They haven't got a clue

[bobk]

"Credit it to you Simon for doing a forecast on a farming forum, you are on a hiding to nothing."
Yeah, I must be mad

Oh dear, the dreaded 'accuracy' stats. And therein lies another tale.

Firstly, what are we measuring? When I make monthly/seasonal predictions I always stress they are guidance and subject to change. These 'guidances' are to be seen completely differently to conventional day to day forecasts. It's the trends in the guidance forecast which should be of most use, rather than the forecasts themselves. So, for example, if a long range forecast (see you have me saying 'forecast'; now ) changes day to day or week to week then confidence in that forecast must be low. If it is a consistent story the confidence in it rises.

Should you be using the longer ranges to plan day to day activities? Certainly not, but what you can do is use them to plan workloads in the coming month, i.e. I am currently thinking (could all change), somewhere mid-June for a 5 or 6 day spell of decent harvesting weather. As a farmer I'd expect you to now be thinking about contractors for that time, with a view to it possible all changing the closer we get. But at least some information is better than no information at all.

I produce monthly 'forecast' maps of temperature and rainfall for some private clients. These are amended and updated, but the originals are always available. They are available and can be used as a rough guide, but can't really be used for true accuracy figures as they aren't that sort of guide. Probably not making much sense, but I hope you understand.

You might think measuring day to day accuracy is easier. Oh, how wrong you would be! Did you know the Met Office use a guide of +/- 2C for maximum and minimum temperatures? From their website "The early morning forecast on our website is used to produce a percentage number of the times when the forecast is accurate to within +/- 2°C". Now, that means that if I predict a maximum temperature for your farm of 20C for tomorrow, if it is anywhere in the range 18 to 22C then I have been 100% accurate! Or, if I predict a minimum temperature of 2C, anywhere between 0C (a frost) and +4C (certainly no frost) is deemed accurate?

To me those figures don't stack up.

So accuracy is certainly a debate that we need to have. Ultimately, I believe it is my customers (you!) who get to say whether I am accurate or not. If we aren't accurate, you don't use us, and certainly won't pay for our forecasts.

Sorry, wittering on, but this is a topic I've struggled with for many, many years.

Best wishes,
Simon

I think your farmers weather sponsored by farmers guardian is excellent .

 

[linga]

Makes you wonder how they can then base a whole raft of taxation and regulation on 'Global Warming' if their models are as accurate as that......

because weather and climate is different.. a good analogy is tides. Its perfectly possible to predict when and what height a high tide is going to be and well in advance too but no one can predict what height each individual wave will be nor exactly when it will arrive.

 

[Simon Keeling]

Thanks bobk, nice to hear.

Yes linga, agree with that!

 

[Goweresque]

because weather and climate is different.. a good analogy is tides. Its perfectly possible to predict when and what height a high tide is going to be and well in advance too but no one can predict what height each individual wave will be nor exactly when it will arrive.

However predicting the high and low tides has a long history of accurate prediction. The method used has proved its predictive capacity over decades if not centuries.

Whereas the history and accuracy of modern computer climate models are.......................



Very short and zero.

 

[bobk]

ffs they've done it again , predicted heavy rain friday all week and now it's changed to dry all day

 

[Simon Keeling]

Hi bobk,

Who are 'they'? suspect BBC?

All week WE have been predicting that there would be rain passing east on Friday. In the west it would arrive as a 'proper' band of rain, but as it heads east it would fracture and break up, turning ever more showery and having virtually nothing left on it by the time it reached eastern England.

Our forecast for the Midlands has always been for totals less than 3mm, most likely nearer the 2mm level.

Don't throw the baby out with the bath water (I know from your previous message that you're not, but please state who 'they' are so we aren't all tarred with the same brush )

Best wishes,
Simon

 

[bobk]

Hi bobk,

Who are 'they'? suspect BBC?

All week WE have been predicting that there would be rain passing east on Friday. In the west it would arrive as a 'proper' band of rain, but as it heads east it would fracture and break up, turning ever more showery and having virtually nothing left on it by the time it reached eastern England.

Our forecast for the Midlands has always been for totals less than 3mm, most likely nearer the 2mm level.

Don't throw the baby out with the bath water (I know from your previous message that you're not, but please state who 'they' are so we aren't all tarred with the same brush )

Best wishes,
Simon

Yes " they " are the bbc, in fact just seen breakfast weather showing huge swath of rain passing over the uktomorrow

 

[linga]

However predicting the high and low tides has a long history of accurate prediction. The method used has proved its predictive capacity over decades if not centuries.

Whereas the history and accuracy of modern computer climate models are.......................



Very short and zero.

Short yes but some would disagree with your statement of zero.
But if one has a position of belief then that is a difficult position to change. No doubt links and figures can be produced to prove anything

 

[Goweresque]

ffs they've done it again , predicted heavy rain friday all week and now it's changed to dry all day

I have this theory that you need to get a handle on the 'theme' of the summer. This year seems to be a 'generally dry, with occasional dampness' one, for me at least. So you can mow grass and not worry about it getting rained on for days on end. Maybe a shower or two, but nothing drastic.

Whereas another summer may have a totally different 'regular heavy rain storms crossing the country' theme. We all know the weather gets stuck into patterns, if you can identify the pattern, it gives you a better chance of deciding if the rain that is forecast 5 days out is actually going to be a massive low pressure crossing the country, or just a few spits and spots here and there.

 

[onthehoof]

The BBC weather app is absolute rubbish. Have hay to bale tomorrow, the BBC telly forecast for tomorrow says rain fizzling out and hardly any here, the weather app says rain all day with thunderstorms all evening !!!

 

[Paddington]

Why is it a meteorologist will struggle to forecast within +/- 2C in a weeks time, but a climatologist can tell you with 100% confidence what the average temperature will be in 30 years time to the nearest 0.1C

 

[Simon Keeling]

Love it Paddington....very good point

BBC weather app is there to know whether to hang out the washing, not whether to make hay. If it's a website, and it's free then it ain't worth looking at - remember, this is your for your business, not the washing!

By the way, models are struggling with Fridays rain. They still want to reactivate the front as it comes eastwards and give some moderate bursts of showery rain along the front. I suspect most of England (away from the west coast) will get away with little rain. Most of it always in the west.

Simon

 

[Simon Keeling]

Latest models for tomorrow now becoming available. Two charts for you....GFS is out now (below), expecting the hi-res UK Met Office rain charts by around 1230, will post here.

1. Here's the GFS rainfall prediction from 10am to 4pm . Front fragmenting as heads east. But notice the heavier rain SW of Isle of Wight?

2. That rain then passes through SE England late Friday afternoon (1pm to 7pm rainfall shown on chart). See what I mean about the front reigniting as it heads east? I suspect this is just due to the cool air aloft overriding the warm air at the surface in southern and southeast England and allowing an odd moderate showery bursts to form along the front. Notice the still fragmented nature of the front to the north?

 

[linga]

Why is it a meteorologist will struggle to forecast within +/- 2C in a weeks time, but a climatologist can tell you with 100% confidence what the average temperature will be in 30 years time to the nearest 0.1C

Don't think any climatologist would quote 100% confidence.
But isn't this just weather versus climate again

 

[bobk]

I have this theory that you need to get a handle on the 'theme' of the summer. This year seems to be a 'generally dry, with occasional dampness' one, for me at least. So you can mow grass and not worry about it getting rained on for days on end. Maybe a shower or two, but nothing drastic.

Whereas another summer may have a totally different 'regular heavy rain storms crossing the country' theme. We all know the weather gets stuck into patterns, if you can identify the pattern, it gives you a better chance of deciding if the rain that is forecast 5 days out is actually going to be a massive low pressure crossing the country, or just a few spits and spots here and there.

Good points , anyway makes no odds really as it's doing nothing today . I'll go and get some more plastic

 

[bobk]

@Simon Keeling , I'm thinking of hanging my washing out for a week , next week or the one after ?

 

[Simon Keeling]

Met Office 10am to 4pm, notice ideas for heavy showers in SE, suspect this is showery bursts ahead of main front, looks overdone

Met Office 1pm to 7pm, not showing the SE showers. Think this confirms models just high;fighting possibility of the front coming alive again briefly as it passes, one or two moderate showers on it?

 

[Simon Keeling]

Good points , anyway makes no odds really as it's doing nothing today . I'll go and get some more plastic

I hear there's a great app for that from BBC Weather

 

[Bury the Trash]

Simon, just interested ,would there be a good career now in meteorology? I have a youngster here who is doing well in Maths and science (stronger on maths).cheers.

 

[The Agrarian]

Seeing as you're on this thread simon, I'd like to add that your look ahead videos have transformed the way I approach forecasting in general. The charts available are superb. I'm the kind that has to be doing everything myself, so I love to get stuck into the charts for a little while each day (while I'm doing weather dependant work, of course!). I've learned an awful lot from those videos. There is plenty that I still have to begin to get a grasp of, like MJO etc. But that will take lots of reading some other time!

Most of all, I appreciate your realistic and honest approach to forecasting. You don't come across as wanting to use it as a 'get out clause', or to cover your back end (like the met office and bbc do) - just to present it as you see it. Farmers like straight talking!