Tropical cyclone track data from the European model now appears at HurricaneCity
Posted by Chris in Tampa on 6/2/2016, 4:54 pm
HurricaneCity's Tropical Cyclone Model & Best Track System:

http://hurricanecity.com/models/



I covered some of this a week ago, but this afternoon the system ran automatically without any issues so I wanted to let people know about it.

ECMWF (European Centre for Medium-Range Weather Forecast) data is available through ten days, when available.

It is only available for depressions and higher, not for invests. It will be available for the Atlantic, East Pacific and Cental Pacific at HurricaneCity, like other model data is.

It appears in the system with the following identifiers. (which are subject to change on the site)

ECMO:
ECMWF global model [GTS tracker] - This is the high resolution (HRES) Euro

ECME:
ECMWF Ensemble Control Member [GTS tracker]

EE01 - EE50:
Fifty ECMWF Ensemble Members [GTS tracker]

The data comes across the GTS (Global Telecommunication System), although I download it via FTP. I store it under the abbreviations I do because it seems to match the model identifiers in the NHC's model listing. The model files from the ECMWF do not actually contain four character identifiers.

The high resolution Euro will be stored under the identifier ECMO.

When available, the high resolution Euro will appear among the other NHC model data at:

3:05am EDT (0Z run)
3:05pm EDT (12Z run)

The ensemble data will appear, when available, at:

4:35am EDT (0Z run)
4:35pm EDT (12Z run)

Sometimes the ECMO may not appear, but some ensemble members will. That occurred earlier for Bonnie when it was weaker. Only some of the ensemble members were available and the high resolution Euro was not.

For Bonnie, after it was downgraded to an invest for a time, there was no ECMWF tropcial cyclone track data available for it. The first run that was available again was for 12Z today.

If Euro data is available, a message will appear on several of the pages that contain the data for that run. You can also look at the model archive page to see which runs have the data:

http://hurricanecity.com/models/models.cgi?basin=al&year=2016&storm=02&display=archive

It's in the "Model Data" column with the text "(with Euro)".

The Euro is a late cycle model. It appears with other late cycle models in the system in real time. The 0Z run will appear with the 6Z early cycle models. The 12Z run will appear with the 18Z early cycle models. Late cycle models are released hours after the forecast base time since they take longer to process. In the NHC's model file, they are released about six hours later when the early cycle models are then released. (The early cycle models are processed quickly, they might be released about 30 to 45 minutes after their forecast base time.)

ECMWF data comes from:

http://www.ecmwf.int/en/forecasts/datasets/public-wmo-and-acmad-datasets

The files are in binary format and cannot be downloaded and opened in a text editor, but if you are curious that is where the data comes from. (And if really curious, you can decode the tropical cyclone track files here.)

This data is considered WMO Essential. While some parts of the Euro are not free, this data is free for both commercial and non-commercial. There is some graphical data that can be used as well, but that is not displayed here.

Eventually the location of the maximum wind (coordinates), which is available for forecast points in the ECMWF data, will be added to the system. I have not done that part yet.

Some improvements will be made to how ECMWF data is displayed. I plan on grouping ensemble members together in the Google Earth file eventually, so they can be turned on and off more easily, as well as toggling members more easily on the Google Maps map.



Best Performing Models Map...

Based on how the best performing models map works, it may take up to 48 hours for the European model to be eligible to appear as the model system will need to compare the ECMWF's first 48 hour forecast position to the storm position 48 hours from that forecast. Once it can do that, it can get the model error for the 48 hour forecast and compare it to other models. To do that model error is calculated for the 48 hour forecast for all the models, using only the 48 hour forecast error over the past 24 hours. For many models, that means up to 4 runs. For the ECMWF, which comes out twice daily, that means up to two runs. Early in the life of a storm earlier forecast hours are used to determine the model error since early on there will not be a 48 hour forecast to compare to a storm position. The forecast hour used in the calculation slowly rises to 48 during a storm's first approximately 7 days.

Then the top 5 best performing models appear on the front page. Ensemble members do not appear in the best performing models map calculation for the ECMWF or other ensemble systems.



Model Names...

The model name list has been updated at Hurricane City. The model name file was updated a few days ago and I went through adding things that were new and updating what was old.

http://hurricanecity.com/models/models.cgi?page=models

If you want to know even more model names, here is some more for worldwide data:
http://www.nrlmry.navy.mil/atcf_web/docs/database/new/techlist.dat.pearl
139
In this thread:
Tropical cyclone track data from the European model now appears at HurricaneCity - Chris in Tampa, 6/2/2016, 4:54 pm
< Return to the front page of the: message board | monthly archive this page is in
Post A Reply
This thread has been archived and can no longer receive replies.