Invasive rodent eradications in French Pacific islands: a critical analysis of past efforts

Introduction

Islands are one of the primary focal points in the current extinction crisis caused by invasive alien species worldwide (Samaniego et al. 2020; Roy et al. 2023). Island ecosystems have experienced early and severe anthropogenic extinction events and currently harbour a disproportionate number of globally threatened or recently extinct plant and animal species (Ricketts et al. 2005; Tershy et al. 2015). This is particularly true for tropical island ecosystems where invasive alien species are considered as the main threat to biodiversity (Blackburn et al. 2004; Lockwood et al. 2013), especially mammalian predator species, which are the principal drivers of native biodiversity loss on islands (Doherty et al. 2016). Among them, introduced rodents (rats, Rattus sp; mice, Mus sp.), which have colonised 80% of the world’s archipelagos (see Atkinson 1985), are the group that has had the most severe and widespread impacts on island birds, mammals, and reptiles (Doherty et al. 2016).

In recent decades, the eradication of invasive rodent species has become a major lever of action implemented for the conservation and restoration of native biodiversity on islands, with very positive effects and major successes on the island biota (Spatz et al. 2022). Indeed, since the first successful eradications of rodents on islands, whether deliberate or accidental, in mainland France and New Zealand in the 1950s and 1960s, eradication methods, protocols, and toxins have gradually improved (Howald et al. 2007; Campbell et al. 2015). This has resulted in a rapid and marked increase in the number and the surface area of islands freed from rodents in most regions of the world; for example, 596 successful island rodent eradications worldwide out of 871 attempts reported in 2019 (Spatz et al. 2022). However, marked differences have been observed in the success rates of operations, which were significantly lower in the tropics (Harper and Bunbury 2015; Russell and Holmes 2015).

In the tropical and subtropical Pacific, islands are home to five major invasive rodent species (Harper and Bunbury 2015; Samaniego et al. 2021). The Pacific rat (Rattus exulans) was introduced by the Austronesian/Polynesian settlers on several waves of introduction about 3000 years ago (Barnes et al. 2006). The other four species (black rat, Rattus rattus; brown rat, Rattus norvegicus; tanezumi rat, Rattus tanezumi; and house mouse, Mus musculus) were introduced in several waves (Harper and Bunbury 2015) with the European colonisers from the 17th century (Matisoo-Smith and Robins 2004; Drake and Hunt 2009). The introduction of these species had disastrous consequences for local ecosystems, leading to the decline or extinction of many endemic species, the disruption of island food chains, and the alteration of natural habitats (Fukami et al. 2006; Holdaway et al. 2007). The impacts of these rodents extend beyond biodiversity, also affecting local economies through the destruction of crops and the transmission of diseases, highlighting the urgency of developing effective and context-adapted eradication strategies.

These eradication strategies can be decided based on information from syntheses of operations, both successful or unsuccessful, which have already been conducted. Such syntheses already exist, covering eradication efforts at a global level or focusing on leading countries in terms of rodent eradication, such as Australia and New Zealand (Segal et al. 2022; Spatz et al. 2022). The advantage of these syntheses is that they enable feedback and knowledge to be integrated into decision-making, facilitating the implementation of optimal protocols or strategies (Pullin et al. 2020). They favour the assessment of the effectiveness of different eradication methods, and their adaptation to the unique context encountered on each island.

Here, we aimed to critique past rodent eradication efforts in the French territories of the Pacific. No compilation of experiences focused on these areas is available (e.g. see Russell and Broome 2016; Segal et al. 2022). The four French tropical Pacific territories (French Polynesia, New Caledonia, Wallis and Futuna, and Clipperton atoll) constitute a collection of more than 3600 islands and rocky or sandy islets according to the census by the UN Environment Programme World Conservation Monitoring Centre (UNEP-WCMC 2010), (~2900 islands in French Polynesia and ~700 in New Caledonia) distributed over a maritime surface area of ~6.8 million km², 19 islands in Wallis and Futuna, and one in Clipperton. This large set of islands is home to an exceptional terrestrial biodiversity, falling within three of the World’s biodiversity hotspots (Myers et al. 2000), encompassing 113 Important Bird and Biodiversity Areas and home to a high number of endemic species with for example, ~3000 endemic vascular plants, 166 endemic land snail species, ~100 endemic lizard species, and >50 endemic bird species (Beauvais et al. 2006; Pippard 2012; Soubeyran et al. 2015). In such a context of vibrant yet threatened island biodiversity, combined with the significant threats posed by introduced rodents, compiling data on and analysing rodent eradication operations carried out in the French territories of the Pacific appears as an important task for future eradication programs. It would also help raising awareness among governmental agencies as well as environmental and conservation organisations on the feasibility of such programs. In this study, we reviewed and summarised all documented rodent eradication attempts conducted in these areas in the past 40 years (1982–2022).

In particular, we aimed to answer the following three questions: (1) how many attempts have been made to eradicate rodents and which species were targeted? (2) what are the outcomes of these operations in the different territories? How large are the treated areas, and did this size change over time? and (3) what methods and biosecurity strategies have been used? We also aimed to compare the results with those of other Pacific countries where such syntheses are already available from New Zealand and Australia. The data analysed will be used to draw up recommendations for future eradication operations. This synthesis will also serve as a basis for decision-making regarding the prioritisation of islands to be treated and the methodologies to be adopted in similar tropical insular environments.

Materials and methods

Study sites

The study was conducted in four French tropical island territories in the South Pacific: (1) New Caledonia (~267,000 inhabitants); (2) French Polynesia (~308,000 inhabitants); (3) Wallis and Futuna (~11,000 inhabitants); and (4) the uninhabited Clipperton atoll (Fig. 1). Due to their geographical position, these four territories are characterised by a tropical to subtropical climate. However, New Caledonia exhibits a slightly milder climate, influenced by the trade winds (Maitrepierre et al. 2012).

Fig. 1.

(a) Location of the four French territories in the Pacific. (b) New Caledonia, (c) French Polynesia, and (d) Wallis. The red dot represents the islands where rodent eradication attempts have been made. The numbers correspond to the number of islands where rodent eradication attempts have been made in French Polynesia. The red lines represent the limits of the Exclusive Economic Zones. For mapping, we used QGIS 3.16.10 open source software.

Results

We found information on 85 invasive rodent eradication attempts that were carried out in the French Pacific territories between 1982 and 2022, on 77 different islands, covering a total area of 2773.36 ha (Table 1). A total of 38 attempts were carried out in New Caledonia (total area of 290.50 ha), 34 in French Polynesia (total area of 2386.81 ha), and 13 operations were recently (2021) conducted in Wallis and Futuna (total area of 96.05 ha, all around Wallis). No operation was conducted in Clipperton. These eradications focused on three different rodent species (R. rattus, Rr; R. exulans, Re; M. musculus, Mm). Most involved a single species but sometimes two co-occurring species were targeted (Rr and Re, Rr and Mm).

A total of 50% of all operations conducted in New Caledonia were considered successful while only 5.26% failed (Table 2). However, 14 of the 38 operations were implemented prior to a clear confirmation of rodent presence on the islands and therefore have an unknown pre-status. In French Polynesia, the success rate was 41.18%, with a 32.35% failure rate and 26.47% of operations still awaiting formal confirmation of their outcome. For Wallis, 79.92% of operations were considered successful while only 7.69% failed. However, 2 of the 13 were implemented despite no evidence of rodent presence on these islets and are therefore categorised as having an unknown pre-status.

Focusing on operations that were monitored for at least 2 years post-treatment, with a known outcome, and where the presence of rodents prior to the operation had been clearly established (i.e. 21 operations in New Caledonia, 25 in French Polynesia and 11 in Wallis), success rates reached 90.48% in New Caledonia, 56% in French Polynesia, and 90.91 in Wallis (Table 3).

There is a 10-fold difference in the size of treated islands between French Polynesia and New Caledonia or Wallis (Fig. 2 and Table 4). The average surface area of treated islands was around 7 ha in Wallis (7.39 ± 9.34 ha; min, 0.34; max, 31.55) and New Caledonia (7.64 ± 8.60 ha; min, 0.25; max, 47.83) compared with 72 ha (72.33 ± 126.37 ha; min, 0.70; max, 442.12) in French Polynesia. In New Caledonia and Wallis, the eradication operations carried out concerned small islands, with a maximum surface area of 47.83 ha for New Caledonia and 31.55 ha for Wallis, whereas in French Polynesia the eradication attempts concerned larger islands, reaching up to 442.12 ha.

Fig. 2.

Surface areas of islands treated for rodent eradications in the French Pacific territories. Blue, French Polynesia; green, New Caledonia; orange, Wallis.

Eradication attempts of introduced rodents began at roughly the same period in New Caledonia and French Polynesia; i.e. the late 1990s and early 2000s (Fig. 3). In New Caledonia, the eradication campaign got off to a flying start, with 26 operations conducted the same first year (Fig. 4). Eradication attempts then continued in a scattered pattern, targeting small and medium-sized islands without a subsequent increase in the number or size of projects planned.

Fig. 3.

Cumulative surface area of islands treated in the past 60 years. Green square, New Caledonia; blue diamond, French Polynesia; orange triangle, Wallis; yellow circle, French Pacific Archipelago; dark green circle, New Zealand; black circles, Australia.

Fig. 4.

Number of eradications carried out annually in French Pacific territories inthe past 40 years. Green, New Caledonia; blue, French Polynesia; orange, Wallis.

In French Polynesia, however, the number of projects has increased steadily over time, as has the surface area treated, progressing in stages, depending on the new methods used to spread baiting.

All recorded eradication attempts have been carried out by poisoning, mainly distributed by hand broadcast (70.59% in French Polynesia), sometimes combined with bait stations (68.42% in New Caledonia) or more recently by aerial broadcasting (helicopter or drone) (Table 5). Brodifacoum at least 0.02 g/kg was used in 97% of eradication attempts. Out of 85 operations, 83 use two baiting sessions, with bait quantities (per ha) varying depending on the method: in French Polynesia, this was 12 kg/ha for bait stations, 2–27.5 kg/ha for hand broadcast, 21.1–32.3 kg/ha for aerial broadcast, and 25–27.5 kg/ha for drones combined with hand broadcast; in New Caledonia, this was 2–7 kg/ha for bait stations, 3.5–25 kg/ha for hand broadcast, and 7–13 kg/ha for hand broadcast with bait stations; and in Wallis, this was 25–48 kg/ha for hand broadcast and 25 kg/ha for drones.

Although often mentioned in the documents reviewed for all the recorded eradication attempts, biosecurity protocols were not necessarily carried out. In New Caledonia, 76.32% of eradication operations were not accompanied by any protocol or biosecurity measures, 18.42% included the implementation of early detection systems combined with communication campaign, while 5.26% included only communication. In French Polynesia, 100% of operations included the implementation of early detection systems combined with communication, while in Wallis, 100% of operations were accompanied by communication campaigns (but none by early detection systems). All (100%) eradications were motivated by the conservation of island biodiversity, particularly seabird colonies. 72 operations aimed to conserve seabird colonies and/or potential nesting habitats, covering a total of 2711 ha across all territories. However, very few targeted a specific endangered species, except for example, the Fairy Tern (Sternula nereis) in the northern Important Bird Area of New Caledonia, which involved six operations treating a total of 49 ha. In French Polynesia, the Polynesian Ground Dove (Pampusana erythroptera) was the focus of four operations across two islands, treating 11 ha, and the Tuamotu Sandpiper (Prosobonia parvirostris) in the Tuamotu Archipelago was the focus of three operations across three islands, treating a total of 7 ha.

In New Caledonia, 68% of the operations were carried out by a foreign private operator, 26% by local NGOs supported by international NGOs (BirdLife International) and local institutions and 6% by a research team. In French Polynesia, 97% of operations were carried out by local NGOs supported by international NGOs (BirdLife International and Island Conservation), while all eradication attempts in Wallis were carried out by a cooperation between local institutions and the international NGO Island Conservation.

Discussion

French territories in the Pacific represent not only a unique set of thousands of islands and islets spread over considerable ocean surfaces, but also a remarkable concentration of endemic and endangered plant and animal species. Preserving this exceptional ecological heritage must include significant actions to counter the impacts of invasive species alongside a resolute strategy to restore island environments by removing pest species, among which introduced rodents are species that should be targeted as a priority.

Our study showed that at least 85 different rodent removal projects have been attempted in the French territories in the Pacific in the past 40 years on 77 different islands representing 2773.36 ha of eradication surfaces. Compared with the 193 and 91 offshore islands where rodent eradication attempts have been carried out in New Zealand and Australia, respectively (Segal et al. 2022), these numbers indicate significant efforts devoted to the eradication of invasive rodents in the French Pacific islands. However, compared to these two leading Pacific countries, the eradications carried out in the French Pacific territories had started later, were less ambitious in terms of surface areas treated, and experienced weaker growth over time, particularly in New Caledonia. Indeed, in Australia and New Zealand, rodent eradication began in the late 1960s and early 1970s. The number of operations and the size of treated areas grew rapidly in the mid-1980s. The ramp-up in eradication operations has been exponential, reaching 25,000 ha for Australia in 2014, and 47,000 ha in New Zealand in 2018 (Segal et al. 2022; Innes et al. 2024; Fig. 3).

Interestingly, the dynamics of eradication projects are markedly different between French Polynesia and New Caledonia. While projects began around the same period, we observed a gradual build-up in French Polynesia only, whereas operations in New Caledonia remained small-scale and limited in number. This clear difference in the scale, complexity, and frequency of eradication programs may partly reflect the earlier and broader adoption of aerial baiting techniques, either by helicopter or, more recently, by drone in French Polynesia. However, other factors such as funding levels and institutional frameworks likely also played a role (Holmes et al. 2015a). Also, projects there have benefit from consistent efforts and involvement of two major international NGOs (BirdLife International and Island Conservation) in close support of a local organisation, the Polynesian Ornithological Society (also known as ‘Manu’) deeply committed to an ambitious rodent eradication effort. The structured approach adopted by these specialised organisations, focusing on the prioritisation, design, and implementation of eradication programs, is more likely to deliver sustained efforts and long-term success compared to successes achieved when projects are implemented solely by governmental management agencies without a clearly defined strategy (e.g. Tershy et al. 2012; Samaniego et al. 2021). We note that where most of the eradication attempts are recent (2021 and 2022) in Wallis, the eradication campaign was commissioned by the local authorities to the Island Conservation NGO, which then relied on local staff for on-site operations.

Global eradication attempts have a success rate of 87% (N = 516) (Holmes et al. 2015b), while in Australia and New Zealand, leaders in this field in the region, the respective success rates were of 58% (N = 126) and 54% (N = 306), respectively (e.g. Segal et al. 2022), this is likely due to the fact that Australia and New Zealand have conducted numerous early trial attempts and pioneering projects compared to other territories, which focused more on applying established methodologies (Towns and Broome 2003; Rauzon 2007; Holmes et al. 2015b). In contrast, eradications carried out in New Caledonia and Wallis showed a success rate of 90.48% and 90.91%, respectively, indicating high efficiency, although most operations there involved relatively small islands. In French Polynesia, the overall eradication success rate was closer to those observed in New Zealand and Australia. This is likely because, as in these two countries, eradications have also been attempted on larger islands (on average eight times larger than in New Caledonia), which typically makes eradications more challenging and thus reduces the overall success rate compared to smaller islands. The likely causes of failure in tropical environments are not yet fully understood. Even under theoretically unfavourable conditions, such as a high abundance of crabs, which compete for access to the bait, abundant food availability, and significant rodent populations, an operation can sometimes be successful (Samaniego et al. 2020). Nevertheless, identifying the real reasons for failure often remains challenging, especially several years later.

Various factors may contribute to the failure of an operation, including the lack of detailed reports. Possible causes are due to operational or human errors such as incorrect bait application or inadequate preparation (e.g. the absence of preliminary studies). Adverse weather conditions (Holmes et al. 2015b) such as heavy rainfall washing away bait or excessive delays between baiting rounds, can also allow rodent populations to recover. Operations must be properly budgeted, with adequate financial reserves for unforeseen expenses (Kappes et al. 2019). Ultimately, the success of an operation largely depends on thorough preparation and meticulous execution (Samaniego et al. 2021), including precise handling, appropriate monitoring supported by effective biosecurity measures (Kennedy and Broome 2019), and the production of comprehensive reports. However, reports of failed projects are often lacking or only available in limited detail (Catalano et al. 2019; Samaniego et al. 2021).

The virtually systematic use of Brodifacoum at 0.02 g/kg (97% of the operations examined) underlines that this has been accepted as an efficient bait for rodent eradication (Howald et al. 2007; Broome et al. 2010; Parkes et al. 2011). The diversity of bait application methods, ranging from simple manual spreading to more advanced techniques such as the use of drones, illustrates an adaptation to new technologies and local specificities (Segal et al. 2022) allowing larger surfaces or steeper islands to be treated (Broome et al. 2017; Hoffmann et al. 2023). Drones have expanded the possibilities for effective aerial baiting methods on small remote islands, particularly in French Polynesia, where expertise was limited and helicopters would be expensive (e.g. drone projects for the islets of Ua Pou and Kamaka). However, as an innovative technology, it remains a very expensive solution compared to ground-based methods. Additionally, ground-based methods allow project managers to employ more local staff and actively involve them in the project. This is advantageous as it helps to raise awareness of biosecurity rules within local communities and encourages their participation in the project. One of the main drawbacks of using drones is their limited capacity to operate under challenging weather conditions. On Rapa Island, where wind and sea conditions are rarely optimal, drone technology would not have been cost-effective. Adapting methods according to the budget is essential. Another specificity of the French Pacific islands is a level of social acceptance that is probably lower as compared to Australia and New Zealand. New Zealand conducts a significant amount of research on social acceptance, for example, in the context of the Predator Free 2050 program (Russell 2014; Dickie and Medvecky 2023).

The documents and reports we reviewed rarely mentioned the education and awareness of local populations to biosecurity protocols. Collaboration between managers, NGOs, local communities, and the private sector is crucial for the successful implementation of biosecurity measures. It allows for integrated management and better social acceptance (Harper 2020). A better integration of local populations early in these projects, with regards to both biosecurity and eradication operations, is clearly needed to improve their success, but also to prioritise projects that are also meaningful for local communities. However, it can be challenging to change the habits of populations, especially when cultural aspects are involved, which requires sustained communication efforts over time. Local social acceptance, as well as institutional acceptance, is not easy to achieve (Crowley et al. 2017) in a tense and ‘fragmented’ context in terms of decision-making and political positioning on biodiversity and its conservation challenges, as is the case in New Caledonia (Rodary 2024). This situation is further complicated by deep cultural differences. Gaining the support of local customary authorities at various levels, as well as that of institutional and administrative authorities, is complex and takes time. It is also essential to understand how local communities perceive invasive species, as these perceptions can significantly influence the success or failure of eradication projects (Kapitza et al. 2019).

This synthesis also revealed that, although biosecurity was systematically mentioned in the feasibility studies, anti-reinvasion measures were not systematically implemented nor maintained over time, particularly in New Caledonia compared with French Polynesia. In New Caledonia and Wallis, only limited communication and public information actions appear to have been systematically implemented. Although communication campaigns can be useful (Haley et al. 2023), they are generally not sufficient in order to efficiently detect and prevent the natural or human-induced reinvasion of islands by rodents; specific surveillance, maintained operational over time to guarantee an early detection of re-invasion, is essential (e.g. Russell et al. 2008b; Davis et al. 2023).

It would be relevant to complete the database by including information such as costs (largely missing information in the analysed documents here) and the number of local people involved. Additional data on cases of multi-invasion, vegetation type and density, rodent abundance, crab abundance, allocated budget, baiting grid layout, quantity of bait used, accessibility challenges of the islands, and the types of islets, could also be incorporated. This would enable multifactorial analyses to identify the causes of both success and failure in these territories.

Supplementary material

Supplementary material is available online.