Multi-hazard risk to global port infrastructure and resulting trade and logistics losses

Ports are crucial for global economies and supply chains, but their coastal locations expose them to natural hazards like extreme waves, cyclones, and earthquakes. Past disruptions, such as Hurricane Ike (2008) and the 2013-2014 UK floods, highlight the significant economic consequences of port closures. The predicted increase in natural hazard risk due to climate change and the projected growth in maritime trade necessitate detailed global-scale risk assessment to guide infrastructure planning and investment. Existing research has limitations: large-scale analyses often lack asset-level detail and economic damage quantification, while detailed local studies lack global scope. This study addresses these gaps by providing the first asset-level multi-hazard risk analysis for 1340 global ports, quantifying both physical damages and associated logistics and trade losses.

Previous large-scale studies primarily focused on port exposure to coastal flooding or multiple operational thresholds, representing ports as single points without quantifying economic damages to assets or associated disruptions. Detailed asset-level analyses exist but only on a local scale. Scaling up this detailed analysis globally faces challenges: the lack of a global port asset database, the omission of the link between port operations and critical infrastructure networks, and the low resolution of global hazard datasets. Existing studies on logistics losses are also limited to a small number of ports. This paper aims to bridge these gaps through the development of a new database and a comprehensive risk framework.

This study creates a new port infrastructure and land-use database encompassing port assets (terminals, breakwaters, cranes), critical infrastructure (road, rail, electricity), and industrial facilities within a 1km buffer around ports. A multi-hazard risk framework incorporating operational disruptions (extreme wind, temperature, waves, overtopping) and physical damages (tropical cyclones, earthquakes, river flooding, pluvial flooding, coastal flooding) is employed. A fault tree methodology captures multiple failure pathways impacting port operations. Port-level trade flows are combined with risk estimates to quantify disrupted trade and associated logistics losses (port operators, carriers, shippers). A sensitivity analysis addresses uncertainties, incorporating port resilience through variation of engineering design standards, reconstruction costs, and recovery duration. The analysis uses various data sources including satellite imagery (Google Satellite, OpenStreetMap), GRIB database for roads, Gridfinder for electricity transmission lines, and various databases for power plants and port-specific data. Hazard data is gathered from UNISDR Global Assessment Report 2015 for earthquakes, synthetic TC paths for tropical cyclones, JBA Consulting for fluvial and pluvial flooding, and a custom-built model for coastal flooding. Fragility curves and recovery duration estimates are drawn from various sources, and logistics losses are estimated using a simplified approach based on previous research. A variance-based sensitivity analysis (Sobol) with 10,000 parameter samples is employed to assess uncertainty and sensitivity.

The analysis reveals that 94.8% of ports are exposed to more than one natural hazard, with 50% exposed to 4 or 5 hazards. The median port-specific risk is $7.6 billion per year, with tropical cyclones ($2.4 billion), fluvial flooding ($1.9 billion), and coastal flooding ($0.8 billion) as the most significant contributors. The top five most at-risk ports are located in Asia and North America. While absolute risk is larger in high-income countries due to larger port areas and higher infrastructure density, relative risk (per square meter) is higher in small ports in low-income countries. Physical damages to port infrastructure contribute 58.6% to port-specific risk, followed by logistics losses (22.2%) and critical infrastructure damages (19.2%). Globally, $66.9 billion of trade is at risk annually (approximately 0.8% of total maritime trade). Small Island Developing States face disproportionately high trade risk (3.7 times higher than non-SIDS). The sensitivity analysis shows that uncertainties are geographically varied, with large uncertainties in Africa, the Middle East, East Asia, and Northern Europe. Engineering standards are the major driver of uncertainty for port-specific risk, while operational and recovery resilience are more important for trade risk. Tropical cyclones are a dominant factor in trade risk due to their impact on port downtime.

The findings underscore the need for a multi-hazard approach to port risk analysis. The significant economic risk to ports, both in terms of direct damage and trade disruptions, highlights the urgency of investment in resilience measures. The disparity in relative risk between high-income and low-income countries highlights the need for tailored solutions considering economic capacity. The importance of critical infrastructure in port functionality is emphasized, highlighting the need for integrated risk management strategies. The high trade risk, particularly in SIDS, underscores the global systemic implications of port disruptions. Climate change projections would exacerbate this risk.

This study provides the first comprehensive global-scale assessment of natural hazard risk to port infrastructure and associated trade. The substantial economic risk necessitates prioritization of investments in resilience. Future research should focus on incorporating compound and consecutive disasters, tsunami risks, and detailed climate change projections to refine risk estimates and inform adaptation strategies.

The study makes several simplifying assumptions, such as treating hazard occurrences as independent and using simplified models for logistics losses. The data used, while extensive, has limitations in terms of resolution and global consistency. The sensitivity analysis, while comprehensive, can't fully capture all sources of uncertainty. Furthermore, the inclusion of only specific hazards may underrepresent the total possible risks.