I wonât waste my time with the numerous studies that show the improvement of NWP accuracy in recent decades as I think it would be dismissed out of hand.
Originally Posted by: doctormog
I think there is an element of trying to be too accurate.
In the old days, we would have had a "sunshine and showers" symbol on the BBC weather map, plonked over a large area, and for just a morning or afternoon time slot.
These days you'll have an exact copy of the UKV output for a given hour, complete with convective shower predictions. This feeds the apps, of course.
Sadly far too many people take those apps as gospel, and get quite annoyed if it rains when the app was showing sunshine! (This happens a lot, as the models still struggle with convective rainfall). It's much less of an issue for frontal rainfall, as that's generally easier to predict.
Similarly in the timespan Jiries is referring too, forecasts were by necessity brief. You would have had something like "scattered showers in the south and east, longer spells of rain over Scotland" - whereas these days you will have a weather app showing an 80% chance of rain at 10AM, a 0% chance at 11AM, and a 30% chance at noon... way too specific for most people's needs!
That said, with time we tend to remember the good (or completely wrong!) forecasts, rather than the mediocre "rain lighter than forecast", "high winds didn't show up" and so forth. There were plenty of times in the 80s and 90s, at least, where snow was mentioned but something cropped up and it didn't come off as planned. And other times, the opposite, of course: "Maybe enough snow for a snowball", said the TVS weather forecaster, then we had a foot and a week-long freeze!
Again, now we have all the ensemble output, model runs four or more times a day, we get to see all those aborted weather things that don't come off. Pre-Web, they simply wouldn't have been mentioned in most cases!
(The "chief forecaster's briefings" were used by the BBC back in the day too, I suspect they no longer have access to these now they've ditched the Met Office and access the data via a third party. A shame, as those briefings were fascinating!
From "Will", who worked at the Met Office in the 90s - this was posted in usw in 1999:
OK time to put a few misunderstandings to rest once and for all.The TV forecasts, indeed all of the forecast output from UKMO is based on the
Chief forecaster's appraisal of the situation. With input from other forecasters
as well including those at the BBC on conference.
The forecasters at NMC Bracknell have access to every single model output from
all countries as well as a whole range of *proper* ensemble products as well as
mesoscale model outputs and frequent satellite data, radar data and all other
types of observations, the wealth of data is staggering. All of this is assessed
for consistency and runs which stand out will be noted but often not followed
even if it is the UKMO model. After coming to a consensus view the charts are
prepared using on-screen modification to ensure 4-D consistency.
The reasons behind the issue of the forecast charts and the weather that is
expected are then sent out to all UKMO forecast Offices as a briefing for all
forecasters including those on the media.
Sometimes alternative scenarios are drawn up with probability of occurrence.
These are for UKMO forecasters so that they can bear in mind when delivering
forecasts to customers.
That is a very general overview of the forecast process. Hope it clears up a few
misnomers.
There is no adherence to one model, all play their part, particularly the full
ensembles at longer range where they can provide more accurate guidance than any
one model. The superb handling of the recent anticyclonic spell with 1040 hPa
pressure is a case in point which was spot on in the ensemble mean which NMC
often used.
I rarely go into such detail in public, but as this is a scientific group I do
not wish mis-understandings to arise.
Â
Edited by user
13 August 2023 15:51:23
|
Reason: Not specified