Decoding the NHC Forecast Advisory (AL092019 Update)

TL;DR: storm is still disorganized, tracks are trending to keep the storm offshore, and NHC may shift their forecast even further east later today.  Bahamas still needs to prepare, given the pre-existing damage.  The US Coast can just watch right now, but AL09 will likely become Tropical Storm Humberto later today.  Dangerous conditions are unlikely in the US, but worth paying attention in case that changes.  But read the rest; it’s Friday, not like you guys are working anyway 🙂

The National Hurricane Center produces two primary products: the Forecast Advisory, and the Watches and Warnings contained within it and the Public Advisory, that are based on the Forecast Advisory.  The other products are derived from these two primary products.  While the PA and FA are pretty clear, it takes some time and digging to interpret what they mean.  This is why I suggest your first step is the “Key Messages” product, which is a nice summary of the big picture.

The Forecast Advisory is at the heart of what NHC does.  It is a forecast for the storm position, intensity, and wind radii for the storm at fixed times in the future, out to five days, and is based on a careful study of the forecast models blended with forecaster experience.  The first 3 days are considered a forecast, the last 2 only an outlook.  That is because of the uncertainty that, while much better than it was even 10 years ago, is still significant.  As a text product it takes some decoding.  One of my key gripes with TV weathercasters is they normally focus on peak winds, and the maximum radius of tropical storm force winds.  The problem is that peak winds generally exist in a very small area of the storm – the intense damage swath of a hurricane is typically less than 30 miles either side of the track, and any significant impacts at all only 100 miles..  And these distances vary tremendously across a storm.  So you will hear things like “tropical storm force winds extend 200 miles from the center!”  That may be true in the NorthEast quadrant (which is generally, but not always, the strongest part of the storm), but elsewhere these distances can be half that far. The FA includes the wind radii in each quadrant.  Let’s take a look at the 5am FA for AL09 using the AWIPS II software used by the National Weather Service, with the forecast wind radii depicted … click to embiggen.  Note that the final two positions (96 and 120 hour, or 4-5 days from now) position generally doesn’t have wind radii estimated.

The blue arcs show the radius of tropical storm force winds (40mph).  Off of Jacksonville, the yellow arc shows 50 knot (56mph, the wind speed that things really start to break).  But there’s more: these are winds over water!  Over land they are typically 20% less due to friction.  So along the Florida coast, while those blue arcs go inland a bit, it is unlikely any areas other than within a couple miles of the beach will actually see winds that high, if the storm follows that exact track and intensity.  Now lets look at the latest advisory:

Not even any blue arcs on land.  That’s the raw data.  What I do is take that raw data and process it through sophisticated damage models.  Here’s the latest impact estimate from my new Stribog model, using this latest (11am Friday 13 September) National Hurricane Center forecast track and intensity.  What Stribog does is take the raw data from the Forecast Advisory (or any other track/intensity model, but I generally don’t show those in public to avoid diluting the NHC message), and using a coupled wind, wave, and storm surge model, it calculates the conditions along the track on a 982 meter grid.  Notice how asymmetric the damage swath is around the track. I then classify that (well, the computers do) in to easy to understand categories and display it using Google Earth to make a pretty map …

2 thoughts on “Decoding the NHC Forecast Advisory (AL092019 Update)

  1. I’m guessing there’s no model to show what 40 mph winds do to an island that’s been beaten and chewed up and spat out by a virtually stationary Cat 5 hurricane.

    • Sadly there is. I developed a model to assess the vulnerability of populations who are displaced by natural disasters. A lot of that work was based on data from populations displaced by conflict zones as well. At 40mph there is significant damage to temporary housing and shelters, supplies are lost, and so forth. Compounds the tragedy … have had lots of unfortunate opportunities to test in, like in Haiti after the earthquake in 2010.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.