Mapping the Hidden Fault Lines: Where Refugee Mothers Face the Greatest Earthquake Vulnerability

When the ground broke beneath southeastern Türkiye on February 6, 2023, the earthquakes that followed did not distribute their damage evenly across the map—or across the families living on it. A new study published in the Journal of Emergency and Disaster Medicine has combined psychometric testing with geographic information systems (GIS) to map, neighborhood by neighborhood, exactly where post-earthquake vulnerability concentrates among Turkish and Syrian refugee mothers in Gaziantep Province. The findings challenge a common assumption about disaster vulnerability: that refugee status itself is the defining risk factor. Instead, the research points to a more nuanced picture in which parental stress, parent–child conflict, housing damage, displacement, and bereavement—not nationality—predict who is most at risk.

The research team, led by investigators at Gaziantep University with collaborators in the United States, recruited Turkish and Syrian refugee mothers of young children through community health centers, refugee support organizations, and municipal contacts in earthquake-affected neighborhoods. Each mother completed a sociodemographic form covering housing damage, displacement, bereavement, children’s sleep and nutrition before and after the disaster, and access to healthcare. Parenting stress was measured with the Turkish adaptation of the Parental Stress Scale (PSS), while the parent–child relationship was assessed using the Child–Parent Relationship Scale (CPRS). Internal consistency was strong across the key subscales, with Cronbach’s alpha values of 0.80 for the PSS Stressor subscale, 0.73 for the CPRS Conflict subscale, and 0.75 for the CPRS Positive Relationship subscale.

From these instruments, the team constructed a composite vulnerability index. The psychosocial component standardized and averaged the PSS Stressor and CPRS Conflict scores, while the disaster-exposure component summed three binary indicators: housing damage, displacement, and loss of a loved one. The final continuous index combined the standardized psychosocial and exposure scores, and for visualization purposes the index was linearly rescaled to a 0-to-20 range while preserving its rank order. Convergent validity was confirmed through correlations with the anchor psychosocial measures, and known-groups validity was demonstrated by significantly different scores across housing damage, displacement, and bereavement categories.

The spatial analysis formed the study’s distinctive technical contribution. Working in ArcGIS Pro 3.6, the researchers created neighborhood boundary polygons and linked them to neighborhood-level summary statistics derived from the participant data. Earthquake intensity was represented by a polygon-based Mercalli intensity layer; neighborhood polygons were intersected with these intensity polygons, overlap areas were calculated, and an area-weighted mean Mercalli intensity value was assigned to each neighborhood. Because only five neighborhoods were available, the geospatial results were interpreted descriptively and exploratively rather than as formal spatial clustering or hotspot analysis. Even so, the mapped outputs revealed visible geographic variation, with certain neighborhoods showing notable outliers of elevated vulnerability against a broadly homogeneous distribution.

The statistical results were striking. In multiple linear regression, higher parental stress on the PSS Stressor subscale (β = +0.208, p = 0.009) and greater parent–child relational conflict on the CPRS Conflict subscale (β = +0.107, p = 0.048) were each significantly associated with higher vulnerability scores. Environmental and social adversities also mattered: housing damage (β = +0.086, p = 0.024), displacement (β = +0.114, p = 0.015), and loss of a loved one (β = +0.092, p = 0.033) all independently predicted greater vulnerability. Yet Syrian refugee status itself was not an independent predictor (β = +0.059, p = 0.352), and logistic regression models confirmed that refugee status did not increase the odds of belonging to the high-risk vulnerability group, defined by the upper quartile of the index distribution.

Mediation and moderation analyses added further precision. The researchers tested whether children’s sleep disturbances and nutrition problems transmitted the effect of parental stress onto vulnerability, but neither pathway reached statistical significance (ACME = −0.064, p = 0.884 for sleep; ACME = −0.070, p = 0.768 for nutrition). Similarly, bereavement did not moderate the relationship between parental stress and vulnerability (interaction β = −0.050, p = 0.151), although bereavement exerted a powerful direct effect on vulnerability (β = +1.24, p < 0.001). In other words, losing a loved one dramatically raises a family’s vulnerability, but it does so without altering how parental stress operates within the household.

The study also documented profound disruption of children’s daily lives in both groups. Regular mealtimes collapsed from 84 percent to 28 percent among refugee families and from 92 percent to 46 percent among Turkish families. The proportion of children described as having very good nutrition fell from 56 to 20 percent in the refugee group and from 49 to 28 percent in the Turkish group, with refugee children showing more selective eating and repetitive food choices. Sleep patterns deteriorated in both communities after the disaster, with refugee mothers reporting ‘inability to sleep’ in 12.2 percent of cases compared with just 2.2 percent among Turkish families. Healthcare access diverged even more sharply: only 14.3 percent of refugee families reported no change in child health monitoring after the earthquake, compared with 59.2 percent of Turkish families, while logistical problems and shortages of doctors were reported far more frequently by refugee households—evidence of systemic barriers that persisted even as the immediate emergency receded.

Interestingly, total parental stress was higher in the Turkish group, driven by the Parental Stressors and Lack of Control subscales, while positive emotional components such as parental reward and satisfaction did not differ between groups. The refugee group scored higher on the total CPRS, reflecting small but similarly directed increases in both conflict and positive-relationship dimensions—a pattern the authors interpret cautiously, suggesting it may indicate intensified family bonds during crisis, in which closeness and conflict rise together, or differing reporting tendencies rather than genuinely lower burden. The theoretical scaffolding of the study drew on the Eco-Biological Developmental model, Bronfenbrenner’s ecological systems framework, and Tronick’s stress-buffer transduction model, which together explain how caregiver stress can propagate behavioral and physiological changes to infants, making the mother–child dyad the critical unit of post-disaster resilience.

The authors are careful to frame the work’s limitations. The modest sample drawn from only five neighborhoods constrains generalizability and the strength of spatial inference, and convenience-based recruitment through health and community channels may introduce selection bias. The cross-sectional design precludes causal interpretation, and several earthquake-related variables rested on caregiver recall. Even so, the methodological message is clear: vulnerability after catastrophe is a spatially patterned, family-centered phenomenon, and effective recovery requires geographically informed, family-oriented interventions. Protecting routines—meals, sleep, healthcare follow-up—alongside psychosocial support and stress management may do more to reduce long-term harm than blanket aid distributed without regard to where and how need is concentrated. As climate pressures and seismic risk keep disasters on the global agenda, this neighborhood-level mapping approach offers a replicable template for directing scarce resources to the families who need them most.

Subject of Research: Neighborhood-level spatial distribution of post-earthquake psychosocial vulnerability among Turkish and Syrian refugee mothers

Article Title: Spatial distribution of post-earthquake vulnerability among Turkish and Syrian refugee mothers: a neighborhood-level study

Article References: Karadağ, M., Çalıkoğlu, E. O., Bulucu, B. B., Ruhm, F., Çalışgan, B., Coban, G. B., Erk, I., Otucu, E., Kurt, U. B., Chaban, Z., & Koga, P. M. (2026). Spatial distribution of post-earthquake vulnerability among Turkish and Syrian refugee mothers: a neighborhood-level study. Journal of Emergency and Disaster Medicine, 2(1), Article 9. https://doi.org/10.1007/s44467-026-00013-0

Image Credits: AI Generated

DOI: 10.1007/s44467-026-00013-0

Keywords: earthquake, Turkiye, Syrian refugees, maternal mental health, parental stress, GIS, vulnerability index, child sleep, nutrition disruption, healthcare access, disaster recovery, Gaziantep