Impacts of typhoons on atoll island vegetation in the Federated States of Micronesia: a remote sensing approach

Abstract

Atoll islands are small, remote and low-lying accumulations of unconsolidated reef-derived sediments located atop coral reef platforms throughout the Earth’s tropical and subtropical oceans. These islands represent the only inhabitable landmass in countries, such as the Maldives and the Republic of the Marshall Islands, and are perceived to be highly vulnerable to the effects of sea-level rise, and to the increased frequency and magnitude of tropical storms associated with climate change. Typhoons, in particular, are a major agent of natural disturbance on atoll islands. However, while the impacts of typhoons on reef structure and atoll island geomorphology are well documented, there has been a paucity of research pertaining to their impacts on atoll vegetation. Atoll vegetation provides many important functions, such as binding and stabilizing sediments, as well as constituting a major source of food and income for island residents. Until now, there has been no research undertaken to investigate the large-scale impacts of typhoons on atoll island vegetation. This study investigates the impacts of Typhoon Maysak in 2015 and Wutip in 2019 on the vegetation of atoll islands across the Yap and Chuuk states in the Federated States of Micronesia using open-source Landsat-8 and commercial Planet Dove imagery. Using vegetation indices, namely the normalized difference vegetation index (NDVI) and the normalized burn ratio (NBR), the impacts of both typhoons on vegetation were measured at 128 atoll islands between 2013 and 2021. This analysis was undertaken using the powerful cloud-based geospatial platform Google Earth Engine, which enabled fast calculations for a large volume of imagery. Between 30 and 148 satellite scenes were collected for each atoll across the eight year period, and stored as image collections on a per-island basis. The vegetation indices were calculated and then averaged across all images, where time-series were generated showing the short and long-term storm impacts. Results showed that moderate resolution Landsat-8 and high-resolution Planet Dove imagery could effectively detect and document storm-induced vegetation damage, and that both the NDVI and NBR were effective indicators of storm damage. Atoll vegetation recovered rapidly after both storms, with most of the recovery occurring within 2.5 years post-storm. The results also revealed that most of the vegetation damage occurred within 80 km of the storm track at wind speeds exceeding 100 km/hr and wave heights above 10-metres high.