We are still evaluating how the new suite of AI models may compare (e.g., AIFS, Google-DeepMind, Graphcast). Thus far, it looks like the same limitations that apply the ordinary NWP modeling also applies to the new suite of AI modeling.

Now for (2) sharing deterministic model output at long lead times:

Forecast skill for individual features meso-to-synoptic scale in size (i.e., #TC) lose predictability fast >5-7 day period.

The Environmental Modeling Center has detailed stats that show this rapid skill drop off in all models. 📉

Yup. It’s always pure entertainment

Numerous studies have shown that radiosondes are a critical input to the computer models that help us forecast the weather accurately on an ongoing basis.

Plus, they provide forecasters quality insight in and immediately ahead of high-impact events like severe storms and hurricanes.

What Nate said.

Ridiculous

Contact your representatives don’t let up they need to hear you.

Didn’t have republican support hoping this bill along with the other things start the conversation.

@ross.house.gov

This also includes getting a radar to Charlotte, which folks have fought for years. This is desperately needed.

More progress to protect and further fund and support NWS/ NOAA but I thank @RepTimMooreNC @RepPatHarrigan and @RepDeborahRoss for understanding how important critical weather infrastructure and information is, and how it should be nonpartisan with widespread support. #ncwx

The spring pattern is very elaborate this year and it’s leading to large swings and temperature and precipitation mostly la Niña related