Getting the National Weather Service Back on Track
Today's Realities and the Future of the National Weather Service
On April 2, an ominous, powerful tornado churned through the Arkansas countryside toward the town of Lake City. As it made its way toward the city, storm chasers posted photos and increasingly dire on-the-scene reports. Per ABC News’ chief meteorologist Ginger Zee, the Lake City tornado siren sounded and local TV was warning the residents of the town to take shelter. According to Ginger and Dr. Craig CeeCee, an expert in public sheltering, at least one of the town’s shelters was filled to overflowing. At least one private sector weather company, AccuWeather’s business group, issued an independent tornado warning for the city.
The strong EF-3 tornado came and went, destroying or damaging more than 50 buildings and temporarily making a mockery of the town’s motto: “Peace, Quiet, Home.” But, there was great news: no deaths and no serious injuries.
The only entity Lake City did not officially hear from during the run-up to the storm’s arrival? The National Weather Service (NWS).
Lake City was not put under an official tornado warning until the storm was on top of the city. The meteorological information is provided after the "St. Louis" footnote, below.
Yet, in spite of the lack of an official, timely warning, residents were well-aware the storm was coming; evidence being the siren, local news, the people in shelters and lack of deaths. In recent decades, television and private sector meteorologists "backstop" the NWS. This is yet another illustration that the private sector is more than capable of handling extreme weather. The significance of this will be discussed below.
Now that the Supreme Court has reconfirmed (at least for the present) President Trump has authority to fire federal workers (note: I am not commenting on the wisdom of that decision as it regards the NWS), it is time to have a full and candid discussion of the reality of weather science in 2025 and the future of the federal weather agency going forward.
In March, I made these points:
What Were the March 5 Suggestions?
- NOAA has lots of fat. Repurpose that money to NWS and give any left over to the Treasury.
- NWS is a vital national resource but it is increasingly having issues providing accurate and timely forecasts and warnings of extreme weather.
- Reverse the layoffs. Get the NWS's (far too many) pre-DOGE open positions filled. Find a better way to hire meteorologists. Taking 9 months to fill a position is absurd.
- Establish a "back to the future" school (as the NWS had from ~1991 to 2004) in extreme weather forecasting and warning techniques. It should be of 4 to 6 weeks duration. Every NWS meteorologist that didn’t attend the school referenced above must attend. Addition: The school should include non-model techniques for extreme weather forecasting.
- Restore all closed upper-air observing stations and, if there are funds, establish limited (2, preferably) additional automatic balloon stations on abandoned Gulf oil platforms and in mainland areas where there are large gaps. These are critical in an era of mesoscale modeling. This has real potential to improve forecast and warning quality.
- Fund gap-filler inexpensive C-band radars, especially since the replacements for the S-band WSR-88D's are not forecast to be available before 2040, if then. The latter are increasingly unreliable.
- Addition: NOAA has been a weight on the NWS. Congress should make the NWS plus NESDIS, NOAA Labs (with some given a new mission statement) and other federal weather and drought programs into an independent agency like NASA.
- To insure the above investment provides returns to the taxpayers who fund it and to provide forecast and warning accountability in an era of increasing mega-disasters, establish a U.S. Disaster Review Board modeled after the hugely successful National Transportation Safety Board. It will study all types of natural disasters and make recommendations for improving our forecast and response abilities.
In today’s National Weather Service’s focus is often not on serving the public through great forecasts and warnings: it is "IDSS" -- off-site and in-person forecasts and briefings to whomever wants them to the extent local resources will allow. This includes everything from the Super Bowl to local school districts.
"The Making of an IDSS Agency"
![]() |
NWS meteorologist providing IDSS services off-site from her office. NOAA photo. |
In 2017, the NWS changed its mission statement to focus on “impact-based decision support services.” The headline below is from its own public information statement:
IDSS, for example, is when four NWS meteorologists leave their office to provide on-site support services for (true example) a Texas NASCAR race -- during a tornado watch, no less. That is nothing but corporate welfare. Obviously, those four NWS employees were not available to serve the public at a critical time.
While correlation does not prove causation, the increase in NWS's public service problems, like the drop in NWS tornado warning accuracy, more or less parallels the rise of IDSS [For the record, I also believe the drop also correlates to the retirement of experienced meteorologists who were graduates of the radar/warning school]. For example, a situation in St. Louis* supports the "more IDSS, the worse public warning service becomes" hypothesis.
While the National Weather Service -- unfortunately -- shut down its school for radar interpretation and severe storm warnings in 2004, they are now in the process of starting a new school to teach their meteorologists how to better perform IDSS services (which, often, are nothing but corporate welfare) away from their offices during high-impact weather events. This takes even more (scarce) resources away from serving the public-at-large.
As members of the taxpaying public, which do you prefer: the NWS launching weather balloons to improve the accuracy of its forecasts and warnings for everyone or using its scarce resources on corporate welfare? Why not let the very capable private sector handle specialized users' needs?
While there are others considering making major changes in the National Weather Service, I believe the presence of the 100+ local offices, fully staffed and focused on serving the public-at-large, is vital to providing the best public service possible. However, the agreement the NWS has with its employees' union, the National Weather Service Employees Organization (NWSEO), which forbids the employee's record of forecast/warning accuracy and customer service from being considered when an employee is being considered for a promotion needs to be eliminated. It is contrary to the public interest.
![]() |
In many cases the problem was not with the sirens themselves but the lack of timely NWS tornado warnings. |
Overwhelmingly, the meteorologists of the National Weather Service wish to do the right thing.
My representative, Sharice Davids, has introduced the WARN Act to learn how to better warn people of tornadoes and other dangerous storms. I endorse this bipartisan proposed legislation. This is an opportunity for DOGE, the Trump Administration, and Congress to repair these problems and move the U.S. National Weather Service back into world leadership. Nothing less will provide optimum service to the public.
*In 2007, an unforecasted wind storm wreaked havoc on the St. Louis Bi-State area. The local electric utility experienced the worst weather-related outage in its history. Because it was not forecast, the local NWS was caught short-handed. Yet, instead of focusing on catching up to the weather conditions surrounding the storm and adequately warning the public, their office spent precious time calling and briefing the St. Louis Cardinals. Video of conditions at Busch Stadium is here. The irony is that the Cardinals were one of our clients at the time, and we (AccuWeather Enterprise Solutions) had already provided an 80 mph wind warning and were in constant contact. This also reflects the issue with NWS culture that causes many of their people to believe "only we can handle extreme weather" and "we must be all things to all people." That culture is one of the things that has led the NWS astray.
The Meteorological Data Regarding the Lake City (non-)Warning
The data showed a powerful tornado moving toward Lake City for quite some time. This is 6:40pm radar data showing the tornado was already on top of Lake City the first time that evening the NWS officially warned the city.
Below is the tornado warning that was issued at 6:34pm. Note that Lake City was not mentioned.
Your comments regarding IDSS are spot on. If people knew how little time and effort is spent on improving forecasts and warnings for the general public they would be appalled. From my experience, the overall majority of NWS employees want to do the right thing and have a good work ethic, however, there is no incentive for improving warning skills or punishment for poor performers. A pervasive mentality of "better safe than sorry", often communicated by so called "partners", is used as an excuse for the high false alarm rates of many meteorologists.
ReplyDeleteAfter focusing on the warning of this episode a little more closely (it is not a “non-warning”), we can see that the original tornado warning for southeast Craighead County issued at 600 pm CDT (TO.49) did indeed extend into the Lake City city limits.
ReplyDeletehttps://mesonet.agron.iastate.edu/vtec/event/2025-O-NEW-KMEG-TO-W-0049/radar/USCOMP-N0Q-202504022335/tab/themap
According to the WFO Memphis storm survey the tornado path extended through the western part of the city near the intersection of AR18 and AR135, where the heaviest damage occurred (high end EF3). It appears from radar data that the circulation core travelled through the western part of the city roughly between 640-642 pm while still in the original polygon. That part of Lake City was under a tornado warning for over 40 minutes prior to arrival and was under a “PDS Warning” for roughly 12 minutes. Downtown Lake City (the dot) was just outside the polygon (about ½ - 1 mile). A new Tornado Warning (TO.55) was issued at 637 pm to extend the warning northeastward along the tornado path. WFO MEG should check that their Warngen Cities list contains Lake City since it also did not appear in TO.55. An easy fix. I also agree that it would have been better to extend TO.49 further northeast a little earlier than 637 pm with TO.55. The tornado was halfway between Bay Village and Lake City when this occurred, and had ramped up in intensity near Bay Village. And, yes MEG issued 36 warnings/follow-ups between 6pm and 7 pm within its CWA !
All in all though, from what I could tell, I thought the weather enterprise performed well. Not perfect, but effective. The sirens were sounded in advance (possibly more than once), the TV Mets appeared to have communicated the threat in an excellent manner, Law Enforcement/EM was aware of the warnings and praised them in interviews, spotters relayed detailed information that was invaluable. Including NWS efforts - this seems more like a team effort than a “backstop”. I think it should be depicted in that way. Advance warning of a near EF4 with no fatals should be praised.
Your comment that “The only entity Lake City did not officially hear from during the run-up to the storm’s arrival? The National Weather Service (NWS)” - is not correct, to put it diplomatically. It seems more designed to support your agenda(s) than fairly describe the event. Hopefully, this type of analysis isn’t in the plan for a NDRB. Constructive analysis and review are welcome - this type approach is not (there is a name for it). Sometimes, I think you are hurting your case more than helping.
Readers: there is a lot of "inside baseball" here but I thought it was important to post this so you could decide for yourselves. MEG = National Weather Service office in Memphis. TO's = are part of a coding system to keep warnings and other messages straight.
DeleteThere are fundamental flaws in Anonymous' (I wish he/she had added their name) reasoning and they are: the 6:37pm warning superseded the earlier warning. Period. The earlier warning is non-effective once a newer warning replaces it. If Lake City was included in the 6:37pm warning, it would have been listed in the "locations impacted include" section and it was not. You can see this for yourself by looking just above the comments at the text of that warning.
As to what might have been the intent of NWS Memphis, I don't know. We can only judge by the accuracy of the warnings.
As to "praising," I wrote, "Overwhelmingly, the meteorologists of the National Weather Service wish to do the right thing." Seems like praise to me, even though the warning was not accurate in this case. And, yes, America's Weather Enterprise provided effective warning to Lake City even if the NWS didn't.
I wrote an entire book full of praise for the modern NWS ( https://www.amazon.com/Warnings-Story-Science-Tamed-Weather/dp/1608320340/ref=tmm_hrd_swatch_0?_encoding=UTF8&dib_tag=se&dib=eyJ2IjoiMSJ9.8emXc5jSpPMnhK1izUGwLA.jGzd6TqDA-6Gqs4RO1jtkJRYYL4ZSRYvR1J4CEl8jbo&qid=1740067464&sr=8-1 ), however, the quality of NWS warnings has deteriorated since then. I "call 'em as I see 'em" and I hope the future NDRB will do so, too.
Note to Readers: I just received a comment from Neil S (retired NWS, thanks for using your name) that makes almost exactly the same case as Anonymous's above. Rather than cluttering the comment section I will post two of his differing points and comment on them.
DeleteNeil S: "If the polygon that expired at 7 PM was just a few miles longer, Lake City would have been [in] the outer edge of the polygon instead of just outside of it."
True, but it wasn't. Either a location is included in a warning or it is not.
Neil S: "Lake city officials and anyone looking at a radar..." It is not the job of non-meteorologists to try to analyze radar data, it is our job. That's why we get paid. We don't ask non-physicians to diagnose cancer from an MRI. A nearly EF-4 intensity tornado is a mortal danger.
Neil S: "The NWS was not negligent..." I completely agree. The Memphis NWS office at that time was overwhelmed by tornado-producing thunderstorms between 6:30 and 7pm. I believe this was a case of something being accidentally overlooked rather than negligence. That's why it is good the NWS has "backstops."
Another comment in general:
It took aviation and medicine a long time to mature to the point where it could soberly and analytically review their successes and, especially, its errors. Weather science's occasional inability to acknowledge errors sometimes holds us back.
Mike, Thanks for your considerate response and I think we agree that we are continuing to work toward the ideal balance of services from the NWS and private sector, even if our perspectives on specific details of an event are a little different. No one is perfect in the NWS or the private sector but we do all work together pretty well and again, continue to adjust little by little to optimize our collective services to all users. AI is certainly adding another dimension to our figuring out how we best serve our users.
DeleteHi Neil, thank you for getting back to me and I agree with you completely.
DeleteYou are right that no one is perfect. I think it is appropriate to write about one of my failures which I am working on, now. I hope to have it up early next week.
Thanks again.
I heard again from our friend Anonymous who continues to argue: "Lake City was indeed not mentioned in the TO.55 locations impacted section. That is because the WFO Memphis Warngen cities database likely needs to be updated. An easy fix." I'm astonished to hear that expressed -- again. Lake City is a town of 2,000+, larger than some of the towns that were listed. If it wasn't in the written list and it wasn't in the polygon of the most current warning, it doesn't count as being in an official tornado warning. Wishing does not make it so. The next tornado warning wasn't issued until the storm was on top of the town (6:40pm). Note: WARNGEN = the NWS's software to issue storm warnings.
DeleteHe further claims the far west part of the city (which was damaged) was in the warning even though the "Lake City dot" was not. It is unreasonable to think people are out with theodolites trying to measure the size of the city limits on a warning map -- especially under the pressure of an approaching tornado. If the NWS wanted to include Lake City, they should have insured the "dot" denoting Lake City was in the polygon and it was listed it in the "locations included" list.
It IS true that at 6:40pm a special weather statement was issued pertaining to Lake City's tornado risk (TE message). But that does NOT trigger weather radios, WEA smartphone alerts, apps or services like StormCall or StormWarn. If the NWS wants a location under a tornado warning, they need to issue a tornado warning.
Finally, I reviewed all of this -- again -- from my Twitter feed and this blog. I posted the warnings in question. Now, go back and look at https://x.com/USWeatherExpert/status/1907578301951914052 . You will find that is my 6:38pm Twitter post (the times are appended to the Tweets by Twitter, not by me) telling Lake City in capital letters to get into shelter! The NWS didn't do it until two minutes later.
We weather scientists need to act like scientists. Facts are friendly! We cannot improve unless we acknowledge our issues, resolve to fix them, and then take measures to improve. Brushing off problems gets us nowhere. Next week, I'm going to write about a major warning failure in my organization and how we handled it.
I heard from a current National Weather Service meteorologist this afternoon who wrote, "But because of the polygonology, Lake City got, for lack of a better word, shafted."
Delete