TL;DR: a tropical depression has spun up in the Western Pacific. It will become a typhoon, but at the moment looks to mostly avoid land. Another system will bring some rain to the southeast, more north Georgia than south.
We’re lurching in to the northern hemisphere hurricane season, and have a storm in the West Pacific. Tropical Depression Five (soon to be Tropical Storm “Hagupit”) looks to track east of the Philippines. Model guidance is fairly tight. CAs usual, cick any image to embiggen.
Will keep watching to see if it behaves.
Another system looks to track across the Southeast over the next five days. The forecast rain rates look to surge across North Georgia on Thursday …
This has prompted flood watches across parts of Georgia and Alabama, including Mos Eisley Spaceport the Atlanta area:
This could dump a lot of rain up there, the GFS totals are over 6″ in places, with a lot coming down in a few hours, so things could get messy and flash flooding is likely. Here is a snapshot of the forecast rainfall between 8pm tonight and 2am tomorrow:
As the system progresses towards the coast it looks to lose a lot of energy, and while the pastoral beauty of the Frogmore Metroplex (Coastal Georgia and the SC Lowcountry) will get some much needed rain, totals are likely to be just at or below an inch.
So why are some of the maps above so blocky? One of my (admittedly many) pet peeves is “upscaling” of weather model data to make the picture look prettier – and more accurate/detailed/reliable than they really are. The above graphics from the GFS model are showing the actual model outputs at the resolution of the model (12.5km spectral, 25km effectively for these global outputs). I could smooth them and make them look prettier (you can see the contour lines “look” better), but I think that’s deceptive.
Here is a simulated radar view from the High Resolution Rapid Refresh model of the North Georgia area for 6pm this evening, which runs at a 3km resolution, but only over the US:
Here’s the current radar view, which is from the MRMS data feed at 1km:
Why aren’t GFS or other global models run at a higher resolution? In computer models there is always a trade-off between available data, resolution, computer resources, and noise. If you make the model too detailed, you either don’t have the data to support it, it starts to get “noisy” and loses accuracy, or it takes too long to run. Not detailed enough, it’s not as accurate or misses small scale features like thunderstorms. .
So when you look at my maps, you see what the underlying sources are generating. If it’s blocky, that’s a signal that the model is lower resolution. It doesn’t mean it is less good.
If you appreciate the blog, the best way to get it is to subscribe to the emails. You can subscribe and also donate to keep it going at this link. Many thanks to those who have contributed to the blog!
You can also follow us on X/Twitter (@EnkiResearch), Telegram, BlueSky, or even Facebook if that is your preferred social media muck. But be aware that social media algorithms don’t show you posts in chronological order, “upgrade” sensationalist material, and “downgrade” posts that have outside links in order to keep you on their platform.