• Bryan Norcross

Reimagining the hurricane forecast cone – and you can help

When people want to know the latest hurricane forecast, for better or worse, they look for the cone. Whether on a weather app or the National Hurricane Center website, people go for the cone when they want to know what’s going on. Nothing else comes close as a first choice for forecast information.

It turns out, the Japanese Meteorological Agency used a cone-like graphic as early as the 1950s, and a brilliant NHC programmer and meteorologist named Charlie Neumann published papers showing error ellipses for hurricane forecasts, which were sort of like a cone, in the late 1970s and 80s. But in those days before the internet, they didn’t get widespread distribution, and forecasts for the public were just points on a map, and occasionally a fan with color-coded probabilities.

In the early 90s, I created a graphic showing the width of a would-be cone based on the National Hurricane Center forecast and a rough estimate of the average errors at 24, 48, and 72 hours in the future. That graphic made its debut on TV for Hurricane Andrew in 1992. At that time, the graphics software we used did not allow the cone to be drawn in the ice-cream-cone shape that we use today, but at least the idea that errors increased with time was clear.

In 1996, TV graphics software finally allowed the cone to be drawn and filled in, much like the cones of today. The fully filled-in cone debuted on Miami TV that summer.

A cone-like graphic was faxed around within the government through this time, but not distributed to the public. That changed in 2002 with the first official public cone from the National Hurricane Center for Tropical Storm Arthur in July.

The following year, NHC forecasts were extend to 5 days, and the new cones reflected that change. The first 5-day cone was issued for Tropical Depression Two, which never amounted to anything, in June of 2003.

Since then, the only big change in the cone has been its width at each forecast period. Since the width is based on the average errors in National Hurricane Center forecasts in past years, and since NHC errors decrease almost every year, the cone has gotten progressively narrower.

Now, the cone is so narrow that most often bad weather occurs outside of cone. “Am I in the cone?” is no longer a good question, not that it ever was a 100% guarantee of safety. Still, when the cones were fat, the most dangerous weather often fell within the cone, which made the messaging fairly simple.

Now that the cone is narrow, more explanation is needed. In an era of cellphones and abbreviated communications, explanation is often in short supply. So the time has come to rethink how the cone is made, and maybe the cone itself.

The National Hurricane Center is leading a large-scale, multi-year effort to imagine how the cone could evolve or what might replace it in the future. Should the same-width cone be used for all storms, or should the width or color vary based on the intrinsic uncertainty in the particular forecast scenario or the forecast strength of the storm? What information should be on the cone graphic? Is a cone the best way to convey the forecast in the first place?

Well, now you can throw in your two cents. A team at Florida State University’s Meteorology Department, in consultation with the National Hurricane Center and others, is asking for the public’s reaction to various ideas on how the cone could change. They also want to know how people interpret the cone as is stands today.

To participate in the FSU survey, click https://fsu.qualtrics.com/jfe/form/SV_232D3MHcaAotE7s

Monkeying with the cone is tricky business. Although the underlying science behind the cone size and the meaning of the details in the graphics are often misunderstood, the simplicity of the current design is a big selling point. Still, if the effort underway can identify modifications that clarify the forecast without confusing people who don’t have in-depth knowledge, all of us who live in the hurricane zone will benefit.

Make your voice heard!