James Doss-Gollin
banner
jdossgollin.bsky.social
James Doss-Gollin
@jdossgollin.bsky.social
Assistant prof @ Rice CEVE. New Haven kid in Houston, husband, dad, romanista, albirrojo. I study and simulate hydroclimate risks to enable adaptive, efficient, and resilient infrastructure.

https://dossgollin-lab.github.io
https://ai4climaterrr.rice.edu
My favorite thing about Thanksgiving week is that all the companies who’s crap I need to unsubscribe from hopefully send me multiple reminders
November 26, 2025 at 4:58 PM
Reposted by James Doss-Gollin
Our free and open source software for estimating property level flood risk under uncertainty now has a peer reviewed software paper!

We would love your thoughts for improvement and contributions! github.com/abpoll/unsafe

@jdossgollin.bsky.social @vsrikrish.bsky.social
Just published in JOSS: 'UNSAFE: An UNcertain Structure And Fragility Ensemble framework for property-level flood risk estimation' https://doi.org/10.21105/joss.07527
November 13, 2025 at 2:23 PM
Reposted by James Doss-Gollin
On Wed, Dec 10 I am offering a comprehensive introduction to Bayesian regression modeling, with exercises in both RStan and PyStan, at a steep discount in an effort to raise funds for World Central Kitchen and United Farm Workers. For registration details see betanalpha.github.io/courses/.
Courses
betanalpha.github.io
October 27, 2025 at 12:33 PM
There seems to be very little coverage, but Hurricane Melissa is about to do terrible damage to millions of people.

gdacs.org/report.aspx?...
October 27, 2025 at 2:25 PM
Climate risk models require DEEP expertise to use well. It is easier every day to download some black-box risk estimate for your house/community/portfolio. However, the common mindset of "some data is better than no data" can lead to trouble!
Absolutely fantastic study comparing anonymized outputs of climate analytics models.

Everyone loves to bag on flood (rightly so), but can we talk about how the spread here is "30mph breeze to Cat 5 hurricane" www.fca.org.uk/publication/... HT @ruarirhodes.bsky.social
October 24, 2025 at 6:57 PM
Thanks to everyone who is out (on behalf of all of us knocked out by daycare germs) to demand and defend accountability, transparency, democracy, and the very concept of morality.
October 18, 2025 at 7:56 PM
I literally only use this app to grumble
September 29, 2025 at 2:42 PM
I knew having kids would change my life, but I wasn’t ready for Spotify to pack my recommendations with 98% white noise and nature sounds
September 19, 2025 at 1:45 PM
Reposted by James Doss-Gollin
📊New research published in #ERCL - in the 'Focus on #ClimateChange Informed Catastrophe Modeling to Support #ClimateRisk Management'

🌧️Bayesian spatiotemporal nonstationary model quantifies robust increases in daily extreme rainfall across the Western Gulf Coast: iopscience.iop.org/article/10.1...
September 11, 2025 at 3:43 PM
Broke: I want AI to do my laundry and dishes so that I can do art and writing

Woke: I want AI to fill out when2meet / doodle / calendly polls so that I can complain about bad papers on social media
September 9, 2025 at 3:57 PM
Controversial take: every statistical analysis you do doesn’t need to be an R package.

Packages should be infrastructure and set a very high bar of software engineering for them. To disseminate your analysis, just make some well documented templates and tutorials
September 8, 2025 at 12:28 PM
Reposted by James Doss-Gollin
What a clear, concise summary of the state and potential of renewable energy technologies. I particularly like how @rebeccasolnit.bsky.social includes the fact that nearly two-thirds of primary energy gets wasted today. Burning fossil fuels is a very inefficient way to provide useful services.
August 1, 2025 at 4:39 PM
How exciting is research according to...
July 29, 2025 at 4:35 PM
Grumpy take: you cannot just write in your methods section "we used X software", you should cite and note the software but you still need to explain what the software is doing and why it's a good choice
July 25, 2025 at 8:14 PM
Talk about a profoundly shredded moral fabric
July 23, 2025 at 12:11 AM
Reposted by James Doss-Gollin
This country won't get serious about climate change until we're honest about flood and fire risk and insurance to match.

Imagine if Camp Mystic had been required to carry liability insurance in the event that its campers were flood victims. Those cabins would have been moved years ago.
July 13, 2025 at 5:21 PM
I understand everyone wants to cover the tragedy, but I would MUCH PREFER to be part of a public discussion of flood warning systems and safety BEFORE the next horrific disaster happens than AFTER.

As long we remain uncomfortable with proactive safety measures, people will keep dying.
July 10, 2025 at 2:24 AM
Reposted by James Doss-Gollin
Early warning systems work. Early warning systems are crucial. This camp lost no one. See next skeet for quote. apnews.com/article/texa...
Quick action by one Texas summer camp leads to timely evacuations ahead of deadly flood
One local summer camp in the path of the disastrous flooding in central Texas was able to avoid any loss of life by closely monitoring weather reports.
apnews.com
July 8, 2025 at 3:28 PM
Reposted by James Doss-Gollin
But this does illustrate a few tragic and uncomfortable truths. The first is that even quite good weather forecasts do not automatically translate into life-saving predictions--there's a lot of other work that has to take place to contextualize the forecast and ensure it gets to right people.
July 5, 2025 at 3:50 PM
#Flood modelers: lots of chatter RE soil profiles in "flash flood alley". Do we really think floods would have been meaningfully different if you dropped that much rain on those hills, but they had more/deeper topsoils? Of course the effect is not zero, but not convinced it's big.
July 8, 2025 at 3:16 PM
Reposted by James Doss-Gollin
Based on some of the reactions and comments I am absorbing by average people and elected officials, I think there needs to be a serious explanation and lesson on what the limits of predictability are for floods like the one in Texas.
July 5, 2025 at 11:37 PM
Reposted by James Doss-Gollin
And don’t remove observational capacities and cut funding to weather forecasting agencies and then blame them for “faulty” forecasts!
July 5, 2025 at 2:47 PM
Horrific things are happening along the Guadalupe River right now. Absolutely devastating for the families of everyone missing.

I'm skeptical that we're going to build our way out of experiencing major flooding with 15+" rain on hilly terrain. Instead, we need...
July 5, 2025 at 1:48 PM
Reposted by James Doss-Gollin
I feel like I'm losing my mind. Here is the transcript of what Zohran Mamdani actually said about the phrase "globalize the intifada." This is his "refusal to disavow it." This is what has people terrified. Just fucking read it.

Source: podscripts.co/podcasts/the...
June 26, 2025 at 7:50 PM
Reposted by James Doss-Gollin
My latest feature is out: As the world has warmed, the regional flaws of climate models have grown apparent. Meanwhile, methods to localize the results of models have proliferated.

It all leads to lots of uncertainty -- within a very certain global trend. And cities need answers now.
Local predictions of climate change are hazy. But cities need answers fast
Scientists are figuring out where “downscaled” climate models struggle—and how they can be improved
www.science.org
June 5, 2025 at 6:41 PM