed at me. Maybe I’m just dense. :slight_smile:

Anyway, back to Chris’ forecasting rules.

According to the limited info that Davis has published the latitude and longitude also play a part in the console forecast.

I suppose if it was using rules like the Zambretti forecaster (mentioned here some time go) then it would need to know latitude. The Zambretti was designed for northern hemisphere use and the rules needed to be turned upside down for the southern hemisphere.

Based on some of the update notes for WeatherLink, the console may be using the lat/long to map the station location to a “zone”. An example would be a coastal california zone. Outside the US they probably have more generic rules.

Steve

Zones would probably give an extra degree of tweaking the result. I’m still interested to see any rules that can be used to turn the available parameters into reasonable forecasts. I’m not trying to pinch the VP console rules…just trying to see what kind of rules there might be for predicting the weather over short periods.

I doubt you will do better than reverse engineering the Zambretti :wink:

[quote author=nikoshepherd