COAMPS GRIBs

[franksingleton]

I am not a fan of meso-scale forecasts because I place more importance on planning rather than on short term predictions. However, I have to give some talks in the New Year and wanted to see how COAMPS was performing. I have been looking at COAMPS because it is a free service that gives data at 0.2 degree spacing if you get it using Saildocs.

Over the past 2 days I have been looking at COAMPS in the area where we were his year – N Spain down to about Cape Finisterre. It has been very poor compared to the GFS, NAE, UKMO, NOGAPS and SWS all available on weatheronlin.co.uk and/or Passageweather.

Is there anyone out there who uses COAMPS regularly? Is this a blip or is it just not good? I understand that it uses NOGAPS as its basis but NOGAPS seems to be pretty similar to the GFS and nothing like COAMPS. I am mystified. Any views?

 

[Roberto]

Availability

Hello Frank,

I have not used them so I cannot comment about performance.
Just a few notes about availability

As of today, if one opens the GRIB window request in Airmail/Sailmail/Viewfax (which is a sofwtare used by most people having SSB email capability + some with satellite), asking for COAMPS, you get this image about data availability, that is only waters around the US

I knew that after a number of requests the US recently agreed to release COAMPS for the Mediterranean as well, so they are available, even if the area is not "shaded" in the Airmail software. That may be a reason why not many people are looking at this data.

Coverage is not worlwide anyway, on the West Atlantic, COAMPS availability stops where the Airmail shaded area ends, I just tried asking for a COAMPS of the NE coast of South AMerica and the file stopped exactly at 60W where the dashed line is located.

On the Eastern Atlantic side, I tried to ask this area

and the answer was this (the clutter it's just wrong display settings)

Availability stops at 15W which appears sensible given the type of model, and at 30N that does not help much someone going towards the Canaries.
North of the UK no problems it goes to Iceland and beyond.


FWIW..

regards

 

[tudorsailor]

Looking at PassageWeather I see that it offers GFS SWS and COAMPs

I like zygrib but it does not specify what data it uses

Could we have lesson on the different sources of data and their acronyms?

Thanks

TS

 

[franksingleton]

Hello Frank,

I have not used them so I cannot comment about performance.

.....

Availability stops at 15W which appears sensible given the type of model, and at 30N that does not help much someone going towards the Canaries.
North of the UK no problems it goes to Iceland and beyond.


FWIW..

regards

Thanks, Roberto.

The latest Saildocs news that I have says

US NAvy's COAMPS ("Coupled Ocean/Atmospheric Mesoscale Prediction System") model, which is available for the eastern Pacific, Central Americal and western Atlantic with a resolution of 0.2 x 0.2 degree. This data is now available on the internet, and is available from Saildocs on an experimental basis. This data comes from a server which is not an official "production" server, and is on an "as-available" basis.

.

However, Passageweather has COAMPS for the Mediterranean, Biscay, waters around the UK and the Baltic.

I used email to query@saildocs.com and sent messages like

send coamps:40N,48N,12W,4W|0.2,0.2|0,6,...,48|wind

. The system worked.

The T=0 looks OK but it seems to go astray after 24 hours and by 48 it was nonsensical. All very odd as the USN starts with NOGAPS. They say that they use a 3D Var data assimilation scheme. I cannot believe that their meso-scale model is so bad.

Mind you, running the GFS with a 25 km grid (although we only get it at 50 km) is making regional models redundant. Although the UK still runs the NAE, they do not regard it as an operational tool. I think that the GFS is as good as anything - as long as you remember that it usually under-forecasts the stronger winds.

Best wishes to you and the family.

 

[franksingleton]

Looking at PassageWeather I see that it offers GFS SWS and COAMPs

I like zygrib but it does not specify what data it uses

Could we have lesson on the different sources of data and their acronyms?

Thanks

TS

zyGrib uses the GFS for its wind forecasts. If you get the latest version you will see that it also gives wave forecasts using WW3.

I am off to bed just now but will try to give some thought to acronyms. Horrid things but everone uses them.

 

[franksingleton]

Looking at PassageWeather I see that it offers GFS SWS and COAMPs

I like zygrib but it does not specify what data it uses

Could we have lesson on the different sources of data and their acronyms?

Thanks

TS

First, zyGrib and Ugrib both use the GFS.

GFS = Global Forecast system, The US NCEP (national Center for Environmental Prediction). The model uses a grid of about 25 km but only gives data on a 50 km grid.

NOGAPS = (US) Navy Operational Global Atmospheric Prediction system. It is an alternative when the GFS is unavailable. No better.

COAMPS = (USN) Coupled Ocean / Atmosphere Musicale Prediction System. A meso-scale model that takes output from NOGAPS to initialise the model, the computes for smaller areas on a grid of about 10 km. Data are issued on a 20 km grid. On Passageweather the data are at about 50 km spacing.

UKMO is sometimes used to describe the UK Met Office global version of its unified model. It runs on a grid of about 25 km. The only output I see from it in grid point form is on Weatheronline.co.uk.

NAE= North Atlantic European model. This is the UK limited area model covering that area and available only on the Weatheronline site.

NAM = North American Meso-scale model. Sometimes known as WRF NMM or Weather Research Forecast Non-Hydrostatic meso-scale Model. The grid spacing used is about 10 km but data are provided on a 20 km grid.

SWS = Speedwell Weather Services. They, like others, provide meso-scale forecasts on repayment. Quite likely, this will be the WRF NMM. Their output is available on Passageweather on a grid that looks like 50 km.

ECMWF = European Centre for Medium Range weather Forecasting. I understand that GRIB data are available but I do not know how to get them – ie am not clever enough. In effect, it uses a grid length less than other models.

In performance, ECMWF comes out on top using the RMSE (root mean square error) of surface pressure at 72 hours over the N Hemisphere. UKMO comes next, followed by all the other models in a bunch. However, there is little difference between them all. I use the FDS and find it to be a very useful planning tool.

Meso-scale models whether those from national weather services or private companies have to start from a global model bases. They are likely to better guidance in the short term, maybe 24 hours, in terms of wind speed.

I hope that helps. I have no doubt missed some. Please come back if you have further queries.

 

[billreilly]

COAMPS = (USN) Coupled Ocean / Atmosphere *Mesoscale* Prediction System. A meso-scale model that takes output from NOGAPS to initialise the model, the computes for smaller areas on a grid of about 10 km. Data are issued on a 20 km grid. On Passageweather the data are at about 50 km spacing.
. . .
SWS = Speedwell Weather Services. They, like others, provide meso-scale forecasts on repayment. Quite likely, this will be the WRF NMM. Their output is available on Passageweather on a grid that looks like 50 km.

COAMPS data is issued on an 18 km grid... on PassageWeather.com the data are at 18 km spacing on the forecast areas that are small enough to show that much detail.

The SWS model is developed and run by Sailing Weather Service (not Speedwell Weather Services), and is a development of the MM5 model... The SWS-CM1 is a coastal model run on a 10 km grid, while the SWS-RM1 is a regional model run on a 25 km grid... on PassageWeather.com the data are at 10 km and 25 km spacing on the forecast areas that are small enough to show that much detail.

As with all the wind forecast charts, sometimes the size of the chart makes it impossible to show all data points (and still be readable), so when necessary PassageWeather interpolates the data to a coarser grid for readablity purposes... the size of this grid varies depending on the size of the area...

Bill

 

[franksingleton]

COAMPS data is issued on an 18 km grid... on PassageWeather.com the data are at 18 km spacing on the forecast areas that are small enough to show that much detail.

The SWS model is developed and run by Sailing Weather Service (not Speedwell Weather Services), and is a development of the MM5 model... The SWS-CM1 is a coastal model run on a 10 km grid, while the SWS-RM1 is a regional model run on a 25 km grid... on PassageWeather.com the data are at 10 km and 25 km spacing on the forecast areas that are small enough to show that much detail.

As with all the wind forecast charts, sometimes the size of the chart makes it impossible to show all data points (and still be readable), so when necessary PassageWeather interpolates the data to a coarser grid for readablity purposes... the size of this grid varies depending on the size of the area...

Bill

Thanks for the corrections.