Irma was our second hurricane since moving to Florida, though we ended up not being home last year for Matthew, as we traveled to North Carolina for a wedding. That storm brought my first lesson about hurricane forecasting, that you really can't be sure what to expect until the last few days. A track just 30 miles more west would have resulted in much stronger winds for us.
That uncertainty was still a thing with Irma, but only in the sense that we didn't know for sure how bad it would be. We'd get a fairly close hit regardless, and the scary variables applied more to the coasts. Being inland, we could have the reasonable expectation that winds at worst would be 80 to 90 mph, but because of the uncertainty, they could be low as 40. Power outages and water loss were probable, though against the odds, our power outage was brief and after the storm. People in neighboring areas still don't have power, almost a week later.
It was Friday night that I heard from a friend, who had a relative in one of the NOAA recon planes, that indicated the northern turn would come later than expected, pushing the storm up the gulf coast instead of the Atlantic. Sure enough, the revised track that night put Tampa at risk. That was a significant change only 48 hours out from the worst of it.
One thing that was clear: The National Hurricane Center forecasts every six hours were the critical source of truth. The text of their forecasts were pretty straight forward and offered explanations around why the storm was strengthening or weakening, and importantly, the bearing and speed of the storm. It was free of the nonsense that the local TV stations engaged in. As the storm was tracking NNW by Lakeland, one of the locals (I'm looking at you, WESH) was instilling fear by insisting that "the eye wall is headed right for Orlando!" The NHC was pretty clear about this: The storm center had been moving at about 330 degrees with every hourly update, not directly north, and "eye wall" was a bit of an exaggeration for a storm that had been dragging across land for hours. The Weather Channel was almost as bad, with about 5% information and 95% nonsense like having those assholes standing out in the wind.
Once the first bands started crossing through, there was some risk for tornados, and that's where a good radar app helped. The National Weather Service issues warnings for entire counties, which doesn't make a lot of sense for the giant counties in Florida. While the text of the warnings describes the locations of the action, the app shows the bounding boxes and tracking cones for individual cells. That helped us see that a tornado warning applied to the northeast part of the county, and we could see an extreme wind warning to the south issued in the neighboring county (which fortunately lapsed without a replacement near us).
Precipitation radar isn't the whole story though, and that's another way that TV and the feds differed. Heavy rain is less concerning when you're not in a flood prone area, but wind is important. Having a radar app that does wind as well is helpful.
We ended up topping out with sustained winds in the mid-50's, gusts up to the 70's. Some folks lost some shingles, but trees took the worst of the damage in the neighborhood. Our location is just about as ideal as you could have while still in Florida.