Lessons learnt from population movements over the atoll’s history
Until the middle of the eighteenth century, the population was scattered over the atoll and all the islands included in this study were inhabited60. Following repeated attacks by nearby Anaa Atoll’s warriors in the second half of the eighteenth century, the scattered communities gathered around the three passes in the north (Avatoru and Tiputa passes) and north-west (where a pass existed at that time) of the atoll to better ensure their security. However, this strategy failed, and inhabitants finally took refuge in Tahiti ~ 1750. When they returned to Rangiroa ~ 1825, they spread out again across the atoll. This traditional settlement pattern was challenged again ~ 1850 when missionaries arrived and gathered the population on the northern and north-western islands in the frame of the Christianization process. However, the southern Otepipi island (no. 12; Fig. 2B) remained settled at that time. In the twentieth century, natural disasters, including the 1906 and 1983 TCs, contributed to the concentration of the population in the Avatoru-Tiputa area (islands nos. 2–9; Fig. 2B). The 1983 TC led to the permanent abandonment of Otepipi and to the out-migration of ~ 10% of Rangiroa’s population to Tahiti48. It is unknown whether these inhabitants returned to Rangiroa later. Moreover, official statistics reveal that since 1946 (600 inhabitants), the population of Rangiroa has increased continuously, with an acceleration since the end of the 1980s44, with 1300 inhabitants in 1988 and 2567 inhabitants in 2017. These findings confirm that internal and out-migration have for long been used to face climate and non-climate pressures on this (and potentially other) Tuamotu atolls.
Is internal relocation physically relevant?
A previous study revealed that the northern settled islands generally have lower elevation indices (from 1 to 3, expect for Tiputa with 4) than the uninhabited rural islands targeted for future development which exhibit indices of 3 and 4 (see details in Ref.21). Most of the northern settled islands exhibit elevations < 3.74 m for all of the five variables considered (averaged island elevation, maximum ocean-side beach ridge elevation, maximum lagoon-side beach ridge elevation, lowest elevation in the inner part of the island, and highest point on the island), whereas the uninhabited rural islands targeted for future development show higher elevations (e.g. maximum ocean-side beach ridge and/or highest elevation > 6.0 m for some islands). The last century’s concentration of people and infrastructure on the northern islands has thus led to increased population and infrastructure exposure to climate pressures.
Island susceptibility to high-tide chronic and cyclone-driven flooding is highly variable among study islands, and comparatively high for most of the northern settled islands (except for islands Nos. 8 and 9) compared to some uninhabited rural islands targeted for future development (Nos. 11 and 12). The extent of high-tide chronic flooding under IPCC SLR scenarios SSP1-2.6 and SSP5-8.5 in 2100 varies across study islands, depending on the height of their lagoon-side beach ridge (Fig. 3; Supplementary Material 5). Under SSP1-2.6, no island is significantly affected by high-tide chronic flooding, as lagoon- and ocean-side beach ridges prevent flooding. Under SSP5-8.5, two islands, including Avatoru (No. 2; which currently concentrates most of the population and public services, and the harbor infrastructure), and the island hosting the airport (No. 5) experience land loss along their low-lying lagoon coast. The flooded area reaches up to 100 m in width from the coastline in the central part of island No. 5. Under SSP1-2.6, the other settled islands exhibit no or very limited high-tide flooding along their lagoon shoreline. The two islands that show the lowest susceptibility to high-tide chronic flooding are the easternmost islands of Ohotu (No. 8) and Tiputa (No. 9; first island that was settled in the northern part of the atoll). The three uninhabited rural islands considered, two of which (Nos. 10 and 11) have relatively high ocean- and lagoon-side beach ridges, all remain flood-free under both SSP1-2.6 and SSP5-8.5 in 2100, with limited differences between these two scenarios.
High-tide water levels under SSP1-2.6 and SSP5-8.5 climate scenarios for different island profiles around Rangiroa Atoll. This figure highlights major elevation and topography differences between islands, which result in high variations in their susceptibility to high-tide chronic flooding under SSP1-2.6 and SSP5-8.5 climate scenarios: islands Nos. 2, 5 and 9 exhibit lagoon-side land loss, whereas island No. 12’s beach ridges prevent land loss. For details on Extreme Sea Levels, see Fig. 7 in “Materials and Methods”.
The analysis of cyclone-driven flood scenarios shows that all islands except islands Nos. 11 and 12 are flooded over their entire land area under both SSP1-2.6 and SSP5-8.5 in 2100 (Fig. 4; Supplementary Material 6). Except for island No. 9 (transect 2), the maximum elevation of these islands is lower than the maximum instantaneous water levels under intense TC conditions for present days sea levels. Given the intense wind, waves, and flow velocity conditions during a TC, flooding due to overtopping will increase risks for human life in the future. Interestingly, islands Nos. 11 and 12 have non-flooded areas under future intense cyclonic conditions. On island No. 11, these areas correspond to the highest parts of the lagoon-side sand dunes and are approximately 110 m wide. On island No. 12, the flood-free area is located along the ocean-side, where the height of the beach ridge exceeds ESLs and extends over 50–100 m in width. These findings confirm that some of the uninhabited rural islands targeted for future development are safer than the northern settled islands.
Indicative cyclone-driven maximum instantaneous water levels, as modelled on Avatoru, under SSP1-2.6 and SSP5-8.5 climate scenarios, for different island profiles around Rangiroa Atoll. As a result of major elevation and topography differences, islands Nos. 2, 5, 9 and 12 are unequally susceptible to extreme cyclone-driven overtopping under both SSP1-2.6 and SSP5-8.5: whereas islands Nos. 2, 5 and 9 are flooded over their entire surface, island No. 12 has non flooded areas along its elevated ocean side. Note that extreme water levels are based on simulations made for Avatoru (No. 2), and that the estimation of flooding on island No. 12 would require a detailed hydrodynamic study. Yet, values provided in ref. 52 are considered conservative enough to support the conclusion that island No. 12 offers the best safety potential during a TC. For details on Extreme Sea Levels, see Fig. 7 in “Materials and Methods”.
In conclusion, internal relocation from currently settled and generally very low-lying northern islands (except for islands Nos. 8 and 9 that are significantly higher than other settled islands) to higher islands, including nearby and distant uninhabited islands of the atoll, can be considered a relevant adaptation option to reduce population and human asset exposure to future marine flooding on Rangiroa Atoll.
Is internal relocation feasible from a political-institutional perspective?
As a preliminary remark, it must be noted that French Polynesia is a French Overseas Territory with the status of a ‘Collectivité d’Outre-Mer’. This means that French Polynesia has a high degree of political autonomy from the French state since its first legal status of autonomy in 1984, which was reinforced in 1996 and 2004. French Polynesia has the competence to regulate all the fields that are not expressly assigned to the French state or to the municipalities headed by a mayor, and is especially responsible for urban and rural planning, risk prevention, and environmental management. As a result, French national relocation policies do not apply to French Polynesia.
First, official planning and legal documents and the semi-structured interviews conducted with institutional actors confirmed the absence of a dedicated policy or legal framework aimed at supporting internal (i.e. within a given island group) or external (i.e. from an atoll to Tahiti) relocation. However, interviews revealed that although it was not conceived with the aim of supporting internal relocation, the Rangiroa PGA project includes population redistribution over the atoll and could therefore contribute to climate-related internal relocation. Indeed, it targets six new islands for economic development in the uninhabited southern part of the atoll, including island No. 12, which has flood-free areas under both SSP1-2.6 and SSP5-8.5 climate scenarios under both high-tide chronic and cyclone-driven flooding (Figs. 3 and 4; Supplementary Material 5 and 6).
Second, neither human resources nor funding are currently dedicated to risk management and climate adaptation on the atoll, which prevents any relocation project to be implemented. Indeed, such projects need effective support from lead institutions, as stated by inhabitants who confessed that it should be led by the local (Rangiroa’s municipality; 47% of responses) or national (i.e. the Government of French Polynesia; 24%) public authorities. It can be underlined that public contributions to the public enquiry concerning the Rangiroa PGA project have denounced the lack of financial support to adapt individual homes to climate change by raising the floor to 1.50 m. If financial support does not exist for these types of targeted measures, it is likely that it will not exist for internal relocation projects either.
Third, no legal tool exists to support land acquisition for relocation purposes. However, article D.131-1 of the urbanism Code of French Polynesia states that General Development Plans can delimit pre-emption zones, especially for building up land reserves. In application of this article, Rangiroa PGA project indicates that the municipality plans to establish pre-emptive rights to acquire land in future development areas. Interestingly, two testimonies were collected in the field, which revealed that under specific circumstances, land issues can be resolved. The first one refers to private land acquisition by the public authorities in the 1980s for general interest purposes, including the establishment of the main administrative area (town hall, school, and fire station) on Avatoru Island (No. 2) and the construction of social housing for coconut farmers on island no. 10. The second example relates to the relaxation of land tenure rules by the French Polynesia Land Tenure Service in Tubuai Island, Australes Islands, in the aftermath of TC Oli (2010) to allow for reconstruction in safer areas. The authorities accelerated the agreement procedures through the limitation of family authorization to only five signatures when ‘undivided land’ generally involves four generations and more than ten owners. Such past experiences could lay the ground for more long-lasting changes in administrative procedures in a way they become climate adaptation compatible, for example to revise building location regulations according to new SLR projections. Although the Rangiroa PGA project (Ref.61) indicates that the municipality plans to establish pre-emptive rights to acquire land in future development areas, it however remains vague in terms of the associated financial means, stating that ‘it will be necessary to put in place a financial plan to respond to the opportunities that will arise‘. As a result, to date, no financial plan has yet been established to allow land acquisition, nor any specific financing mechanism exists to raise funds for the creation of new development centers and settlements. A rather case-by-case approach is planned that is not compatible with the basics of ‘managed and strategic retreat’33.
Fourth, the analysis of planning documents (Ref.65) revealed that internal relocation is not included in the vision of climate adaptation of institutional actors. This aligns with recent findings that institutional actors from the Government of French Polynesia did not view climate change as a danger, nor as a priority, because they face much more urgent problems, such as unemployment or waste management, daily62,63.
Collectively, the absence of a dedicated policy or legal framework aimed at supporting internal relocation, of dedicated human resources and funding, of legal tools aimed at supporting land acquisition for relocation purposes, and of internal relocation in institutional actors’ vision of climate adaptation indicate that today the political-institutional feasibility of internal relocation is rather limited.
Is internal relocation feasible from a socio-economic perspective?
The Rangiroa PGA draft document recognizes the major role that tourism has played in both socio-economic development and the construction of major infrastructure (e.g. international airport in 1966): ‘tourism has been a strong catalyst for the economic growth of the municipality’ (Presentation report of the General Land Use Plan of Rangiroa, 2017). It therefore uses tourism as a lever for socioeconomic development through job provision in the six southern targeted islands (including No. 12), where guest houses and hotels, physical infrastructure (including a harbor, causeways between concerned islands, service roads, etc.), and the provision of public services are planned. Although this PGA project is not a relocation plan, these findings show that public actors are aware of the need to equip future settled areas with all facilities, which could help to overcome socio-economic barriers related to internal relocation.
Is internal relocation socially acceptable?
The focus groups revealed that land tenure conflicts, which are widespread on Rangiroa, might act as a social barrier to internal relocation. A participant explained: ‘if some people need land, they should make a request to their own family [but such requests are not always successful, in particular because] today, elders mistrust the youth, which increases land-related tensions; [as a result] when you get a land plot from your family, you necessarily cling to it as if it was a treasure, because you might never get another one, [and] when you have a place, you won’t move, because it is too complicated, due to undivided land’. This view was supported by another participant who mentioned the possibility to buy land, which was confirmed by our own observations that land plots were for sale in the northern settled islands. Except for a few focus group participants who freely expressed themselves, other participants refused to approach the land tenure taboo issue. Focus groups also revealed that despite participants observed significant climate (e.g. warmer sea, warmer climate, change in seasonality and rainfall) and environmental (e.g. decreased fish stocks, increased coral bleaching, increased coastal erosion) changes over the past decades, most of them considered climate change not a major problem for the moment. Some of them confessed that this would not make sense to consider internal relocation, because some environmental changes are caused by inadequate human practices and climate change impacts are perceived still limited on the atoll. Last, other participants proved fatalistic: ‘Anyway, this is God who decides, so’.
The social survey allowed us to dive deeper into climate change perception and relocation. Seventy-four percent of 102 interviewees agreed that SLR might become a threat for the atoll in 2050. When asked if the first line residents who would be affected would at that time accept to move to another area, 43% of interviewees responded yes, whereas 9% set conditions and 48% were uncertain. According to interviewees, if relocation was required and was easily doable, inhabitants would preferably move to another area on the atoll (28% of responses), to the mountainous islands of Tahiti, Moorea, or the Marquesas Islands (29%), or to the nearby elevated reef island of Makatea (16%). However, 27% of interviewees did not answer or did not know where people would prefer to move.
Between 50% (for island No. 12) and 62% (for island No. 1) of interviewees have already been to the distant unsettled rural islands targeted for future development in the PGA, with ~ 30% of them having family land there. When asked if inhabitants would accept to move to these islands if SLR would force them to move, 43%, 38% and 30% responded positively for island Nos. 1, 10 and 12, respectively, if jobs and facilities are provided. However, 21% of interviewees did not know what to respond and another 20% responded that inhabitants would refuse to move to these islands.
When asked about their personal position about relocation, only 34% of interviewees provided a response. Twenty-three out of the thirty-five respondents (66%) said they would accept to move to islands Nos. 1, 10 and 12 if it was required and easily feasible. When asked how to make relocation possible, 48% of respondents said that their family would provide them with a land plot, whereas 29% declared that they could buy or rent a piece of land and 43% admitted that they did not know.
Collectively, these findings suggest that internal relocation could be implemented in 2050, provided that the project is led by the public authorities (71%) and involves inhabitants (87%), and that the land-related barrier is overcome.
