More About the Model Guidance Site

    I have fashioned a script system to generate plots of track and intensity guidance for current tropical cyclones.  This page contains information about the product availability, data sources, forecast aids, and terms of use.

How to Use the Guidance Page

The guidance page contains links to the last five plots generated from the various forecast aids in each basin. These can be accessed by clicking on the frame links below (frame 1 being the newest plot and frame 5 being the oldest plot). If there is more than one storm in the basin, then the current plots for each storm will be in frame 1, frame 2, frame 3, etc. up to the total number of storms in the basin. The remaining frames will contain the previous plots from those storms (which could be quite old if nothing has been happening recently). The current active storms are displayed in the small text box near the top of the page. The systems are designated by their A-deck filename, e.g. aal052005.dat. The first letter (a) just indicates that these are the A-deck files, the next two letters designate the basin (al for Atlantic, ep for Eastern Pacific, cp for Central Pacific, wp for West Pacific, io for North Indian Ocean, and sh for Southern Hemisphere). The next two digit number is the storm number (05 in this case). Numbers 90-99 are issued to invests when it appears that a tropical disturbance may develop into a tropical depression. Finally, the last number is the 4-digit year (2005). 

•  Remember that the current official NHC forecast (OFCL) is NOT shown on early guidance plots. Instead, the 6-h old interpolated forecast is shown (OFCI).
•  Be aware that you should not make life or death decisions based on model guidance - refer to your local emergency management officials and statements from their local National Weather Service Forecast Office.
•  Remember that models are frequently prone to LARGE track and intensity errors.
•  Do not fixate on any one model forecast. All models are wrong to some extent. The likelihood that any single model forecast track will actually verify (say within 10 miles) is quite small in most cases.
•  The model track guidance can be useful for getting a general idea of which way the storm will go. The spread of the model guidance can often give a useful estimate of the uncertainty associated with the forecast (but not always).
•  In general, the uncertainty of a forecast increases with lead time. A 5-day forecast has much more uncertainty than a 2-day forecast. Please familiarize yourself with the updated verification statistics of both the official forecasts and the various models: National Hurricane Center Forecast Verification and 2005 Verification Report.

Product Availability and Timeliness

The guidance system should update every six hours at or a few minutes before the following times:  0200, 0800, 1400, and 2000 (UTC).  

•  Late guidance products will be available at the same time as the early guidance, but the valid time is delayed by 6 hours. 

•  Intensity guidance products are currently shown for the early-cycle aids, but some of the models shown are actually interpolated late-cycle models. 

Sources of Data

    Nearly all of the operational forecast aids for the Atlantic, Eastern Pacific, and Central Pacific basins are obtained from the NOAA's high speed public ftp server. The raw forecast data are contained in files called 'adecks'. The UKMet Office forecast model track forecasts (UKM) are obtained by special arrangement from the UK Met Office. From time to time, experimental forecast aids may also show up in the plots.

About the Forecast Aids

    Each model or forecast aid has a unique 4-character 'TECH' identifier. For an index of the 'TECH' identifiers of the various forecast aids, click here. For more information about the various forecast aids, please see the newly updated (as of Sep 2007) model description page from the National Hurricane Center.

    The forecast aids are produced by various numerical weather prediction centers around the world. The forecast aids fit into two categories: early-cycle products and late-cycle products. Early-cycle products are forecast aids which only take a short time to produce and are available at the synoptic time. Most of the statistical and simple dynamical models fit into this category. Late-cycle products are the forecast aids which take longer to run and hence, are not available at the synoptic time. These are generally available 6-h after the synoptic time, so by the time you get them, the forecast is already 6-h old. The global and regional  models, and the products which depend on them are generally late-cycle products. To allow comparison between the early- and late-cycle products, the late-cycle products are interpolated in time to the same times as the early-cycle products. Thus, in the early-cycle plots, all the forecast aids with a 'TECH' identifier ending in 'I' are interpolated late-cycle products and represent a forecast at least 6-h old.

    GFS ensemble plots are also generated for the Atlantic and Eastern Pacific basins. These plots show the uncertainty inherent to the data assimilation system used in NCEP's Global Forecast System (GFS) model (the same model that makes the 'Aviation' forecast). The ensemble mean track is AEMN, the standard Aviation track is AVN0. For comparison, the hybrid consensus ensemble (CONU) is also shown (this consensus is a combination of the NOGAPS, UK Met Office, GFS, and GFDL models).

About the Plots

    The track and intensity forecasts for the various forecast aids are parsed and converted to a convenient format, then passed to a plotting script along with a few pieces of information about the basin, etc. The plotting script uses NCAR Command Language (NCL), a versatile (and free!) scientific data processing and visualization language that has been developed specifically for the atmospheric sciences.

Acknowledgements

    Nearly all the forecast data is provided courtesy of the National Hurricane Center.

    The UK Met Office model is provided by special arrangement with the UK Met Office.

    I owe special thanks to Mary Haley for her helpful assistance with several aspects of the plotting script.   

    I am grateful to my advisor, Wayne Schubert, for allowing me to develop and maintain this page on work time.

    Finally, my research group provides the computer platform that the products are generated on, and CSU provides the Internet bandwidth for this site.

Terms of Use

    DISCLAIMER: By using this web site, you agree to release and indemnify the owner of this page (Jonathan Vigh) from any and all liability, damage, or loss arising from your use of, or any decisions based on, this web site. The model guidance products on this page are derived from various numerical weather prediction models or statistical methods. They come from multiple numerical weather prediction centers around the world. These models are prone to LARGE errors. Absolutely no guarantees are made as to the accuracy or timeliness of this data. None of the information here should be considered as official guidance -- DO NOT USE for life and death decisions. For official storm information including watches and warnings, rely on the National Hurricane Center, your local National Weather Service office, and your local emergency officials. NO MEDIA may mention this page or feature any products from this page without obtaining express permission from the web site owner. Please also view the general disclaimer for this web site.

Please direct comments or questions about this page to Jonathan Vigh. (Please do not write or call to ask for a personalized landfall forecast, as I do not have time to respond to such requests).