iiinova — where your digital self keeps evolving

Looking at the 0.5”/50th% cluster analyses from @gingerwx.bsky.social from the 28/1200 UTC model cycle, it seems to be an ECMWF (wetter) vs. GFS/GGEM (drier) battle. Subtle changes will have an outsized influence on the final outcome.

Rainfall cluster analyses, showing three cluster-mean forecasts and the full ensemble mean rainfall forecast. Covers data 1-2 (through Thursday morning).

050

The current (29/0100 UTC) NBM probabilistic rainfall forecast for Erie, CO for the next three days has one of the weirdest distributions I can recall seeing, particularly between the 25th and 50th percentiles: 10th %: nothing 25th %: nothing 50th %: 1.22” 75th %: 2.12” 90th %: 4.13” #COwx

180

But anyway, public products’ utility is derived from high skill *plus* their trustworthiness (for the developing agency and/or product). Kalshi et al. will always lack that trustworthiness aspect.

If that’s true, I don’t think it should be surprising that traders could beat the NBM or other singular products on a consistent basis. Their livelihoods depend in part on doing so!

Speculating for a moment, my guess is that Kalshi weather traders are predominantly experts (possibly including some who do that work for a private company!) who have developed their own predictive models blending many sources of information and seek to cash in on their performance.

I think what they’re getting at is that weather-sensitive stakeholders should consider using Kalshi predictions to power their decisions since they’re more accurate than any single public-facing product. There is more to it than accuracy, though, for those decision makers (more on this at the end).

Rich Stearns @richstearns.bsky.social · 28/09/2026

Reposted by Clark Evans

I finally closed my account on X- it’s just too toxic. Would love to pick up some more followers here on Bluesky. Follow me if you have a passion for ‘the least of these’ and like thoughtful Christian dialogue.

053

Clark Evans @clarkevanswx.github.io · 28/09/2026

Dislike. That lowest color level is no better than 1.1-1.3”, so it’s not as if it’s even been close to 2”!

020

Clark Evans @clarkevanswx.github.io · 28/09/2026

This position closes on Monday (tomorrow!) at 11:59p MDT!

031

Clark Evans @clarkevanswx.github.io · 28/09/2026

This position closes on Wednesday at 11:59p MDT!

072

Reposted by Clark Evans

Casey Drottar @caseydrottar.bsky.social · 26/09/2026

My goodness, what an incredible moment. After Justin Verlander secured the final out of his career, Tigers manager A.J. Hinch came out of the dugout...to summon Verlander's daughter out to embrace her dad on the mound.

642334457

On the bright side, my son told me tonight that he loves watching lightning with me as we watched the storms 50 miles northeast of us at his bedtime. ❤️

080

Just noticed that this week’s Drought Monitor moved portions of the northern Denver metro area into the D4 Exceptional Drought category. It’s exactly where I would’ve expected it - and I’m smack dab in its center. Sigh.

Drought Monitor for 9/24/26, showing D4 exceptional drought in the northern Denver metro area.

152

OAR’s lawyers would probably encourage that SSEOv2 not be public, but if there would be some way of sharing that with us (to help as a baseline against which to compare REFSv2 prototypes), I know we’d greatly appreciate it!

High Plains Library District @myhpld.bsky.social · 25/09/2026

Any plans for an externally viewable SSEOv2 viewer as well?

130

Reposted by Clark Evans

Spokescat Woody here: Please allow my floofy tummy to grab your attention and remind you to apply to be a member of HPLD's Board of Trustees. Application closes at 5 PM today (Friday the 25th), so get on it!

121

Clark Evans @clarkevanswx.github.io · 24/09/2026

Our @cires.colorado.edu colleagues are seeking to hire a postdoctoral associate to be based at GSL focused broadly on microphysics parameterization: jobs.colorado.edu/jobs/JobDeta... Applications are due by October 8th!

jobs.colorado.edu

CIRES / NOAA Global Systems Laboratory, Post Doctoral Associate in Numerical Weather Prediction (Microphysics)

052

Clark Evans @clarkevanswx.github.io · 24/09/2026

We appreciated the detailed final report (and the ongoing presentation)!

Clark Evans @clarkevanswx.github.io · 24/09/2026

Glad you got some deep-dish pizza! Great catching up with you this week.

Clark Evans @clarkevanswx.github.io · 21/09/2026

I don’t like how Gmail handles those reactions. It makes it seem like a new message on both desktop and mobile when…it’s very much not.

Clark Evans @clarkevanswx.github.io · 21/09/2026

The 45 kt threshold is over a longer time period (36 h in that case).

Clark Evans @clarkevanswx.github.io · 20/09/2026

Congrats on your rain dump. Two miles to your north and all I’ve got is slightly damp concrete.

020

Clark Evans @clarkevanswx.github.io · 19/09/2026

I look forward to seeing how rainfall will miss us today. Will it be, “outflow from a dying storm just to our west that initiates a new storm just to our east?” Or maybe a replay of yesterday? Or possibly, “it begins clouding up too early, so we don’t build enough instability (but do just east)?” 🤔

130

Clark Evans @clarkevanswx.github.io · 19/09/2026

Narrator: the very light rain in the general area at the end dissipated in place.

Clark Evans @clarkevanswx.github.io · 18/09/2026

I think just about everyone in the metro area except me is going to get rain today, ensuring that we don’t finally cross the 3” mark since June 1st. It’s almost comical at this point.

590

Clark Evans @clarkevanswx.github.io · 16/09/2026

I bet @mattlanza.bsky.social would immediately recognize where this video was filmed…

UCAR Center for Science Education @ucarscied.bsky.social · 15/09/2026

Reposted by Clark Evans

Captain’s Log. Stardate 80317.5. The 30th anniversary of Super Science Saturday approaches. These are the voyages of @ncar-ucar.bsky.social. Its continuing mission: to explore our planet Earth; to seek new answers to complex questions; and to boldly go where no one has gone before.

Super Science Saturday. 30 years of boldly going. Saturday, November 7, 2026, 10:00 am - 4:00 pm. NSF NCAR Mesa Lab. 1850 Table Mesa Drive, Boulder, CO, 80305

143

Clark Evans @clarkevanswx.github.io · 15/09/2026

The state really needs to expedite the Arapahoe modernization project - and Boulder County the Boulder-to-Erie regional trail!

Clark Evans @clarkevanswx.github.io · 13/09/2026

Pure joy. This year’s @lavuelta.bsky.social was full of pure ciclismo.

131

Clark Evans @clarkevanswx.github.io · 13/09/2026

I think they’re reading the first five (rather than six) letters as the first word, leading to the remaining letters…well…

Kenzie Krocak @mjkrocak.bsky.social · 11/09/2026

Reposted by Clark Evans

Long time, no post, but I have to share a cool job opportunity: Social Science Research and Testbed Support Specialist in Norman, OK. Please pass along to folks who may be interested! Happy to answer questions or point you to someone who knows more. careers-fedwriters.icims.com/jobs/2456/so...

careers-fedwriters.icims.com

Careers: Unlocking Your Path to Success | FedWriters

Unlock a world of opportunities with FedWriters. Our active recruitment is your gateway to dynamic careers where talent and innovation come together.

01119

Clark Evans @clarkevanswx.github.io · 11/09/2026

This position’s deadline will be extended to Sept. 18th at 11:59p MDT. Please share with anyone you know in the planetary boundary layer modeling community.

043

Clark Evans @clarkevanswx.github.io · 10/09/2026

Yeah, any MPAS implementation that replaces HRRR is in 2028 at the earliest, and not necessarily early that year either.

040

Clark Evans @clarkevanswx.github.io · 09/09/2026

Do you love designing and running weather prediction model experiments, particularly to help improve understanding & make models better? A new position through @cira-csu.bsky.social and collaborative with NOAA/GSL may be of interest to you! Details: csusystem.wd12.myworkdayjobs.com/en-US/fortco...

csusystem.wd12.myworkdayjobs.com

26-121 Numerical Weather Prediction Analyst

Job Description The Cooperative Institute for Research in the Atmosphere (CIRA) at Colorado State University (CSU) seeks to fill a full-time research associate position designed to conduct collaborati...

0102

Clark Evans @clarkevanswx.github.io · 08/09/2026

Silver lining: the turnover on the final play of the game made FSU’s turnover margin only +3, meaning that stat ESPN showed (teams were 40-1 last year with a +4 TO margin) became irrelevant. If I’m going to start the workweek on low sleep, at least I only have to go into work 4 days this week.

BUM CHILLUPS AKA SPENCER HALL AKA PRESIDENT FOOTBALL @edsbs.bsky.social · 08/09/2026

Reposted by Clark Evans

lol no way I'm not using something called THE TORNADO KICK every time we punt

420811

Clark Evans @clarkevanswx.github.io · 07/09/2026

I have friends in Port Washington who did that on their cul-de-sac - starting with their own curb, then asking all of their neighbors if they’d be willing to let them do their curbs, too. Most said yes, and it’s coming along nicely.

121

Clark Evans @clarkevanswx.github.io · 06/09/2026

The dashed black line is the 2021-25 average improvement of the NHC official forecast for E. Pacific TCs vs. the no-skill baseline. I have it there for context - in this case, showing that the NHC has found Lowell difficult to predict, but Google DeepMind has done exceptionally well despite that.

020

Clark Evans @clarkevanswx.github.io · 06/09/2026

Atrocious review to put one second back on the clock for Michigan. Absolutely atrocious.

030

Clark Evans @clarkevanswx.github.io · 05/09/2026

Sam’s post prompted me to look up the track verification stats for Lowell - and sheesh, Google DeepMind (GDMI)…

Left: track forecast errors for Hurricane Lowell, right: track forecast improvement vs. a no-skill baseline for Hurricane Lowell. Both tell the same story: Google DeepMind’s forecast has been vastly superior to those of NOAA’s operational models and the NHC official forecast.

1203

Reposted by Clark Evans

American Meteorological Society @ametsoc.org · 04/09/2026

Students: Looking to attend #AMS2027? As an AMS member, you can apply for our student opportunities: ✈️ Travel Grants: Funding for hotels, registration & travel (non-presenters) 🤝 Student Assistants: On-site conference support & networking Learn more & apply by 18 September: bit.ly/4qVUkbq

Student assistants talking to attendees at the 106th Annual Meeting.

042

ALT: DJ Khaled: Another One

Another job opportunity with @cira-csu.bsky.social to work collaboratively with GSL is now available! This one’s for a model developer, focusing on code development, implementation, management, and optimization. For more details: csusystem.wd12.myworkdayjobs.com/fortcollins_...

static.klipy.com

DJ Khaled: Another One

077

One would think that this would prompt some internal reflection and potentially change at ESPN, but I’m not holding my breath. They long ago ditched their ombudsman, and their modus operandi has continuously shifted to the hot-take/breaking news model for years.

More generally, we’re trying to build a strategy to center model and product development around end-user needs. End users knowing what to expect from the models is one of many parts of that. More soon!

Agree 100% with “we are the ones who need to see the products regularly before we can just adopt them and take everyone at their word.” We’re trying to work out how we can do that better and in a sustainable way given the challenges with RRFSv1.

Clark Evans @clarkevanswx.github.io · 02/09/2026

Agreed. In addition to what you shared, while there is some sand (with high heat conductivity) in the soil in far eastern Texas, the diurnal cycle favors heavier rainfall near an inland vortex’s center during the overnight and early morning hours.