Posted by nmm1 on March 7, 2010, 1:57 pm
>>
>>The long-term ones are based on pattern matching, and are merely
>>a more technological form of augury. A month ahead, there is some
>>chance of getting things right, but it's still a bit iffy.
>Like throwing a die or reading the chicken bones then?
That's a little unfair :-) More like buying the Oldest Inhabitant
a drink and listening to him rabbit on. "Ah, I remember a spring
like this in 1897. A right sodden summer there was that year,
too. We used a rowboat to carry the taters back when we dug them."
But done scientifically, you know. With lots of numbers, tables,
charts and graphs.
The point is that it isn't obviously invalid - i.e. it MIGHT work
better than pure guesswork. But, equally well, it might not. Now,
it the Met. Office had sponsored a quarterly television program
"Let's guess what the weather will do", explained the problem and
the various approaches, and (post hoc) compared their estimate with
the two most plausible crude ones ("the same as last year" and "the
average for the UK"), it would both have been interesting and not
attracted the condemnation.
Regards,
Nick Maclaren.
Posted by Sacha on March 7, 2010, 1:32 pm
>
>
> "Sacha" wrote
>> Probably wisely, the Met. office has decided to scrap its seasonal
>> forecast. Following the wishes of people surveyed, they will now start
>> offering forecasts a month forward. So we'll put the barbecues away
>> for now....
>
> When one considers the cost of/and the computer power they have at
> their disposal one wonders why they cannot be a bit more accurate now
> than they used to be.
Not enough windows in the new offices? Seaweed died off? Joking
aside, theirs is a thankless task so I think this is a very sensible
move. Does anyone truly believe they can predict 5 or 6 months ahead,
given what British weather is actually like? Someone only has to
sneeze on Dartmoor for there to be a gale in Plymouth. ;-)
--
Sacha
www.hillhousenursery.com
Shrubs & perennials. Tender & exotics.
South Devon
Posted by prb on March 8, 2010, 4:11 am
>
> When one considers the cost of/and the computer power they have at
their
> disposal one wonders why they cannot be a bit more accurate now than they
> used to be.
A forecast for all seasons,
Sunny intervals with scattered showers, prolonged in places. Winds
variable, temperatures about average for the time of year.
Try it, its never far wrong.
>>The long-term ones are based on pattern matching, and are merely
>>a more technological form of augury. A month ahead, there is some
>>chance of getting things right, but it's still a bit iffy.
>Like throwing a die or reading the chicken bones then?