Weather Forecasts

[Glow worm]

All weather forecasters seem to have enormous difficulty accurately predicting the timing, intensity and track of weather systems. And they don't seem to be getting any better at all. Take yesterday for example, the Met Office predicted torrential downpours from 5am, up to 3 inches of rain and thunder here. We had been warned of this severe weather for 5 days and even had one of them yellow warnings (whatever that is!).

I planned accordingly, got my knackered old panniers out that I don't mind getting wet, and even cycled to work in lycra (shudder ) as I figured this would dry quicker at work. What happened? A bit of drizzle around midday and a bit more at 7pm and two totally dry rides there and back!

That's just one example- it seems to happen all the time- rain is also predicted tonight- it'll never happen - I'd put £ on it. It's best just to look out of teh window and check the weather radar yerself and make up your own mind.

 

[Matthames]

Amazes me how they can be so different - presumably they are interpreting the same data? I use XC - tends to be optimistic for weather- but Wind Observation Map good for wind direction and strength, BBC - often pretty accurate, Met office - useful rain radar, Rain Today - very good, but limited forecast time. Just checked TWO - new to me - to be told there should be 85% cloud cover here - actually there isn't a cloud in the clear blue sky!

Quite a few of the providers have their own in house models that they use, so there would be some variability caused by the differences in the models. What is also interesting is that you can eventually get two different results running exactly the same model using exactly the same initial conditions. This is because most models are made of a series of non-linear differential equations, which due to the nature of chaos will produce different results every time.

The best forecasts are produced using deterministic forecasting, this is where they run the same model with the same data several times to produce an average. If you use several different models you can get a pretty good picture of what is going on. You can get charts with deterministic forecast runs for a particular area, looking at them will give you a good idea of how reliable a forecast is going to be. You can see if it is going to be reliable or not by the spread of the different runs, if they are far apart then the forecast is not going to be reliable, if they are close together then you will find the forecasts will be pretty much spot on.

I tend to produce my own forecasts from charts that are freely available on the internet. Usually from Westwind and the expert charts at weatheronline. Weatheronline do their own set of forecasts, although I haven't really taken much notice of them and can't really comment on their accuracy.

 

[dellzeqq]

they are all completely s***e. (He says with some feeling). Even the Met Office.

 

[Matthames]

All weather forecasters seem to have enormous difficulty accurately predicting the timing, intensity and track of weather systems. And they don't seem to be getting any better at all. Take yesterday for example, the Met Office predicted torrential downpours from 5am, up to 3 inches of rain and thunder here. We had been warned of this severe weather for 5 days and even had one of them yellow warnings (whatever that is!).

I planned accordingly, got my knackered old panniers out that I don't mind getting wet, and even cycled to work in lycra (shudder ) as I figured this would dry quicker at work. What happened? A bit of drizzle around midday and a bit more at 7pm and two totally dry rides there and back!

That's just one example- it seems to happen all the time- rain is also predicted tonight- it'll never happen - I'd put £ on it. It's best just to look out of teh window and check the weather radar yerself and make up your own mind.

To get an idea of how difficult it is to accurately predict the weather, try doing a 1000 piece jigsaw with only 500 pieces. Because that is exactly what the weather services have to do.

I think that quite a few weather reports are all delivered wrong. Weather forecasts should really be delivered stating the probability of a certain event happening, for instance there is a 1/5 chance of rain.

A game that is usually played by meteorologists is that they make a forecast for a certain area for a certain time period, which is then compared to what actually happens. The one that is closest wins. Playing it gives you a great appreciation of how much of an art forecasting is.

 

[Glow worm]

To get an idea of how difficult it is to accurately predict the weather, try doing a 1000 piece jigsaw with only 500 pieces. Because that is exactly what the weather services have to do.

I think that quite a few weather reports are all delivered wrong. Weather forecasts should really be delivered stating the probability of a certain event happening, for instance there is a 1/5 chance of rain.

A game that is usually played by meteorologists is that they make a forecast for a certain area for a certain time period, which is then compared to what actually happens. The one that is closest wins. Playing it gives you a great appreciation of how much of an art forecasting is.

I think what surprised me this week was the complete certainty with which the storms appeared to be predicted here, right up until about 6 hours after they had failed to appear. So in that sense, I agree it is a presentational issue in terms of how certain they had appeared to be. It really was a spectacular balls up !

Suggesting a 1/5 probability is a good idea, but you'll always get people saying 'do I need a brolley or not?'! I do sympathise with forecasters as grappling with the future of unstable, rapid moving air masses is, as you say the art of judging probability. Perhaps then, broadcast forecasters should just be more honest and say our confidence of these events is low. I have heard this, but only infrequently.

Maybe us cyclists notice these things more than Joe Bloggs and are more critical as we often need to plan in terms of kit etc, for weather.

 

[briantrumpet]

What's annoying is that the Met Office actually calculate the percentage probability of each of their forecasts being right - there's obviously going to be increasingly lower probability the further ahead they forecast, but also some systems are inherehnty difficult to predict. They know all this, calculate the probability ... but don't publish it anywhere. This is because they polled the public, and the public said that they didn't want to see the probabilities, they just wanted certainty.

So as usual, the public are hypocrites - we say we want higher animal welfare standards on farms, but go to the supermarket and buy the cheapest chickens. We say we don't approve of Rupert Murdoch, but we subscribe to Sky and bought the News of the World in large numbers. We don't want to see when a forecast has a probability of only 30%, but complain when the forecast isn't 'accurate'.

Anyway, rant over, I use a mixture of the Met Office, Rain Today and XC Weather (and the Mountain Weather Information Service if I'm somewhere that convers, sach as the Lake District) - using the those in combination, you can can normally make some pretty good guesses.

 

[Glover Fan]

I am finding increasingly that the BBC's long term forecast are often wildly inaccurate. I find www.weather.co.uk to be the best.

If I am looking for a couple of hours before a ride www.raintoday.co.uk and looking out the window is a good option.

 

[skudupnorth]

I stick my head out of the door and dress acordingly or take spare stuff if it's touch and go !

 

[mangaman]

All weather forecasters seem to have enormous difficulty accurately predicting the timing, intensity and track of weather systems. And they don't seem to be getting any better at all. Take yesterday for example, the Met Office predicted torrential downpours from 5am, up to 3 inches of rain and thunder here. We had been warned of this severe weather for 5 days and even had one of them yellow warnings (whatever that is!).

I planned accordingly, got my knackered old panniers out that I don't mind getting wet, and even cycled to work in lycra (shudder ) as I figured this would dry quicker at work. What happened? A bit of drizzle around midday and a bit more at 7pm and two totally dry rides there and back!

That's just one example- it seems to happen all the time- rain is also predicted tonight- it'll never happen - I'd put £ on it. It's best just to look out of teh window and check the weather radar yerself and make up your own mind.

I agree GW

I think the BBC always predict the worst case scenario - they're still tarnished by rubbishing the 1987 "hurricane" and if you read their predictions for 4 or 5 days away it's hail / lightning / snow etc. As you get closer, they gradually reduce the severity of the forecast.

Come the day in question it's still and sunny.

I tend to look at the BBC / accuweather / met office and weather uk.

Come the day I rely on rainfall radar. It's the only way.

It's amazing how you can watch a huge splurge of rain heading your way by looking at half hourly rainfall radar images from the met-office, yet they predict no rain in their forecasts.

(I won't mention MK Dons BTW Glow worm)

Private grief GW and I are sharing and not relevant to the thread BTW