Are MeteoGroup using ECMWF model data at the BBC?

123 views
Skip to first unread message

xmetman

unread,
Mar 8, 2018, 9:47:38 AM3/8/18
to weathera...@googlegroups.com

Keith (Southend)

unread,
Mar 8, 2018, 10:29:32 AM3/8/18
to Weather and Climate
Certainly looks the case from your example Bruce.

Keith (Southend)

Jack Frost

unread,
Mar 8, 2018, 1:43:51 PM3/8/18
to Weather and Climate
Well on their website it says "A combination of models such as real-time ECMWF, GFS and our large network of observation stations power our sophisticated forecasting model." Does this mean they don't actually have their own in-house model, and just use the output from other models?

Liam

Julian Mayes

unread,
Mar 8, 2018, 4:16:07 PM3/8/18
to Weather and Climate
Evening all. The easy way to find the answer would just be to ask me! So, here's the answer -yes. My colleagues develop the daily weather story by examining all major models. For the synoptic scale, EC is often used as the basis, but with our frontal analysis and by editing the weather elements according to our interpretation of all models. The latter point can be important. An example might be when a model misses or is slow to acknowledge a developing feature - for example EC was a bit slow to highlight this morning's trough over central Britain. As this looked like giving another snow event, it would have been a serious omission.

To answer Liam's point, we need to distinguish between NWP models that run on a handful of supercomputers around the world, usually developed by the largest national met agencies, and other types of model which are more commonplace which we do have, such as road forecasting energy balance models. It is up private sector forecasting companies to develop a business case for the purchase and use of global model data. There's nothing new about this - us and others have been doing this for decades. We could hardly compete otherwise now could we? I hope this clarifies.

Cheers. Julian Mayes, MeteoGroup.

Jack Frost

unread,
Mar 8, 2018, 7:08:16 PM3/8/18
to Weather and Climate
Ah, okay. Thanks Julian!

I guess it's like AccuWeather over here in the US, who seem to use the GFS and ECMWF output (along with some US mesoscale models) for their forecasts. Is the Met Office the only big NWP centre who doesn't sell their data to other organizations?

Liam

John Hall

unread,
Mar 9, 2018, 2:17:39 PM3/9/18
to Weather and Climate
I hope that AccuWeather's forecasts for the US are better than those for the UK that appear in the Daily Telegraph, which tend to be pretty poor. 

Jack Frost

unread,
Mar 9, 2018, 2:40:46 PM3/9/18
to Weather and Climate
Surprisingly, they do seem to be quite good.

I usually watch the forecasts on the local Chicago TV station, which are in association with AccuWeather. They are presented by people who actually have BSc's in meteorology, and have lived in Chicago for years, so they know the area and often disagree with what the model predicts! I also like it that when rain/snow is predicted they show the outputs from different models, to get an idea of the spread of possible totals.

In general, if they warn of thunderstorms, or heavy rain or snow up to about 5 days in advance, it's usually correct. I was impressed last summer how well forecast thunderstorms actually were. Maybe Chicago's weather is a bit more predictable than in the UK. The weather stories on the AccuWeather website to tend to sensationalize upcoming events, but I just ignore those bits, and look at the data.

Liam

John Hall

unread,
Mar 9, 2018, 3:42:59 PM3/9/18
to Weather and Climate
I suspect that when you say "and have lived in Chicago for years", that's a major factor. I imagine that their forecasters for the UK are making the forecasts from 3,000 miles away. Also forecasting for a city must be easier than forecasting for a substantial area with sub-regions that can have very different weather from one another, especially when you have only limited space in the paper to cover then all. There's also the point that, to accord with the newspaper's deadlines, the forecast has to be made perhaps twelve hours before people will read it.

All but the first point would have affected the Met Office forecasters as well of course, back before they decided to stop providing papers with daily forecasts. But my memory is that their forecasts were a good deal better, even though this must have been ten or more years ago, and one would expect forecasts to have improved since then.


Reply all
Reply to author
Forward
0 new messages