iiinova — where your digital self keeps evolving

3. What does the fact that only a small fraction of stations show statistically significant trends imply for signal/noise ratio and fractional attribution of precip extremes to warming? (Model-based fractional attribution is not very trustworthy for precip extremes) 4/4

200

Not sure what the best dynamical indices would be, but composite/regression analysis between precip extreme and the flow could give a clue. Since the strongest trend signal seems to be in elevated regions, flow structure over mountain could be important. 3/n

120

2. Instead of correlating with modes (ENSO, PDO, AO) to detect dynamical (+thermo) influence, wouldn't it be better to dynamical indices analogous to thermo indices (TCWV, DEW), because the modes conflate different dynamical effects and have low degrees of freedom. 2/n

110

R. Saravanan @sarava.net · 29/05/2026

Very nice study. After a quick read (I might have missed stuff), I have these questions: 1. Why did the thermo trends switch from weak -ve to strong +ve going from 1980-2024 to 2000-2024 (Fig9/10)? Contiguous US temperatures basically increased during this period, right? 1/n

100

I agree that climate science should distance itself from the storylines. Also, I don't think climate sensitivity is built into concentration scenarios, but crude assumptions about carbon cycle are used to convert emissions to concentrations. (Wouldn't it be circular to build in climate sensitivity?)

000

Reposted by R. Saravanan

Nature @nature.com · 13/05/2026

Yulu Hou and her partner experimented with using ChatGPT to automate marking of undergraduate assignments. Here’s what they learnt go.nature.com/3P2M9M5

go.nature.com

Can AI tools assess coding assignments?

Yulu Hou and her partner experimented with using ChatGPT to automate marking of undergraduate assignments. Here’s what they learnt.

1157

Reposted by R. Saravanan

ClimateDynamics @climatedynamics.bsky.social · 12/05/2026

If anyone wants to read an excellent oped in the New York Times by Matt Huber, let me suggest you start with this one. www.nytimes.com/2015/06/07/o...

nytimes.com

Opinion | The Deadly Combination of Heat and Humidity (Published 2015)

0183

Reposted by R. Saravanan

Erik R. Nielsen @eniels13.bsky.social · 11/05/2026

And we are off with the 2026 @tamuatmo.bsky.social @tamuartsci.bsky.social @tamu.bsky.social SEA-Met study abroad program in collaboration the Caribbean Institue for Meteorology and Climatology (CIMH) in Barbados!

061

Reposted by R. Saravanan

Aditya Narayanan @adityan.bsky.social · 11/05/2026

1. Antarctic sea ice seemed to resist global warming for decades. While Arctic sea ice declined rapidly, Antarctic sea ice showed no clear overall loss and even expanded between 2007 and 2015. Then, after mid-2015, it changed abruptly. www.science.org/doi/10.1126/...

science.org

Compound drivers of Antarctic sea ice loss and Southern Ocean destratification

Recent sea ice loss was initiated by ocean upwelling in East Antarctica and by radiative flux anomalies in West Antarctica.

1105

That some models, often simplified, show AMOC collapse does not worry me as much as the steady warming. It's not uncommon for comprehensive GCMs in early stages of development to show AMOC collapse for pre-industrial control runs. Once re-tuned, this collapse goes away and model becomes stable. 2/2

020

I support strong rational action to mitigate climate change and am sad that that's not happening. I only worry about slow harm of steady warming, not abrupt climate change. I have done research in the past on this topic. Abrupt change is possible but not inevitable. It is also hard to predict. 1/2

120

I have noticed that climate scientists are not that pessimistic, in part because they are focused on their research and also because they (rightfully) ignore much of the apocalyptic talk about tipping points and catastrophes that animates climate activists. Media amplifies this and spreads pessimism

320

Ben Sanderson @benmsanderson.bsky.social · 08/05/2026

I've always wished social media was like email: Large corporations can make money off email, but the protocol has no technical barriers to entry for small players. Yes, there are differences between social media and email, e.g., the network effect, but hopefully federated social media can handle it.

010

Reposted by R. Saravanan

pyCMAT is a python port of (some of) John Fasullo's CMAT climate diagnostic package. It produces nice comparative graphics and html summariees of ESM performance between versions, and can download CMIP6 data for comparison. github.com/benmsanderso...

172

Models generally agree (I think) on the clear sky radiative GH effect, for a given CO2 level (with some uncertainty on H2O levels?), but don't agree on cloud feedbacks which are sensitive to complex microphysics parameterizations. That means "simple physics" cannot explain recent warming amplitudes.

131

Glen Peters @glenpeters.bsky.social · 06/05/2026

I generally agree with the above, and I'm not sure satellites can measure longterm EEI trend very accurately. However, an issue I've been concerned about is that increased ASR, not decreased OLR, is driving recent increases in global T. That's poorly known SW cloud feedback, not classical GH effect.

220

Reposted by R. Saravanan

India’s economy expanded 7.5% in 2025, but India’s fossil CO2 emissions rose by just 1.1%—far lower than the 4.1% average over the previous decade. What is going on? www.ecpulse.in/snapshots/ap... 1/

35717

I did a more careful calculation. Solar radiation comes at 1360 W/m2. Solar panels are more efficient in space, say 30% => 400 W/m2. An acre is 4000 m2 => 1.6 MW per acre or 600 acres to generate 1 GW. Assuming more inefficiencies => a 1000 acre panel to generate 1 GW. A lot less than needed on land

021

I guess I need to put a lot of effort into re-balancing my retirement portfolio to avoid hype stocks, if my retirement plan even allows that. Perhaps the WSJ should publish a guide to doing that ...

150

Christopher Mims @mims.bsky.social · 03/05/2026

It does give an order of magnitude estimate. The panel area in space will be significantly less (factor of 4 from geometry plus more due to lack of clouds), because the sun always shines.

000

Reposted by R. Saravanan

Who's going to buy into the gigantic IPOs of companies with no profits and questionable paths to profitability, such as SpaceX and OpenAI? You will, through your employer's retirement plan, writes @jmackin2.bsky.social www.wsj.com/finance/stoc...

wsj.com

Stock Indexes Are Contorting Themselves to Include SpaceX and OpenAI

Index funds holding trillions of dollars of 401(k)s and other investments are rushing to change their rules as the IPOs of SpaceX, OpenAI and Anthropic draw closer.

16368144

An estimate from @hausfath.bsky.social posting on the other site:

491