John Bistline
@bistline.bsky.social
Energy systems modeling, economics, policy | IPCC, NCA, Stanford/CMU alum | Views my own
Economics
30%
Engineering
28%
Moral of the story: models are only as good as their data, definitions, and constraints. Also, good models don't replace judgment but organize it. Dantzig's wife studied the menus and said, "I'll put you on my diet." And he lost 22 pounds.
Then: two pounds of bran per day. Anne threatens to take him to the hospital, imposes a cap, and swaps in blackstrap molasses. New constraint, similar nutrients. Classic LP substitution.
Next day the model prescribes 200 bouillon cubes/day. Anne Dantzig drops a legendary pun: "What are you trying to do, corner the bullion market?" Doctor confirms: no upper bound on salt in the requirements. So Dantzig adds one. Behold, the origin story of upper bounds.
First menu from the IBM 701 includes... 500 gallons of vinegar. Because the source listed vinegar as "very weak acid" with water content = 0, the model treats it like dense fullness. "I decided vinegar wasn't a food."
While at RAND, Dantzig's doctor tells him to go on a diet. So he models his own diet as an LP to maximize "feeling full." Objective ≈ food weight minus water weight. What could go wrong?
Background: Dantzig invents the simplex algorithm for solving linear programming problems but needs a test case. In a Pentagon bull session, Marvin Hoffenberg suggests: "Try Jerry Cornfield's diet problem for GI rations" (i.e., finding a low-cost diet meeting nutritional needs).
Modeling Monday: In honor of George Dantzig's 111th birthday, everyone knows the "late to class/solves two open problems thinking they were homework" story. Fewer know the hilarious backstory of the diet problem. Buckle up for an operations research thread... 🧵
Ultimately, what matters is the emissions intensity of electricity generation. On that front, you're right that CA is currently lower than TX, but TX has made much faster progress (from a higher starting point) and is catching up to California.
Ultimately, what matters is the emissions intensity of electricity generation. On that front, you're right that CA is currently lower than TX, but TX has made much faster progress (from a higher starting point) and is catching up to California.
Updated data on renewables and energy storage deployment. The race between California and Texas is... how shall we say... no longer close.
20+ years of climate impacts research in one figure. Climate change hits everything, everywhere, all at once.
From Solomon Hsiang's new NBER paper on empirical methods and economic impacts of climate change: www.nber.org/papers/w34357
From Solomon Hsiang's new NBER paper on empirical methods and economic impacts of climate change: www.nber.org/papers/w34357
Updated UNEP Emissions Gap just dropped. Current policies are projected to lead to peak warming of 2.8°C (with a range of 2.1-3.9°C). NDCs can push this down to 1.9-2.5°C.
Everyone's talking about AI's electricity demand. But new EIA data show planned gas builds look... pretty normal. Additions through 2030 track the last decade, not a 2000s style surge. Is this turbine supply constraints, lagging data, or something else?
Reposted by Aaron Sojourner, Greg Linden
Big EV gap between China and the USA.
- China has 52 BEV models with >400-mile range with many under $50k; USA has 4 all priced above $75k
- EV sales share in 2024: China 47%; USA 10%
Great paper in Science by @jhelvy.bsky.social.
- China has 52 BEV models with >400-mile range with many under $50k; USA has 4 all priced above $75k
- EV sales share in 2024: China 47%; USA 10%
Great paper in Science by @jhelvy.bsky.social.
Postdoctoral Scholar - Environmental Science & Policy
University of California, Davis is hiring. Apply now!
recruit.ucdavis.edu
🚨The Moore Lab at UC Davis is hiring!🚨
Post-doc for a project with @adamsobel.bsky.social on the valuation of climate information for adaptation
Could be a good fit for an environmental economist or a climate scientist - flexible start date and location
Apply by Dec 1st: recruit.ucdavis.edu/JPF07346
Post-doc for a project with @adamsobel.bsky.social on the valuation of climate information for adaptation
Could be a good fit for an environmental economist or a climate scientist - flexible start date and location
Apply by Dec 1st: recruit.ucdavis.edu/JPF07346
Interesting takeaway from a new ESIG report: Half of recent planning studies assume at least one emerging technology by 2030. Do these timelines seem realistic for advanced nuclear, e-fuels, direct air capture, and the ever-popular "generic resource?"
Be careful out there. They're putting RCP8.5 in candy bars.
Celebrate Bat Week by reading about bat bombs or listening to my colleague Christian Newman talk about how the electric sector is working with conservationists to protect bats from white-nose syndrome on EPRI Current: epricurrent.podbean.com/e/63-a-bat-w...
Reposted by Katharine Hayhoe
The Wikipedia article on bat bombs is full of crazy details. My favorite is the section euphemistically labeled “Setbacks” with a photo of a towering inferno caused by an errant bat carrying an incendiary bomb that escaped, parked itself under a fuel tank, and blew everything up.
Eight AP1000s would be big news.
But quadrupling U.S. nuclear capacity by 2050 means building 1.5x that every year for the next 25 years.
But quadrupling U.S. nuclear capacity by 2050 means building 1.5x that every year for the next 25 years.
Reposted by John Bistline
This adds to a growing body of literature that very clearly shows the use of short run marginal emissions rates is NOT a credible measure of consequential emissions impacts of various actions. Here, it's rooftop solar... 🔌💡
“When things have to go quickly.” The Louvre heist brought to you by a “whisper-quiet 230V electric motor.”
Clean energy meets clean getaway.
Clean energy meets clean getaway.
Load projections for ERCOT should called the "Loch Ness Monster" curve. Demand is projected to nearly triple over the next 10 years.
Here's the key takeaway from our new Nature Climate Change piece.
Emissions reductions of rooftop solar are overstated by approaches that inadequately capture substitution effects
Nature Climate Change - Emissions reductions of rooftop solar are overstated by approaches that inadequately capture substitution effects
rdcu.be
Rooftop solar is still an important part of future mixes, but its credited CO2 impact depends on how you count. You can read our article here: rdcu.be/eL8iE
The tl;dr summary is here: restservice.epri.com/publicattach....
The tl;dr summary is here: restservice.epri.com/publicattach....
Does Rooftop Solar Actually Help the Climate?
Inside episode five of Shift Key.
heatmap.news
For more context, there's a great Shift Key episode with @jessedjenkins.com, @robinsonmeyer.bsky.social, and @emilypont.bsky.social on these interactions and why they matter: heatmap.news/podcast/shif...
Modeling Monday: How clean are EVs, really?
Short answer: it depends on the question. Attributional vs. consequential LCA, average vs. marginal, short- vs. long-run... the metric can flip the answer. 1/🧵
Short answer: it depends on the question. Attributional vs. consequential LCA, average vs. marginal, short- vs. long-run... the metric can flip the answer. 1/🧵
You may have spotted the head fake. This isn't just an article about rooftop solar; it's about emissions accounting. These issues are key when accounting for electric sector interventions, including electrification, energy storage, data centers, and energy efficiency: bsky.app/profile/bist...
Using hourly, long-run system modeling (or long-run marginal rates) captures dispatch, prices, investment, and substitution between rooftop and utility-scale solar. That leads to lower CO2 benefits compared with short-run rates or the UNFCCC method.
Counting hourly emissions matters. Rooftop solar output often lands in low-emissions hours and overlaps with utility-scale solar, which leads to substitution.
New in Nature Climate Change: @asawatten.bsky.social and I show how common marginal emissions methods can miss system effects and substitution. Using hourly, long-run modeling can materially alter emissions benefits, which we show for rooftop solar. Paper + summary 👇