Guide to Plots

Guide to Plots

The TCGP site features several types of plots, each of which feature a collection of forecast aids. This guide provides information on how to interpret and use each type of plot. The current page gives an introduction to the concept of a 'TECH' identifier, explains the differences between the early and late model cycles, and discusses the role of vortex trackers. Additional guide pages provide a list of all the forecast aids that can appear in plots of that type.

The guide pages can be accessed using the following links:

What is a 'TECH' identifier?

With the proliferation of numerical models and statistical prediction techniques, there are now scores of forecast aids. Historically, each forecast aid has been given a 4-letter 'TECH' identifier so that forecasters can easily recognize which aids are which. In some cases, projections from a single regional NWP model (such as the GFDL Hurricane Model) may be represented by multiple TECHs. In some cases, these TECHs may represent an alternative configuration of the model, while in other cases, they represent an older forecast which has been interpolated to the current time for comparison with the most recent forecast. In some cases, different agencies in the various basins may use different TECHs for the same model configuration, so there may be no single "key" to matching a specific TECH to a reference model configuration. The TCGP has chosen to use the TECH naming scheme used by RSMC-Miami (National Hurricane Center) for all forecast aids produced for the area overseen by RSMC-Miami and RSMC-Honolulu. For the other global basins, the same TECH identifiers will only be used if the model configuration and tracker are the essentially the same as the model configuration and tracker used in the North Atlantic/Northeast Pacific/Central Pacific basins. TCGP will assign a new, unique TECH to any forecast aid that uses a different model configuration or prediction method in these other basins.

The TECHS which can appear in each plot type are described in more detail in the guide to each plot type. For sake of comparison to NHC's TECHs, the following links are provided:

TCGP now has a global tech that combines the above information and links to additional model documentation when available.

What is the difference between "early" and "late" forecast aids?

Forecast aids can be categorized as either "early" or "late" depending on when they become available relative in the forecast cycle. At NHC, the forecast aids are initialized at the synoptic times of 00, 06, 12, and 18 UTC. The official forecast is due 3 hours after the synoptic time (so at 03, 09, 15, or 21 UTC). 

The simple forecast aids take just minutes to run and are therefore considered as "early" aids since they are available almost immediately and can be used by the forecaster in the current cycle. Early aids include the beta and advection models (BAMS, BAMM, BAMD), the Statistical Hurricane Intensity Prediction Scheme (SHIPS) forecast aids (DSHP and LGEM), the simple barotropic models (e.g. LBAR), and the statistical climatology and persistence aid (e.g. CLP5).

The more complex forecast aids, on the other hand, normally take several hours to run (even on supercomputers!). Because the projections from these forecast aids generally become available several hours into the forecast cycle (and very often after the forecast has gone out), they are termed "late" since they can't be used in the current forecast cycle. To account for this chronic lateness, the output from a 6-hr old late forecast aid is adjusted 6-hr into the "future". As an example, the projected track and intensity from the GFDL model for hours 6-126 is adjusted 6-hr ahead so that it would match the times of an early cycle aid running from 0 to 120 hrs. This allows the forecaster to readily compare its output to the early forecast aids. For historical reasons, the adjustment is often called interpolation (the method of adjustment may differ from a basic linear interpolation however). Thus, the early cycle guidance plots also contain the "interpolated" late models. Any late forecast aids which has been interpolated 6-hr ahead is given a TECH identifier that ends in 'I'. The late models include the global models (e.g. GFS and the UKM) and the high resolution regional models (e.g. HWRF and GFDL). Note: it is also possible to interpolate a late model from 2 forecast cycles ago (12 hrs old). Such models are given a TECH identifier ending in '2'. 

What is a vortex tracker?

Because the dynamical models simulate the future atmosphere in three dimensions over a very large domain, it is not always an easy task to determine where the model is projecting the storm to go or what the model is projecting the intensity to be. To obtain the locations of the model's projected track, as well as the intensity of its vortex, it is necessary to run a vortex tracker on the model's gridded output. There are a variety of methods to track a vortex in a model, but unfortunately there is no perfect method. Each vortex tracker has its pros and cons. Difficulties can arise when the model's simulated storm is very weak, when its vortex is tilted in the vertical (as might happen under strong environmental vertical wind shear), or when it is interacting with another vortex. Many of the U.S. regional models use a standard vortex tracker that was designed by Dr. Timothy Marchok. Many of the other global or regional models may use their own vortex tracker (e.g. ECMWF, UK Met Office). In some cases, the same model may utilize two different vortex trackers. Normally, the tracks from two different trackers will be quite close together, but in the difficult situations described above (e.g. for a sheared storm), they can differ markedly. Once in a while, a tracker may experience an error in which it jumps to another vortex some distance away.   

Additional resources

For much more information about tropical cyclone models and forecast aids, please see the Technical Summary of the National Hurricane Center Track and Intensity Models.

For technical details about the various regional and global dynamical models, please see MetEd, 2007: Operational Models Matrix, University Corporation for Atmospheric Research.

For technical details about the global ensemble systems, please see MetEd, 2007: Operational Models Matrix, University Corporation for Atmospheric Research.