Maldives is among the most climate-exposed countries in Asia and the Pacific. Its transport system is shaped by geography: small islands, low elevation, limited land area, and a road network that often serves as the main link between homes, schools, health services, ports, airports, and local economic activity.¹

A new screening analysis by the Asian Transport Observatory (ATO) examines how much of the Maldives road network may be exposed to future coastal flooding by 2050, and more importantly, where these might be. The analysis combines open road-network data from OpenStreetMap with global coastal flood hazard layers from the Deltares Global Flood Maps dataset.²

The results suggest that around 75 kilometers of the assessed road network (~1,656 km in total) may be exposed to coastal flooding under the 2050 scenarios analyzed. Exposure is nearly identical for the 50-year and 100-year return period events, at 74.88 km and 74.97 km, respectively, equivalent to about 4.5% of the assessed network.

At first glance, this share may appear modest. But in the Maldives, even relatively short exposed road sections can matter. Roads on small islands often provide essential local access. Disruption to a few links may affect access to schools, health facilities, ports, ferry terminals, emergency services, and tourism-related activity.

What was assessed

The analysis used a local Maldives OpenStreetMap extract dated 26 May 2026. It retained the main strategic road classes, including primary, secondary, tertiary, and their link classes.³

Residential and unclassified roads were also included. This is important in small island contexts such as the Maldives, where many roads are not tagged as higher-order roads in OpenStreetMap but still provide important local access.

Flood exposure was assessed using the Deltares Global Flood Maps dataset.⁴ Two future coastal flood scenarios were processed for 2050: the 50-year and 100-year return period events. These layers provided modelled coastal flood depth, with inundation depth estimated from extreme coastal water levels and underlying terrain elevation. The selected product used MERIT-DEM⁵ at approximately 90 m resolution, meaning the analysis accounted for broad elevation differences but may not have captured fine-scale island features such as road embankments, seawalls, drainage structures, reclaimed land levels, or local coastal protection. For each flood scenario, the flood-depth raster was clipped to the Maldives boundary and reclassified using a minimum flood-depth threshold of 0.15 m then overlaid to the road network. The main outputs were total selected road length, exposed road length, and the percentage of road length exposed under the 2050 50-year and 100-year coastal flood scenarios.

National exposure results

The difference between the 50-year and 100-year return period scenarios is small in the national results. The exposed length increases by only about 0.09 km between the two scenarios.

This suggests that, at the national scale and using this global dataset, most of the identified exposure is already captured under the 50-year return period event. The 100-year event adds limited additional exposed length. This should not be interpreted to mean that flood risk does not increase under more extreme events. Rather, it reflects the screening method, the resolution of the hazard data, and the specific threshold used in the analysis.

Which road classes are most exposed?

The exposure pattern differs by road class.

Tertiary roads show the highest exposed share, with about 16% of the assessed tertiary network exposed under both scenarios. Secondary roads follow at around 11%, while primary roads show an exposed share of about 9%.

Residential roads account for the largest exposed length in absolute terms, at around 55 km. This is expected because residential roads make up most of the assessed network. Although their exposed share is lower, their importance should not be underestimated. In island communities, residential roads often provide the last-mile access needed for daily movement and emergency response.

Population proximity to exposed roads

The analysis also used a WorldPop 2030 population distribution layer to understand how many people live near exposed road links.⁶

The proximity assessment found that a quarter (25%) of the population lives within 250 meters of exposed road links, and 41% lives within 500 meters of exposed road links.

This does not mean that these people are directly exposed to flooding. It means that a significant share of the population lives close to road sections that may be affected by coastal flood events.

This adds an important access and service-continuity dimension to the analysis.

What the results mean

The results point to three main insights.

First, national exposure appears concentrated rather than widespread. Around 4.5% of the road network is exposed under the two 2050 coastal flood scenarios. But the affected links may still be important because of the Maldives’ island geography and limited redundancy in local networks.

Second, local and lower-order roads matter. Residential and unclassified roads are often left out of strategic infrastructure analysis. In small island settings, this can miss important access routes. Including these roads gives a more realistic picture of local mobility exposure.

Third, exposure should be interpreted together with people and services. The finding that one quarter of the population lives within 250 meters of exposed road links suggests that further analysis should examine which communities, services, and economic activities depend on these links.

How the results should be used

This analysis should be seen as a rapid assessment meant to identify road links which may require closer investigation, validation, and prioritization considering flood exposure and population proximities. Please note that other elements of identifying critical links (e.g. traffic volumes, proximity of services, how the links play roles in the movements, etc…) are not captured by this rapid assessment process.

It can support early-stage questions such as:

• Which road sections may need more detailed flood-risk assessment?
• Which islands have exposed road links close to populated areas?
• Which exposed links may be critical for access to ports, health facilities, schools, or emergency services?
• Where could resilience investments have the greatest access benefits?

The analysis is not an engineering-grade flood assessment. The flood hazard data are based on a global model and an elevation dataset with approximately 90-meter resolution. This means the analysis may not capture fine-scale island features such as road embankments, seawalls, drainage structures, reclaimed land elevations, local coastal protection, or small variations in terrain.

More detailed local analysis would require higher-resolution elevation data, road elevation or embankment height, drainage information, coastal protection data, and locally calibrated flood observations.

Why this matters for transport resilience

For countries such as the Maldives, climate resilience is not only about protecting infrastructure assets, but also about ensuring access.

A road segment may be short in length but critical in function. It may connect a community to a ferry terminal, a school, a clinic, a port, or an evacuation route. Losing access even temporarily can disrupt daily life, service delivery, and local economies.

This is why open data and screening tools are useful. They cannot replace local engineering studies, but they can help governments and development partners identify where to look first. They can also support more targeted data collection, better project preparation, and stronger integration of resilience into transport planning.

Maldives case shows how combining open road-network data, global flood hazard layers, and population information can provide an initial view of where transport systems may be vulnerable. ATO is coming up with new sets of tools which will consolidate accessible geospatial information (bolstered by bespoke analysis such as this) which is aimed at uncovering insights, and opening up the data and information in a geospatial manner.