disasters that multiply, models that don't

Notes made while writing and checking this brief, shown as written. They use working shorthand.

• The EC2H DCL lists four participating GEO units — the Divisions of Atmospheric and Geospace Sciences (AGS), Earth Sciences (EAR), and Ocean Sciences (OCE), and the Office of Polar Programs (OPP) — and strongly encourages proposals that align with programs in two or more of them.
• Cross-domain connection: shares structure with `wildfire-infrastructure-coupled-modeling` (which addresses one specific hazard chain) and `infrastructure-cascading-failure-modeling` (which addresses cascading failures in engineered systems rather than natural systems). This brief addresses the general framework problem underlying both.
• The `failure:disciplinary-silo` tag is central — the DCL's emphasis is on interactions among hazards and on work that crosses disciplines and GEO units, rather than on any single hazard.
• The `failure:not-attempted` tag reflects the DCL's call for fundamental research on compound and cascading hazards; the letters do not say whether operational compound-hazard modeling frameworks exist.

Reconciliation 2026-09-30: Both sources were read in full on 2026-09-30: the EC2H Dear Colleague Letter NSF 23-125 (https://www.nsf.gov/pubs/2023/nsf23125/nsf23125.jsp, which now redirects to https://www.nsf.gov/funding/opportunities/dcl-extreme-compound-cascading-hazards-ec2h/nsf23-125, fetched directly, full text, dated June 29, 2023, last updated August 8, 2025) and the "Build a Resilient Planet" Dear Colleague Letter NSF 24-022 (https://www.nsf.gov/pubs/2024/nsf24022/nsf24022.jsp, fetched directly, full text, dated November 24, 2023). NSF marks both as archived. The letters are short calls for proposals: they name hazard types, four focus areas, the four participating GEO units, and NSF's FY2024 "Build a Resilient Planet" initiative, and they give no case studies, numbers, or statements about how hazard assessment is currently done. Much of the write-up went beyond them, so the following changes were made. (1) The Source line gave the first letter's title as "EC2H: Extreme, Compound, and Cascading Hazards"; the page gives "Extreme, Compound, and Cascading Hazards (EC2H)." The Source line now carries the titles, publication numbers and dates copied from the two pages, the redirect address, the second letter's address, and the archived status. (2) The Problem Statement gave three hazard chains (earthquake to submarine landslide to tsunami; climate change to wildfire to destabilized slopes to post-fire debris flows; a volcanic eruption producing atmospheric waves, a tsunami and ionospheric disturbances) and said current assessment and emergency management "treat each hazard independently, missing the amplification and interaction effects that often cause the worst damage." The DCL's examples are "submarine landslide-triggered tsunamis" and "climate change and wildfires," with no earthquake trigger, no debris flows and no volcanic cascade, and it says nothing about how hazard assessment is currently done. The paragraph now gives the DCL's two examples, says the DCL calls for fundamental research on cascading or compound hazards and does not describe how hazard assessment is done today, and reports the DCL's system-of-systems and long-term-trend focus areas as things the letter asks for; the earlier "We lack system-of-systems understanding" wording was a conclusion the letter does not state. (3) The whole of Why This Matters (the 2022 Hunga Tonga eruption and that "no existing hazard framework would have predicted this cascade"; California post-fire debris flows that "killed dozens and caused billions in damage" where "residents were told the fire risk had passed"; compound flooding "increasing in frequency along the US Gulf and Atlantic coasts"; insurance models and building codes that "systematically underestimate risk") appears in neither letter and was removed. The section now reports what the letters say: that climate change and solar activity can strengthen hazards or make them more frequent, that hazards "exact a heavy toll on natural and human systems," the list of hazard types, the societal-benefit requirement, and the FY2024 budget-request origin of "Build a Resilient Planet." (4) What's Been Tried said cross-divisional proposals "have historically been difficult to fund because reviewers are specialists in one hazard type," that process models "lack coupling mechanisms between domains" (with volcanic ash and hillslope hydrology examples), that multi-scale processes "require fundamentally new modeling frameworks that do not yet exist," and that space weather hazards "(geomagnetic storms affecting power grids) are rarely considered alongside terrestrial hazards." None of this is in the letters. The opening sentence that hazard research "is siloed by discipline" (seismologists, volcanologists, hydrologists, atmospheric scientists) is also not in the letters. The section now quotes the DCL's call for multiple disciplines and for proposals spanning two or more GEO units, says the letter is silent on how such proposals have fared in review, lists its four focus areas, its listing of space weather alongside terrestrial hazards, and its statement that it is not a new competition or program. (5) What Would Unlock said the EC2H DCL "has no deadline." The letter states no deadline, but NSF's page now marks it archived; the sentence says both. "System-of-systems modeling frameworks that couple atmospheric, solid earth, ocean, and cryospheric processes" became the DCL's own wording, system-of-systems approaches to the drivers of extreme events. The Hunga Tonga "natural experiment" item and "compound hazard metrics that go beyond single-hazard return periods" are not in the letters and were replaced with items the DCL does list (existing NSF/GEO facilities and data resources, new research infrastructure, stakeholder partnerships, workshops, research networks, planning proposals, and the Concept Outline route). "Research coordination networks" became the DCL's "research networks." (6) The entry point's examples (2018 Kilauea eruption to laze plume; 2021 Marshall Fire to debris flow risk) are not in the letters; the example is now the DCL's own, a submarine landslide-triggered tsunami. The second entry point (a simplified wildfire-to-debris-flow model from public fire perimeter and elevation data) is a project suggestion rather than a source claim and was left as written. (7) In Source Notes, "one of the few NSF mechanisms that explicitly encourages cross-divisional proposals" is not in the letters and was removed; "all four GEO divisions" became the DCL's "GEO units," since OPP is listed as an office. The tag notes now say only that the DCL emphasizes interactions across hazards and disciplines and calls for fundamental research, without the claims that coupling "is unexplored" or that frameworks "do not yet exist in operational form." The Hunga Tonga "benchmark case" note was removed. Confirmed in the sources and left as written: the four GEO units (AGS, EAR, OCE, OPP); the system-of-systems and land-ocean-atmosphere-ice language; increased solar activity and climate change as drivers; the encouragement of workshops and research networks; the "Build a Resilient Planet" name. No title changed; the H1 and Display title state the write-up's framing of the gap rather than a claim found in the letters, and are reported for Beth rather than altered. Genome tags untouched. A second reader went over these repairs against both letters on the same day (2026-09-30), found none where the original text was supported, and tightened the wording of the Problem Statement, the societal-benefit sentence, the opening of What's Been Tried, and the deadline sentence so that they say no more than the letters do.