Reality check: will federal budget cuts REALLY mean we can’t forecast hurricanes?

Photo above - believe it or not, this is how hurricane severity is predicted - bunch of guys staring at PC monitors. Could AI possibly offer an edge?

If I were asked, I’d vote for more air traffic controllers, instead of weathermen. There have been a number of well publicized airport glitches and outages recently. And apparently a half dozen more that barely made a ripple in the media. Experts are still trying to sort out if this is hacking, bad software, or rookies in the control tower pressing the wrong buttons. We await their findings.

Let’s turn our attention to weathermen, instead. The link below (“States and cities fear a disaster season”) is hilariously misleading. The article features plenty of rants from federal agency big shots. They warn that there could be a disaster if their budgets are reduced, even a smidgeon. And a bunch of ordinary homeowners who got 2022 FEMA checks were interviewed. Which proves exactly nothing when it comes to forecasting 2025 hurricanes. FEMA is an agency that helps you rebuild in that same dangerous location right after you got flooded or flattened.

I yield to no one in my admiration for rainbow colored satellite images pinwheel shaped hurricanes, or iPhone videos of alligators swimming alongside Ford 150 Platinum Lariat pickups on flooded Florida cul-de-sacs. But I’m skeptical that culling a few steers from the herd of 12,000+ NOAA employees is going to mean that Hurricane Wendy will sneak up on us. (Wendy is one of the official names for 2025 hurricanes, when they reach the starting grid).

But there’s hope on the horizon. Heads of NOAA, FEMA, the National Weather Service, and NEMA (the National Emergency Services Association) should probably take a look at the second link below: “AI weather tool surpasses current forecasts”. That is going to send a chill down the spine of a LOT of GS 12 weather forecasters, no?

AI software programs excel because they use “machine learning”. This is the same approach being used to diagnose and cure cancer and otherl diseases. The “Aurora” program, had it been listened to, accurately predicted the path of typhoons like Doksuri. Aurora and it's brethren are apparently FAR more accurate than a platoon of weathermen with plastic pocket protectors puzzling over satellite snapshots.

Every spring the folks at the National Weather Service tell us the same thing: Expect a more active hurricane season this year. Because of climate change, El Nino, La Nina, sunspots, Canada’s jet stream stepped out of bounds or whatever. No agency will ever get a bigger budget and more headcount by predicting FEWER hurricanes and tornados. But in fact, we’ve had a multi-year lull in hurricane activity for much of the past several decades. 2024 was forecast to be one of the “worst ever” by NOAA. Instead, it was below average. The same “worst in memory” predictions are being recycled for the 2025 season, as a defense against budget cuts. See CBS link at bottom.

I don’t expect that AI weather forecasts like Aurora will be perfect. But I do think they will be better than whatever NOAA has been using to serve up the same hurricane predictions year after year.

And my apologies in advance to homeowners who rebuilt in storm surge areas after previous hurricanes. If your place is destroyed again, don’t blame NOAA, FEMA, NEMA, or the National Weather Service budget cuts. It was your own decision to rebuild in a high-risk zone, after getting federal disaster relief funds. But you're not alone. California homeowners keep rebuilding in the fire zones, too.

I’m just sayin’ . . .

States and Cities Fear a Disaster Season Full of Unknowns Amid Federal Cuts – DNyuz

A new AI-based weather tool surpasses current forecasts

Why South Florida can't let its guard down for the 2025 hurricane season, no matter the forecast - CBS Miami

ElwoodBlues · M

@SusanInFlorida says

believe it or not, this is how hurricane severity is predicted - bunch of guys staring at PC monitors.

DEAD WRONG!!!
Hurricane path and severity are predicted by supercomputers running numerical models of weather. See below.

@SusanInFlorida says

here's a crazy idea: instead of the USA owning all the weather satellites, and having most of the worlds weather forecasters at NOAA and NWS headquarters, suppose we make this a global cooperation thing?

Actually, good weather prediction also depends on thousands of ground stations all across the US.

The NWS also operates 158 WSR-88D Doppler radars across the country. Additionally, the NWS utilizes a network of 10,600 Cooperative Weather Stations.

If your place is destroyed again, don’t blame NOAA, FEMA, NEMA, or the National Weather Service budget cuts.

True, but late warnings also cost LIVES. Five minutes can make a YUGE difference in fatalities.

“AI weather tool surpasses current forecasts”. That is going to send a chill down the spine of a LOT of GS 12 weather forecasters, no?

Nope. The big iron in weather forecasting is supercomputer numerical simulations of weather (fed with as much local data as possible). These were last upgraded in 2023.

These supercomputers, Dogwood and Cactus, are part of the Weather and Climate Operational Supercomputing System (WCOSS). They are located in Virginia and Arizona and have been upgraded to 14.5 petaflops each. In addition to WCOSS, NOAA has research and development supercomputers in other states.

The simulations are run multiple times with slightly different assumptions, producing bundles of storm tracks known as ensemble forecasts. Its those supercomputer outputs that are interpreted by both AI and human weather predictors.

Currently our highest resolution weather models operate on a 55 layer 3km grid for local hourly predictions. Global models that cover oceans use a 28km grid (I don't know how many layers).

But I’m skeptical that culling a few steers from the herd of 12,000+ NOAA employees is going to mean that Hurricane Wendy will sneak up on us.

You can't get good local predictions without good local measurements. You can't get good measurements without people to maintain weather stations, communications networks, and supercomputers. Most weather service personnel are keeping the hardware running, NOT making predictions. The NWS is a data farm and factory. They continually harvest, refine, and reduce data; converting it into ensemble forecasts. Until you analyze the whole farm and factory, you don't know what's essential.

But in fact, we’ve had a multi-year lull in hurricane activity for much of the past several decades.

DEAD WRONG!!!
Do you have any data to support that claim? Because all the data I can find tells the OPPOSITE story!

@Jokersswild @Johnson212