Mike says:

Thanks for the update!
Im sure you’ve answered this before, but where do you get your ECMWF and UKMET imagery, I cant seem to find it anywhere on the internet

Chris says:

Levi-

Great blog, and I really appreciate your scientific background and forecasts.

I just wanted to add a note (clarification) of a reference you made in this post to the “cone of uncertainty” posted by the NHC in their advisory. You stated that the diameter of the cone (which in your slide was the 5-day cone) was illustrative of the uncertainty in the model track forecast. That statement implies that the cone was particularly wide at any given time-step relative to, say, Humberto’s cone, because the models had a wide geographic scatter for track at each time-step. This is not entirely accurate.

It should be noted that for each time-step the cone diameter is the same for every storm in a given season, and reflects long-term accuracy of NHC track forecasts. The diameter of the cone is actually based not on the scatter of the model runs in the case of a given storm, but rather on the *average error* in the predicted track at each time-step during the *previous* 5 years. More specifically, the cone is a series of circles (one per time-step) with a diameter that is 2/3 of the average error (i.e., +/- 2 S.D.) in position at that time-step for forecasts over the previous 5 years. Therefore, for this entire 2013 season, the cone at forecast time-step t is always the same diameter. The NHC posts the following table for this year:

Time step Diameter (nm)
12 33
24 52
36 72
48 92
72 128
96 177
120 229

See the NHC explanation for more info:
http://www.nhc.noaa.gov/aboutcone.shtml