A global analysis of how human infrastructure squeezes sandy coasts

Sandy coasts, comprising about one-third of the world’s ice-free shorelines, host interconnected habitats that deliver essential ecosystem services such as flood defence, carbon storage, freshwater supply, recreation, and biodiversity. Rapid human development in the coastal zone, compounded by global drivers like sea-level rise and extreme weather, is degrading these ecosystems. Locally, pollution, eutrophication, and salinization add stress, but infrastructure development near the shoreline is arguably the most significant local driver of habitat loss and fragmentation. Such infrastructure constrains the accommodation space needed for landward migration of beaches and dunes under sea-level rise, a process termed coastal squeeze. Although shoreline change and erosion have been studied extensively at global scales, a comprehensive global assessment of landward squeeze due to human infrastructure had been lacking. This study addresses that gap by quantifying the proximity of infrastructure to sandy shorelines globally, identifying socio-economic correlates, assessing the role of protected areas, and evaluating future losses of infrastructure-free space under sea-level rise projections.

Prior research emphasizes the critical services provided by coastal ecosystems and documents widespread degradation driven by human activities and climate change. Studies have examined shoreline change, erosion risk, and coastal wetland dynamics under sea-level rise, with debates around the persistence of sandy beaches under future scenarios. However, multiple works identify infrastructure as a primary constraint on shoreline migration, highlighting the absence of a global dataset for accommodation space. Evidence from Mediterranean and Gulf of Mexico dune systems links infrastructure to significant habitat and species losses, and prior geomorphological research indicates that adequate beach width (often >300 m) supports dune development and erosion resistance. The current study builds on these insights by supplying a global, infrastructure-focused analysis of coastal squeeze on sandy shores.

The study quantified infrastructural squeeze along sandy coasts by constructing 235,469 transects, each 25 km long and perpendicular to the shoreline, spaced at 1 km intervals, covering about 29% of the world’s ice-free shoreline. Coastlines and sandy shore classifications were derived from OpenStreetMap (OSM) and a global sandy shores dataset, respectively. Infrastructure data combined paved roads from OSM and building footprints from Global Urban Footprint. Two proxies were computed per transect: (1) infrastructure-free width, defined as the distance from the shoreline to the nearest paved human-made structure (buildings, freeways, car roads, paved bike and pedestrian paths); and (2) heavy infrastructure-free width, defined as the distance to the nearest building or freeway (excluding smaller paved roads and paths). To avoid confounding by natural barriers, the team used CoastalDEM v2.1 to identify steep slopes or cliffs. Transects with maximum seaward slopes steeper than 34° (angle of repose of dry sand) within the infrastructure-free zone were excluded, assuming such slopes indicate non-sandy, retreat-limiting conditions. Because CoastalDEM is limited to latitudes below 60°N, 7% of transects above this limit were excluded. Transects that intersected the shoreline before any infrastructure (e.g., narrow land strips, small islands/peninsulas) were also identified to avoid misattributing limited widths to infrastructure. Socio-economic influences were analyzed via a multiple linear regression using country-level GDP per capita and coastal population density (WorldPop, 1-km resolution) sampled along full transects, with both predictors log-transformed; correlation between predictors was weak (R = -0.22). The role of protection was assessed using the World Database on Protected Areas: a transect was considered protected if within a protected area in the first 2 km inland of the sandy shore. Infrastructure-free and heavy infrastructure-free widths were compared between protected and non-protected transects using Wilcoxon rank-sum tests. Urban vs rural comparisons used a threshold of 300 people/km², with group differences tested using Kruskal-Wallis and post hoc Wilcoxon tests (Bonferroni-corrected). Future accommodation space loss was estimated by comparing current infrastructure-free widths to projected shoreline retreat by 2100 from Vousdoukas et al. (2020) under RCP 4.5 and 8.5 (50th percentile), matching projections within 0.05° of transects. The fraction of transects where projected retreat exceeded current infrastructure-free width was computed globally and by continent.

- Of 235,469 transects, 28% were limited first by natural barriers (coastal geometry or steep cliffs) before crossing any infrastructure. Among the remaining 168,654 transects, 93% encountered buildings and/or paved roads within 25 km inland. - The distribution of infrastructure-free widths is positively skewed, with a global median infrastructure-free distance of 392 m from the shoreline to the nearest structure. - 33% of sandy shores have less than 100 m of infrastructure-free space, indicating infrastructure directly on or near the beach. - Considering only heavy infrastructure (buildings and highways), the median width increases to 1.6 km; nevertheless, 28% of shores have heavy infrastructure within the first 100 m. - Regional patterns: Between 32–45°N, shores have a median infrastructure-free width of 70 m; countries in this band (e.g., Japan, South Korea, Lebanon, Syria, Turkey, Italy, France, Spain, USA) rank among the most squeezed. By continent, median infrastructure-free widths are: Europe 131 m, Asia 151 m, North America 402 m, South America 764 m, Africa 1.6 km, and Oceania 2.8 km. - Socio-economic model: A multiple regression with coastal population density and GDP per capita explains 35% of variance in country median infrastructure-free width (39% when focusing on heavy infrastructure). Both higher population density and higher GDP are associated with reduced infrastructure-free width; population density has the slightly larger effect. - Protected areas: Only 16% of the world’s sandy shore is protected. Protected shores exhibit a fourfold greater median infrastructure-free width (1.4 km) compared to non-protected shores (302 m). For heavy infrastructure, protected shores have a sevenfold greater width (8.2 km) versus non-protected (1.1 km). Most protected areas are rural (95%); protected areas have a 3.0× larger width than non-protected rural areas, and 1.7× larger in urban areas. - Future exposure: Projected shoreline retreat by 2100 exceeds current infrastructure-free space along 23% of sandy shores under RCP 4.5 and 30% under RCP 8.5. By continent under RCP 4.5: Europe 31%, Asia 28%, North America 22%, South America 18%, Oceania 12%, Africa 11% of shores could lose all remaining infrastructure-free space. - Ecosystem risk threshold: 46% of the world’s sandy shores currently have less than 300 m of infrastructure-free space, a threshold below which erosion risk increases and dune development may be impeded.

The study provides the first global quantification of landward coastal squeeze driven by infrastructure on sandy coasts, directly addressing the absence of accommodation-space datasets highlighted by prior research. Findings reveal that infrastructure is typically very close to the shoreline, especially in developed and densely populated regions, leaving limited room for natural shoreline migration and sediment exchange under sea-level rise. The strong associations with population density and GDP indicate that as countries develop and coastal populations grow, infrastructure pressure will likely intensify. Although protected areas substantially increase infrastructure-free widths, their prevalence is low and their capacity to offset squeeze is limited in urban contexts, where anthropogenic pressures constrain the establishment and effectiveness of protection. Projected sea-level rise will exacerbate squeeze, with a substantial fraction of sandy shores expected to lose their remaining accommodation space by 2100, particularly under higher emissions scenarios. The implications include reduced capacity for natural dune formation, increased erosion susceptibility, and losses in ecosystem services such as biodiversity support, carbon storage, recreation, freshwater supply, and coastal protection. Overall, the results underscore the importance of integrating accommodation space preservation into coastal management and spatial planning to maintain ecosystem functioning and resilience.

This work delivers a global assessment of how human infrastructure confines sandy coasts, quantifying infrastructure-free and heavy infrastructure-free widths and linking squeeze intensity to socio-economic drivers and protection status. Key contributions include: (1) evidence that infrastructure is typically within hundreds of meters of sandy shorelines (median 392 m), with one-third of shores having <100 m of space; (2) demonstration that population density and GDP together account for a substantial share of cross-country variation; (3) clear benefits of protection in increasing accommodation space, though protection remains limited in coverage and effectiveness in urbanized settings; and (4) identification of regions at heightened risk of losing remaining space to projected shoreline retreat by 2100. Policy and management should move beyond shoreline fixation strategies to approaches that retain or create accommodation space, including strategic protection, setback policies, and managed retreat where feasible. Future research should refine global estimates by incorporating local sediment budgets, dynamic beach-dune responses to sea-level rise, and additional infrastructure types (e.g., unpaved coastal defences), and evaluate socio-economic pathways that minimize squeeze while maintaining coastal ecosystem services.

Key limitations include: (1) infrastructure data did not account for unpaved coastal defences (e.g., seawalls and dikes), and areas where beaches have already disappeared due to hard structures may be underrepresented; (2) exclusion of transects above 60°N due to CoastalDEM coverage, removing about 7% of transects; (3) identification of natural barriers based on slope thresholds may simplify complex geomorphology; (4) the regression analysis is correlative, with limited socio-economic variables (country-level GDP, gridded population density) and potential unobserved confounders; and (5) future exposure estimates rely on global shoreline retreat projections that carry uncertainties and do not incorporate site-specific sediment budgets or detailed coastal dynamics.