How SCEM Works

SCEM maintains a conflict roster of 8–12 active armed conflicts assessed to meet a minimum threshold of direct combat casualties, territorial dispute, or state-actor involvement. The roster is reviewed weekly. Conflicts approaching inclusion or retirement are flagged in the Roster Watch section.

The roster is intentionally limited to allow genuine depth per conflict rather than shallow breadth. Conflicts that fall below the inclusion threshold (sustained combat has ceased or reduced to incident-level) are retired; new conflicts that cross the threshold are added after a minimum two-week watch period.

As of the inaugural issue (30 March 2026), the roster comprises eight conflicts across Eastern Europe, the Middle East, Sahel/Horn of Africa, Southeast Asia, Caribbean, Central/East Africa, East Asia, and Northeast Asia.

Each conflict is assessed weekly against six indicators. Each indicator is scored on a 1–5 ordinal scale. The score is compared against that conflict's rolling baseline to derive the deviation.

Code	Name	What it measures
I1	Rhetoric	Official statements, propaganda intensity, dehumanisation language, and threat framing from primary belligerents and their state sponsors.
I2	Military Posture	Force positioning, operational tempo, weapons system activation, troop movements, and infrastructure mobilisation consistent with offensive or defensive staging.
I3	Nuclear / Strategic	Nuclear posture changes, strategic weapon deployment or testing, doctrine statements, and command-and-control signals from nuclear-capable actors in or adjacent to the conflict.
I4	Economic Warfare	Sanctions intensity, blockades, resource denial, trade weaponisation, and financial-sector targeting as deliberate instruments of conflict pressure.
I5	Diplomatic	Formal negotiation activity, multilateral engagement, back-channel signals, and the presence or absence of functioning mediation frameworks.
I6	Displacement	Internally displaced persons (IDP) and refugee flows generated by the conflict, including acute events (single-event mass displacement) and structural accumulation.

Most conflict monitoring uses absolute severity scales — a conflict is "high" or "critical" based on activity volume. This approach is poorly suited to chronic conflicts: a war that has produced 10,000 casualties a month for three years has a different analytical significance to a conflict that produces 10,000 casualties in its first month.

SCEM measures deviation from each conflict's own rolling baseline. A conflict where indicators have been at a sustained wartime norm for six months registers zero deviation, even at high absolute levels. A conflict where one indicator moves sharply upward from a stable baseline registers a positive deviation — and that is the genuine escalation signal.

This approach has a credibility advantage: it is harder to manipulate. Absolute-level assessments require subjective thresholds. Deviation-from-baseline requires only that the observed data change meaningfully from prior observations — a more falsifiable claim.

Baselines are contested for the first 13 weekly observations. During the contested period, all bands are marked CONTESTED — this signals that insufficient data exists to distinguish structural baseline from genuine deviation. The baseline locks at week 13 as the rolling median of observed scores.

SCEM draws primarily from independent, open-source primary sources with documented methodologies and field presence. State media and government-adjacent sources are used where no independent corroboration exists, but are flagged with an F4 flag to indicate source limitation.

Source	Role
ISW (Institute for the Study of War)	Military posture assessment, battlefield map data, order-of-battle analysis.
UNHCR (UN Refugee Agency)	Displacement and refugee data — I6 primary source.
OCHA (UN Office for the Coordination of Humanitarian Affairs)	IDP data, humanitarian access, and displacement event reporting.
ACLED (Armed Conflict Location & Event Data)	Conflict event data, fatality tracking, actor mapping.
IOM (International Organization for Migration)	Migration and displacement supplementary data, particularly for border crossings.
Reuters	Breaking event reporting and wire-service factual baseline for all theatres.

These sources are chosen for independence from belligerent governments, documented field presence, and methodological transparency. SCEM does not use social media as a primary source — it may appear as a corroborating flag but is never the sole basis for a scored indicator.

Each indicator is assigned a colour band reflecting its current state relative to baseline. Bands are definitional categories — they describe what the score means, not how it was arrived at.

Band	Meaning
CONTESTED	The baseline is still building (fewer than 13 observations). The indicator level cannot yet be assessed as above or below baseline because no baseline exists. All week-1 indicators are contested.
Green / Normal	Indicator is at or below its established baseline. Activity in this domain is within the conflict's structural norm — no escalation signal.
Amber / Elevated	Indicator is above baseline by a meaningful but not acute margin. Requires monitoring. May become a lead signal if sustained or combined with other elevated indicators.
Red / Anomalous	Indicator is materially above baseline. A genuine escalation signal is present in this domain. Where multiple indicators simultaneously reach Red, a multi-indicator spike is flagged.

F-flags are source or confidence qualifiers applied to individual indicator scores when the underlying intelligence warrants a caveat. They do not change the score — they contextualise it.

Flag	Meaning
F1 · Single source	The indicator score rests on reporting from a single source with no independent corroboration. The event or activity is reported, but the score carries higher uncertainty than usual.
F2 · Adjacent escalation	An escalation signal in a nearby conflict or domain is influencing this indicator without direct causal evidence. The signal is real but the attribution to this specific conflict is inferential.
F3 · Unverified	The reported activity or event has not been independently verified at the time of scoring. The score reflects a provisional assessment that may be revised in a subsequent issue.
F4 · State media only	The source for this score is state media or a government-controlled outlet from one of the belligerents. No independent corroboration is available. Treat with elevated scepticism.