A new, high-resolution global reef island database (GRID) with implications for coastal vulnerability

Abstract. Low-lying reef islands (LRIs) are among the most climate-vulnerable landforms on Earth, yet no globally consistent and spatially comprehensive inventory of these islands has previously existed. This is especially important to underpin assessments of future risk and resilience. To address the absence of a unified global dataset, this study develops the first Global Reef Island Database (GRID), a detailed spatial product that identifies the global distribution of LRIs and collates their key geomorphological, environmental and exposure characteristics. Using the UNEP-WCMC Global Island Database (GID2.1), global coral reef distribution layers and the AW3D30 digital elevation model, all landmasses < 30 km² located within 1 km of coral reef and below a conservative mean elevation threshold of 16 m were identified. A total of 34,404 LRIs were mapped accounting for nearly 11,000 km² of land area and over 60,000 km of shoreline. The vast majority (77 %) are very small (< 0.1 km²) highlighting their geomorphological sensitivity. GRID demonstrates strong agreement with observed island counts (R²= 0.939; ICC = 0.984) and outperforms commonly used shoreline datasets in identifying small islands while preserving accurate geospatial alignment. LRIs are concentrated in Asia and Oceania, with Indonesia (14 %), Japan (12 %) and French Polynesia (10 %) containing the largest numbers. For each island, eight intrinsic variables (including land area, elevation, shape, population and isolation) and ten extrinsic variables (including wave and tidal climate, bathymetry, tropical storm exposure and projected sea-level rise) were compiled to build a globally standardised dataset suited for vulnerability analysis. The GRID demonstrates clear spatial clustering in external forcing variables: wave exposure is highest across mid-ocean atolls, tropical storm density peaks in the northwest Pacific and Caribbean, and projected sea-level rise rates are greatest around the northern Philippines, Japan and Central America. Finally, this study develops an island vulnerability index using the GRID. Approximately 16 % of LRIs are classified as very highly vulnerable, particularly throughout the northwest Pacific and northern South China Sea. National-scale assessments reveal that countries such as China, Japan and the Northern Mariana Islands are at highest risk. It is anticipated that thus far the most comprehensive yet unified database can support a variety of reef island research including climate adaptation planning, conservation and hazard mitigation efforts and coastal erosion studies. The final dataset has been made publicly available at https://doi.org/10.1594/PANGAEA.986811 (Dawson 2025).