07/06/2025
Anyone who wants to blame the Texas floods on the National Weather Service cuts is totally off base. I have several pertinent points, which those of you with short attention spans will have trouble staying with me to the end. Sorry. Here we go. :
1) the region of Texas has terrain that is conducive to flooding if extended heavy rain hits. The residents are aware of this fact;
2) Many NWS forecasters have, for the past two decades, relied very heavily if not exclusively on numerical weather prediction models. Some of us who have been in the field for a long time argued against this trend even two decades ago. Unfortunately, the public's desire for pinpoint weather at their doorstep means NWS forecasters spend excessive amounts of time tweaking automated graphical grids, based on relatively limited analysis of not just the forecast models, but the observations. This is a huge waste of their time that hopefully AI can be at least partially tasked with, eventually.
3)These prediction models have significantly under forecast this system through its entire lifespan and continue to do so today. (I looked at predictions even from this morning's High Resolution Rapid Refresh models and most of the models seriously underpredicted the rainfall that is now ongoing farther north than the rains Friday and yesterday.)
4) Improving the models requires dedicated research and testing. Research dollars in both NOAA and the National Science Foundation in the past 10-20 years have tended to shift away from improving weather prediction (which involves research on phenomena to improve the models as well as development of improved model equations) in favor of climate-oriented research and applications, fueled in part by what's been referred to as "climate hysteria". I'm aware of this from my own experience in the past trying to get money for this type of research and being told "agency priorities have changed." When researchers can't get money for what they want to do, they usually become resigned to going to the areas/priorities where the funding is.
5) As a result, NWS is still using largely the same suite of numerical models it had available 10 years ago but trying to add value by statistical techniques, increasing resolution but with mostly incremental changes to the model physics.
6) This event will most certainly go down in the books as an outlier for which statistical methods applied to the existing model suites will invariably fail. The fact that most of the models this morning still underpredicted the ongoing rainfall now in North Texas, causing the DFW NWS office to have to scramble and add Flash Flood Warnings on the fly to several counties they did not anticipate based on the model forecasts, speaks to the need for a thorough analysis of this event, the model failures for this event, and dedicated research to improve the entire NWP *and* operational forecast process. That research does not have to be the exclusive domain of NWS, NOAA, NSF or NASA. There are now quite a few private meteorological research companies with very talented people that might better user Federal (or State, or Local or even private) dollars for these purposes.
7) The only good that might come out of this tragedy is that there is still a hell of a lot of work to do to improve day to day weather prediction. Is climate predication important? Sure though I've never been hysterical about it. Protecting life and property better in the here and now should never be sacrificed for it.
