For the Pacific Islands, community-based adaptation activities are crucial, and yet it remains uncertain whether they are effectively promoting long-term adaptive capacity. Here we evaluate the performance of 32 community-based adaptation initiatives across 20 rural communities in the Pacific. We find that initiative appropriateness was a strength while sustainability was a consistent issue, locally funded initiatives and those implemented by non-governmental organizations were more likely to perform better, and climate awareness-raising initiatives and those integrated with ecosystem-based adaptation performed best. We also identify four multidimensional and interdependent optimization points for future community-based adaptation initiatives: local approval and ownership, shared access to and benefit from initiatives, integration of local realities, and systems-thinking and forward planning. Our analysis suggests the need for a praxis shift whereby adaptation is locally led, communities drive their own agendas, and donors and implementers become facilitators that resource the diverse capacities of communities and help achieve local objectives equitably.
Despite contributing little to global carbon emissions, the impacts of climate change on small islands are very real1. With projections showing that island communities will continue to face worsening climate change impacts over the remainder of this century and beyond, understanding what successful adaptation looks like for island countries such as those in the Pacific is urgent2. This imperative is furthered as the substantial funding that has flowed into the Pacific Islands region for climate change adaptation has made little progress towards intended objectives to date, and funding may dry up3.
With growing understanding of diminishing returns resulting from top-down climate change responses, assistance is increasingly being delivered at the local scale through bottom-up responses such as community-based adaptation (CBA)4,5. CBA is a small-scale, place-based and grassroots driven approach that has synergies with broader development aspirations6. In principle, the local-scale focus of CBA provides an opportunity for adaptation to better acknowledge and integrate existing local knowledge, capabilities, priorities and context of the community and for impacts to be addressed at the scale at which they are experienced7,8. Achieving effective and sustainable adaptation that promotes reflective engagement with the community (‘ideal CBA’)9, however, is not as straightforward as often implied10. Negative impacts can occur if CBA fails to adequately represent vulnerable populations and generate long-term social resilience11,12.
Despite the range of CBA activity in the Pacific Islands region, uncertainty remains around whether these communities are becoming better prepared to cope in the long term13,14. Bottom-up approaches such as CBA are important in Pacific islands as they can support and use traditional governance systems to help mobilize resources and better use traditional knowledge to support sustainable adaptation15,16. Further, the complementarity that bottom-up approaches often exhibit with cultural norms and connections to land and place in most Pacific islands is critical for socio-ecological resilience and influencing behaviour17. Given the importance of this approach to this region, good practice for bottom-up approaches such as CBA must be further developed and shared as a matter of urgency7,18.
While high-performance adaptation stories are championed to ensure they can be scaled up and out in future programmes, stories of adaptation requiring work and improvement are rarely reported or exploited as a source of learning19. This is regrettable because these lessons are as salient as success stories for optimizing CBA, which has been increasingly implemented despite research on its efficacy being in its infancy20,21. Previous studies that have evaluated CBA have largely been context-specific case studies18,20,22,23,24 offering in-depth insights in specific locations and for particular CBA initiatives but not always providing transferable learnings (although exceptions do exist25,26). Further, adaptation research in small island developing states tends to focus on core or near-core areas of a country rather than rural, peripheral areas that are more difficult to access27,28. By undertaking site evaluations on the success of CBA in increasing adaptive capacity across rural communities in the Pacific, this study tracks progress at a larger scale and addresses these shortcomings.
To evaluate CBA progress, we explore local perspectives and embedded experiences of rural community members across four Pacific Island countries: Fiji, Federated States of Micronesia (FSM), Kiribati and Vanuatu. While the selection of these countries to some extent reflected researchers’ existing networks, it was intentional in that it included one atoll country (Kiribati), one mixed atoll and high-island country (FSM) and two high-island countries (Fiji and Vanuatu), with contrasting economic rankings, to sample a representative range of contexts among Pacific Island countries. Within each island country, rural communities were sampled along core–periphery gradients to capture differences in adaptation contexts, as earlier work shows exist27,29,30. In Pacific Island countries, most rural communities are largely subsistence based, occupying land to which they have title and have done so for generations, and engage in livelihoods that are at least partly culturally grounded29. While there are variations in the nature of such rural communities attributable to peripherality29,30, their commonalities dominate, making site selection for this study more straightforward than would be the case elsewhere.
There is no consensus on, and a lot of ambiguity around, what ‘successful’ adaptation is and how to measure it31,32. On the basis of a review of literature about what constitutes ‘successful’ adaptation, five key evaluation components were deemed appropriate for this study: appropriateness, effectiveness, equity, impact and sustainability (see Table 1 for definitions). This study is based on participant perspectives of CBA performance considering each of these components. Perspectives were derived from 415 participants in 44 focus groups and 62 in-depth interviews between April 2017 and July 2019 (see Methods). Thus, this study relies on embedded and subjective experiences of participants involved. Understanding how individuals and communities perceive their own adaptation experiences is critical as cognitive aspects (for example, perceived self-capacity or perceived adaptation efficacy) are crucial to adaptation intention, community buy-in and adaptive capacity33,34,35. Further, measuring perspectives of individuals towards their own situation allows for more meaningful comparisons within contexts and through time, especially when the metrics are meaningful to local people36. Comparisons through time will be critical as it is too early to accurately determine whether long-term adaptive capacity and vulnerability reduction will be achieved (even by initiatives perceived as high performing at the time of this study). Therefore, although this study still provides valuable insights, there remains a need for ongoing monitoring and evaluation to assess long-term impacts.
CBA in rural communities across the Pacific
Twenty rural communities participated in this study, which gave rise to 14 case study sites as some sites included multiple communities of place (that is, based on ties to a physical space37) that were simultaneously involved in the same or similar initiatives. Exemplifying the miscellany of adaptive responses, this study evaluated 32 diverse CBA initiatives related to enhancing food security (n = 9) or enhancing water security (n = 8) or both simultaneously (n = 1), prevention of land loss (n = 5), relocation (n = 3), climate change awareness-raising (n = 2), marine resources protection (n = 2), and enhancing financial security (n = 2). Table 2 provides a summary of case study sites and CBA initiatives. Supplementary Table 1 describes the intended contributions to adaptive capacity of each type of CBA.
These initiatives are not each stand-alone projects but rather subset initiatives that can be aggregated into 15 overarching, broader projects. This paper is structured around subset initiatives rather than broader projects to capture the diverse experiences and outcomes that emerged for local participants, even within a single project. These initiatives were funded through 8 different funding bodies (3 international donors and 5 local sources) and implemented by 13 different agencies (9 government-related institutions/departments, 3 non-government organizations (NGOs) and 1 local university), some of which partnered to fund or implement initiatives jointly.
Overall initiative performance
We found that initiatives generally performed well in terms of appropriateness while sustainability required substantial improvement across almost all initiatives (see Extended Data Fig 1). Some initiatives had high performance across all five components of adaptation success except sustainability, highlighting how even when other components performed well, sustainability can be the most difficult to achieve because it needs to stand the test of time. On the basis of participant perspectives on CBA performance in terms of the five components, evaluated initiatives were placed along a scale from high to low performance (Fig. 1, based on Extended Data Fig 1).
Although there were no particularly strong patterns in terms of the spread of variables across the scale (and thus no particularly strong influences from the variables on initiative performance), there were some noteworthy patterns. Locally funded initiatives were, for example, proportionately more evident among high- and medium-performing initiatives. Although internationally funded initiatives had a more even spread across the scale, they dominated the low-performing category due to locally funded initiatives performing proportionately higher. Similarly, initiatives implemented by NGOs (whether with local or international funding) were proportionately more present within high- and medium-performing groups while those implemented by governments or universities had a more balanced spread across the scale or largely occupied medium-performing groups, respectively. Climate awareness-raising and marine resources protection (integrated with ecosystem-based adaptation (EbA)) initiatives also tended to have higher performances than others, while initiatives focused on preventing land loss had lower performances overall. There was a largely scattered spread of countries on the scale, except for FSM, which had proportionately more high- and medium-performing initiatives. Through latent content analysis, we also found that a series of factors related to initiative design and implementation shaped the perceived performance of initiatives; these common characteristics are discussed in detail in the following sections and summarized in Tables 3–5.
Initiatives with high performance had several attributes in common (Table 3), one of which was the effective production of outputs linked directly to adaptive capacity. Improvements in climate change awareness and knowledge, for example, enhanced abilities of community members to interpret experienced changes, thereby supporting disaster preparedness and enhancing perceived capacities to cope: “The awareness [sessions] are one thing that we learn a lot [from] … it’s [be]coming clear that, yes, our weather is changing” (Vanuatu-based participant, 2017). Many initiatives also proved appropriate to local context by being tailored to cultural specificities as well as community priorities, resources and livelihoods. In several cases, this nurtured a sense of local approval and positive sentiments around initiative sustainability. Despite many of these high-performing initiatives achieving local approval and appropriateness, the tendency to be top-down in nature emerged as they were largely designed and implemented by external actors (rather than grassroots-driven as in ‘ideal CBA’). The climate awareness-raising initiatives appeared to be somewhat co-driven by local communities as there were presentations from community members in local dialects. These initiatives, however, were also not wholly driven or shaped by local communities and their contexts as some participants (referring to V4b) pointed to the lack of focus on drought as their most perturbing livelihood concern.
Another commonality among high-performing initiatives was the tendency to move away from archetypal parameters of ‘communities’, which are usually based on ties to physical space. These approaches enhanced the equity of initiatives by minimizing internal disputes and having more widespread benefit (for example, by being open to community members across an island or through encouraging information/skill transfers beyond those directly involved). This is not to say that initiatives based on geographic boundaries proved incapable of equity. The success of climate change awareness-raising initiatives (V4b, V5a), for example, was partly dependent on equitable modes of information transfer and delivery for those with low literacy and disabilities (for example, use of a projector for animations, pictures and presentations).
Several initiatives were also able to simultaneously address climatic and non-climatic pressures or issues for rural livelihoods, thereby supporting an overall reduction in vulnerability across a relevant system (systems-thinking) and, in one case, fostering a sense of local approval and satisfaction. The positive sentiments that emanate from this characteristic are exemplified by the following statement related to an initiative that simultaneously addressed non-climatic concerns related to housing (F2): “We were very happy with the relocation … it solved the problem where there was three, four generations [crowded] in one house [in the old village]” (Fiji-based participant, 2017).
The two top-performing initiatives were similar in their integrated approach where, although focused on a community scale, they were also based on the protection of a local ecosystem, which overlaps with EbA. Through this integration, these initiatives simultaneously occasioned several of the aforementioned common characteristics: appropriateness to local context (that is, based on ecosystems critical to rural livelihoods), moving away from geographical parameters (that is, basing the ‘community’ around those that use the ecosystem) and simultaneously responding to livelihood pressures that are not always directly linked to climate (for example, income and food security).
Although these are high-performing initiatives, several issues arose. Poor sustainability emerged and was largely due to poor future planning, where “no one factored in the future” (FSM-based participant, 2018), especially in terms of local trends in climate change and population growth. The tendency for some short-term benefit but poor long-term impact is depicted by one woman: “Water tanks have helped but there isn’t enough for anything except cooking and drinking—we are still bathing in saltwater … drought sees the water source dry up and the tank empty” (FSM-based participant, 2018).
Other issues that arose were more specific to certain initiatives and included delays in implementation processes, which engendered community resentment and disapproval (F5), the need for regular reinforcement of infrastructure (V2d) and some issues with equity and access to decision making (F2). The nature of the first issue further demonstrates how CBA tends to be top-down rather than grassroots driven. In terms of the latter issue, one of the community relocations (F2) was perceived as performing lower than the other (F5) as it had less-equitable decision-making processes for women. Positive aspects, however, outweighed these characteristics, resulting in these initiatives being perceived as high performing. It is important to note that F5 is still an ongoing initiative and the community has yet to relocate despite appropriate, inclusive and sustainable planning; time will tell whether this initiative has a sustained positive impact in the long-term.
Medium performing initiatives were perceived as having mixed positive and negative aspects. Some key characteristics (Table 4) echo those discussed in the preceding section. First, the importance of producing outputs is re-emphasized here as one initiative demonstrated how the inadequate generation of outputs impedes community motivation to maintain activities (that is, they will return to traditional ways) (V4a). Second, the importance of simultaneously catering to climatic and other (potentially non-climatic) livelihood pressures is negatively reinforced in two initiatives that lacked a systems-thinking approach, meaning that other sources of vulnerability (and any associated externalities) were overlooked. In one initiative, this impeded effectiveness and sustainability (V5c): “with the continuous disaster of the volcano … it’s hard to achieve the [adaptation] goals” (Vanuatu-based participant, 2017). This was because ash fall continuously destroyed provided climate-resilient crops and equipment. Similarly, although F3 (as the lowest-performing relocation initiative) was initially approved of by the community, the tendency among implementers to overlook other sources of vulnerability hampered perceptions around the long-term viability of the new site: “We were delighted with the move to the new houses, but we are still worried about the landslide because the houses are on the hill” (Fiji-based participant, 2017). Parallel hazards can perpetuate vulnerability in the system regardless of climate scenario.
Most medium-performing initiatives tended to be appropriate to the local context in terms of community priorities. Three initiatives, however, demonstrated how adaptation can be appropriate in terms of catering to priorities while having other contextual applicability issues (K1, V5b, V5c). These initiatives only partly integrated contextual aspects by focusing on the broad livelihood priorities but then also either relying on external resources (V5b, V5c) or overlooking more-specific yet related needs such as improved market access and water for irrigation (that is, the wider vulnerability context) (K1). These generated issues for sustainability as external resources were difficult for communities to replace for maintenance (V4b, V5c), and crops did not grow well (K1): “The biggest failure is the provision of the water. They said they would provide water that would spray the plants but they didn’t, so most of the crops die” (Kiribati-based participant, 2017). In Kiribati, having another poorly performing initiative has increased people’s stress and community scepticism about externally sponsored adaptation activities. These kinds of issues point to the problematic tendency for these initiatives to be designed and implemented by external actors, rather than being grassroots driven (as in ideal CBA).
In terms of equity, there was a nearly equal number of initiatives that had high performance as those that had issues. Several initiatives in FSM were perceived as equitable in terms of members being informed and consulted (FSM1a, FSM1b, FSM1c). This is a likely result of substantial outreach work on equity by NGOs but also the existence of matriarchal lineages, which ensures decision making around cultural, political and social community dynamics is largely gender neutral. A more-specific characteristic to note was the ability of some initiatives (V4a, V5b, V5c) to advance women’s empowerment and gender equity. As one female participant stated, “At first, I was shy and scared of talking but when [implementing agency] did more activities and training and involved us, this is where I started to build up confidence and become involved … I began to speak up for myself” (Vanuatu-based participant, 2017).
Initiatives with poor equity and exclusive tendencies in this group were largely a result of overlooking social and/or cultural considerations and dynamics. Two initiatives (V2g, V4a), for example, did not implement appropriate ‘community’ parameters for the cultural context and social dynamics, thereby giving rise to exclusion. One initiative (V2g) that merged two villages as one ‘community’ (for initiative purposes), for example, gave rise to internal disputes that resulted in one village being excluded: “it [initiative] was working for a couple of months [but] is now only serving the other community … we’re all related, both villages, but they’re a bit strong-headed” (Vanuatu-based participant, 2018). One participant (from V4a) suggested that establishing ‘community’ parameters on social characteristic boundaries, such as gender or religion rather than geography, may be more equitable. This reiterates the aforementioned potential for different ‘community’ parameters to be more equitable and impactful, but also highlights how the nature of these actions has been shaped by what may be more convenient and suitable for external actors. Local communities were not involved in designing and shaping the initiative to what might be most suitable to them and their local contextual factors, especially in terms of social and cultural considerations and/or dynamics when delineating ‘community’ boundaries. It is important to also note here that V4a illustrates how holistic equity is hard to achieve: although there was exclusion of a religious group, there were also progressions in gender equity and women’s empowerment for those involved.
Using the appropriate ‘community’ parameters to have more impact is not enough on its own. We found performance varied among initiatives that were all implemented through a district training centre as an atypical ‘community’. Two initiatives (V1b, V1c) were ranked lower than V1a (high-performing initiative), and this was largely because the former two required ongoing inputs and upkeep, which became difficult when the leader of the training centre changed. Further, a drought and concomitant lack of water rendered the fish farm (V1c) difficult to maintain, connoting the limited contextual appropriateness of the initiative.
Despite initiatives being appropriate to local priorities (for example, FSM1a, FSM3c) or effective in the short term (for example, FSM1b, FSM1c, V5b), several initiatives were not designed in ways (due to being designed by external actors) to promote and empower the self-sufficiency of communities and their abilities or motivations to sustain initiatives beyond project lifespans (that is, when funding and external support ceased). This included the reliance on external resources, which made it difficult for communities to source and replace equipment parts, as well as poor design elements, which rendered maintenance more difficult than necessary. In terms of the former, one participant asserted that it would “be better if you could take something that is already inside of the community and use that” (Vanuatu-based participant, 2017). Promoting community self-sufficiency is critical for better forward planning and initiative sustainability.
Low-performing initiatives represent initiatives that local communities perceived as having more issues than not, albeit most were still appropriate to local context in terms of community priorities and aspirations. Among these initiatives, some characteristics emerged (Table 5) that reiterate key findings presented in the preceding sections. Low-performing initiatives, for example, further underscore the problem with overlooking future trends as several initiatives became ineffective because of poor planning around, and consideration for, future climatic changes at the local scale. Two initiatives also further illustrate the importance of ensuring community self-sufficiency and/or integrating locally available resources for sustainability: V2e had unmaintainable high costs, which rendered the community unable to achieve self-sufficiency; V3b had a heavy reliance on external resources (for example, expertise and construction material) so that when difficulties emerged, there was a lack of local technical capacity and funds to sustain the initiative. One community member stated: “we should use local experience in design because we know the context … instead of getting someone to come in and then it doesn’t work” (Vanuatu-based participant, 2018). It is apparent that using existing labour and skills is critical for forward planning. This again demonstrates the top-down tendencies of CBA, where initiatives are being designed and implemented by external actors with external resources rather than driven by the community and their situated resources (as in ‘ideal CBA’).
Initiatives focused on preventing land loss generally had lower performances, and this largely stemmed from the tendency to be ineffective in preventing encroachment or inundation in the long term (FSM2b) and/or sidelining community approval and ownership in several ways. Lacking approval and/or ownership emerged from poor alignment with contextual factors (for example, perceived climate risk) due to being a top-down approach with little to no community input or consultation (F4), being privately managed (V2c) and having maladaptive outcomes (F1, F4). One initiative (F1) demonstrated how a lack of forward planning (overlooking future intensification of local climate impacts) resulted in maladaptation, thereby creating new vulnerabilities and disapproval in the community: it became “a liability to our children’s safety, it damages our nets [and] our livelihoods are being affected” (Fiji-based participant, 2017).
It is not solely prevention-of-land-loss initiatives that lacked local approval, ownership and/or consultation. Due to being designed and driven by external actors rather than the community, the lowest-performing initiative (V2f) also overlooked contextual factors and community desires, which generated local disapproval and poor ownership: the “idea was not welcome” and “people were not comfortable” (Vanuatu-based participant, 2018). The community consequently lacked any motivation to maintain the initiative long term, and there was no desire to implement similar initiatives in the future.
Four interdependent points for optimization
Analysis of the characteristics of high-, medium- and low-performing initiatives suggest four points for optimizing future CBA. Across our sample, these optimization points, which are multidimensional and interdependent in nature, were both negatively and positively reinforced as key factors supporting initiative success.
Local approval and ownership
The importance of local approval and ownership is that it encourages higher levels of participation7 and ensures activities are not resisted but sustained beyond the end of the (funded) initiative24,38. Factors that contributed to local approval and ownership included the ability of initiatives to protect livelihoods, respond to direct requests from communities, blend in with local realities and produce benefits beyond reducing exposure to climate risk in the system (that is, address other critical livelihood pressures simultaneously). The latter two demonstrate the interdependencies with other optimization points and how they can influence each other. In particular, the importance of local legitimacy (through integrating local realities) for fostering community support and participation has been noted elsewhere7,17,24. The top-performing marine resources protection initiatives (integrated CBA–EbA initiatives), for example, fostered local approval by being based on, and appropriate to, ecosystems that are critical and relevant to the livelihoods of rural communities. Systems-thinking, in which the wider vulnerability context is recognized33, can similarly enthuse local approval and ownership over initiatives by simultaneously catering to the community’s other pressing livelihood concerns and baseline aspirations that are not climatic (thereby being ‘pro-poor’24,39). We found that some low-performing initiatives did not acquire genuine local approval and ownership because of a range of issues: inadequate consultation and tendency for initiatives to still be top-down in nature, attempts to ‘buy’ ownership, maladaptation and indifferent alignment with contextual realities. In several cases, this resulted in initiative demise after funding ceased, thereby negatively reinforcing the importance of this optimization point and demonstrating its criticality to sustainability and forward planning (as another optimization point)24,38. Instead of external actors attempting to foster local approval and ownership (as seen in this study), initiatives need to be driven by the community, where approval and ownership are inherent.
Shared access to and benefit from initiatives
CBA that challenges entrenched intracommunity inequalities and supports shared access to and benefit from initiatives is critical for sustainability40. In this study, aspects that supported shared CBA included adequate consultation with all community members, the focus on an ecosystem that can benefit and is accessible to all, and improving gender equality in decision making. The latter is crucial as social transformations have been emphasized as critical to CBA25 and gender-based exclusion was a key concern that has emerged in CBA literature4,7,23,24,41. Another important consideration that emerged here was the need to reconsider the geographic scale of ‘community’ as the most appropriate entry point for shared and equitable adaptation. This is because alternative scales may be critical for building social capital and optimizing shared access and/or benefit (for example, approaches based on landscape, ‘whole-of-island’, ecosystems, local institutions or particular population groups)42,43. Basing initiatives on geographic community parameters (which was most common) occasionally excluded certain groups and/or inflamed internal disputes. This reminds us of the elusiveness of ‘communities’ and how the ‘framing’ of a ‘community’ imposed by external actors may not equally benefit intended participants42,43. ‘Community’ boundaries and scales should be understood as complex and contextually dependent24,40, and uncritical assumptions about ‘community’ homogeneity should be avoided4,30,40. External actors must strive for CBA that is driven by the local community as this would ensure that it is communities themselves who delineate their own ‘community’ boundaries on the basis of nuanced social/cultural dynamics and other contextual factors (thereby integrating local realities as another optimization point). This could help avoid the creation of new inequalities (for example, through internal disputes, exclusions or marginalization as seen in this study). External actors should, however, be wary of, and be prepared to help communities address, any other entrenched inequalities within the chosen ‘communities’ to avoid exacerbating them40.
Integration of local realities
Integrating local realities and existing social forms (for example, culture and its shared beliefs, shared metaphors and folkore) improves the likelihood of success for CBA initiatives7,25,44. In this study, as noted elsewhere7, compatibility with local context in terms of culture, priorities, resources, knowledge and livelihoods ensured that activities were not resisted and thereby sustainable. The importance of this optimization point was negatively reinforced in this study by initiatives involving the use of non-local resources as, regardless of any local approval or ownership, this generated difficulties for long-term maintenance and sustainability (that is, forward planning, as discussed in the following section). Undue dependence on external resources has previously been noted as unsustainable and potentially harmful to livelihood futures in such contexts3. Further, overlooking contextual realities more generally can also result in poor ownership, outcomes and sustainability as initiatives may not adequately integrate local processes or be appropriate to key livelihood risks and concerns (for example, constructing a river wall when a sea wall was needed). Local knowledge is critical here as it ensures a better understanding of local processes in the system15. These kinds of issues underscore the importance of having adaptation that is wholly driven by the community rather than external implementers who may overlook these contextual nuances and situated resources. In summary, this optimization point cuts across all other optimization points by playing a crucial role in ensuring community approval and ownership7,24, supporting shared adaptation in terms of benefit and access (for example, by ensuring a better understanding of social dynamics and appropriate community ‘parameters’), as well as being critical to systems-thinking and forward planning (for example, by integrating local resources and understanding the local processes in the wider vulnerability context).
Relevant here is the higher performance among locally funded initiatives compared with those funded internationally. This may be because locally funded initiatives are more likely to be designed and driven by people familiar with local contexts and sociocultural nuances, as opposed to outsiders who may have prescriptive funding conditions43,45,46. This is still, however, one step away from being driven by communities themselves.
Systems-thinking and forward planning
The last optimization point describes the ability of initiatives to have impacts beyond project aims, particularly in terms of addressing vulnerabilities across the relevant system and having long-term impacts (involving changes to accustomed practice). It became evident that bypassing a system approach in which dynamic pressures are viewed as part of a larger, interrelated system or context47,48 can contribute to ineffectiveness and unsustainability. Two initiatives, for example, demonstrated how overlooking parallel sources of vulnerability in adaptation can perpetuate overall vulnerability irrespective of climate scenario because rural livelihoods are challenged by multiple dynamic pressures49.
Other initiatives positively reinforced this optimization point by producing development-related benefits that were beyond climate-related aims, thereby actualizing a key attribute of ‘ideal CBA’ (that is, a ‘no-regrets’ approach embedded within broader development goals6,39,50). A key characteristic that contributed to the high performance of the two top-performing integrated CBA–EbA initiatives, for example, was the abatement of environmental degradation, which then minimized indirect negative socio-economic impacts (for example, in terms of finance and food security) from climate change. The capacity for CBA–EbA integration to combine adaptation and sustainable development has been recognized elsewhere51,52. This is critical as most adaptation-related needs may not be directly related to climate but involve laying the foundations for empowering individuals and communities to pursue adaptation outcomes that they value (for example, through developing capabilities)36. An understanding of local context, including priorities and knowledge, will ensure that the most critical adaptation-related needs in a specific context are addressed. In summary, there is a greater need for systems-thinking approaches to ensure that externalities are foreseen, that adaptation is delivered in a way that encourages recipients to act on their own33 and that activities are sustainable47.
Having impact over time emerged as a key weakness. Several initiatives focused too closely on present-day conditions and hazards, which, although often effectively reduced short-term vulnerability, overlooked the potential for change. This tendency to apply adaptation to everyday climate-related challenges rather than future issues is a problem recognized across most documented adaptation in small island developing states28. These initiatives were not ‘future-proofed’ to deal with the dynamism of vulnerability and enhance long-term resilience53, thereby occasionally resulting in maladaptation. One initiative that was not flexible to future changes involved the construction of sea walls. This highlights one of the major issues with infrastructural developments that sometimes unintentionally create trajectories that are path dependent and difficult to change54. Having adaptation that is community driven and integrates local resources also proved important here as the inability to replace or maintain non-local resources meant communities were not able to be self-sufficient in the long term, especially after funding and external support ceased. Approaches that are flexible to both present conditions and longer timescales (forward planning) are critical55 for building long-term resilience and for avoiding these short-term (sometimes maladaptive) implications.
We found mixed performance among CBA in 20 communities across four Pacific Island countries. While appropriateness was largely positive, sustainability was an area requiring considerable improvement. Initiatives that are locally funded, implemented by NGOs, focused on climate awareness-raising and integrated with EbA were also more likely to be high performing. The latter illustrates why CBA–EbA integration is often promoted5,52. Collaborative learning between the two approaches may be critical for optimizing adaptation5,51, especially in rural communities where livelihoods are inextricably linked to local ecosystems. Despite high performances among these initiatives at the time of this study, it is critical to remember that long-term improvements in adaptive capacity will become clearer over time.
Four multidimensional and interdependent optimization points were identified as key contents of the tool box for improving future CBA. They also point to the need to rethink CBA as not merely being ‘based’ in communities with some level of consultation or input but, instead, as something that is wholly ‘led’ by local people and local institutions43. This is critical as, despite genuine attempts to better integrate contextual specificities, initiatives across high- to low-performing groups demonstrated a tendency to still be driven by external actors (that is, top-down in nature). Despite being based in communities, initiatives are not driven or co-designed by them. This gave rise to a range of issues that hindered CBA performance: community stress and resentment, initiatives having contextual applicability issues (for example, relying on unsustainable external resources), initiatives being designed in ways that do not empower or promote community self-sufficiency, imposed ‘community’ boundaries and local disapproval or lack of ownership.
Rather than experiments in communities to improve their adaptive capacity deficit, communities would, in community-led adaptation, build on their strengths and design/drive their own adaptation aspirations on the basis of local knowledge, experiences and coping mechanisms. With communities leading their own adaptation agendas, numerous aspects of the optimization points would probably be autonomously incorporated: the wider vulnerability context and its local processes will be accounted for; local context will be underpinned, which means important priorities, resources and knowledge will be more appropriately used; and a stronger sense of local approval and ownership will be fostered. These factors, in turn, support sustainability (as the key area requiring improvement) as initiatives, being rooted in community aspirations and local context, will be self-sufficient over time. Related to this, a community-led approach may be better ‘future-proofed’ through the integration of local resources and local knowledge (for example, in terms of observed trends in local processes) that were often overlooked.
The role that implementers and donors should and can play is to become ‘facilitators’ of the desired adaptation aspirations for communities, rather than ‘doing’ adaptation ‘to communities’ under the guise of CBA43. External actors should resource and equip communities with any additional skills, resources and knowledge (that is, develop capabilities) that may optimize self-sufficiency and help communities achieve local objectives equitably and effectively36. This means moving beyond imposed standardized outcomes and, instead, focusing on empowering individuals and communities to pursue the adaptation outcomes they value36. Through such a process, implementers would support and empower ‘communities’ in their diversity of expressions but also guide them in certain processes (for example, supporting shared adaptation in terms of access and benefit or empowering women in leadership positions). Through local facilitators ‘guiding’ the resources and supporting participatory processes with holistically targeted equity framings, deeper understanding of any contextual idiosyncrasies related to inequality and power may overcome critiques of community-driven development56.
Ultimately, the role of implementers and donors is to prepare and support communities to lead their own adaptation because, as we have seen, implementers have left not long after the project has been implemented, and adaptation funding may ultimately dry up3. Despite the supposed development deficit that is projected on communities from the outside, such communities have always been resilient, and there is still much that the CBA field can learn from traditional governance systems, coping strategies and Indigenous knowledge15. Our findings suggest that it is time to support these communities’ diverse capacities and allow their situated and tacit resources to flourish.
This study involved a field-based qualitative approach using focus groups and interviews to explore the multiple perceptions, attitudes and values that people hold in relation to implemented CBA initiatives. Focus groups were used because of the rich understandings that emanate from their ability to mitigate researcher influence, allowing local participants to own discussion spaces63,64. Group sessions are also particularly useful when working with communal societies like those in the Pacific Islands65. Focus groups were split by gender or age where possible to limit influences from entrenched hierarchies. Individual semistructured interviews, structured to focus conversation yet also flexible to optimize the potential for knowledge production66, were also conducted to acquire information that may not have been disclosed in group settings67. This multimethod enquiry facilitated data completeness and confirmation in support of more-coherent understandings68. Where researchers were not competent in preferred vernaculars, gatekeepers were used for translations and for providing trusted access to rural communities69,70.
Focus group and interview design
To elicit a robust understanding around the performance of initiatives in reducing vulnerability and to ensure community sites were evaluated according to the same criteria, an evaluative framework was developed. This framework comprised of five key components that constitute ‘successful’ adaptation: appropriateness, effectiveness, equity, impact and sustainability. These were selected on the basis of a review of literature on monitoring and evaluation indicators and rubrics50,57,59,71,58,60,61,62,72 that was conducted at the initial stages of this research project. Although a range of components was found, only those that were considered relevant and appropriate for the study participants as well as the specificities of CBA were selected (Table 1). The focus group guides and interview schedules were structured by these components and, through a research project inception workshop, were also developed in consultation with eight partner organizations that have long histories of working with communities at the grassroots level. This ensured contextual relevance and the optimization of overall impact by aligning with the needs of their work.
The focus group and interview guides had clear and logical structures: introductions and verbal or written consent, an icebreaker to foster rapport, questions/activities around livelihood threats, climate experiences and responses, questions/activities relevant to each evaluation component, and opportunity for final remarks69. Activities involved creating ranked lists of livelihood threats and creating time lines showing major climatic events and details of what communities did to respond to these changes and impacts.
Forty-four focus groups and 62 semistructured interviews were conducted by seven researchers, who are authors on this paper, across 20 rural communities in Fiji (n = 6), FSM (n = 3), Kiribati (n = 2) and Vanuatu (n = 9) between April 2017 and July 2019. The selection of communities was dependent on two factors: (1) the research team’s networks and partner organizations as this was deemed the most appropriate avenue for trusted access to communities and (2) the ability to best represent different points along core–periphery gradients.
The total number of focus groups can be disaggregated into 18 all-male groups, 20 all-female groups, four mixed groups and two youth groups. All 415 participants across the focus groups and interviews were involved to some extent in CBA initiatives implemented by external agencies (that is, by government, NGOs, university or a combination of these). A majority of the participants were local community members, although government officials and representatives from donor or regional organizations were also interviewed. As our sampling strategy was largely based on expediency, this study may be limited in its capacity to be representative of all diverse groups. Attempts to capture diverse viewpoints, however, have been made by selecting communities across core–periphery gradients and implementing some level of quota sampling to ensure that participants of diverse age and gender were included.
Gatekeepers, who were employees of partner organizations or from our networks in these countries, were used for translations in preferred vernaculars. Gatekeepers were additionally critical for providing necessary introductions and inviting participants to be involved. Most focus groups and interviews were recorded using a digital recorder, or in cases where recordings were not taken, detailed notes were taken. As a study involving human participants, a series of ethics approvals was provided by the School of Earth and Environmental Sciences committee at the University of Queensland (nos. 20170302, 20170503 and 20170601), the University of Queensland central committee (no. 2018001985) and the University of the Sunshine Coast central committee (no. A/16/796). All participants were provided with detailed information about the study before providing verbal or written informed consent to participate in these voluntary focus groups and interviews. We also obtained permits to undertake research in each of the four countries.
The recordings of focus groups and interviews were transcribed verbatim, and all primary researchers were involved in two layers of latent content analysis, which is an interpretive technique employed for its ability to code social data and derive both surface and underlying meanings73. First, to understand participant judgements on the performance of CBA initiatives against the chosen evaluation components, a coding system was used. This involved analysing the language and sentiments around each initiative and categorizing them by four broad analytical codes74 under each evaluation component: high, neutral (that is, no perceived effect), medium (that is, with both high- and low-performance aspects) and low performance (see Extended Data Fig 1). Each initiative was given a total score by receiving three points for every high-performing component, two points for every neutral or medium-performing component and one point for low-performing components. From this, the overall performance of each initiative could be derived, and these fell neatly into ten separate groups, with each group containing the initiatives that had the same total scoring count. These results have been placed along a scale from high to low performance (Fig. 1). The second layer of analysis involved transforming findings under all evaluation components into key ‘themes’, which are groups of content that share common underlying meanings on an interpretive level75. This gave rise to the common characteristics outlined in Tables 3–5 and the four optimization points.
Simple measures of proportion were also used to identify any potential patterns in the spread of variables across the performance scale (Fig. 1). This involved calculating how many initiatives of a specific country, adaptation type, funding body or implementing agency were present in the high- (Groups 1–3), medium- (Groups 3–6) and low- (Groups 7–10) performing groups in relation to their total count (across all groups). For example, there were 13 locally funded initiatives in this study, of which five were high performing, six were medium performing and two were low performing compared with the 18 internationally funded initiatives, which saw five high performing, seven medium performing and six low performing.
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
The data that support the findings of this study are not publicly available due to them containing information that would compromise research participant confidentiality and anonymity.
Nurse, L. A. et al. in Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part B: Regional Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (eds Barros, V. R. et al.) Ch. 29 (Cambridge Univ. Press, 2014).
Nunn, P. & Kumar, R. Understanding climate–human interactions in Small Island Developing States (SIDS). Int. J. Clim. Change Strateg. Manag. 10, 245–271 (2018).
Nunn, P. D. & Kumar, R. Cashless adaptation to climate change: unwelcome yet unavoidable? One Earth 1, 31–24 (2019).
Dodman, D. & Mitlin, D. Challenges for community-based adaptation: discovering the potential for transformation. J. Int. Dev. 25, 640–659 (2013).
Reid, H. Ecosystem- and community-based adaptation: learning from community-based natural resource management. Clim. Dev. 8, 4–9 (2016).
Schipper, E. L. F., Ayers, J., Reid, H., Huq, S. & Rahman, A. Community-based Adaptation to Climate Change: Scaling It Up (Routledge, 2014).
Ensor, J. & Berger, R. Understanding Climate Change Adaptation: Lessons from Community-Based Approaches (Practical Action, 2009).
Kirkby, P., Williams, C. & Huq, S. Community-based adaptation (CBA): adding conceptual clarity to the approach, and establishing its principles and challenges. Clim. Dev. 10, 577–589 (2017).
Pelling, M. Adaptation to Climate Change: From Resilience to Transformation (Routledge, 2011).
Adger, W. N. & Barnett, J. Four reasons for concern about adaptation to climate change. Environ. Plan. 41, 2800–2805 (2009).
Ensor, J. Adaptation and Resilience in Vanuatu: Interpreting Community Perceptions of Vulnerability, Knowledge and Power for Community-Based Adaptation Programming (Oxfam, 2015).
Taylor Aiken, G., Middlemiss, L., Sallu, S. & Hauxwell-Baldwin, R. Researching climate change and community in neoliberal contexts: an emerging critical approach. Wiley Interdiscip. Rev. Clim. Change 8, e463 (2017).
Hay, J. & Mimura, N. Vulnerability, risk and adaptation assessment methods in the Pacific Islands region: past approaches, and considerations for the future. Sustain. Sci. 8, 391–405 (2013).
McNamara, K. E. Taking stock of community-based climate-change adaptation projects in the Pacific. Asia Pac. Viewp. 54, 398–405 (2013).
Nalau, J. et al. The role of indigenous and traditional knowledge in ecosystem-based adaptation: a review of literature and case studies from the Pacific Islands. Weather Clim. Soc. 10, 851–865 (2018).
McLeod, E. et al. Lessons from the Pacific Islands—adapting to climate change by supporting social and ecological resilience. Front. Mar. Sci. 6, 289 (2019).
Dacks, R. et al. Developing biocultural indicators for resource management. Conserv. Sci. Pract. 1, e38 (2019).
Jamero, M., Motoharu, O., Miguel, E. & Nicholson, T. Community-based adaptation in low-lying islands in the Philippines: challenges and lessons learned. Reg. Environ. Change 18, 2249–2260 (2018).
Catalano, A. S., Lyons-White, J., Mills, M. M. & Knight, A. T. Learning from published project failures in conservation. Biol. Conserv. 238, 108223 (2019).
Ashley, L., Zhumanova, M., Isaeva, A. & Dear, C. Examining changes in local adaptive capacity resulting from climate change adaptation programming in rural Kyrgyzstan. Clim. Dev. 8, 281–287 (2016).
Piggott-McKellar, A. E., McNamara, K. E., Nunn, P. D. & Watson, J. E. M. What are the barriers to successful community-based climate change adaptation? A review of grey literature. Local Environ. 24, 374–390 (2019).
Simane, B. & Zaitchik, B. The sustainability of community-based adaptation projects in the Blue Nile Highlands of Ethiopia. Sustainability 6, 4308–4325 (2014).
Vardakoulias, O. & Nicholles, N. in Handbook of Climate Change Adaptation (Ed. Filho, W. L.) Ch. 3 (Springer, 2015).
Remling, E. & Veitayaki, J. Community-based action in Fiji’s Gau Island: a model for the Pacific? Int. J. Clim. Change Strateg. Manag. 8, 375–398 (2016).
Ensor, J. E., Park, S. E., Attwood, S. J., Kaminski, A. M. & Johnson, J. E. Can community-based adaptation increase resilience? Clim. Dev. 10, 134–151 (2018).
Ford, J. D. et al. Preparing for the health impacts of climate change in indigenous communities: the role of community-based adaptation. Glob. Environ. Change 49, 129–139 (2018).
Nunn, P., Aalbersberg, W., Lata, S. & Gwilliam, M. Beyond the core: community governance for climate-change adaptation in peripheral parts of Pacific Island countries. Reg. Environ. Change 14, 221–235 (2014).
Klöck, C. & Nunn, P. D. Adaptation to climate change in Small Island Developing States: A systematic literature review of academic research. J. Environ. Dev. 28, 196–218 (2019).
Nunn, P. D. & Kumar, R. Measuring peripherality as a proxy for autonomous community coping capacity: a case study from Bua Province, Fiji Islands, for improving climate change adaptation. Soc. Sci. 8, 1–26 (2019).
Korovulavula, I., Nunn, P. D., Kumar, R. & Fong, T. Peripherality as key to understanding opportunities and needs for effective and sustainable climate-change adaptation: a case study from Viti Levu Island, Fiji. Clim. Dev. https://doi.org/10.1080/17565529.2019.1701972 (2019).
Ford, J. D., Berrang-Ford, L., Lesnikowski, A., Barrea, M. & Heymann, J. How to track adaptation to climate change: a typology of approaches for national-level application. Ecol. Soc. 18, 40 (2013).
Ford, J. D. et al. Adaptation tracking for a post-2015 climate agreement. Nat. Clim. Change 5, 967–969 (2015).
Kuruppu, N. & Willie, R. Barriers to reducing climate enhanced disaster risks in Least Developed Country–Small Islands through anticipatory adaptation. Weather Clim. Extrem. 7, 72–83 (2015).
Warrick, O., Aalbersberg, W., Dumaru, P., McNaught, R. & Teperman, K. The ‘Pacific Adaptive Capacity Analysis Framework’: guiding the assessment of adaptive capacity in Pacific island communities. Reg. Environ. Change 17, 1039–1051 (2017).
Hagedoorn, L. C. et al. Community-based adaptation to climate change in Small Island Developing States: an analysis of the role of social capital. Clim. Dev. 11, 723–734 (2019).
Dilling, L. et al. Is adaptation success a flawed concept? Nat. Clim. Change 9, 572–574 (2019).
Duane, T. P. Community participation in ecosystem management. Symposium: the ecosystem approach. Ecol. Law Q. 24, 771–797 (1997).
Richmond, N. & Sovacool, B. K. Bolstering resilience in the coconut kingdom: improving adaptive capacity to climate change in Vanuatu. Energy Policy 50, 843–848 (2012).
Heltberg, R., Gitay, H. & Prabhu, G. Community-based adaptation: lessons from a grant competition. Clim. Policy 12, 143–163 (2012).
Buggy, L. & McNamara, K. E. The need to reinterpret “community” for climate change adaptation: a case study of Pele Island, Vanuatu. Clim. Dev. 8, 270–280 (2016).
Ensor, J. Resilience Realities: Resilience and Development Practice in Vanuatu (Oxfam, 2016).
Titz, A., Cannon, T. & Krüger, F. Uncovering ‘community’: challenging an elusive concept in development and disaster related work. Societies 8, 71–99 (2018).
Westoby, R., McNamara, K. E., Kumar, R. & Nunn, P. D. From community-based to locally led adaptation: evidence from Vanuatu. Ambio https://doi.org/10.1007/s13280-019-01294-8 (2019).
Waylen, K. A., Fischer, A., Mcgowan, P. J. K., Thirgood, S. J. & Milner-Gulland, E. J. Effect of local cultural context on the success of community-based conservation interventions. Conserv. Biol. 24, 1119–1129 (2010).
Barnett, J. & Campbell, J. Climate Change and Small Island States: Power, Knowledge and the South Pacific (Earthscan, 2010).
Nunn, P. D. The end of the Pacific? Effects of sea level rise on Pacific Island livelihoods. Singap. J. Trop. Geogr. 34, 143–171 (2013).
Maani, K. Decision-Making for Climate Change Adaptation: A Systems Thinking Approach (NCCARF, 2013).
Sanneh, E. S. Systems Thinking for Sustainable Development: Climate Change and the Environment (Springer, 2018).
Martin, P. C. M., Nunn, P., Leon, J. & Tindale, N. Responding to multiple climate-linked stressors in a remote island context: the example of Yadua Island, Fiji. Clim. Risk Manag. 21, 7–15 (2018).
Hedger, M. M. et al. Evaluation of Adaptation to Climate Change from a Development Perspective Desk Review (IDS, 2008).
Girot, P., Ehrhart, C. & Oglethorpe, J. Integrating Community and Ecosystem-Based Approaches in Climate Change Adaptation Responses (Ecosystems & Livelihoods Adaptation Network, 2012).
Roy, J. et al. in Special Report on Global Warming of 1. 5 °C (eds Masson-Delmotte, V. et al.) Ch. 5 (WMO, 2018).
Ranger, N., Harvey, A. & Garbett-Shiels, S. Safeguarding development aid against climate change: evaluating progress and identifying best practice. Dev. Pract. 24, 467–486 (2014).
Barnett, J. & O’Neill, S. Maladaptation. Glob. Environ. Change 20, 211–213 (2010).
Conway, D. & Mustelin, J. Strategies for improving adaptation practice in developing countries. Nat. Clim. Change 4, 339–342 (2014).
Mansuri, G. & Rao, V. Community-based and –driven development: a critical review. World Bank Res. Obs. 19, 1–39 (2004).
Villanueva, P. S. Learning to ADAPT: Monitoring and Evaluation Approaches in Climate Change adaptation and Disaster Risk Reduction: Challenges, Gaps and Ways Forward SCR Discussion Paper 9 (Institute of Development Studies, Christian Aid and Plan International, 2010).
Adger, W. N., Arnell, N. W. & Tompkins, E. L. Successful adaptation to climate change across scales. Glob. Environ. Change 15, 77–86 (2005).
Sterrett, C. L. Bringing Innovation to Scale: Resilience to Climate Change (CARE International, 2015).
The State of Adaptation under the United Nations Framework Convention on Climate Change (UNFCCC, 2013).
de Loë, R., Kreutzwiser, R. & Moraru, L. Adaptation options for the near term: climate change and the Canadian water sector. Glob. Environ. Change 11, 231–245 (2001).
Lane, M. B. & McDonald, G. Community-based environmental planning: operational dilemmas, planning principles and possible remedies. J. Environ. Plan. Manag. 48, 709–731 (2005).
Kamberelis, G. & Dimitriadis, G. Focus Groups from Structured Interviews to Collective Conversations (Routledge, 2013).
Leavey, P. Oxford Handbook of Qualitative Research (Oxford Library of Psychology, 2014).
Lata, S. & Nunn, P. Misperceptions of climate-change risk as barriers to climate-change adaptation: a case study from the Rewa Delta, Fiji. Climatic Change 110, 169–186 (2012).
Brinkmann, S. in Oxford Handbook of Qualitative Research (ed. Leavey, P.) Ch.14 (Oxford Library of Psychology, 2014).
Creswell, J. W. Research Design: Qualitative, Quantitative and Mixed Methods Approaches 3rd edn (Sage, 2009).
Seawright, J. Multi-Method Social Science (Cambridge Univ. Press, 2016).
Hennink, M. M. Focus Group Discussions (Oxford Univ. Press, 2014).
McFadyen, J. & Rankin, J. The role of gatekeepers in research: learning from reflexivity and reflection. GSTF J. Nurs. Health Care 4, 82–88 (2016).
Onestini, M. Final Evaluation Report for the Community Based Adaptation Programme (UNDP, UNV, GEF and UNOPS, 2013).
Faulkner, L., Ayers, J. & Huq, S. Meaningful measurement for community-based adaptation. New Dir. Eval. 2015, 89–104 (2015).
Bengtsson, M. How to plan and perform a qualitative study using content analysis. NursingPlus Open 2, 8–14 (2016).
Cope, M. in Qualitative Research in Human Geography 3rd edn (ed. Hay, I.) Ch. 14 (Oxford Univ. Press, 2010).
Graneheim, U. H. & Lundman, B. Qualitative content analysis in nursing research: concepts, procedures and measures to achieve trustworthiness. Nurse Educ. Today 24, 105–112 (2004).
We are very grateful to the 415 participants for providing such valuable, important and meaningful insights in this study. Without you, this study would not have been possible. We also thank and acknowledge the various local gatekeepers who were instrumental in helping with fieldwork logistics, introductions and access to communities, and translation. This research was funded through an Australian Research Council Linkage grant (number LP160100941).
The authors declare no competing interests.
Peer review information Nature Climate Change thanks James Ford, Carola Kloeck and E. Lisa F. Schipper for their contribution to the peer review of this work.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
White is high performance, light grey is neutral performance (that is no perceived effect), dark grey is medium performance (that is with high and low performance aspects) and black is low performance.
About this article
Cite this article
McNamara, K.E., Clissold, R., Westoby, R. et al. An assessment of community-based adaptation initiatives in the Pacific Islands. Nat. Clim. Chang. 10, 628–639 (2020). https://doi.org/10.1038/s41558-020-0813-1
Nature Climate Change (2021)
Generational retreat: locally driven adaption to coastal hazard risk in two Indigenous communities in Fiji
Regional Environmental Change (2021)
Natural Hazards (2021)
Identifying and mitigating risks to completion of small grant climate change adaptation projects: evidence from the Pacific
Regional Environmental Change (2021)