Many weather scientists have noted that NOAA's global weather prediction model, the GFS, is now in fourth place, behind the European Center, the UK Meteorology Office, and the Canadians.
This is pretty depressing considering the U.S. spends more on weather prediction research and development than all those groups combined. This NOAA global model is the foundation of U.S. operational weather prediction efforts; thus, Americans are experiencing inferior weather forecasts as a result.
But it is worse than that.
NOAA global predictions have a severe "drop out" problem in which there are sharp, precipitous declines in forecast skill. Major declines in skill not shared by other major weather prediction centers.
Let me show you.
Below are the skills of various modeling systems from April over the northern hemisphere.
It evaluates the ability of models to get things right in the middle of the troposphere...around 18,000 ft (500 hPa pressure)--for a day 6 forecast. 1 indicated perfect score. Above about .8 the forecast is quite useful. Below ~.65 not so much.
The best forecast is the European Center (red line), while the US model (black line) is generally much less skillful.
Note that sometimes the US model skill drops like a rock to below .7 and on one date to below .6. These are drop outs...and represent severe loss of skill.
Note that the European model almost never does the same.
During the past few days (May 3-4), the U.S. model had another loss of skill (day 5 is shown in this graphic, with red being the US model, blue and black the European Center). Very bad.
Certain atmospheric flow patterns appear to give the US model a hard time. One of them is an omega block, in which a ridge (high) has two troughs (lows) on both sides.
Important: this dropout problem has been going on for years and has nothing to do with budget cutbacks, fired personnel, or some weather balloons not being launched.
Let us be clear. The U.S. needs a vital, state-of-science NOAA, with weather prediction capabilities that are the best that weather science can provide.
Many of us in the weather community understand what is wrong with NOAA and have concrete ideas on how to fix this unfortunate situation. I have written two published papers on the topic and have served on several national committees that provided strong recommendations.
The current administration wants to fix NOAA and make American weather prediction "great again." But during the first months of their tenure, it has made serious errors, such as mindlessly firing junior staff.
Will they talk to the meteorological community to develop a science-informed plan that could greatly improve U.S. environmental prediction and do so at a lesser cost than today?
I hope so. It would be a home run for the American people.
_______
Announcement
25% projected cuts to NOOA - no one in the current regime has plan to make weather forecast "great again".
ReplyDeleteIt really is a shame how much we rely on government, and how often we are tricked into thinking the government has the best interests of the common people at heart.
ReplyDeleteI can't fathom how it's a shame to have the government supply consistent, reliable, and free data that literally saves lives every day and is critical to the successful operation of our entire economy as well as our military. It would be a huge mistake to entrust something like that to the private sector, which is something individuals like Elon Musk would love to see. Imagine that the NWS is privatized and makes their weather data subscription-based and/or decides that there's not enough profit in providing reliable, up-to-date data to rural areas (which actually describes massive areas of the continental US), and/or is a publicly traded company that is eventually bought out via stock acquisition by foreign investors.
DeleteThese aren't crazy ideas either, BAMCIS. They're pure capitalism in America. Be very careful what you wish for, you just might get it.
Its a shame when the government itself is NOT consistant and reliable. We sort of have counted on there always being subject matter experts or at least those with clear thinking as well as dedication to the service provided unfettered by political bias. Now the precedent is set where loyality and idealogical adherance comes first. Not saying that loyality tests have not tainted our history in the past, but now we have examples were spite and spite alone decides policy that can impact many lives.
DeleteWhat it really comes down to now is that trust of the government is broken. It was wobbly before but now its just ...heh..stick a fork in it. Its done.
Why not save some money and simply use the European forecasts?
ReplyDeleteRenting is rarely cheaper than having you're own service, once we had to but European data they would raise the price. Also the European model needs data all over the world including our NOAA weather data to be accurate. The l9ss of NOAA data will decrease the accuracy of all models.
DeleteIt is a national security issue. We need to have our own meteorological infrastructure in case of war or other serious national issue.
DeleteNOAA and the NWS have many problems that existed six months ago, before President Trump assumed office. The Biden Administration had four years to repair the GFS and did nothing. They appointed a NOAA Administrator who didn't care about weather until Trump assumed the office and he could make political hay over it.
ReplyDeleteBefore you say, "Trump also had four years..." keep in mind that he could not get his nominee for head of NOAA confirmed.
Like Cliff, I am appalled how poorly the cuts at NWS and NOAA have been handled. Trump's administrator for NOAA is, again, is hung up in the Senate regardless of seemingly everyone believing Neil Jacobs is a great candidate.
And - here's to more weather stations, real data! This has come up year after year, despite "who's in the white house." Crikey! The "politicking" and "press" about this is ridiculous.
DeleteSounds like you may need to get a UW-WRF-ECMWF. ;) Love your blog, thanks for all you do!
ReplyDeleteI've heard several of your blogs on the problems with our system and would like to , respectfully, ask what you think should be done by Neil Jacobs to fix this? Can you provide a link to your papers on this?
ReplyDeleteA chilly 65 today. still waiting for our first heatwave of the season hopefully by the end of the month the 80s will return for the first time this year.
ReplyDeletePrivatize NOAA. Maybe Elon will buy it.
ReplyDeleteCliff - My first question is whether the NOAA model forecasts are improving over time "as we speak" or are they getting worse?
ReplyDeleteSecond question is whether NOAA model forecasts are improving at a significant rate, but just not as fast at the other models?
Third question is if all models are incrementally improving, but some faster than others, where is the fire?