Disappearing cities on US coasts

The paper addresses how near-term sea-level rise (SLR) combined with coastal land subsidence will drive inundation hazards in major US coastal cities by 2050. Contextualizing within accelerating global mean SLR (currently about 3.7 mm per year) and growing coastal populations and assets, the authors emphasize that relative SLR (geocentric SLR plus vertical land motion, VLM) determines local flooding risk. Given that sea levels across emission scenarios diverge more by 2100 than by mid-century, short-term, high-resolution local assessments that incorporate VLM are crucial for policy and adaptation planning. However, comprehensive VLM measurements are sparse, leading to uncertain risk estimates. The study aims to quantify exposure of land area, population and properties to high-tide inundation in 32 major US coastal cities by integrating high-resolution InSAR-derived VLM, LiDAR-based elevation, and IPCC geocentric SLR projections, and to evaluate the role of subsidence, differences with IPCC relative SLR datasets, and implications for adaptation and climate equity.

The authors synthesize evidence on accelerating global and regional sea-level rise, increasing coastal hazards (heatwaves, hurricanes, storm surge, tidal flooding, erosion), and the significant socioeconomic exposure of US coastal populations. Prior work shows global mean SLR ~0.17 m over the last century with accelerating rates; US coasts are experiencing above-global-average SLR. Literature highlights that land subsidence, driven by natural processes (e.g., GIA) and anthropogenic activities (e.g., groundwater withdrawal, drainage, hydrocarbon extraction), exacerbates relative SLR and flood risk, yet is often underrepresented in planning. IPCC AR6 projection products and NOAA assessments provide geocentric and relative SLR scenarios, but tide-gauge-based VLM may not capture urban-scale spatial variability. Previous studies demonstrate the need for evaluating flood resilience strategies and adaptation pathways for coastal cities and note disproportionate climate impacts on vulnerable communities, with limited focus on within-country inequities. This study builds on InSAR-based VLM mapping advances on US coasts to refine exposure estimates and assess discrepancies with IPCC-derived relative SLR products.

• Study domain: 32 major US coastal cities across Atlantic (11), Gulf (11), and Pacific (10) coasts.
• Datasets:
  • Vertical land motion (VLM): High-resolution InSAR-derived VLM maps (millimetre-level accuracy) across US coasts, capturing spatially variable uplift/subsidence within urban areas.
  • Elevation: High-resolution LiDAR digital elevation models (DEMs).
  • Sea-level rise: IPCC AR6 geocentric SLR projections (baseline 2020), primarily SSP2-4.5 for 2050 assessments; comparisons with IPCC relative SLR dataset for sensitivity/benchmarking.
  • Tides: High-tide estimates to delineate present-day and future high-tide inundation extents.
  • Exposure data: 2010 US Census for population; properties and home values from ZIP Code Zillow Home Value Index (ZHVI).
• Scenarios and assumptions:
  • Relative SLR computed as geocentric SLR plus projected VLM to 2050, assuming linear continuation of current VLM rates.
  • Undefended scenario: ignores effects of flood-control structures.
  • Defended scenario: models cities with existing levees/floodwalls (131 structures across 32 cities) to estimate protective capacity through 2050.
  • Bounds: Report lower, median, and upper estimates using equation (3) (details in Methods), presented as ranges.
• Inundation modelling:
  • Combine DEMs with projected relative SLR to map areas at or below future high-tide levels (static inundation approach) for 2020 baseline and 2050 projections.
  • Quantify additional exposed land area and intersect with census blocks and property datasets to estimate exposed population, property counts, and home values.
• Attribution analyses:
  • Partition total inundated area into contributions from subsidence-only versus combined subsidence + geocentric SLR by 2050.
  • Compare InSAR-based citywide VLM averages with IPCC tide-gauge-derived VLM at PSMSL stations; assess differences in exposure estimates (area, population, properties) between InSAR-derived and IPCC-derived relative SLR.
• Inequality analyses:
  • Evaluate racial disparities using eight US Census racial/ethnic categories for exposed populations by coast and city.
  • Assess economic inequality via estimated home-value exposures; highlight cities with disproportionate impacts (e.g., New Orleans, Port Arthur).

• Overall exposure (undefended scenario, SSP2-4.5, 2050):
  • Additional exposed land area across 32 cities: 1,334–1,813 km² (median 1,570 km²).
  • Exposed population: 176,000–518,000 people (median 272,000).
  • Exposed properties: 94,000–288,000 (median 145,000).
  • Home-value exposure: US$32–109 billion (median ~US$50 billion).
  • Maximum exposure equates to ~1 in 50 people and ~1 in 35 properties within the 32 cities by 2050.
• By coast (undefended, 2050):
  • Atlantic (11 cities): Additional 773–951 km²; 59,000–263,000 people; 32,000–163,000 properties; US$14–64 billion in home values. Miami dominates exposure: 38–44% (340–360 km²) of area, 38–46% (22,000–122,000) of population, and 41–49% (13,000–81,000) of properties.
  • Gulf (11 cities): Additional 528–826 km²; 110,000–225,000 people; 58,000–109,000 properties; US$14–21 billion. New Orleans already has substantial present-day exposure (318–426 km²; 386,000–448,000 people; 176,000–209,000 properties below high tide) due to low-lying areas.
  • Pacific (10 cities): Additional 20–40 km²; 6,000–30,000 people; 3,000–15,000 properties; US$4.5–22 billion; lower hazard attributed to higher topography, lower subsidence, and lower geocentric SLR rates.
• Defended scenario (accounting for existing flood-control structures): By 2050, relative SLR threatens an additional 1,006–1,389 km², 55,000–273,000 people, and 31,000–171,000 properties, with 61–63% of area, 79–89% of people, and 80–89% of properties on the Atlantic coast; many Atlantic cities lack adequate flood-protection systems.
• Role of land subsidence in 2050 inundation (fraction of total area when both subsidence and SLR considered): Atlantic 11.9–15.1%; Gulf 22.9–35.4%; Pacific 4.8–8.1%.
• At several Gulf tide-gauge stations, current local subsidence outpaces geocentric SLR through at least 2045 and 2070 (assuming linear subsidence), underscoring subsidence as a dominant near-term driver.
• Comparison with IPCC-derived relative SLR exposure:
  • Broad agreement within uncertainty across many cities; however, notable city-level divergences, especially along the Gulf.
  • Where IPCC tide-gauge VLM underestimates within-city subsidence (e.g., Atlantic City, Charleston, Savannah, Mobile, Biloxi), IPCC-based exposure is underestimated; where it overestimates (e.g., Virginia Beach, Jacksonville, Miami, New Orleans, Slidell, Lake Charles, Port Arthur, Huntington Beach), exposure is overestimated.
  • InSAR provides higher-resolution VLM and improves exposure estimates versus tide-gauge proxies.
• Climate inequality findings:
  • Gulf coast: Minoritized groups constitute 43% of the total population but 64.2–71.5% of the exposed population (all minorities combined); Black or African American residents alone account for 50.0–57.7% of exposed. Asians are overrepresented among the exposed on Gulf (2.6–4.4%) and Pacific (21.4–26.3% at median/upper bounds) coasts relative to their population shares.
  • Economic disparities: Disproportionate home-value exposure in New Orleans and Port Arthur parallels higher exposure in predominantly minoritized communities.

The study demonstrates that integrating spatially resolved VLM with geocentric SLR and high-resolution topography substantially refines near-term inundation risk estimates. The results show that not accounting for spatial variability in subsidence leads to systematic under- or overestimation of exposure in many cities, especially on the Gulf coast where anthropogenic subsidence is significant. Even with existing flood-control structures, sizable exposures remain by 2050, indicating that current protection is insufficient or absent in many places, notably along the Atlantic coast. The relatively modest exposure on the Pacific coast reflects higher elevations and lower subsidence but does not eliminate other hazards such as cliff retreat or increasing high-tide flooding. The findings highlight the critical role of subsidence management in coastal resilience and the policy relevance of high-resolution InSAR-derived VLM. The documented racial and economic disparities in exposure underscore climate-change inequality within the US, implying that adaptation strategies must address both physical risk and social vulnerability. Overall, the results stress the urgency for proactive, long-term, multi-faceted adaptation and infrastructure upgrades to contend with accelerating relative SLR.

By combining InSAR-derived VLM, LiDAR elevation, and IPCC geocentric SLR projections, the study quantifies mid-century inundation exposure for 32 US coastal cities and reveals the substantial contribution of land subsidence to near-term relative SLR. It shows that many cities face significant increases in exposed land, population, and properties by 2050, with particularly large exposures along the Atlantic and Gulf coasts. The comparison with IPCC tide-gauge-based VLM highlights the value of high-resolution VLM for accurate urban exposure assessments. The work also documents disproportionate impacts on minoritized and lower-income communities, emphasizing climate inequity. The authors advocate for proactive, sustained adaptation that includes upgraded and new structural protections, nature-based solutions, land-use planning, and, critically, mitigation of anthropogenic subsidence, alongside broader climate mitigation to limit long-term SLR. Future research should refine dynamic flood processes beyond static inundation, incorporate infrastructure functionality and cascading failures, improve socioeconomic exposure datasets, and monitor evolving VLM drivers and adaptation effectiveness.

• Undefended analyses do not incorporate existing or future adaptive measures; the defended scenario accounts for current levees/floodwalls but not potential future upgrades, and effectiveness is uncertain.
• Static inundation approach focuses on high-tide levels; dynamic processes (storm surge, waves, compound flooding, drainage capacity) are not explicitly modelled.
• VLM is projected linearly to 2050; potential nonlinearity or changes in subsidence drivers are not captured.
• Exposure estimates use 2010 Census population and ZHVI-based property/home values; demographic and market changes since 2010 may alter actual 2050 exposure.
• Home-value exposure excludes critical infrastructure valuation and indirect economic impacts; thus, values are conservative.
• IPCC tide-gauge VLM may not represent intra-urban variability; while InSAR improves resolution, coverage, coherence, and error structures can vary locally.