Four on Francine

A Friday Hodgepodge

Once again, dealing with a hurricane taught me a few things...

1. Hurricane Francine was the first storm we've had to deal with since moving to New Orleans. It was also as close to a surprise as a storm has ever been to me, and I plan to keep it that way.

There had been hints that something could develop in the southwestern Gulf of Mexico the week before, but I didn't appreciate how close that really is. The storm had barely organized Monday, but it was clear that we might need to prepare and leave before Wednesday.

We did alright considering, but I will never again fail to top off our gas on any weekend with something brewing down there, and weather patterns that will send it in our general direction.

Gas stations were closed or running out that Monday and we were lucky to be able to fill up as quickly as we still did.

2. One thing that will help me know to pay attention will be the weather. A cold front had caused the weather to be noticeably cooler in the morning the week before, and part of the southern trailing edge of that front lingered off Mexico and contributed to the development of this storm from a tropical wave that drifted into the same area.

I'm no meteorologist, but... I've seen storms develop from trailing edges of cold fronts over warm water enough times over the years that I'll take cool weather at this time of the year as a cue to pay more attention.

Back on the Atlantic coast of Florida, I saw this occur to our northeast, and it wasn't generally a concern. Ends of cold fronts now land in hot water southwest of us, so it's a signal to start paying closer attention.

3. The National Hurricane Center is experimenting with a new format for its famed cone graphic, which extends watch/warning shading inland from the coast.

Overall, I like the idea, but there might be some work yet to do:

The Experimental Cone (Image by the National Hurricane Center, public domain.)

Lastly, a word about continuity. We try to refrain from making frequent changes to the watches or warnings in effect for a specific area, and that includes not discontinuing watches and warnings soon after they were issued, or changing between tropical and non-tropical watches and warnings in the middle of an event. This can mean that even when there are shifts or changes to the forecast, watches and warnings for an area may remain in effect for a while longer until we're absolutely certain the risk of strong winds has diminished. Why do we do this? Imagine if each time you go to your doctor, you're given a different diagnosis for a health problem, or prescribed a different medication to treat it. You might start to lose trust in your doctor, or at the very least not know what to believe. A consistent message from your doctor, with a gradual evolution of your treatment over time, is likely to instill more faith in his or her expertise. Our philosophy is the same. We tend to make incremental and gradual changes to our forecasts and the watches and warnings in effect, which in the long-run makes them more trustworthy. [bold added]

Because Francine formed somewhat chaotically, forecasters had to adjust where they thought the center of the storm was, meaning that there was a big-enough shift in the predicted track of the storm that I went from mildly to very concerned overnight. The experimental cone pictured above confused me at first, because it looks a little like there is still a more westerly prediction of landfall due to the old inland warnings being left in place.

I an now inclined to chalk that up to inexperience with the format, but I wonder if that was a common problem.

4. The Windy app/website has added a new model that will help hurricane trackers make better decisions. I noticed ICON, but didn't know what it was until I read the storm postmortem at The Eyewall:

AI modeling & ICON scores a big win

I went through and assessed the 15 model runs leading up to landfall from various models with Francine. Keep in mind that the landfall point is not the only variable that matters, but it's an important one. And one thing you cannot tolerate as a forecaster is a lot of whipping around within the models...

While the ICON kept western Louisiana in play for a while, it too corrected east in time. This, combined with its performance during Beryl (as well as last week with Invest 90L) made it another valuable tool. Will these models fail in the future? Yes, all models do. And the ICON remains prone to spurious tropical systems that never end up happening. But at this point, I think there's a lesson to be taken from this as a meteorologist. particularly when the models show this sort of stability. The GFS and Euro operational model (and the ensemble guidance) also performed respectably, but those models showed less consistency run to run than did the ICON and AIFS.

The Windy site explains what this high-performing model is.

-- CAV

P.S.: I would like to thank again the various readers who have told me about go-to storm tracking sites over the years, including at least Windy (brief review), the Tropical Tidbits blog (brief review), and The Eyewall, another blog.