OceanStWx
-
Posts
19,778 -
Joined
-
Last visited
Content Type
Profiles
Blogs
Forums
American Weather
Media Demo
Store
Gallery
Everything posted by OceanStWx
-
Let me just grab a quick flip phone photo of the passage.
-
Well the lore on the rockpile is two weeks after MWN's first snow the valleys will see flakes. And after the 3rd frost/freeze the next snow up top sticks and stays for the season. Both of which happened on the 22nd/23rd.
-
Those ballpark numbers sound right. Gauge has data back to 1965.
-
It's going to be right up there in terms of recent historic surge events. Looks Katrina-esque from the MS coast around the Ft. Myers Beach area.
-
Pretty sure he also said "could be" not "is".
-
Regarding track forecasts, keep in mind too that with northward moving TCs around FL that an east/west miss of 30 miles can result in landfall 100 miles farther away along the north/south oriented coasts. Again, specifically talking landfall location here, not cone forecasts.
-
It's also a difference between a deterministic forecast (watch/warning) vs. a more probabilistic forecast (past track forecast error/cone).
-
I think overall it was a pretty good forecast, unfortunately most people only focus on the center of the cone and not the edges. I just wish cones could be forward looking, instead of backward looking. If we had a good way to measure or index forecast uncertainty they could probably fairly easily adjusted cone radius to capture 75% or 90% of past errors rather than 67% for example.
-
Overall I have a ton of respect for the NHC crew, my one nitpick revolves around the construction of the cone. It's only designed to capture 2/3 of the average track error over the last 5 years. As forecasts have improved the cone gets smaller. What it doesn't do is factor in situations where there is larger than normal forecast uncertainty. I would love to see a little more of that uncertainty factored into forecasts.
-
Already 1.5 feet above forecast at Charleston.
-
Could be a bit of slower burn like Katrina too. Lack of services may lead to fatalities in the coming days, like how even after the worst had passed the levee failures did the bulk of the damage in New Orleans.
-
The worst of it is likely just beginning for Fort Myers proper and Cape Coral. Winds are finally starting to push water up the river (along with the pressure related surge), and as winds come around to more westerly it should really start to shove water into those canals.
-
What's interesting is that the CDO actually looks like it's improved and become more symmetrical the last couple of hours.
-
It's all fun and games until the machines rise up and start rounding up weenies for detention centers.
-
Yes and no. A human trains the machine learning system for specific forecasts (e.g. severe or flooding). And while the machine learning system uses model input to go through the tree, it also randomizes which variables are being used to get a diverse dataset. And you can continue to train the system over time, so while the model itself may not be explicitly forecasting severe, the system may know that sometimes severe still results.
-
I still need to really read into the under the hood stuff, but theoretically it is machine learning so it should know about all types of set ups.
-
3% chance of seeing a scary looking cloud near EEN.
-
I loved the slap down of Mass when he used 12z 850 mb sounding data instead of 00z to say that we were getting all-time record highs from non-record 850 temps.
-
Mine are actually root sprouts, but I'm realizing that the peach tree was grafted already (Redhaven on Lovell) so I don't want to graft the rootstock.
-
How hard is it to graft? I have a couple suckers growing up around my peach tree and figured rather than pruning them I could try and save them.
-
Kids Christmas presents tossed into Mass Bay.
-
-
Seasons in seasons.
-
Looks like that convection entering SW CT is slowing down. It's still moving for now, but could stall for a bit between DXR and BDR.
