One thing I think this event (and many others over the last few years) highlights is that the overwhelming focus on the cone by political officials, emergency managers, and the media needs to stop. In many ways I think the NHC wishes they had not started the cone graphic nearly 2 decades ago, because it is a fairly poor messaging device (since it only focuses on the center track and has a static width based on historical error rather than a dynamic width based on the uncertainty surrounding a particular storm) and has become so popular they can't simply discontinue it. To highlight this fact, portions of Collier county which received significant storm surge were not in the cone for most of the event. A lot of people would chalk this up to track error... but even with a perfect forecast some of these areas (particularly like Marco Island/Everglades city) STILL wouldn't have been in the cone for at least the last few days of the event really highlighting the point that impacts can extend well away from the cone.
This problem is only going to get worse because NHC track forecasts keep improving, this means that in future years the cone width will continue to shrink and consequently it becomes statistically more likely to see significant impacts outside the cone. Now the NWS is aware of the cone's limitations for hazard messaging and has been producing probabilistically based Hurricane Threats and Impacts (HTI) graphics for a few years now which actually give emergency managers/public officials and idea of the "reasonable worst case scenario" (which is what you should plan for). I wish these would get more traction than the cone... or to be honest even the traditional watches and warnings.
Here is an example from early Tuesday morning when the center of the cone was still focused towards Tampa... the surge HTI graphic actually depicts the surge threat quite well, yet only looking at the cone from that advisory you wouldn't be too worried in Collier.
