OceanStWx

Not great, Bob. That's going to be my biggest loss with the NAM.

Clearing out some of the computing resources also potentially leads to more model runs. There are ways to get more data out of this.

It was like 2017ish. It was a bit of a Frankenstein "model". We really wanted some sort of probabilistic convective scale guidance, but it takes a lot of computing to run a large CAM ensemble. So somebody smarter than me decided to take what we already had running (ARW, HRRR, NSSL, etc) and turn them into a poor man's ensemble along with the 12 hour old versions of the same models. Pros: it provides some probability-like forecast parameters, it has a range of models/convective cores that can provide insight when biases are known. Cons: it is not a true ensemble, as you would prefer one core and the entire range of possibilities for that.

That's the theory of the case anyway. Instead of pouring resources into multiple models, we can focus all our energy on making the RRFS really good. Of course the FV3 core was really not great for convection and had to be scrapped. So now we're starting over with MPAS for a core. In an ideal world you would have the GFS/GEFS for longer range, RRFS/REFS for the inside 60 hours, and WoFS for storm scale, event driven cases.

I could have if I had been at work. As far as I know it is still 5 min. Temps are taken every min, and the high temp is the highest 5 min average. The only 2 min check I know on temps is to ensure they aren't more than 10 deg different on the 1 min obs (if they are it is set to missing). Winds are 2 min, so maybe that's the confusion.

I'm excited to marinate in the 90 degree garage tonight building a mud kitchen for the kids. At least I have fresh Tree House.

I was preparing myself for it, but they've seen my play too many times now to think I'm anything close to a 10.

Played a little Ryder Cup style tourny down in PA this weekend (Sunday singles rained out by the MCS). Had my first birdie since 2023 in the practice round Friday, but the real fun was going down 3 through 7 in our best ball match - only to come roaring back winning 7 of the next 8 holes to close out the match on 15. I'm currently a 21.5 index, but I had a middle 9 hole stretch of +3. That will do a lot of damage when you are getting strokes.

And more to come. Starting in July policy will be for the forecast to be populated with the NBM and largely left alone except for the near term. But we are wasting human resources generating dewpoint grids on day 7 given the current staffing situation.

At least something is loaded.

Before it clouded up, you could see some pillars on the Katahdin cam.

I know that because of our tree blockage issues that trips the rain rate algorithm over to a special attenuation version, BUT it only works below the freezing level. So everything above the freezing level is suspect at best. So for any precip accumulation product that's a factor for us. I'm confident someone right now is working on the contract to get our radar raised to mitigate the issue.

And it's like almost 10 dBZ off too. I know our beam blockage does through some things off above the freezing level, but Z shouldn't be one of them.

4 inches and another cell dropping 6"/hour headed their way.

Swift water rescue for the chickens.

It isn't often you can reach the 50 dBZ Donovan height around these parts, but that storm managed it.

It was more or less tracking right on the edge of the steeper lapse rates. Most impressive that it was able to become surface based despite the time of day.

One of our techs is there now, but I'm not hopeful for recovering it.

50 dBZ over 30,000 ft at 7 am. Not too shabby.

Live look...

Friday has that look, where we don't overturn the air mass on Thursday and end up doing the marginal risk on Friday instead.

Yes. There were two with the supercell by Great Barrington, one in CT, another in NJ, and several across parts of PA.

Woof. I'm glad I never had to use 8 bit radar data. I'm also pretty sure ALY was still operating the 74C for this event, this was the summer they installed the 88D. Can definitely infer some things about it with OKX data. It's roughly 40 kt Vrot over 80 nm from the radar. So considering it's already smoothing the details because it's 8-bit data, and the range from the radar means the large beam volume smooths it even more, this would've been impressive with today's super res.

There is a pretty good signal on the EPS for a wet weekend after next.