cancel
Showing results for 
Search instead for 
Did you mean: 

Met Office Rainfall Forecast for April/May/June 2012

Community Veteran
Posts: 38,251
Thanks: 937
Fixes: 56
Registered: 15-06-2007

Met Office Rainfall Forecast for April/May/June 2012

http://www.metoffice.gov.uk/media/pdf/p/i/A3-layout-precip-AMJ.pdf
Quote
SUMMARY - PRECIPITATION:
The forecast for average UK rainfall slightly favours drier-than-average conditions for April-May-June as a whole, and also slightly favours April being the
driest of the 3 months.
With this forecast, the water resources situation in southern, eastern and central England is likely to deteriorate further during the April-May-June period.
The probability that UK precipitation for April-May-June will fall into the driest of our five categories is 20-25% whilst the probability that it will fall into
the wettest of our five categories is 10-15% (the 1971-2000 climatological probability for each of these categories is 20%).
I think they got the probabilities a bit wrong
17 REPLIES
Community Veteran
Posts: 5,342
Thanks: 609
Registered: 23-09-2010

Re: Met Office Rainfall Forecast for April/May/June 2012

Makes you wonder what they use their ginormous computer for?
They should stick to seaweed. Probably just as "accurate" and it would save the tax payer a fortune.
Community Veteran
Posts: 18,545
Thanks: 191
Registered: 12-08-2007

Re: Met Office Rainfall Forecast for April/May/June 2012

And there are still hosepipe bans in force Roll eyes Roll eyes
Community Veteran
Posts: 16,836
Thanks: 1,124
Fixes: 13
Registered: 06-11-2007

Re: Met Office Rainfall Forecast for April/May/June 2012

I thought Michael Fish had retired from forecasting...... Cheesy
David_W
Rising Star
Posts: 2,293
Thanks: 29
Registered: 19-07-2007

Re: Met Office Rainfall Forecast for April/May/June 2012

The UK is very hard to predict, weather wise, over long periods, even over shorter periods as so much can possibly affect it.  Even the most powerful computers can't predict our weather as the variables can change quickly and constantly, what they can do is give a rough estimate based on previous readings and current readings but like I said, if just one variable changes then the entire prediction can become totally wrong.
Community Veteran
Posts: 18,545
Thanks: 191
Registered: 12-08-2007

Re: Met Office Rainfall Forecast for April/May/June 2012

It's raining cash for this guy:
cash bonus
alanf
Aspiring Pro
Posts: 1,931
Thanks: 77
Fixes: 1
Registered: 17-10-2007

Re: Met Office Rainfall Forecast for April/May/June 2012

Quote from: Oldjim
I think they got the probabilities a bit wrong

Unscientific! If I predict that on a toss of a unbiased coin six times it will probably come down heads 3 times and tails 3 times that does not mean I was wrong in my prediction if on one occasion it comes down heads 6 times.
IanSn
Grafter
Posts: 492
Thanks: 5
Registered: 25-09-2011

Re: Met Office Rainfall Forecast for April/May/June 2012

With the jet-stream flipping randomly away (and back again sometimes) from where its supposed to be (by a few hundred miles), for reasons they don't really understand, I'd imagine its even harder to predict what the weather's going to do one minute to the next.
Speaking of which, we haven't actually had a summer for about 4 or 5 years?... getting a bit much  Cry
Dressed for November everyone?
Community Veteran
Posts: 5,322
Thanks: 467
Fixes: 1
Registered: 21-03-2011

Re: Met Office Rainfall Forecast for April/May/June 2012

In our area "the ash came before the oak" so were were due  "a soak"  according to the country saying. We didn't need no massive expensive confuter to work out what the weather would be this summer.
Coupled with the fact that the local Mayor's side were supposed to be  playing cricket on official opening of the local council vanity project (cricket pitch) last weekend. It was obvious there'd be heavy rain.
Now Zen, but a +Net residue.
MJN
Aspiring Pro
Posts: 1,067
Thanks: 32
Fixes: 2
Registered: 26-08-2010

Re: Met Office Rainfall Forecast for April/May/June 2012

Quote from: Oldjim
I think they got the probabilities a bit wrong

Not at all; you just don't understand probabilities.
That's one of the biggest problems with weather predictions and the general populus - people expect to be told what the weather is going to be like and are not able/willing to accept that such forecasting is not feasible. Climate and weather is horrendousbly complex and verging on chaotic and thus must always be caveated with liklihood of occurance. Even then, as explained by another poster with the coin tossing, the fact a single event subsequently occurs or not has little bearing on the validity of that probability, particularly for long range forecasts (greater than a couple of days) where confidence levels will inevitably be low.
A typical forecaster nevertheless has to cater for this demand and dumb it down to a selection of symbols representing sun, rain and snow etc loosely scattered across the map whilst also being forced to give one word answers to the question 'Will we get a BBQ summer?' or 'Will it be a white Christmas?'. It is therefore not surprising that they often seemingly get it 'wrong'.
Given the nation's preoccupation with the weather I wish they taught more of it, particularly that relating to forecasting, at school then our collective understanding and expectations might well be a lot better.
Mathew
Community Veteran
Posts: 38,251
Thanks: 937
Fixes: 56
Registered: 15-06-2007

Re: Met Office Rainfall Forecast for April/May/June 2012

I do understand probabilities and a probability of 10-15% for the wettest is a bit wrong in the event
If their super computers and models were any good I would have expected them to get closer to the actual result
Note that from the linked paper the probability distribution tails off completely at about 330mm for April/May/June whereas from the actual records http://www.metoffice.gov.uk/climate/uk/anomalygraphs/ ; the anomaly for April was 182mm and for May was 101mm which, if added to the 1971-2000 average makes about 200mm + 182mm + 101 mm = 483mm and that ignores the rain in June
Looking at the distribution graph that number is way beyond any expected probability
Also just considering April the highest expected rainfall (the extreme high value on the probability) would be about 140mm and in fact it was approximately 250mm
MJN
Aspiring Pro
Posts: 1,067
Thanks: 32
Fixes: 2
Registered: 26-08-2010

Re: Met Office Rainfall Forecast for April/May/June 2012

Quote from: Oldjim
I do understand probabilities and a probability of 10-15% for the wettest is a bit wrong in the event

No, you clearly don't understand them as illustrated by that sentence. The fact that it may have been the wettest does not make their 10-15% prediction probability invalid whatsoever. You cannot observe the outcome of a single event and then retrospectively assess the validity of the predicted probability, as Alan's example illustrates.
Quote
Note that from the linked paper the probability distribution tails off completely at about 330mm for April/May/June whereas from the actual records http://www.metoffice.gov.uk/climate/uk/anomalygraphs/  the anomaly for April was 182mm and for May was 101mm which, if added to the 1971-2000 average makes about 200mm + 182mm + 101 mm = 483mm and that ignores the rain in June
Looking at the distribution graph that number is way beyond any expected probability

Give over Jim; for you to claim a greater knowledge than the Met over statistical analysis is nothing but laughable. Do you seriously think their forecasting is done by extrapolating a couple of lines on a graph and that's that?
Mathew
Community Veteran
Posts: 38,251
Thanks: 937
Fixes: 56
Registered: 15-06-2007

Re: Met Office Rainfall Forecast for April/May/June 2012

Frankly judging by the forecast probabilities they gave which was basically just a standard distribution with the centre moved slightly towards the drier end I have no faith whatsoever in their so called computer models.
That forecast wasn't a forecast at all it was actually a standard distribution of how the rainfall in that period has been in the past with a slight baseline shift and with the wetter tail reduced
Based on historical records and person with a piece of paper and a calculator could have produced that forecast.
I think that either you are deliberately misunderstanding my point or just stirring it up.
The basic point is that their so called forecasts are showing little or no skill despite all the super computers costing vast amounts of money running the GCM programs
Going back to the 10-15% that was actually based upon a rainfall somewhere between 280mm and 330mm approximately and so the actual rainfall was probably outside the 99.9% probability as generated by the model.
Of course if you are going to claim that the computer models are no more accurate than coin tossing even when only looking 1 month ahead then I will stop arguing as clearly you have as much faith in the computer models as I have
Community Veteran
Posts: 5,342
Thanks: 609
Registered: 23-09-2010

Re: Met Office Rainfall Forecast for April/May/June 2012

Quote from: Oldjim
I have no faith whatsoever in their so called computer models

Maybe that's where the met have gone wrong. They spent all our millions on a computer model instead of getting a real one. Smiley
Community Veteran
Posts: 2,284
Thanks: 219
Fixes: 1
Registered: 04-08-2009

Re: Met Office Rainfall Forecast for April/May/June 2012