Monitoring Internal Displacement During the Syrian Civil War via Satellite-Based Vehicle Detection

Abstract Reliable data on human mobility is critical for understanding population displacement during armed conflict, yet such information is often scarce, delayed, or inaccessible in crisis-affected regions. In this study, we present a case study of the Syrian Civil War, where we use satellite-derived vehicle detections as a proxy for monitoring internal displacement patterns. We develop an approach that integrates a region-adapted vehicle detection model with a partial coverage correction method, enabling robust estimation of car counts from incomplete satellite observations. Using IDP reports from official sources, we evaluate the relationship between vehicle dynamics and displacement patterns across Syrian cities. Our validation results show that when the official reports and satellite images overlap in terms of time period and geography covered, then there is a good directional agreement between changes in car counts and reported internal displacement. However, we also observe that these two sources are often complementary in terms of exact geography and time periods covered, indicating that our remote sensing approach could help fill data gaps. These findings demonstrate that satellite-based vehicle detection can provide a valuable and scalable complement to traditional displacement monitoring methods in data-scarce conflict settings, as illustrated through the Syrian case. Dataset Syria-IDP-Car-Counts dataset (TSV (Tab-Separated Values) file) includes mapped humanitarian reports and vehicles detected from satellite imagery. Each column's description is provided below: Column Description Year Year in which the displacement or mobility data was reported Period Reporting time range within the year, represented as FromMonth–ToMonth (e.g., January-August) Location Name Name of the location where the displacement event or assessment occurred Admin Level Administrative or geographic aggregation level of the reported data (e.g., city, governorate) Human Population Variation Change in the reported number of IDPs between the start and end of the reporting period, categorized as Increase or Decrease #IDP Number of internally displaced persons (IDPs) reported for the corresponding location and period Baseline Population Baseline population estimate obtained from the WorldPop repository Baseline Car Population Baseline car population estimated from pre-war car-count observations using the mean value across the baseline period Car Population Variation Change in the reported vehicle count between the start and end of the reporting period, categorized as Increase or Decrease. C_LF Difference in car counts between the last observation and the first observation within the time period including optimal temporal buffer of three months around the report dates (k=3) C_MM Ordered difference between the maximum and minimum observed car counts within the time period including optimal temporal buffer of three months around the report dates (k=3) C_LR Regression-based estimation of car count change within the time period including optimal temporal buffer of three months around the report dates (k=3) Raw Car Counts Dictionary containing monthly car-count observations for the analysis period, including a maximum temporal buffer of four months before and after the report dates. Missing observations are represented as NaN Reported By Humanitarian organization responsible for generating or publishing the original report or assessment Source Platform or repository where the report was collected from, typically ReliefWeb URL Web link to the original source, report, or dataset used for data collection and verification