Model/Run updates

Discodog · 14-02-2014 11:37am #1

Are the latest runs issued to all the sites at the same time or is there one site that gets them first.

I am trying to guess when the rain will clear Galway. The Hirlam, GFS & NAE updates were almost 12 hours ago. This seems too long for a typical fast moving Atlantic storm.

Is Weatheronline the best source?
ME's new short forecast updates every hour but they fail to mention that HIRLAM only updates every 6 hours.

octo · 14-02-2014 1:47pm

Discodog wrote: »

Are the latest runs issued to all the sites at the same time or is there one site that gets them first.

I am trying to guess when the rain will clear Galway. The Hirlam, GFS & NAE updates were almost 12 hours ago. This seems too long for a typical fast moving Atlantic storm.

Is Weatheronline the best source?
ME's new short forecast updates every hour but they fail to mention that HIRLAM only updates every 6 hours.

I think the new forecast runs every hour.

Discodog · 14-02-2014 8:37pm

ME update every hour but surely HIRLAM isn't being rerun. I also can't see how ME could redraw it every hour.

maquiladora · 14-02-2014 8:41pm

Discodog wrote: »

ME update every hour but surely HIRLAM isn't being rerun. I also can't see how ME could redraw it every hour.

It shows the next 9 hours based on the latest HIRLAM. Them model itself updates 4 times a day. Met Eireann simply updates the page to show you the next 9 hours of the current model.

Discodog · 14-02-2014 8:45pm

maquiladora wrote: »

It shows the next 9 hours based on the latest HIRLAM. Them model itself updates 4 times a day. Met Eireann simply updates the page to show you the next 9 hours of the current model.

Yes that's what I assumed. It's no different to yr.no.

So why do the latest model runs take so long to appear on websites?

maquiladora · 14-02-2014 9:03pm

Discodog wrote: »

Yes that's what I assumed. It's no different to yr.no.

So why do the latest model runs take so long to appear on websites?

Same model as on Yr.no yes. Higher res output on the ME site though, pressure and rainfall charts are more detailed.

Why does it take so long? Even for supercomputers it takes time to do the huge amount of calculations needed to model the atmosphere. The higher resolution of the model, the longer it takes.

Discodog · 14-02-2014 9:17pm

maquiladora wrote: »

Same model as on Yr.no yes. Higher res output on the ME site though, pressure and rainfall charts are more detailed.

Why does it take so long? Even for supercomputers it takes time to do the huge amount of calculations needed to model for the atmosphere. The higher resolution of the model, the longer it takes.

So is the quoted run time when they started or finished :-) ?

At 10am this morning Weatheronline were still showing the 00.00 charts. Do they appear earlier any where else?

maquiladora · 14-02-2014 10:15pm

Discodog wrote: »

So is the quoted run time when they started or finished :-) ?

At 10am this morning Weatheronline were still showing the 00.00 charts. Do they appear earlier any where else?

Which model? They all roll out at different times.

Harps · 14-02-2014 10:16pm

The time quoted on the model run is the time it was initialised, it takes a while to run and I'd assume also takes a while for the NOAA, UKMO etc to make the model data available to the public and for various websites to then convert the numerical data to the graphical charts we see.

If you're looking for the most accurate rainfall predictions the hirlam and ukmo hi-res are probably the two best to check though I've never found any model particularly accurate for precise rainfall predictions

octo · 15-02-2014 4:19pm

Discodog wrote: »

ME update every hour but surely HIRLAM isn't being rerun. I also can't see how ME could redraw it every hour.

Email them to find out, but I've been told that it's a "mini-model" of sorts that takes its boundary conditions from the latest HIRLAM run, but updates every hour from the latest synoptic reports.

maquiladora · 15-02-2014 4:36pm

octo wrote: »

Email them to find out, but I've been told that it's a "mini-model" of sorts that takes its boundary conditions from the latest HIRLAM run, but updates every hour from the latest synoptic reports.

I did a little digging and you're right. The HIRLAM on the short range forecast is not just the main HIRLAM that is run 4 times a day. It does take that output but it is then run hourly out to 9 hours, it used to be 6 hours. I didn't even know that existed!

Discodog · 15-02-2014 4:49pm

maquiladora wrote: »

I did a little digging and you're right. The HIRLAM on the short range forecast is not just the main HIRLAM that is run 4 times a day. It does take that output but it is then run hourly out to 9 hours, it used to be 6 hours. I didn't even know that existed!

Does this mean that the ME short forecast should be more accurate than, say yr. no., which is only updated every 6 hours?

Discodog · 15-02-2014 4:54pm

Harps wrote: »

The time quoted on the model run is the time it was initialised, it takes a while to run and I'd assume also takes a while for the NOAA, UKMO etc to make the model data available to the public and for various websites to then convert the numerical data to the graphical charts we see.

If you're looking for the most accurate rainfall predictions the hirlam and ukmo hi-res are probably the two best to check though I've never found any model particularly accurate for precise rainfall predictions

I assumed that NAE was supposed to be the best but updates are often really slow.

Our weather conditions can change rapidly so frequent updates must help.

The only thing that I trust & rely on is the radar.

maquiladora · 15-02-2014 5:50pm

Discodog wrote: »

Does this mean that the ME short forecast should be more accurate than, say yr. no., which is only updated every 6 hours?

Yep!

Model/Run updates

Comments