The research question
When a disaster becomes a federal declaration, what warning record and observed event sit behind it? The records need to agree on place, time, and hazard before they become part of the same history.
Three records, one evidence chain
Warnings
Historical National Weather Service watches, warnings, and advisories show what people were told before or during hazardous conditions.
Observed events
Maintained disaster records describe the flood, tornado, severe storm, or other event that occurred in a specific place and time.
Declarations
FEMA declarations document the federal response to an incident and the communities included in that response.
A published chain follows this order: warning, event, declaration. The observed event is the bridge. A warning is linked only when it fits the event's location and time, and the event already has a strong connection to the declaration.
How the record is built
The work starts with FEMA's incident dates and declared areas. Those records are matched to observed disaster events using the affected places, hazard type, and timing. Historical warnings are then compared with the matched event, not with every declaration in the same state or season.
- Place first: county-level overlap is the main test, with broader geography kept separate as weaker evidence.
- Time first: warnings must fall in the period surrounding the observed event.
- Evidence stays graded: high, medium, and supporting links remain visible as different levels of confidence.
- No automatic causality: a chain records supported context for study; it does not claim that one warning caused a declaration.
What the first national pass shows
The declaration record contains 5,192 FEMA disasters. Of those, 3,367 already connect to a primary observed event. The warning layer adds evidence for 1,545 declarations through 1,304 matched events, producing 4,618 warning-event-declaration records.
Those counts are a starting research record, not a claim that every declaration has a weather-warning predecessor. Earthquakes, many wildfires, and events outside the historical warning coverage need a different source trail. Keeping that boundary visible is part of the method.
Why the chain matters
The chain lets researchers move from a federal response back toward the local conditions that preceded it. It supports questions about warning coverage, repeated impacts, event severity, and the gap between a local hazard and a later declaration. Counties can be compared without losing the records that make the comparison meaningful.
Read the evidence at the right level
A warning records communicated risk. An event records what was observed. A declaration records public response. They are related, but they are not interchangeable. The project keeps each record distinct so researchers can inspect the evidence and test different interpretations.