CERIFR Research — Migration Scenario Engine

Glossary

Concise definitions of the technical terms used throughout the Migration Scenario Engine. 21 terms grouped by category.

Indices & outcomes

MPI
Migration Pressure Index — composite index (0–1) from 5 stress components (climate, conflict, disasters, governance, economy). Higher = more migration pressure. Normalised by percentile rank within each SSP group.
TPI
Trapped Population Index — combines migration pressure (MPI) with limited mobility. Flagged at P95 within SSP. High TPI = high stress AND low mobility.
Net Migration
Immigration minus emigration. 5-year cumulative. Positive = net inflow.
Flow Rate
Migration rate per 1,000 origin population per 5-year period.
Corridors
Bilateral origin–destination pairs (e.g. MEX → USA). The panel covers ~52,670 directed corridors.
Diaspora Stock
Cumulative stock of migrants from origin in destination country. The strongest single predictor in the historical panel (r ≈ 0.33).

Climate scenarios

SSP
Shared Socioeconomic Pathways — IPCC scenarios: SSP1 (Sustainability), SSP2 (Middle of the Road), SSP3 (Rivalry), SSP5 (Fossil-fueled).
Scenario
Narrative future scenario crossed with each SSP: Baseline (ML only), Baseline+, Adaptation Success, Fragmentation, Climate Extreme.
CMIP6
Coupled Model Intercomparison Project Phase 6 — climate model ensemble for projections through 2100. The MSE uses a 5-GCM ensemble (ACCESS-CM2, GFDL-ESM4, IPSL-CM6A-LR, MIROC6, MPI-ESM1-2-LR).
Temperature Anomaly
Deviation of mean temperature from the WMO 1961–1990 reference, in °C.
Precipitation Anomaly
Deviation of precipitation from the WMO 1961–1990 reference, in mm/year.
ND-GAIN
Notre Dame Global Adaptation Initiative — climate adaptation score (0–100) per country.

Governance & structural variables

WGI
Worldwide Governance Indicators — World Bank index. The MSE uses Government Effectiveness (GE) and Rule of Law (RL). Scale: −2.5 to +2.5.
Gravity Model
Migration flows as a function of mass (population) and distance — analogous to physical gravity. The MSE uses CEPII GeoDist for the static gravity backbone.

Modelling & statistics

Stacking Ensemble
Combination of three models — GAM (weight 31.7%, OOF R² = 0.795), Random Forest (33.3%, R² = 0.804) and XGBoost (35.0%, R² = 0.813) — via a Ridge meta-learner on out-of-fold predictions. Pooled OOF R² = 0.826. Reaches 99.9% of the temporal autocorrelation ceiling.
OOF
Out-of-Fold — predictions on data not used during training (cross-validation). Avoids overfitting bias. The MSE uses 5-fold expanding-window OOF predictions.
IPF
Iterative Proportional Fitting — calibration method that aligns model predictions to known marginal totals (UN WPP origin and destination totals).
CPI
Conformal Prediction Intervals — distribution-free uncertainty intervals (50% and 90% coverage) based on out-of-fold residuals. Mondrian-binned by flow magnitude; multiplicative bootstrap with N = 500 replicates.
SHAP
SHapley Additive exPlanations — explainability method that quantifies each feature's contribution to the prediction. The MSE publishes SHAP attributions for both Random Forest and XGBoost components.

The full glossary is also accessible inside the interactive dashboard at migrationengine.org (Methodology & Glossary tab).