Adam Pollack
crispapoll.bsky.social
Adam Pollack
@crispapoll.bsky.social
Policy analysis for problems defined by pluralistic values and deep uncertainties. Most of my research is on flood-risk management problems.

Research scientist @ Dartmouth College. Incoming assistant professor @ Univ. of Iowa

https://abpoll.github.io/
So many to choose from. Might be pardoning the drug trafficking former president of Honduras
December 6, 2025 at 7:38 PM
What an embarrassing self-own to try and get the attention of dear leader
December 6, 2025 at 7:35 PM
Any word on the ~$7.6B Army Corps preferred plan for the NJ back bays? (www.nap.usace.army.mil/Missions/Civ...)
www.nap.usace.army.mil
December 5, 2025 at 8:44 PM
Have this on too!
December 4, 2025 at 1:34 AM
Great resource, thank you
December 3, 2025 at 6:50 PM
Hey, with the $ we raise maybe we can fund our public counterpart!
December 3, 2025 at 6:24 PM
Helpful, thank you! Do you know if this is re: homeowners policies or others as well (e.g., separate flood)?
December 3, 2025 at 6:24 PM
Reposted by Adam Pollack
I am continually working to get as many different actors in this space as possible to actually describe, in meaningful detail, not just *what* data was used but *how* it was actually used. Often that's not even very clear from the peer-reviewed papers that are coming out in this space!
December 3, 2025 at 5:54 PM
Reposted by Adam Pollack
Or any journal, honestly. Saying “peer review” as a defense against questions about transparency is a ploy for people who don’t know better. Any of us on this thread, who published regularly know that it is a very imperfect system and not some stamp of approval.
December 3, 2025 at 5:41 PM
Flood disclosures def not perfect but can include info beyond (rel. brief) direct flood history and the best climate risk estimates can do is capture what these disclosures tell us. Other hazards I sit out, haven't studied those! I hope people are tracking those parcel fire histories!
December 3, 2025 at 3:45 PM
Love that!
December 3, 2025 at 3:24 PM
Agree! I think disclosures give you that and we don't really need climate-risk estimates for that. Adding climate change I'm not even sure we can get the direction right yet (maybe negligible flooding under historical conditions to non-negligible in distant future)
December 3, 2025 at 3:24 PM
I could see some kind of tiered map provision where data assimilation frameworks could create a relatively high lowest tier (esp. relative to the 50 year old maps today). The question for me is what decision are the maps for vs. other risk representations for different (and more) decisions
December 3, 2025 at 3:06 PM
I model risk for most of my research so this is weird from me, but I just don't know that for many decisions it's all that useful to have out-of-the-box risk estimates. So many questions about assumptions, uncertainty, interventions, etc. that require a decision analysis framework - a real service!
December 3, 2025 at 3:04 PM
This is a great article. I still think there's an interim issue where people want this kind of information and FS being relatively best in class is a low absolute bar. Usable information is not always useful
December 3, 2025 at 2:49 PM
That made me spit out coffee lol
December 3, 2025 at 2:45 PM
I really don't see what they provide as closing information asymmetries. To the extent they do, disclosures already would do that. They just add noise to markets, overall randomly creating winners & losers (but systematically at local scales as suggested in Schubert et al)
National‐Scale Flood Hazard Data Unfit for Urban Risk Management
Flood risks are concentrated in urban areas, where national-scale hazard models are less accurate Flood exposure estimates become increasingly uncertain at finer scales and may misrepresent the s...
agupubs.onlinelibrary.wiley.com
December 3, 2025 at 12:39 AM
With 6 feet (even 1 foot) of systematic inundation error on an observed event with known forcings, adding climate change (among other things they do to get estimates) only adds uncertainty. So adding CC can misinform more and their model is never better than observations, so why not stick to those?
December 3, 2025 at 12:37 AM