What an Event Severity Level Rests On: Public Thresholds and Where They Run Out
Quick Answer: No public body publishes one severity scale for all events. Earthquakes, floods and disasters have published thresholds (USGS PAGER, GDACS, EM-DAT); single attacks, explosions and protests do not. Off-Nadir Delta's Delta Impact Scale gives an event the highest public threshold its reported figures cross, from 0 to 5, and assigns no level when no figure was reported. It is Off-Nadir Delta's own scale, not one issued by a public body.
A severity number next to a news event invites one question: severe according to what? This article lists the public thresholds that exist for single events, the kinds of events for which none exists, and how Off-Nadir Delta's Delta Impact Scale builds a level from them.
There is no single public severity scale for events
Public scales are written per hazard, by the body responsible for that hazard. Each one scores something specific.
| Scale | Published by | What it scores | Thresholds |
|---|---|---|---|
| PAGER | U.S. Geological Survey | One earthquake, by estimated impact | Yellow 1–99, Orange 100–999, Red 1,000+ estimated fatalities; the higher of the fatality and economic-loss alerts sets the level |
| GDACS flood alert | United Nations and European Commission | One flood | Orange: more than 100 dead or 80,000 displaced. Red: more than 1,000 dead or 800,000 displaced |
| EM-DAT entry criteria | CRED / UCLouvain | Whether a disaster is recorded at all | At least 10 deaths, or at least 100 people affected, or an emergency declaration or call for international assistance |
| UCDP Georeferenced Event Dataset | Uppsala University | Whether an incident of organized violence is recorded as an event | At least 1 direct death at a specific place and date |
| Copernicus EMS damage grading | European Union | Damage to one structure, read from imagery | Destroyed, Damaged, Possibly damaged |
Two of these are worth reading closely. EM-DAT's criteria and UCDP's definition are inclusion rules: they say when something counts as a disaster or as an event, not how severe it is. And PAGER's levels are set from estimated ranges of fatalities and losses, not observed counts.
Attacks, explosions and protests have no public event-level scale
For a single strike, explosion or clash, we did not find a public scale that grades severity. UCDP records such an event once it has at least one direct death, and keeps a best, low and high estimate of deaths. Its better-known numbers — 25 battle-related deaths for an armed conflict, 1,000 for a war — apply to a conflict over a calendar year, not to one event. Applying them to a single incident would misuse them.
This matters for anyone comparing scores across products. A severity figure for a missile strike is, in every product we know of, that product's own scale.
What "GEOINT" does and does not define
The term has a legal definition. U.S. law (10 U.S.C. § 467), repeated in the National Geospatial-Intelligence Agency's GEOINT Basic Doctrine, Publication 1.0 (2018), defines geospatial intelligence as "the exploitation and analysis of imagery and geospatial information to describe, assess, and visually depict physical features and geographically referenced activities on the earth." That is a definition of a discipline. The doctrine does not set out a method for scoring events.
For imagery itself there is a public scale: the National Imagery Interpretability Rating Scale (NIIRS). The Civil NIIRS Reference Guide, prepared by the Imagery Resolution Assessments and Reporting Standards (IRARS) Committee, rates an image from 0 to 9 by the tasks it supports — detecting large buildings at level 2, individual houses at level 3. NIIRS rates images, not events.
How Off-Nadir Delta labels an event: the Delta Impact Scale
Off-Nadir Delta gives each event a level from 0 to 5 on its own Delta Impact Scale. The level is the highest public threshold that a reported figure crosses. Nothing is weighted or added up.
| Level | The event reaches it when | Threshold from |
|---|---|---|
| 0 | It is not a physical event: a statement, a procedure or a plan | Our own criterion |
| 1 | Something physical happened; the reports give no deaths and no damaged structure | Our own criterion |
| 2 | 1 or more killed, or a structure or facility damaged | UCDP; Copernicus EMS |
| 3 | 10 or more killed; 100 or more injured, displaced or evacuated; or a structure destroyed | EM-DAT; Copernicus EMS |
| 4 | 100 or more killed, or 80,000 or more displaced | USGS PAGER; GDACS |
| 5 | 1,000 or more killed, or 800,000 or more displaced | USGS PAGER; GDACS |
Three rules keep the level from being a matter of taste:
- No figure, no level. When the reports give no casualty count and no damaged structure, and it is not known whether anything physical happened, the event is shown as "Not established". No default is filled in.
- The highest threshold crossed sets the level. This is how the public scales work: PAGER takes the higher of its fatality and loss alerts, EM-DAT asks for at least one of its criteria, and GDACS raises a flood alert at a number of dead or a number of displaced.
- Order inside a level is the reported number itself. Deaths first, then injuries. There are no weights.
Dividing the ladder into six levels is our choice. So is applying fatality thresholds written for earthquakes and floods to attacks and accidents, for which no public event-level scale exists. The figures themselves are read from the reports by an AI model and can be wrong or missing; the event detail shows each figure with the report it came from.
Whether a satellite can check it: Delta GEOINT Readiness
How serious an event is and whether imagery can check it are different questions, so they get different labels. Delta GEOINT Readiness is a class from A to E, looked up from three recorded facts: whether the event leaves a physical mark a satellite can read, how precisely it is located, and whether freely available 10 m class imagery is enough. Class A means it can be checked now with free imagery; class E means nothing would show from orbit. There is no score and no weighting.
The exact tables are on the scales page.
What this does not claim
None of the bodies named here issues, reviews or endorses the Delta Impact Scale or Delta GEOINT Readiness. The thresholds are cited because a reported figure either reaches them or does not — a statement anyone can check against the source. Where no public threshold exists, the product says the criterion is its own.
The AI readings shown beside these labels, and their limits, are described on the methodology page.

Remote sensing specialist with 10+ years in satellite data processing and AI. Founder of Off-Nadir Lab. Ph.D. in Informatics (Yokohama National University, 2026). Master's in Earth System Science and Technology (Kyushu University). Co-author, Remote Sensing Encyclopedia. More about the author →