Global-scale changes in the area of atoll islands during the 21st century: Since 2000, land area on 221 atolls examined had increased by 61.74 km2 (6.1 %); most of the increase in land area occurred in the Maldives & South China Sea

Global-scale changes in the area of atoll islands during the 21st century. Andrew Holdaway, Murray Ford, Susan Owen. Anthropocene, Volume 33, March 2021, 100282. https://doi.org/10.1016/j.ancene.2021.100282

Highlights

• Since 2000, land area on 221 atolls examined had increased by 61.74 km2 (6.1 %).

• 153 atolls increased in size and 68 decreased in size between 2000 and 2017.

• Most of the increase in land area occurred in the Maldives and South China Sea.

• Land reclamation was primarily responsible for land area increases on atolls.

Abstract: The long-term persistence of atoll islands is under threat due to continued sea level rise driven by anthropogenic climate change. One widely discussed potential impact of sea level rise is the widespread, chronic erosion of atoll islands. Despite concerns of erosion driven by sea level rise, no published evidence exists of pervasive erosion of atoll islands at a global scale. Existing studies of atoll island change have been based on small, temporally sparse samples of islands on a limited number of atolls. As a result, the global response of atoll islands coincident with sea level rise remains uncertain. Using rich collections of Landsat imagery, this study analyses changes in land area on 221 atolls in the Indian and Pacific Oceans. Results show that, between 2000 and 2017, the total land area on these atolls has increased by 61.74 km2 (6.1 %) from 1007.60 km2 to 1069.35 km2. Most of the change in land area resulted from island building within the Maldives and on atolls in the South China Sea. Since 2000, the Maldives have added 37.50 km2 of land area, while 16.57 km2 of new islands have appeared within the South China Seas Spratly and Paracel chains. Understanding the extent of land area change at the global scale improves insights into the variation in responses, and how the manipulation of land is shaping the potential habitability trajectory for some atoll communities. Results highlight the anthropogenic imprint on the size of atoll islands, thereby providing a better understanding of variations in the future trajectories of human settlement and adaptation within atoll settings.

Keywords: Atoll islandsErosionSea level riseIsland buildingLand reclamation

5. Discussion

5.1. Global scale island change

This study has examined, for the first time, land area change on atolls at a global-scale. The study measures land area change in the major atoll nations including the Maldives, Marshall Islands, Kiribati and Tuvalu, as well as large numbers of atolls in Micronesia and French Polynesia. This examination of island change covers a spatial extent far greater than previous efforts to monitor land area change on atolls, which have not exceeded ten atolls in any given study (Table 1). Likewise, by utilising Landsat imagery, we resolved land area changes occurring at larger time intervals than in previous studies.

Results from this study showed that, at a global-scale, no major reduction has occurred in total landmass on atolls in the Pacific and Indian Oceans. To the contrary, due primarily to the construction of artificial islands, considerably more land on atolls is present now than in the year 2000, an increase of 61.74 km2. This increase represents an area over twice as large as the entire land area of the nation of Tuvalu. Most of the increase in land area resulted from the construction of islands on a small number of atolls in the Maldives and South China Sea (SCS) (Figs. 4 & 5). When we exclude the ten atolls with the largest land area increases, all of which are in the South China Sea and Maldives and showed evidence of island building, we noted only a 19.59 km2 increase in land area. Of those 211 atolls, the average increase in land area was 0.09 km2, with a median value of 0.03 km2 per atoll. When we excluded the atolls in the South China Sea and considered all other atolls in the Pacific Ocean, we saw a 6.00 km2 increase, with a per atoll average increase of just 0.04 km2 with a median value of 0.01 km2. Our results showed that, aside from atolls that are heavily engineered, the land area on most atolls has been stable with no evidence of loss over the study period. This observation is consistent with recent studies that have also shown a predominantly stable or accretionary trend in the area of atoll islands worldwide (McLean and Kench, 2015; Duvat, 2019).

5.2. Reclamation and Island building

The most significant total changes to the planform area of atoll islands have resulted from local-scale land reclamation and island building in the Maldives and South China Sea, which we observed before, during and after construction in our record. The observations of island building on atolls is consistent with remotely sensed observations of coastal cities worldwide, where land reclamation has been widely adopted as an approach to increase land availability relatively efficiently (Sengupta et al., 2018; Tian et al., 2016; Chee et al., 2017), underpinned by extensive investment to economic and political agendas.

Within the South China Sea, widespread and rapid island building has occurred to support the construction of military bases (Barnes and Hu, 2016; Asner et al., 2017). Examination of two separate island chains within the South China Sea (Spratly Islands and Paracel Islands) showed that the Spratly Islands underwent a rapid increase in area between 2014 and 2016, with 14.70 km2 land added over this period. The increase largely resulted from the construction centred on Mischief, Fiery Cross and Subi Reefs, where 6.14 km2, 2.82 km2 and 4.29 km2 of land has been added respectively. The land area in the Paracel Islands increased steadily between 2000 and 2017, with planform land area increasing from 8.00 km2 to 9.86 km2. Much of this growth has been the result of >1km2 of reclamation on Woody Island. Island building in the Maldives, which eclipses the amount constructed in the South China Sea by 37.50 km2 to 16.57 km2, has been to provide additional land for the local population and to expand tourism facilities and associated infrastructure, such as airports (Fallati et al., 2017). More recent observations, outside of our study period, show large islands under construction on South Malé atoll, which appear likely to add >2 km2 of land.

5.3. Implications of island building

The majority of atolls examined in this study are sparsely or unpopulated atolls. They have little existing or likely future large-scale construction similar to that observed in the Maldives and South China Sea. That atoll nations have demonstrated the desire and ability to manipulate densely populated urban atoll islands is evident, however, to foster economic growth and in cases adapt to sea level rise. For example, island building in the Maldives has, in places, been designed not only to increase the planform area of the islands, but also to raise the land levels to potentially mitigate future impacts of sea level rise (Brown et al., 2020). In the case of Hulhumalé, an artificial island in the Maldives with a population of ∼17,000, incorporating future sea level rise as a design criterion was deemed economically viable, with revenue generated from sales of real estate likely to cover the costs of construction (Bisaro et al., 2019). Similar planning to expand an island and build vertically are underway on Tarawa Atoll, the most populated atoll in Kiribati, with a proposal to building a raised section of island, ∼300 ha in size, to accommodate 35,000 residents and withstand two meters of sea level rise (Walters, 2019). Raising existing islands equally poses unique challenges and additional costs, however, relative to constructing new islands (Brown et al., 2020), and may result in unanticipated maladaptive trajectories.

5.4. Monitoring future islands using remote sensing

We showed that freely available imagery and processing and classification techniques within a cloud-based analytical platform can detect massive changes of atoll island planform area over the 21st century. We caution, however, that small-scale changes in atoll area, the styles of which are often detected using local-scale assessments, are unlikely detectable accurately using this approach. Consequently, our approach does not replace existing island-scale assessments of changes needed to inform local-scale coastal management and climate adaptation strategies. Rather, findings of this study represent a promising avenue for a systematic global-scale program for monitoring changes in atoll islands that can better inform where high-resolution monitoring might be necessary. Our approach is highly automated and uses freely available imagery and analytical toolsets. Similar efforts using Landsat imagery have been employed at a global scale to monitor changes in global surface water, forest cover and shoreline change (Hansen et al., 2013; Pekel et al., 2016; Luijendijk et al., 2018). The accuracy of our classification is quite high, however, given the pronounced difference between land and sub-tidal land cover classes. This accuracy arises because, with the 30 m resolution, even a small number of misclassified pixels can greatly affect the relative changes in atoll landmass, particularly on smaller atolls. Our approach is consistent across space and time, enabling comparison between the rates of change within different atoll chains, which at present is problematic due to the fragmentary nature of small-scale high-resolution studies.