Planned relocation may reduce communities’ future exposure to coastal inundation but effect varies with emission scenario and geography

The study investigates whether planned relocation of entire coastal communities effectively reduces exposure to sea level rise and coastal flooding over time and under different emissions scenarios. Planned relocation is increasingly considered where other adaptation measures (protection or accommodation) may leave residual risk, but it can be disruptive, costly, and sometimes maladaptive if destination sites are also hazard-prone. Existing examples from places such as the Maldives, Fiji, Sri Lanka, India, and the Philippines show that relocations have sometimes resulted in exposure to similar or new hazards at destinations. There is limited quantitative evidence assessing exposure reduction, particularly regarding future risk across multiple time horizons and emissions pathways. This paper addresses that gap by comparing projected inundation exposure at origin versus destination sites for completed or ongoing planned relocations, providing essential insights for prioritizing adaptation investments to ensure people are safer over the 21st century.

Prior research has largely focused on socioeconomic outcomes of planned relocations and conceptual frameworks for adaptation success. Quantitative assessments of hazard exposure reduction are rare. A study in the U.S. Midwest found over 95% reduction in riverine flood damage costs after relocations (non-coastal). In Tacloban, Philippines, 12 of 13 relocation sites were less exposed to coastal storm surge. Modeling work in French Polynesia suggested potential future SLR exposure reductions from relocation within an atoll. However, no prior study has systematically assessed future coastal hazard exposure reduction across a global set of completed or in-progress planned relocations, across multiple emissions scenarios and time horizons. The broader adaptation literature also notes a paucity of evaluations linking actions to reductions in exposure, with exceptions such as dikes.

The authors conducted a five-phase analysis to quantify exposure reduction from planned relocation for coastal communities. Phase 1 (Case selection): From a new database of over 400 planned relocations, 17 cases were selected that met criteria: initiated due to coastal hazards; origin within 1 km of coastline; single-community move from one origin to one destination; sufficient information to geolocate sites; and coverage by Climate Central’s CoastalDEM v2.1. Alaska cases were excluded due to DEM coverage limits. Phase 2 (Geolocation): Using QGIS 3.22.1 and Google Earth imagery, the team delineated polygons for origin and destination sites. Where maps with predefined site boundaries existed, they were traced and georeferenced. Otherwise, perimeters of inhabited areas (houses, community centers) were traced based on photos and narrative descriptions, excluding agricultural lands and airstrips. Phase 3 (SLR and flood projections): Inundation surfaces were created from CoastalDEM v2.1 elevations, IPCC AR6 sea level rise projections under SSP1-2.6 (low), SSP2-4.5 (mid), and SSP5-8.5 (high), and local annual flood heights (once-per-year water levels) from the GTSR dataset. Projections were evaluated for 2030, 2050, and 2100 using median (50th percentile) values. Inundation layers represent areas with elevation less than or equal to projected SLR plus annual flood height. Phase 4 (Exposure calculation and reduction metric): For each case, scenario, and time horizon, the percentage of site area inundated was computed for origin and destination by intersecting site polygons with inundation layers. Exposure reduction was defined as the difference in percent inundation between origin (counterfactual without relocation) and destination. Phase 5 (Assessment comparison): The authors searched for planning/assessment documents used in site selection to determine whether and how future hazard exposure, time horizons, emissions scenarios, and alternatives were considered. Across the 17 cases, only two identifiable assessments (both in 2016) evaluated current hazards; neither incorporated future SLR scenarios across multiple horizons or emissions pathways. Data and code are available at the cited public repository.

- Across 17 planned relocation cases spanning the Pacific, South and Southeast Asia, and the Americas, planned relocation generally reduced exposure to projected SLR plus a once-per-year flood, but the magnitude varied by geography, emissions scenario, and time horizon. - In all cases, origin sites are projected to be exposed to inundation, with exposure increasing toward 2100 and under higher emissions scenarios. - Destination sites in 9 of 17 cases are projected to be exposed under at least some scenarios, with exposure tending to increase later in the century. - Illustrative 2100 SSP5-8.5 results: Gardi Sugdub (small island to mainland) reduced inundation from 100% at origin to 0% at destination; Vunidogoloa (mainland to mainland) reduced from 60% to 0%; Taholah (mainland to mainland) reduced from 59% to 0%; Kandholhudhoo (small island to small island) reduced from 100% to 66%. - By 2100 under mid-emissions (SSP2-4.5), average exposure reduction across cases is 47%, ranging from 100% (Gardi Sugdub) to 2% (Jakarta). - Spatial pattern differences (2100): • Small island to small island: high inundation at origins (≈78% low emissions to 87% high) and substantial inundation at destinations (≈36% low to 45% high), yielding relatively modest reduction potential, especially under high emissions. • Small island to mainland/large island: high origin inundation (≈83% low to 90% high) but much lower destination inundation (≈13% low to 14% high), producing the greatest average reduction potential across scenarios. • Mainland/large island to mainland/large island: moderate origin inundation (≈39% low to 46% high) and low destination inundation (≈5% across scenarios), indicating substantial but smaller absolute reductions due to lower origin exposure. - Some destination sites show considerable projected inundation (e.g., Kandholhudhoo, Gemendhoo, Madifushi, Isle de Jean Charles, El Choncho, Narikoso, Denimanu), while others have marginal exposure (e.g., Enseada, Vunisavisavi ≤4% destination area). - The effectiveness of relocation is time- and emissions-dependent; benefits can erode by late century under higher emissions pathways.

Findings demonstrate that planned relocation can reduce exposure to coastal inundation, but effectiveness varies widely and is dynamic over time, with stronger benefits in lower-emissions futures. Geographic context is critical: small island developing states (particularly atolls like the Maldives, where much land lies <1 m above mean sea level) have limited availability of truly safe sites, leading to smaller reductions when relocating between small islands. Moves to larger islands or mainland sites tend to yield greater long-term exposure reduction. The analysis highlights a major gap in practice: the near absence of systematic forward-looking risk assessments guiding site selection; only two of the 17 cases had assessments, both focusing on present-day conditions and not future SLR across scenarios or time horizons. Relocation decisions often reflect multiple drivers beyond environmental risk—such as land availability, political interests, livelihoods, demographic pressures, and cultural factors—shaping both the choice to relocate and destination selection. The results underscore the need for anticipatory, scenario-based assessments that integrate community priorities and knowledge to ensure relocations maintain reduced exposure over time, and raise climate justice concerns for countries lacking viable safe internal destinations.

This study provides a first systematic, global, quantitative assessment of how planned relocations affect communities’ future exposure to SLR and coastal flooding across multiple emissions scenarios and time horizons. It shows that while most relocations reduce exposure, the extent of reduction depends on emissions pathways and geographic context, and in many cases destination sites still face nontrivial future inundation risk. The work underscores the importance of incorporating forward-looking, scenario-based hazard assessments into relocation planning, alongside socioeconomic and cultural considerations, and highlights climate justice implications where safe destinations are scarce. Future research should expand to multi-hazard contexts, explore a wider range of flood frequencies, integrate socio-economic vulnerability and equity metrics, include complex relocation patterns (multiple origins/destinations), and use improved elevation and hydrodynamic models to better capture changing extremes and connectivity.

Key limitations and uncertainties include: (1) reliance on median SLR projections within each SSP; (2) potential underestimation of extreme high waters as storm intensity/frequency may increase; (3) use of a simplified static inundation (“bathtub”) approach without full hydrodynamic connectivity; (4) uncertainties in CoastalDEM vertical accuracy and downscaling, and in GTSR flood statistics; (5) focus on a single annual-flood threshold rather than a spectrum of event frequencies; (6) exclusion of other hazards (e.g., storms, landslides) that can alter overall risk; (7) inability to fully account for confounding adaptation measures, social capital, economic opportunity, and politics at sites; (8) exclusion of cases with multiple origins/destinations; and (9) lack of integration of socio-economic vulnerability, livelihoods, and equity measures, which affect outcomes even for similar exposure levels. These factors suggest results should be interpreted cautiously and paired with localized, participatory assessments.