Because of the increased interest internationally (non-U.S.) in the GFS model and the high resolution of the data, the NCEP data servers, which we have historically retrieved much of our GFS high resolution data, has become extremely congested and too slow for operational use during the GFS update cycle. However, we do have means of access to this higher resolution GFS data, but they are in different formats and have their own issues.
We are in the process of doing a much larger rewrite of how process the GFS model data and where we get the various data pieces. This will ensure a fast, reliable GFS model output going forward. This is a complicated process as we have a mixture of legacy 32bit code and newer 64bit code that we are trying to convert everything to.
This was not something we anticipated we would have to do, but sometimes you have to make adjustments on the fly. We hope to have running smoothly within the next several days.
Thanks for your patience.
Unexpected problems with the new GFS Surface Flux fields with some of legacy conversion software. The extremely high resolution of the new data is causing memory issues with the 32bit software. We are working to migrate this portion of the code over to 64bit software to handle the larger memory footprint these datasets require. In testing we didn’t run into this problem for some reason.
We apologize for the delay and hope to have this sorted shortly.
If there are delays or issues with the High Resolution Version of the GFS. You can use our
NOAAPORT delivered GFS model which is still the same upgraded GFS model, but has fewer variables, but is generally faster and very reliable.
The upgraded GFS model went live today. Over the next several days we will be tuning our systems to optimize how we process the new GFS output. The are a number of new variables and higher resolution fields available. We want to make the best use of our resources. You may notice some slight changes in the model graphics or timing of the products for a few days as we make changes in how we process this data.
We are currently working hard to bring on much higher resolution GFS products on our new Linux servers that we just setup. We hope to have the new GFS products available soon for our customers.
Thank you for your patience as we continue increasing our computing capacity and product offering here at Wright-Weather.com
There was a change in the way NCEP coded the delivery of the new GFS output of the 12Z data. We have corrected our system to handle this. We are rerunning the 12Z GFS now and actively working to get the 18Z running on the new cluster. We apologize for the delay.
This is a major upgrade for NCEP and us and there were a number of changes going from the old GFS to the new GFS model with a much different file naming scheme, as well as internal variables names.
The GFS model will be upgraded tomorrow, 1/14/15. It is under going a major upgrade. The below PowerPoint link will explain in detail the model physics and initialization changes that have been made, as well as, the effect on model performance.
This is a major change to the model and there will likely be new biases and features that we didn’t see in the old GFS due to the higher model resolution. The model is now running at about 13km (near the equator).
We now have several HP DL380 server’s clustered to handle the higher resolution GFS, as well as, HRRR data.
We will have the GFS data in 2 different resolutions. The standard 0.5 degree which have been displaying for years now and soon we will have surface and other 2D fields 0.25 degree.
Look for the higher resolution data in the coming days. The new GFS based model data at 0.5 degrees will begin tomorrow (01/14/15) however, unless NCEP decides to delay the implementation.
NCEP’s servers which we receive some of our model data have been degraded the past 24 hours. Some products will be delayed or possibly incomplete.
Model products which receive via satellite will be on time, which are not affected by the NCEP data servers/network issue.
The GFS NOAAPORT link is:
The NAM NOAAPORT link is:
The latest message sounds encouraging from NCEP. So hopefully any delays will be short-lived:
THE NCEP SUPPORT FOLKS HAVE BROUGHT THE PRIMARY FTP SERVER BACK
UP..FINAL CHECKOUTS HAVE BEEN PERFORMED ON THE PRIMARY FTP SERVER
AND IT HAS BEEN PLACED BACK IN SERVICE AT 0400Z..A PRELIMINARY
CHECK OF FTPPRD AND NOMADS INDICATED THAT DATA IS UP TO
DATE..QUEUES ARE BLEEDIND DOWN..SOME DATA MAYBE STUCK IN QUEUE
BUT FOR THE MOST PART DATA APPEARS TO BE UP TO DATE.. THANKS FOR
YOUR ON-GOING PATIENCE WITH THIS ISSUE TODAY.
A strong upper level low pressure system will move out of the Rockies into the Southern Plains tonight. This will intensify a surface low pressure system over Western Oklahoma that will move eastward through Oklahoma during the morning on Monday.
Returning Gulf of Mexico moisture across Central and Eastern Texas with dewpoints in the middle 60s to middle 70s will continue to advect northward overnight. Showers and thunderstorms are expected to develop across Central & Eastern Texas later this evening, ahead of a cold front that will sweep into North Central Texas around midnight. A triple point region will develop and move near the Red River with other smaller mesoscale boundaries south of the warm front where convection will develop. A strong line of convection is expected to develop along the cold front as it moves east of the dry line region around 4-5Z. Storms will likely become severe as they intensify under a strong 60-70kt cyclonic mid-level jet.
Parameters suggest all modes of severe convection are possible. Linear convection is likely to be the most predominant mode, with embedded line echo wave patterns (LEWP’S) with severe wind gusts, but embedded & discrete supercells are also quite likely given the high shear values, very high 0-3km helicities. Bulk Richardson Numbers also favor some discrete supercells along and south of I-20.
With the low lifted condensation levels, very high 0-1km helicities, shear values and moderate instability. Some tornadoes are quite possible with the stronger mesocylones.
The region at most risk of supercells is south of the Red River in Texas with an enhanced risk of tornadoes along and south of I-20 where the combination of helicities, shear, CAPE and upper forcing will be maximized. South of Waco, convective inhibition values will likely be too high to allow for much activity, but if a cell develops may be severe/tornadic.
The severe threat works eastward across Northeast Texas into Arkansas ans Louisiana during the day on Monday.
A spring like cold front will move through the Mid-West and combine favorable wind shear with strong instability Saturday into Sunday afternoon and will likely produce widespread severe thunderstorms with large hail, damaging winds with the possibility of a few tornadoes.
An unseasonably strong upper level short wave will drop southward into the Ohio-Valley on Sunday and will increase the mid and upper level flow across the region. Favorable instability, bulk shear & moderate helicity will support organized severe thunderstorms beginning on Saturday across IA/MO/IL/IN. The threat shifts eastward on Sunday into IN/OH/KY/TN/WV. A tornado threat exist along the quasi-stationary front across northern IN/OH and a conditional threat also exist within the low level theta-e advection zone across eastern Ohio/Kentucky & West Virgina where 0-1KM EHI’s & 0-1Km bulk shear are highest.
Should be an active 48 hour period. Additional severe storms are likely into Monday across the Northeast and Mid-Atlantic.
Added the ECMWF 500mb winds to the ECMWF regular resolution graphics. These can be useful in projecting the mid-level winds that can aid in organized severe thunderstorm development. These can be found in the drop down menu, as well as, the individual time frames. These are available for the Northern Hemisphere at the moment, but will include all sectors soon.
NCEP has recently upgraded the WRF-ARW/NMM Hires windows for the Eastern U.S. to a CONUS ARW and NMMB ~4km WRF runs. They are run at 00Z and 12z and produce hourly output out to 48 hours at roughly 5km.
We are in the process of making the necessary changes within our software and systems to accommodate the increase in files sizes and processing requirements to handle the larger domains. We should have these new data sets available and finalized later this week. Look for them where the old WRF-ARW/WRF-NMM EAST model pages were located.
Here is the NCEP technical bulletin describing all the changes to the model & physics in the upgrade.