Coastal Cities With the Highest Modeled Flood Damage by 2050
Coastal Cities With the Highest Modeled Flood Damage by 2050
Among the 15 high-risk cities reported in the primary study, Guangzhou has the highest modeled expected annual coastal flood damage in 2050 under RCP8.5 at $330.8 billion per year. New Orleans follows at $209.2 billion, while Mumbai ranks third at $112.4 billion after the published 2050 values are sorted from highest to lowest.
The metric is expected annual economic damage from coastal flooding under uncertain relative sea-level rise. It combines future socioeconomic exposure with city-specific flood-damage relationships and probabilistic sea-level projections. It does not rank cities by population exposed, inundated land area or centimeters of sea-level rise.
Abadie et al. analyze 136 major coastal cities. Their main published table reports detailed 2050 RCP8.5 expected-damage values for 15 high-risk cities, and those confirmed values form the ranking below. The original table is ordered by 2100 RCP8.5 damage; this page re-ranks the same 15 cities specifically by their published 2050 values.
The table is a compiled research dataset based on 4 sources, with row-level source and method notes shown in the ranking table. All 15 entries are modeled projections. There are no official values and no official forecasts in this ranking.
Unit: billions of U.S. dollars per year. Target year: 2050. Direction: higher modeled expected damage ranks higher. Scenario: RCP8.5. Main Abadie estimates assume no adaptation.
How to read the ranking: a higher position means a larger modeled economic loss from coastal flooding under the selected scenario. It does not mean the city is expected to experience the world's largest physical rise in mean sea level.
Guangzhou has the largest expected annual damage among the 15 published cities.
Shanghai has the smallest 2050 value within this confirmed 15-city cohort.
The study assesses 136 cities; this ranking uses the 15 city-level values reported in its main table.
2050 RCP8.5 projections in US$ billions per year; 0 official values and 0 official forecasts.
What the coastal flood damage ranking measures
Expected annual damage estimates the average economic loss generated by the modeled distribution of coastal flood outcomes. It is broader than the cost of a single storm and different from simple exposure. A city can contain large amounts of property in a coastal floodplain but still have a different expected-loss profile because protection standards, extreme water levels and the depth-damage relationship also matter.
The Abadie model combines probabilistic relative sea-level projections with damage functions derived from the coastal-city framework developed by Hallegatte and colleagues. Future socioeconomic conditions affect the amount and value of development at risk, while sea-level change alters the distribution of damaging coastal water levels.
This distinction explains why the ranking should not be interpreted as a list of cities with the greatest projected sea-level rise. The research compares economic consequences. Asset concentration, urban growth, coastal protection and vulnerability can produce large differences in damage even when physical sea-level changes are closer together.
Across all 136 cities, Abadie et al. estimate aggregate expected damage of about $1.6 trillion per year by 2050 under RCP8.5. The 15 values reported in the study's main table total $1.2428 trillion per year, illustrating the strong concentration of modeled losses within the published high-risk group.
Top 10 by modeled coastal flood damage in 2050
Re-ranking the study's published 2050 RCP8.5 values places Guangzhou first, New Orleans second and Mumbai third. Osaka-Kobe and Tokyo complete the top five within the 15-city cohort.
Top 10 confirmed 2050 RCP8.5 expected annual damage values
| Rank | City | Damage | Source / method note |
|---|---|---|---|
| 1 | Guangzhou, Guangdong | $330.8B/yr | China · Abadie et al. (2020) · modeled projection · target 2050 · RCP8.5 · no adaptation. |
| 2 | New Orleans | $209.2B/yr | United States · Abadie et al. (2020) · modeled projection · target 2050 · RCP8.5 · no adaptation. |
| 3 | Mumbai | $112.4B/yr | India · Abadie et al. (2020) · modeled projection · target 2050 · RCP8.5 · no adaptation. |
| 4 | Osaka-Kobe | $101.1B/yr | Japan · Abadie et al. (2020) · modeled projection · target 2050 · RCP8.5 · no adaptation. |
| 5 | Tokyo | $79.8B/yr | Japan · Abadie et al. (2020) · modeled projection · target 2050 · RCP8.5 · no adaptation. |
| 6 | Kolkata | $65.9B/yr | India · listed as Calcutta in Abadie et al. (2020) · modeled projection · target 2050 · RCP8.5. |
| 7 | Nagoya | $64.1B/yr | Japan · Abadie et al. (2020) · modeled projection · target 2050 · RCP8.5 · no adaptation. |
| 8 | Alexandria | $50.0B/yr | Egypt · Abadie et al. (2020) · modeled projection · target 2050 · RCP8.5 · no adaptation. |
| 9 | Guayaquil | $49.1B/yr | Ecuador · Abadie et al. (2020) · modeled projection · target 2050 · RCP8.5 · no adaptation. |
| 10 | Tianjin | $45.5B/yr | China · Abadie et al. (2020) · modeled projection · target 2050 · RCP8.5 · no adaptation. |
Values retain the one-decimal precision published by Abadie et al. The rank is calculated from the 2050 RCP8.5 expected-damage column, not from the paper's printed 2100 ranking order.
Chart: 15 confirmed 2050 projections
Guangzhou and New Orleans form a distinct upper tier. The remaining published values range from $112.4 billion per year for Mumbai to $15.7 billion for Shanghai.
Methodology
The compiled research dataset uses Abadie et al. (2020) as the controlling source for every ranked 2050 value. Hallegatte et al. (2013) supplies the underlying coastal-city damage framework and the socioeconomic reference component. IPCC sea-level scenarios provide the RCP8.5 climate context, while the associated Zenodo record documents the broader research dataset.
Metric and unit
Expected annual economic damage from coastal flooding, expressed in billions of U.S. dollars per year.
Ranking direction
Higher modeled expected damage ranks higher. The order is recalculated from the published 2050 values.
Target year and scenario
Target year: 2050. Climate pathway: RCP8.5. The main ranked estimates assume no adaptation.
Coverage
The research analyzes 136 cities. The ranking includes the 15 cities whose detailed 2050 RCP8.5 values are printed in the main study table.
Projection method
City-specific probabilistic relative sea-level projections are combined with coastal flood damage functions and future socioeconomic conditions.
Simulation
The model generates one million relative sea-level outcomes for each city, year and scenario and evaluates the associated distribution of economic damage.
Reference input
The row-level reference value records the 2050 socioeconomic-only, no-sea-level-rise damage estimate from the Hallegatte framework. It is a counterfactual model input, not a historical observation or a CAGR starting point.
Source hierarchy
Abadie et al. controls ranked values; Hallegatte et al. controls the reference damage framework; IPCC sources control scenario context; Zenodo provides supporting research data.
Formula: expected damage in 2050 is the mean economic loss across the modeled 2050 damage distribution generated from city-specific socioeconomic conditions and simulated relative sea-level outcomes. No fixed annual growth rate or CAGR is applied.
Growth and trend parameter: the relevant physical input is the city-specific probability distribution of relative sea-level change under RCP8.5 rather than a single constant growth rate. The model combines this distribution with the city damage function to estimate future economic losses.
Assumptions: the ranked estimates use RCP8.5 and assume no new adaptation in the Abadie damage calculation. Future socioeconomic conditions are incorporated through the underlying city damage framework. The sea-level model includes glacial isostatic adjustment, while city-scale land motion beyond that component is not fully represented in the main ranking.
Missing values and inclusion: only the 15 detailed 2050 RCP8.5 values printed in the primary study table are included. No additional city value is inferred from a figure, reconstructed from an unrelated dataset or added to create an artificial Top 100.
Conflicts: estimates from different coastal-risk studies are not averaged. Differences in socioeconomic pathways, protection assumptions, adaptation, subsidence and flood-damage methods can produce materially different results.
Official values, forecasts and projections: an official value is a published observation from an authoritative statistical source; an official forecast is a forward estimate issued by an official body; a modeled projection is conditional on a research model and stated assumptions. All 15 rows here are modeled projections.
Rounding: the ranked values retain the one-decimal precision published in Abadie et al. Ranking is based on the numeric 2050 value shown in the table.
What the metric does not measure: it does not measure exposed population, permanent inundation area, fatalities, centimeters of local sea-level rise, adaptation cost, infrastructure downtime or the probability that a city will be submerged.
Ranking of 15 confirmed coastal-city projections
The table compares the 15 city-level RCP8.5 expected-damage estimates reported in the study. Guangzhou leads the 2050 ranking, followed by New Orleans and Mumbai.
2050 RCP8.5 expected annual coastal flood damage
| Rank | City | Damage | Source / method note |
|---|---|---|---|
| 1 | Guangzhou, Guangdong | $330.8B/yr | China · Abadie et al. (2020) · modeled projection · target 2050 · Hallegatte socioeconomic reference: $11.928B/yr · no adaptation. |
| 2 | New Orleans | $209.2B/yr | United States · Abadie et al. (2020) · modeled projection · target 2050 · Hallegatte socioeconomic reference: $1.583B/yr · no adaptation. |
| 3 | Mumbai | $112.4B/yr | India · Abadie et al. (2020) · modeled projection · target 2050 · Hallegatte socioeconomic reference: $6.109B/yr · no adaptation. |
| 4 | Osaka-Kobe | $101.1B/yr | Japan · Abadie et al. (2020) · modeled projection · target 2050 · Hallegatte socioeconomic reference: $0.261B/yr · no adaptation. |
| 5 | Tokyo | $79.8B/yr | Japan · Abadie et al. (2020) · modeled projection · target 2050 · Hallegatte socioeconomic reference: $0.058B/yr · no adaptation. |
| 6 | Kolkata | $65.9B/yr | India · listed as Calcutta in the source · modeled projection · target 2050 · Hallegatte socioeconomic reference: $2.704B/yr. |
| 7 | Nagoya | $64.1B/yr | Japan · Abadie et al. (2020) · modeled projection · target 2050 · Hallegatte socioeconomic reference: $0.564B/yr · no adaptation. |
| 8 | Alexandria | $50.0B/yr | Egypt · Abadie et al. (2020) · modeled projection · target 2050 · Hallegatte socioeconomic reference: $0.199B/yr · no adaptation. |
| 9 | Guayaquil | $49.1B/yr | Ecuador · Abadie et al. (2020) · modeled projection · target 2050 · Hallegatte socioeconomic reference: $2.813B/yr · no adaptation. |
| 10 | Tianjin | $45.5B/yr | China · Abadie et al. (2020) · modeled projection · target 2050 · Hallegatte socioeconomic reference: $1.810B/yr · no adaptation. |
| 11 | Shenzhen | $42.0B/yr | China · listed as Shenzen in the source framework · modeled projection · target 2050 · Hallegatte socioeconomic reference: $2.929B/yr. |
| 12 | Bangkok | $38.7B/yr | Thailand · Abadie et al. (2020) · modeled projection · target 2050 · Hallegatte socioeconomic reference: $0.596B/yr · no adaptation. |
| 13 | Abidjan | $20.9B/yr | Cote d'Ivoire · Abadie et al. (2020) · modeled projection · target 2050 · Hallegatte socioeconomic reference: $0.826B/yr · no adaptation. |
| 14 | Hai Phong | $17.6B/yr | Vietnam · Abadie et al. (2020) · modeled projection · target 2050 · Hallegatte socioeconomic reference: $0.320B/yr · no adaptation. |
| 15 | Shanghai | $15.7B/yr | China · Abadie et al. (2020) · modeled projection · target 2050 · Hallegatte socioeconomic reference: $0.063B/yr · no adaptation. |
Abadie et al. (2020) controls the ranked RCP8.5 values. Hallegatte et al. (2013) supplies the socioeconomic coastal-damage reference inputs. These reference values are model components rather than observed historical losses or simple growth bases.
Key findings from the 2050 ranking
Key Insight
Guangzhou's $330.8 billion projection is about 58% higher than New Orleans and almost three times Mumbai's 2050 value.
Notable Pattern
Guangzhou and New Orleans together account for about 43.4% of the total modeled damage represented by the 15 published 2050 values.
Regional Concentration
Eleven of the 15 confirmed cities are in Asia-Pacific, including seven of the Top 10. China, Japan and India are especially prominent in the published cohort.
Outlier
New Orleans is the only U.S. city in the confirmed 15-city group and ranks second at $209.2 billion per year, well above Mumbai in third place.
What the results mean
The ranking identifies where the selected model produces the largest expected annual economic losses under RCP8.5 in 2050. It is therefore most useful as a comparison of modeled financial exposure to coastal flooding rather than a direct measure of physical sea-level rise.
Economic scale matters. Cities with valuable and densely concentrated assets can generate very high expected losses even when another location experiences comparable or greater physical sea-level change. Protection standards and flood vulnerability also affect the relationship between water levels and economic damage.
The projections are conditional, not predictions of exact future losses. Actual outcomes will depend on emissions, urban development, coastal protection, adaptation, local subsidence and other forms of vertical land motion that can differ from the model assumptions.
For climate-risk planning, the ranking is best used alongside separate measures of population exposure, critical infrastructure, poverty, adaptation capacity, current defense standards and locally observed relative sea-level change.
FAQ
Which coastal city has the highest modeled flood damage in 2050?
Guangzhou ranks first among the 15 published cities at $330.8 billion per year under the RCP8.5 expected-damage scenario.
Why are there 15 cities instead of a Top 100?
The underlying research covers 136 cities, but the primary article's main table provides detailed 2050 RCP8.5 expected-damage values for 15 high-risk cities. The ranking is limited to those directly confirmed values rather than filling the remaining places with estimates from incompatible datasets.
Are these official 2050 forecasts?
No. They are modeled projections from peer-reviewed research. The values depend on the selected climate scenario, socioeconomic inputs and damage model.
Does the ranking show which city will experience the most sea-level rise?
No. The ranked metric is expected economic flood damage. Physical sea-level change is one input to the model, but economic exposure and vulnerability also affect the result.
Does the ranking measure the number of people exposed?
No. Population exposure is a separate metric. A city can have high modeled financial losses without ranking in the same position by exposed population.
Why is New Orleans second here when the original paper places Mumbai ahead of it?
The original table is ordered by its RCP8.5 values for 2100. This ranking sorts the same cities by the published 2050 column, where New Orleans has $209.2 billion and Mumbai has $112.4 billion.
What does expected annual damage mean?
It is the average economic loss across the modeled annual damage distribution. It is not the expected cost of one specific flood and does not mean the same loss would occur every year.
Does the model include future adaptation?
The Abadie estimates used for the ranking assume no adaptation. New coastal defenses, land-use changes or other resilience measures could therefore alter realized future losses.
Is local land subsidence fully represented?
No. The main sea-level framework includes glacial isostatic adjustment, but city-scale land motion such as strong local subsidence is not fully incorporated in the principal ranking and remains an important source of uncertainty.
Sources
This is a compiled research dataset. Abadie et al. (2020) is the controlling source for every ranked 2050 RCP8.5 value. Hallegatte et al. provides the underlying coastal-city loss framework and socioeconomic reference inputs. Supporting sources document the climate scenario and associated research dataset.
Abadie et al. (2020) — Comparing urban coastal flood risk in 136 cities
Primary source for the ranked 2050 RCP8.5 expected-damage values, model design and no-adaptation assumption.
Hallegatte et al. (2013) — Future flood losses in major coastal cities
Underlying coastal flood-loss framework, socioeconomic scenarios and city-level damage inputs used by later modeling work.
Hallegatte et al. — Supplementary Information
Detailed city-level socioeconomic and no-sea-level-rise reference values used to document the model inputs in the ranking table.
https://courses.ems.psu.edu/earth107/sites/earth107/files/Unit1/Mod1/nclimate1979-s2.pdf
Zenodo — Abadie et al. additional coastal flood-risk dataset
Associated research-data archive covering additional sea-level scenarios, expected damages and decadal model outputs.
