The research team, led by Md. Najmul Kabir of Hokkaido University and Shahjalal University of Science and Technology, focused specifically on communities living outside the coastal embankments that were constructed in the 1960s under the Coastal Embankment Project. These embankments were designed to shield low-lying deltaic land from tidal surges and cyclone-driven flooding, but they have repeatedly failed during extreme events, and sedimentation and land subsidence continue to reduce their effectiveness. The people who live beyond these structures face the full force of coastal hazards without comparable structural protection, yet they remain among the most understudied populations in Bangladesh’s extensive disaster-risk literature.

To build their index, the researchers integrated three complementary components: a Natural Hazard Vulnerability Index derived from cyclone wind speeds, storm surge heights, flood depths, salinity levels, and rainfall patterns; a Physical and Geological Vulnerability Index capturing elevation, slope, geomorphology, shoreline change, tidal range, and sea-level rise; and a Socio-Economic Vulnerability Index built from face-to-face interviews with 320 households across four mouzas. Principal Component Analysis provided data-driven weights for the hazard and socio-economic indicators, ensuring that the final composite index reflected the empirical structure of the local data rather than subjective judgments about which factors matter most.

The natural hazard component revealed a stark escalation of risk with longer return periods. Using Log-Pearson Type III distributions fitted to annual maximum series, the team found that the Natural Hazard Vulnerability Index rose from 0.35 at the 2-year return period to 0.75 at the 100-year return period. Cyclone wind speeds increased from approximately 92 kilometres per hour at the 2-year event to roughly 264 kilometres per hour at the 100-year scenario, while storm surge heights climbed from around 2.75 metres to nearly 6 metres at the 50-year return period. Salinity intrusion, which averaged below 1.5 decisiemens per metre annually but peaked at 3.49 decisiemens per metre during the pre-monsoon season, showed clear seasonal variability and a rising trend in recent years, reaching approximately 2.50 decisiemens per metre in 2017.

Population exposure expanded dramatically across these scenarios. The proportion of people exposed to cyclone winds rose from 13.31 percent at the 2-year return period to 66.84 percent at the 200-year scenario, while storm-surge exposure increased from roughly 24 percent to 71 percent over the same range. Across all hazard components combined, the exposed population grew from approximately 23.78 percent at the 2-year return period to 70.45 percent at the 200-year scenario, demonstrating that even moderate increases in hazard intensity translate into substantial increases in the number of people at risk.

The physical and geological assessment painted an equally sobering picture. The terrain across the study area averaged just 1.09 metres in elevation with a mean slope of only 0.25 degrees, conditions that favour water accumulation and prolonged inundation by limiting natural drainage. Shoreline change analysis using eight Landsat images acquired between 1980 and 2017 revealed end-point rates ranging from minus 4.8 to plus 18 metres per year, with some areas gaining land through accretion while others retreated. Perhaps most alarmingly, relative sea-level rise measured from tide-gauge records between 1958 and 2017 reached 15.08 millimetres per year, a rate nearly five times the global mean of approximately 3.2 millimetres per year, potentially reflecting regional deltaic subsidence processes that compound the threat of rising oceans.

The socio-economic dimension told a more nuanced story. Social vulnerability scored relatively low at 0.31, and economic vulnerability similarly registered 0.32, reflecting the adaptive capacity embedded in community networks, diversified livelihoods, and the resourcefulness of fishing households who have long adapted to seasonal variability. However, housing and infrastructure vulnerability stood out at 0.65, the highest of the three socio-economic domains. Ninety percent of houses were single-story structures built from bamboo, tin, wood, or palm leaves, nearly half were in poor condition, and more than 60 percent lacked disaster-resistant design features. Drinking water came primarily from tube wells located inside the embankment, requiring an average journey of 2.08 kilometres, while the absence of paved roads left residents dependent on storm-vulnerable embankment pathways for evacuation and daily transport.

When the three components were integrated into the composite Coastal Vulnerability Index, the mean value across the study area was 0.47, corresponding to low vulnerability on the classification scale. However, this average concealed substantial spatial variation. Nearly half the study area, 49.22 percent, fell into the moderate vulnerability category, while 15.81 percent was classified as high and 2.47 percent as very high vulnerability. Only 3.63 percent registered as very low vulnerability. Correlation analysis showed that the Natural Hazard Vulnerability Index had the strongest positive association with the composite index at a correlation coefficient of 0.83, followed by the Physical and Geological Vulnerability Index at 0.75, while the Socio-Economic Vulnerability Index showed a weak negative association at minus 0.12, indicating that stronger household coping capacity modestly reduced the composite vulnerability in areas where hazard intensity and physical fragility were otherwise high.

The researchers subjected their index to rigorous validation and uncertainty analysis. Monte Carlo simulation with 500 runs of random Gaussian perturbations showed median class-flip rates near zero for all indices, confirming that the vulnerability classifications were stable under small perturbations. External validation against reconstructed hazard maps of the 2002 riverine flood and the 2007 Cyclone Sidr storm surge demonstrated significant positive correlations, with the index discriminating high-intensity storm-surge zones with an area under the receiver operating characteristic curve of 0.895, indicating excellent performance. The stronger correspondence with Cyclone Sidr surge patterns compared to the 2002 flood suggests that the index is particularly well suited to capturing the coastal hazard processes that dominate this landscape.

The study’s most consequential finding may be its demonstration that uniform adaptation strategies would be less effective than interventions tailored to community-specific conditions. Gazi Mahmud, with the highest mean composite vulnerability at 0.61 and the greatest concentration of high-vulnerability land, requires priority attention to resilient housing, safe-water access, and evacuation connectivity. Kumirmara, with active shoreline dynamics and the largest proportion of very high vulnerability land, needs erosion management and restrictions on settlement expansion onto unstable newly accreted surfaces. Padma requires drainage maintenance and reliable evacuation routes, while Paler Baliatali, despite its larger proportion of comparatively low-vulnerability land, still needs continued shoreline monitoring and improved basic services. As relative sea level continues to rise at this extraordinary local rate and cyclone intensity increases under a warming climate, the window for targeted, evidence-based adaptation in these unprotected communities is narrowing with each passing season.

Subject of Research: Geospatial assessment of coastal vulnerability using a composite index integrating natural hazards, physical-geological conditions, and socio-economic factors in outside-embankment communities of Barguna, Bangladesh

Article Title: Geospatial assessment of natural hazard risks in coastal Bangladesh with a focus on communities outside embankments in Barguna

Article References: Kabir, M. N., Islam, S., Sikder, M. B., Shahnewaz Sayem, S. M., Barua, E., & Islam, M. M. (2026). Geospatial assessment of natural hazard risks in coastal Bangladesh with a focus on communities outside embankments in Barguna. Discover Geoscience, 4(1), Article 405. https://doi.org/10.1007/s44288-026-00774-1

Image Credits: AI Generated

DOI: 10.1007/s44288-026-00774-1

Keywords: coastal vulnerability, Bangladesh, Barguna, natural hazards, sea-level rise, cyclone, storm surge, embankments, geospatial analysis, socio-economic vulnerability, Log-Pearson Type III, adaptation planning