Climate threats to coastal infrastructure and sustainable development outcomes

Climate hazards, including floods, cyclones, and erosion, pose significant threats to development outcomes in coastal regions globally. These hazards disproportionately impact poor and vulnerable populations due to factors like exposure bias – the tendency for poorer communities to reside in more exposed locations. Repeated climate impacts can create poverty traps, hindering asset accumulation and negatively affecting health, education, employment, and potentially leading to migration. National development plans risk being derailed by such impacts, underscoring the need for targeted interventions. However, scarce resources necessitate informed spatial analyses of climate risks and vulnerabilities to ensure effective allocation. While social vulnerability has been mapped and integrated with flood risk assessment, high-resolution geolocated household-level data is often lacking. This limitation hinders accurate assessment of local vulnerability differences and accurate targeting of adaptation efforts. The Sustainable Development Goals (SDGs) offer a comprehensive framework for assessing development progress, and several studies have examined how interventions can contribute to multiple SDGs. However, a more granular, spatially explicit understanding of how climate hazards impact SDG achievement through infrastructure service disruptions is needed. This study addresses this gap by analyzing the impact of climate hazards on infrastructure service delivery and its subsequent effects on SDG attainment in coastal Bangladesh, a region acutely vulnerable to multiple climate hazards, where inadequate infrastructure services hamper human development and resilience efforts.

Existing research highlights the disproportionate impact of climate change on the poor and vulnerable (Winsemius et al., 2018; Hallegatte et al., 2017; Verschuur et al., 2020). Studies have documented poverty traps stemming from repeated climate impacts in various regions, including coastal Bangladesh (Borgomeo et al., 2018; Barbour et al., 2022). The potential for climate extremes to disrupt national development plans is also recognized (Casado-Asensio et al., 2016; Ishiwatari & Surjan, 2019). While geospatial mapping of social vulnerability and its integration with flood risk management exist at various scales (Koks et al., 2015; Englund et al., 2022), the high-resolution integration of household-level data with hazard, exposure, and SDG objectives remains incomplete (Ferreira et al., 2018; Murshed et al., 2021). The SDGs provide a framework for comprehensive development assessment, with studies examining the contribution of cross-cutting interventions to multiple SDGs (Thacker et al., 2019; Fuso Nerini et al., 2017, 2019; Adshead et al., 2019, 2021; Fuldauer et al., 2022). However, the need for measurable indicators and spatially explicit approaches for more targeted planning and adaptation strategies is widely acknowledged (Fuldauer et al., 2021, 2022).

This study uses a multi-hazard spatial assessment approach combining various datasets to understand the impact of climate hazards on SDG progress in coastal Bangladesh. The methodology involves several key components: 1. **Spatial Database Creation:** A database was created integrating climate hazard data (coastal and riverine flooding from the World Resources Institute's Aqueduct Floods platform, cyclone data based on historical cyclone simulations, and erosion data from the Deep WaterMap model) with point and network data for critical infrastructure assets (electricity substations, market centers, healthcare facilities, cyclone shelters, and educational institutions) sourced from various government and public sources. The spatial resolution varied across datasets, with some having higher resolution (30m x 30m) than others (4.4km). 2. **Household Data Integration:** A high-resolution synthetic household dataset from the World Bank (8.2 million households), validated against independent household surveys, was used. This dataset provided socioeconomic characteristics (access to services, wealth quintiles, etc.) associated with household locations. 3. **Nearest-Neighbor Analysis:** Nearest-neighbor analysis linked each household to its nearest service-providing asset in each subsector, enabling the assessment of household dependence on specific infrastructure. A simplified radius-based approach was used due to limitations in road network data and the clustered nature of household positions in the synthetic dataset. 4. **Wealth Quintile Assignment:** A wealth index was constructed using household survey data to assign households to wealth quintiles (poorest to wealthiest), offering a more objective measure of wealth than income or consumption data in this context. Principal component analysis was employed after data normalization. 5. **Exposure Bias Calculation:** Risk ratios were calculated to quantify the likelihood of service disruption for different wealth groups compared to the median group. Statistical significance of poverty and wealth biases was tested using a hypothesis test comparing binomial distributions. 6. **SDG Progress Assessment:** Current progress towards five SDG indicators directly linked to infrastructure services (SDG 3, 4, 7, 8, and 13) was assessed at the upazila level using the household dataset, considering access to services and distance to assets. The "progress at risk" was calculated as the share of households with current access facing potential service disruptions due to hazard exposure under specified scenarios (2030 timeframe, 50-year return period, RCP 4.5 for floods; historical data for cyclones and erosion).

The study revealed several key findings: 1. **Widespread Infrastructure Service Disruption:** All 8.2 million households in coastal Bangladesh are potentially exposed to some infrastructure service disruption from climate hazards. For a baseline 1-in-50-year hazard, coastal flooding, riverine flooding, and cyclones are projected to disrupt significant portions of the population. Erosion has cumulatively disrupted a large share of the population since 1987. 2. **Poverty Bias:** A statistically significant poverty bias was observed in 69% of coastal subdistricts (upazilas), indicating poorer households are more exposed to service disruptions than wealthier households across various infrastructure types (cyclone shelters, education facilities, market centers, healthcare facilities). The exception was electricity substations, where a wealth bias was observed in some areas due to the location of substations serving wealthier communities. 3. **Threats to SDG Progress:** Climate hazards pose significant threats to achieved progress towards several SDG indicators. Coastal flooding most impacts SDG 7 (electrification), river flooding affects SDG 3 (health) and 4 (education), erosion mainly impacts SDG 3, and cyclones broadly impact several indicators (SDG 3, 4, 7, 8, and 13). 4. **Prioritization of Adaptation Measures:** Focusing on a limited number of the most vulnerable upazilas can yield substantial gains in safeguarding SDG progress. Targeting 33% of the most vulnerable areas could safeguard 50–85% of progress towards the studied SDG indicators. 5. **Underserved Populations:** A substantial share of households lack access to key services (e.g., 46% lack electricity access). Integrating high-resolution hazard data with areas of low service access can inform spatial planning of new infrastructure to better serve these populations and minimize exposure to risks.

This study demonstrates the significant threat climate hazards pose to sustainable development outcomes in coastal Bangladesh, highlighting the disproportionate impact on poor and vulnerable populations. The findings underscore the limitations of national-level SDG reporting and the need for high-resolution spatial analyses to inform targeted interventions. The results show that even with limited investment focusing on highly vulnerable areas can generate substantial benefits, safeguarding significant portions of progress made towards several SDGs. The integration of high-resolution geospatial data with household characteristics and SDG targets offers a powerful tool for identifying exposure bias and tailoring adaptation strategies. The methodology can be adapted to other regions and contexts, emphasizing the importance of locally informed, pro-poor adaptation plans.

This study highlights the critical need to integrate high-resolution spatial hazard analysis into national development planning to address the disproportionate impact of climate hazards on vulnerable populations. The findings emphasize the potential of targeted adaptation measures to safeguard progress toward the SDGs in coastal Bangladesh and provide a replicable methodology for data-scarce regions. Future research should explore the effectiveness of various adaptation measures, assess the economic costs and benefits of different prioritization strategies, and further refine methods for integrating various data sources to enhance the accuracy and precision of vulnerability assessments.

While the study uses a comprehensive dataset, some limitations exist. The differing spatial and temporal resolutions of the climate hazard datasets present a challenge, and future work should strive for greater consistency. The simplified radius-based approach to assessing service access could be improved by incorporating road networks and more detailed accessibility modelling. The study focuses on five SDGs and five infrastructure types; future work could expand this to encompass other SDGs and infrastructure types. Lastly, the study uses a synthetic household dataset, and future research could benefit from integrating data from additional household surveys for enhanced accuracy and precision.