How accurate was the forecast for Nine?

Tropical Depression Nine · 4 advisories archived · 2 forecasts resolved

We record every National Hurricane Center advisory as it is issued. Once the storm has moved past a forecast’s valid time, that forecast can be checked against where Nine actually was. This page is that comparison - 4 advisories archived, 2 forecasts resolved so far.

Every forecast, against what happened

The white line is where Nine actually went. Each amber line is one forecast NHC issued, newest brightest - so the spread between them is how much the outlook changed as the storm evolved.

Where Nine actually went4 forecasts, newest brightest

Intensity: predicted against actual

Intensity is the hard part. Track forecasting has improved enormously over thirty years; predicting how strong a storm will get - especially one that intensifies rapidly - has improved far less.

2 forecasts checked · 2 over-forecast it. A line sloping down to the right means the storm was weaker than predicted; up means stronger.

Track error by lead time

How far the forecast centre was from where the centre actually turned out to be, in nautical miles.

Each dot is one forecast. Below the dashed line is better than the National Hurricane Center’s published 48-hour average.

Every forecast checked

Advisory	Lead	Track error	Forecast	Actual
001	+9h	25.2 nm	40 mph	40 mph+0 mph
002	+9h	19.5 nm	45 mph	45 mph+0 mph

How this is worked out, and what it is not

Each advisory carries the storm’s observed position at the moment it was issued. Strung together, those observations are the storm’s actual path. A forecast made twelve hours earlier for this moment can then be compared against it. Where a forecast is valid between two observations, the position is interpolated linearly; forecasts valid outside the observed window are left unscored rather than extrapolated.

This is not NHC’s official verification. They score against post-season best-track data, reanalysed with aircraft and satellite information that is not available while the storm is running. Our numbers use the operational positions as issued, which is a slightly different - and slightly harsher - question: how well did the forecast match what was known at the time?

With 2 forecasts resolved, these figures describe Nine only. They are far too few to say anything about NHC’s skill in general, and we do not present them as though they were.

Tropical Depression Nine is one storm. The point of keeping this archive is that after a season of them, the question stops being anecdote.