Model forecasts are generally very good, but . .

[Graham Easterling]

Translating them into weather forecasts for a particular location, though improving, have some way to go. Yesterday & today are a good example.

The MetO (I'm not picking on them, thay are certainly better for this location than the BBC) forcasts for

Yesterday (checked late morning.)

Dark cloud all day, maximum of 17C.

It was 19C by 13:00, reached 19.4C, good spells of sunshine from late morning until 16:00 or so. With little wind and high humidity it felt very warm.

Today (also checked late morning)

Half suns all day, full sun early afternoon, maximum of 20C

It was actually cloudy, quite dull at times, from 10:00-13:00. Temperature at 13:00 17C. Also quite breezy, so it felt a fair bit cooler. 

It has now brightened up with warm sunny spells, and the breeze has dropped somewhat. However, maximum so far ((now near 15:00) 18.1C, so down on yesterday.

It has been somewhat sunnier at NW facing Sennen, but a big difference in the sea conditions.

Yesterday, lovely swell in near calm conditions.

Today, blown almost flat by the offshore wind, and none of yesterday's deep colours.

I fully realise west Cornwall is a difficult geographical location for a post code forecast. However, if I could spend a few minutes glancing at the satellite & local wind conditions, and then decide that yesterday's forecast was helplessly pessimistic for the north Cornwall coast and set off for a surf in warm sunshine, then surely that information can be fed into the supercomputer to update the forecasts. After all it's probably more powerful than my 8 year old PC or iPad.

Still, we'll soon see what AI can do.

Oh, and bring back the awesome Magic Seaweed, boy they were accurate.

Graham

Penzance