[Better Than The Models ?]

Osmposm, you're absolutely right, I had been at Auntie's sherry last night and the Bristol Cream always brings out the worst in me. RJS's input is always scientific and well reasoned, I think I was having a moment due to some of the mocking and lets face it, insulting (quackery comes to mind) contributions from some posters. Apologies to Roger for my drunken misunderstanding, I think it was the right ammunition but most certainly the wrong target.

BTW, you should have seen that post before OON got hold of it, something about my good woman and a telephone box! Kids, don't do sherry!!!

[Better Than The Models ?]

There was a question about the method being used to assess the maps. It actually is fairly simple already, so I can't dumb it down all that much, but here's what is being done:

Looking at MB's maps, they cover an area mainly between 20 deg W and 20 deg E, and from about 45 N to 70 N. So to simplify comparison, I recorded the pressures (on the 3 Feb map) at each 5 deg intersection on that grid, for example, 55N 10W which would be in Donegal Bay. I added a map of my own with only this pressure grid. Now that we are into the 10-day period with reasonably reliable model runs, I am recording each day's output from GFS and ECM on these same grids. Later today (Thursday) we will be looking at the day 7 forecast maps for this validation.

So that leaves the question of how the maps are compared. This is a simple correlation, which is a statistical measure of similarity. You would have to check a statistics textbook for the definitions. But this is widely used in statistics, and runs from values of -1.0 (completely different data) to 1.0 (similar data). By "similar data" we infer a similar pattern, so for example if MB's map looked exactly right but was all 20 mbs lower, it would still correlate at 1.0 -- and presumably the winds would correlate highly too. A correlation of about 0.5 is quite significant in this sort of grid comparison. Even so, the standard definition of the meaning of correlation is this: when you square the correlation co-efficient, that gives you the percentage of variance explained. So even at this relatively high correlation of 0.5, only 25% of variance is explained.

Many highly-complex climate studies investigating temperature or rainfall and solar activity have been done over the years and you quite often see correlation stats of about 0.2 or 0.3 in those, so I've learned over the years in climatology that 0.2 is about the thresh-hold correlation at which anyone will get interested in a linkage, something lower than that is essentially random. A correlation of 0.1 after all explains 1% of the variance.

Highly negative correlations can be useful in hypothesis building -- if you tested your hypothesis and found a correlation of -0.98 you would realize that your guiding theory was reversed from reality. For example, if you were particularly thick and chose to research the notion that beer consumption would increase in cloudy, wet weather and then got a correlation of -0.9 on your study, then you might clue in that beer consumption would increase in sunny, warm weather. So a large negative correlation is not necessarily the worst news in a research study.

But before you ask for the next grant, you might want to change the hypothesis.

Man, you definitely know your stuff, but do you not think that your last paragraph is patronising or condescending? You have already made your point about the narrow minded scientific community and the alleged conspiracy to stifle new and radical ways of thinking, but the great irony is that you have joined their gang!!!

Can't we just wait and see? Is there any need for someone as you to denigrate MB's forecast? When it goes all tits up, pull it to pieces by all means, but for now as an intelligent man, I am shocked by your attitude and would hazard a guess that MB would be offended by this, after all, its only a forecast. BTW.... I fancy a weeks surfing at Woolacombe Bay at the end of April, surfs up? Sunshine? Anyone???? METO or MurcieBoy or RJS or GFS or....... Toss a coin...... What needs to be remembered is when the forecast was made.........

[Better Than The Models ?]

I read this thread and wonder...... What is it that makes intelligent people attempt to ridicule someone for the only reason that they disagree with what they say? He says "big storm", they say "no......" for many different and sometimes, but not always valid reasons, and thats the point...... But when Einstein's alleged IQ is brought in to this it all turns into a lot of ugly and very unhelpful dog biscuits.

For Christ's sake, judge the man on his forecast and nothing more.... for now.

Not that it matters, but I've been a member and an avid reader of this website for over six years now, and I've found the information here more than helpful on a number of occasions (nod to Steve Murr and Yamkin!!!!).

Anyway, can we please judge MurcieBoys forecast after the proposed dates and not before? I never post so I have to say a big thankyou to the hosts and mods of this site, first call for weather. I raise my glass to you.

Edit........ Dave? The Eye in the Sky...... Really enjoy your posts mate, best thing about this site is seeing you lot making monkey's out of the media!!!! And you do.

[Better Than The Models ?]

I don't understand the sarcasm directed towards MurcieBoy, he makes a VERY bold forecast which will stand or fall by the 5th of February. He provides no evidence or explanation for the forecast so we have no way to debunk it, yet. Let us see.

Steve Murr's contribution (minus the pub wind up) is certainly the most pertinant, "where does the energy come from?". Nevertheless, lets just wait and see and then let the dogs off the lead!!

[Sunday 19th December 2004]

3.0C, 1003hPa. Stopped raining about an hour ago. Nothing more than rain here tonight. :huh:

<{POST_SNAPBACK}>

Stratford, East London, light rain, 7 degees. For the record I ain't been sledging since 83. We need snow.