<![CDATA[Indonesia, as the world's largest archipelagic nation, faces significant disaster risks due to its position at the convergence of three major tectonic plates. This study employs Generalized Linear Latent Variable Models (GLLVM) to analyze relationships among 12 Hazard, Exposure, and Vulnerability Assessment (HEVA) indicators across 34 Indonesian provinces. The HEVA dataset used in this study was obtained from the United Nations University – Institute for Environment and Human Security (UNU-EHS), which provides harmonized global risk indicators for hazard intensity, exposure levels, and socioeconomic–environmental vulnerability. Unlike conventional approaches assuming variable independence, GLLVM captures complex dependency structures through latent variables, providing deeper insights into multidimensional disaster risk patterns. Model-based ordination analysis reveals distinct spatial risk patterns. Eastern provinces (Papua, Maluku) demonstrate high physical vulnerability and exposure despite lower hazard levels, while Java provinces show moderate hazards but lower vulnerability due to better infrastructure and governance. A notable negative correlation (r < -0.70) between hazard levels and vulnerability indicators suggests that regions frequently exposed to disasters develop stronger adaptation capacity. Conversely, vulnerability indicators show very strong positive correlations (r > 0.90), indicating interconnections requiring holistic interventions. Incorporating geographical covariates such as population, number of islands, and provincial areas reveals significant relationships with HEVA indicators. Population shows negative associations with physical and environmental vulnerability but positive relationships with climate and geophysical hazards, i.e., the corresponding 95% CIs do not contain zero, reflecting urbanization's dual nature. The number of islands positively correlates with multiple vulnerability indicators, highlighting structural challenges in archipelagic disaster management, including limited accessibility and infrastructure connectivity. Provincial areas demonstrate positive relationships with vulnerability indicators but negative associations with economic exposure, indicating concentrated economic activities in urban centers. These findings emphasize differentiated spatial approaches for disaster mitigation.]]>
