Long-Term Visualization of Hurricane Forecasts 3


When a storm is active in the Atlantic, the latest and greatest hurricane forecasts from the National Hurricane Center get almost all of the attention, whether from media, enthusiasts or other meteorologists.

Forecast swaths for all official forecasts made for Sandy.  Darker tracks: most recent, lighter tracks: oldest

Forecast swaths for all official forecasts made for Sandy. Darker tracks: most recent, lighter tracks: oldest

However, the most “current” forecast doesn’t always tell the full story. The Hurricane Center doesn’t like to make big forecast track shifts from advisory to advisory when large changes in model guidance occur. For example, if the models shift from a recurving system to one that moves 20 degrees further west in a 6 hour forecast cycle, the NHC will gradually make track adjustments to reflect the trend over several forecasts. This is to ensure forecast tracks don’t “windshield-wiper” between forecasts.

Remember, their goal is simple…minimize track and intensity forecast errors. They have learned “flip-flopping” causes higher forecast errors, and there is value in persistence.

The above graphic is a longer-term view of the 5 day NHC forecast tracks issued for Sandy, with lighter colors representing the early forecasts, gradually becoming darker over time (latest forecast is the darkest shade of blue).

This analysis can reveal the longer term forecast thinking, either in real-time (how are trends changing) or in retrospect. In this example, notice how the early forecasts (lighter paths) had an easterly bias, reflecting the early model trends of recurving Sandy in most of the models.

However, notice how tightly clustered these forecasts were when Sandy bent back toward the US east coast. That tight grouping showed the confidence they had over many forecast cycles. By that time, Sandy was very well-behaved, and the model guidance was consistently showing an east coast landfall.

I am considering making these type of graphics/products available during the 2013 Atlantic hurricane season, updated after each advisory is issued.

Please feel free to pass along any comments or questions about this graphic.


About Michael Watkins

Mike Watkins is the founder of Hurricane Analytics, a private organization specializing in data visualization and predictive analytics, with a special focus on tropical meteorology. They analyze complex meteorological data and communicate that information in easy-to-understand terms, to help clients prepare and anticipate the disruptive impact of Atlantic hurricanes.


3 thoughts on “Long-Term Visualization of Hurricane Forecasts

  • David Van Bergen

    Interesting analysis Mike. I’m assuming you think that these might give a clue as to the confidence NHC has in their forecast? It could be interesting to watch and see how they cluster. I would find it especially interesting if there was a major change during the forecasting cycle such as Ike, originally forecasted to go up the East coast and I know for sure where he ended up.

    • Michael Watkins Post author

      Exactly, Dave! It will be interesting to load up the 2012 Isaac data for exactly that reason: I know there are a few forecasts that show Isaac heading toward eastern Texas, which will probably stick out like a sore thumb in this kind of view. Thanks for reading along and for the valuable feedback, it’s always great to hear from you. See you over at Storm2K soon.