During hurricane season, you frequently see what are known as the "spaghetti plots" of tropical models. They show the predicted path of tropical systems run by a variety of different computer model schemes. What you may not know is what each of those models is and what biases they may have. Here's a brief overview of what you're looking at when Tom posts the hurricane model plots from the South Florida Water Management District.
XTRP: Extrapolation. Not a model. It takes the current storm motion and extrapolates it out through the time period, meaning that if the storm never changed speed or direction, that's where it would end up. Not very useful, and frankly I personally wish this would not be published on these charts (and others), but it is what it is.
TVCN: Tropical Variable Consensus Model. It's an average of anywhere from two to seven different other tropical models. According to former National Hurricane Center director Max Mayfield, this model performed best back in 2008. As it is a consensus model, it takes some of the stronger points of a few different other models and hopefully smooths out some of the biases.
NHC: This is just the NHC's official forecast.
|South Florida Water Management District Tropical|
BAMD/BAMM/BAMS: Beta and Advection Model (BAM). Each one of these is a different version of the BAM model. D for Deep, M for Medium, and S for Shallow. The BAM model takes the upper level winds from the GFS, takes their vertical average, and then corrects the trajectory to account for the Coriolis Force's variation with latitude. So it biases storms further north and west because of this. Each version is for an average of winds at different atmospheric levels.
The BAMS averages winds from 850 mb to 700 mb and is probably most useful in weaker systems.
The BAMM averages winds from 850 mb to 400 mb (or just under 25,000 feet).
The BAMD averages winds from 850 mb to 200 mb (or between 45 and 50,000 feet).
What you can take home from the three BAM models is that if they're clustered together, confidence in that storm's forecast is likely better than normal. If the three groups of BAM models is spread out, there are a number of other things that may be at play.
GFDL: Geophysical Fluid Dynamics Laboratory Model. The GFDL is based in Princeton and runs one of the more skilled hurricane models. This model was developed specifically to predict tropical systems and is run four times per day, and it is nested within the GFS model. The GFDL is one of the more respected hurricane models, but it too has occasional issues.
UKM: This is the UKMET model, which is a global model run by the United Kingdom's Meteorological Office. This model is only run twice per day. This model, like most global models, works well for track, but not so skillful with intensity.
NGPS: NOGAPS Model: Naval Operational Global Atmospheric Prediction System. Another global model, similar to the GFS or UKM model. This one occasionally outperforms some of the other main ones...but it's very streaky...running a hot hand and quickly going cold. It too is a good model to use for track, but like all global models, struggles with intensity.
AVNO: Operational GFS model. The GFS used to be known as the AVN (Aviation) model, hence the acronym. The GFS is obviously one of the main global models...despite its frequent flaws.
AEMN: Aviation Ensemble Mean Model. This is just the GFS Ensemble mean track...taking an average of each individual GFS ensemble member.
HWRF: Hurricane Weather Research and Forecasting Model. Probably the most advanced of all the hurricane models. This model is specifically developed for hurricanes. It has a very sophisticated initialization (so the base starting point of the model is strong), and it has a physics scheme and high resolution built to best model tropical cyclone conditions. This model has been rather disappointing in my own opinion thus far. But research and development on improving it continue and the model should improve considerably with each hurricane season.
CMC: Like the UKM and AVNO, this is a global model, but this one is run from Environment Canada (the Canadian equivalent of NOAA/NWS). Same caveats apply.
APxx: These gray lines indicate the various GFS ensemble members (averaged in the AEMN above). They are all plotted however to give you an idea of the "spread" within the model itself.
CLP5: Clipper model (Climatology Persistence Model). This model is general used to gauge how the others are doing, moreso than being of use for actually forecasting the storm. In theory, you should ignore this model when using these plots. In general, it just uses climatological history and the current trajectory to predict the future path.
There are many other hurricane models used at various websites you may frequent. To get more details on those and a few more details on what you read above, here's a list of some resources: