SOCIOENVIRONMENTAL VULNERABILITY AND DISASTER EARLY WARNING: A SPATIAL AND MULTIVARIATE ANALYSIS OF PUBLIC ENGAGEMENT IN CURITIBA (PARANA, BRAZIL)

Abstract

This study investigates the relationship between socioenvironmental vulnerability and the number of mobile phones voluntarily registered in the National System of Protection and Civil Defense to receive disaster alerts via SMS in Curitiba, Brazil. Multivariate analysis and spatial statistics techniques were applied using a Geographic Information System to identify registration patterns across different socioenvironmental contexts. Spatial data distribution, performed using ArcGIS Pro, revealed that while registrations are concentrated in more vulnerable peripheral areas, they are proportionally more frequent in central regions that are less exposed to risks. Autocorrelation and clustering (k-means) analyses indicate that socioeconomic variables, such as sanitation, income, and education, negatively influence the number of registrations, suggesting structural barriers to accessing risk information. Furthermore, the elderly population stands out as more likely to register for alerts. The findings highlight inequalities in the dissemination of early warnings and emphasize the need for inclusive strategies to enhance the resilience of vulnerable populations, especially in face of climate change impacts.