me 59 Numbers 11-12— Routledge Journal of Environmental Planning and Management 2103 Tyler aan aup Journal of Environmental Planning and Management, 2016 Vol. 59, No, 12, 2102-2123, http://dx.doi.org/10.1080/09640568.2015,1126241 is now also commonly used to describe the capacity of social systen stressors Brooks, Adger, and Kelly 2005; Cutter et me 2008; Alder eatin Resilience may be conceptualised with reference to ‘basins of attraction’ that describe the tendency of systems to settle toward equilibria. Most systems have numerous equilibria to which the social system may Converge depending on the nature and ; a fa a haa it i 1 Does et fe resilience Maines ed oe damage from climate magnitude of a given perturbation. Walker et al. (2004) define resilience as the capacity related disasters? Evidence based on survey data . A eo to maintain the current equilibrium, ie. to withstand perturbations. David Gawith*, Adam Daigneault™ and Pike Brown® ‘hee Klein, Nicholls, and Thomalla (2003) suggest that resilience should be used in 4 limited manner to describe the amount of stress that a system can withstand while “Department of Land Economy, University of Cambridge, Cambridge, United Kingdom; °Landcare femaining in the same basin of attraction. This conservative view of. resilience is echoed Research, AucHland, New Zaalende “Landeae Resterch Ciscoth New eed by Pelling and Manuel-Navarette (2011), Pearson and Pearson (2012), and Wilson et a | (2013), and is most widely applied in the literature on environmental hazacdk (Cutter (Received 30 April 2015; final version received 25 November 2015) et al. 2008). In contrast, Adger e7 al, (2005), Folke (2006), Gallopin (2006), Cutter et al (2008), and Brown and Westaway (2011) emphasise the i : : Policy-makers view community resilience as fundamental to mitigating loss and organise and to change their pia in = ae Se | damage from climate-related disasters, Although energy has been devoted to defining equilibria. Indeed, Gallopin (20 ti ean te Sor ebktt to thor resilience, less effort has been devoted to analysing the effects of resilience on loss. paar » Gallopin (2006) notes that the existence of multiple equilibria heme ties the concept of resilience from that of vulnerability. Damhofer (2014 466) and damage, which is critical in places with limited capacity for adaptation, We use survey data to develop a composite index of community resilience in Fiji and then evaluate the extent to which community resilience mitigates loss and damage. We find that community resilience is negatively correlated with damages over which human social systems.” In the literat : that com 5 ges over which h : re on ae intervention may be effestive, but not vit damages pa letervention i bt commonly conceptualised in order Laie ee is effective, suggesting that community resilience may limit impacts. We further - : gesting ty y y etal. 2008; Brown and Westaway 2011), . ae that this result holds for a cyclone (about which communities had substantial advance : warning) but not for river flooding (for which communities had little advanced In this paper, we adopt the definition used by Chapin et al. (2009, 24), who see warning), suggesting that early warning is necessary for community resilience to resilience as the capacity of a social system “to absorl ‘trum become responsive. o ’b a spectrum of shocks or Keywords: natural disasters; early warning; Fiji; flooding; cyclones 1. Introduction a nested adapti i iti iptive cycle in which antecedent conditi id i i Natural disasters affected 232 million people, killed over 100,000 people, and caused disaster impacts that may or may not lead to paeenlie ariel peas —- more than US$100 billion worldwide in damage each year between 2001 and 2010, on continuous, and the antecedent conditions change through mitig stig and me | average (Guha-Sapir et a/. 2011), Moreover, the number of natural disasters recorded per In the context of this model, we focus on assessing ae eee year has increased markedly since 1940 (Munang et al. 2013), and population growth and interacts with event characteristi i ie feapiniod sail clone ee riskeprone areas have increased the Acteristis fo enable or constrain coping responses. We analyse a tisk of loss and damage from natural disasters (IPCC 2012; Munang et al. 2013). It is likely that anthropogenic climate change will amplify the number and severity of such gine disasters over the coming century (Preston et al. 2006; Bates et al. 2008), although Anlecedent Condtions §=—§ C=——— pest event community resilience has been touted as a key strategy for managing increased disaster risk under climate change (¢.g., Tompkins and Adger 2004; Tompkins 2005; Boyd et al. 2008; Cutter et al. 2008; Tompkins ef al. 2010; Leichenko 2011; Nelson 2011; Wise Wh et al. 2014), q™ a fostiect, Stage Z The term ‘resilience’ originated in the physical sciences, where it is used to describe a JEN ne tee > at material’s ability to “store strain energy and deflect elastically under a load without breaking or being deformed” (Gordon 1978).' The term was first used in the ecology literature by Holling (1973) to describe the capacity of species and ecosystems to endure stress (Klein, Nicholls, and Thomalla 2003; Norris et al. 2008; Nelson 2011), Resilience ‘Absorplve Capacity Exceeded? Short-Term ===> Longer Term & Preparedness *Corresponding author. Email: daigneaulta@landcareresearch.co.nz Figure 1. The disaster resilience of place (DROP) taken from Cutter et al, (2008). © 2016 University of Newcastle upon Tyne2104 D. Gawith et al. snapshot of resilience represented by the first four elements in the DROP model. As such, while the DROP model is ‘process-based,’ our approach can be regarded as ‘outcome- based’ (Cutter et al, 2008; Akter and Mallick 2013). We acknowledge that resilience is dynamic and characterised by multiple adaptive cycles interacting across a range of scales and dimensions (Brown and Westaway 2011). However, adaptation cannot suppress all climate-related impacts (Parry et al. 2009; Nicholls et al. 2011; Warner and van der Geest 2013; Warner, van der Geest, and Kreft 2013; Mathew and Akter 2015), particularly in small island states that are disproportionately vulnerable to climate-related natural disasters (e.g., Nurse and Moore 2005; Douglas 2006; Alliance of Small Island States 2008; Monnereau and Abraham 2013). As such, we focus on loss and damage from climate-related disasters. Furthermore, we take a targeted view of resilience, emphasising the indicators considered most important in the context of this study. Aalbersberg (2011) suggests that social and human capital are the most important determinants of adaptive capacity (used synonymously with resilience in this context) in the Pacific. Social and human capital are also relevant aspects of resilience, when advanced warnings are provided, because they enable communities to work together to mitigate loss and damage. Further to this, Adger (2000) suggests that resilience is best understood at the community level, while Berkes and Jolly (2002) emphasise that community resilience measures adaptive capability rather than individual resilience, which more accurately reflects coping ability. Therefore, we focus specifically on community (i.e., social) resilience. While a number of studies have attempted to measure resilience empirically (Kaly, Pratt, and Mitchell 2004; Akter and Mallick 2013; Cutter, Ash, and Emrich 2014; Angeon and Bates 2015), little is known about the benefits that resilience affords. In a policy environment driven by cost—benefit analysis, the lack of information on such benefits hampers our ability to operationalise assessments of resilience. This investigation thus seeks first to measure community resilience, then to assess whether this aspect of resilience affects loss and damage stemming from climate-related disasters, and if so, by how much. Specifically, our approach is to use survey microdata to construct indicators of community resilience in 36 Fijian communities that were exposed to multiple severe natural disasters in 2012. We then compare average loss and damage incurred in more and less resilient communities to calculate the role of resilience in mitigation. Because some forms of loss and damage are essentially unavoidable in hydro- meteorological natural disasters (Parry et al. 2009; Nicholls et al. 2011; Warner and van der Geest 2013; Warner, van der Geest, and Kreft 2013; Mathew and Akter 2015), we evaluate several categories of loss and damage, including crop loss, direct damage (e.g., damage to housing and assets), and indirect damage (e.g., lost wages and purchased cleaning supplies). Our central research question is ‘does community resilience mitigate loss and damage incurred during natural disasters?’ We hypothesise that community resilience is most important when communities have time to take concerted action, and we test this hypothesis by comparing the effectiveness of community resilience in reducing the loss and damage associated with flooding (which came with little warning) with that in reducing the loss and damage associated with a tropical cyclone (about which residents were notified several days in advance). If resilient communities are able to limit loss and damage from natural disasters, then undertaking steps to foster community resilience is a viable strategy for mitigating residual loss and damage from climate-related natural disasters (Craeynest 2010). However, if a resilient community’s response to pending Journal of Environmental Planning and Management 2105 natural disasters is time dependent, then community resilienct i oe u 4 € 1S not a panacea. As the inquiry’s findings are broadly applicable. aes To answer these questions, we first introduce the i i i msm 8, physical, social, and economic characteristics of the study sites on Viti Levu, Fiji. The impacts of a large flood in January 2012 and Cyclone Evan in December 2012 are then described. Details of the household surveys are provided in the following section, along with the calculation of event loss and damage and the construction of the community resilience index. The empirical model follows. Results of the study are th i empirical model y are then presented, followed by discussion 2. Study sites The Fiji Islands are remote and prone to climate-related disasters. According to Neumayer, Pliimper, and Barthel (2014), countries that experience frequent natural disasters are likely to invest more in disaster reduction and damage mitigation, implyin; that such communities may be more resilient than similar communities elsewhere Maru etal. (2014) propose a counter-narrative suggesting that disaster-prone communities are chronically disadvantaged and may, in fact, be less resilient than other communities. Barnett and Campbell (2010) and McMillen er al. (2014, 1) suggest that the findings of Neumayer, Pliimper, and Barthel (2014) may characterise the situation in Fiji, noting the Pacific’s “long history of resilience to environmental variability and unpredictability”; however, given the complex and dynamic interplay between communities and risk, Maru et a (201 a caution against interpreting this narrative in isolation. (Our study sites are the Ba and Penang river catchments locate iji’: i Viti Levu (Figure 2). The Ba River runs north from its nan ey Fiji lslands Catchments Viti Levu I Fy catchments Figure 2. Locations of the Ba and Penang river catchments in the Fiji Islands,2106 D. Gawith et al. mountainous parts of Viti Levu, spilling into the Pacific near the village of Nailaga. ‘Ba’ is also the name given to the province, a tikina (an administrative area comprising several towns and/or villages), and a prominent town. The population of the Ba district is predominantly rural and generally poor, with 34% of residents living below the poverty line (Narsey 2008). Some 45,879 people live within the boundaries of the Ba River catchment, most in Ba Town and downstream, where flooding is a particular risk. Flooding was recorded on the Ba River in 1871, 1892, 1918, 1931, 1938, 1939, 1956, 1964, 1965, 1972, 1986, 1993, 1997, 1999, 2009, and 2012 (McGree, Yeo, and Devi 2010). Tropical cyclones have also caused substantial loss and damage to crops, property, and life in the province. Bordering Ba Province on the east, Ra Province is comparatively small, with just 29,464 residents at the time of the 2007 census. Approximately 15% of the population live in Rakiraki Town, its only urban settlement, with the remaining 85% living in scattered rural settlements and villages. Overall, 53% of the population of Ra Province live below the poverty line (Narsey 2008), suggesting that this population may be especially vulnerable to disasters (Maru et al. 2014). The Penang River is the district’s main waterway and flows approximately 1 kilometre outside Rakiraki Town. While the Penang River is considerably smaller than the Ba River, flooding and forced evacuations in recent years have prompted the Rakiraki provincial administrator to call for proposals to divert the river and/or to relocate Rakiraki Town (Fiji Ministry of Information 2012). One-third of the population in the Ba River catchment and nearly 70% of the population in the Penang River catchment are iZaukei (ie., indigenous Fijians) (Fiji Bureau of Statistics 2012). iTaukei have complex social structures that provide membership in multiple groups. For example, they are born members of tokatoka (family clans). Each totatoka is part of a matagali (clan); each matagali is part of a yavusa (tribe); and each yavusa is part of a vanua (a community of people associated with a specific land area). Belshaw (1964, 123) notes that iTaukei society is “collectivist in organisation and spirit” and that this characteristic has remained unchanged despite significant political change and development. iTaukei hold 87% of the land in Fiji via inalienable customary title, either farming the land themselves (often with important cash crops such as yagona) or leasing the land out to be farmed by others. Wage employment is lower than among other ethnic groups in Fiji (Kumar and Prasad 2004). The remaining population in the Ba River and Penang River catchments is almost exclusively of Indo-Fijian ethnicity (Fiji Bureau of Statistics 2012), largely descended from indentured labourers brought to Fiji to work on colonial sugar plantations during the nineteenth century. Fiji’s sugar production continues to be dominated by Indo-Fijians, who often live in scattered settlements close to cane fields that are generally leased from iTaukei owners (Kumar and Prasad 2004). Compared with iTaukei, Indo-Fijian society is individualistic and self-reliant (De Vries 2002; Rao 2005). In contrast to collectivist iTaukei communities, Indo-Fijian households self-insure against risk via wealth accumulation. Weather-related disasters caused considerable loss and damage in Fiji in 2012. Between 21 January and 12 February 2012 the districts of Nadi, Ba, and Ra recorded 1- in-50-year flooding that killed 11 people, temporarily displaced 1,300 people, and caused US$20.7 million and US$6.9 million in loss and damage for the Ba and Penang river catchments, respectively (Daigneault, Brown, and Gawith 2016). Difficulties in flood forecasting compounded by sparse monitoring and poor topological information (Daigneault, Brown, and Gawith 2016) meant that the average community was alerted only 5 hours before the arrival of the January 2012 flood (Figure 3). Poor early warning ee lls Journal of Environmental Planning and Management 2107 0 10 20 30 40 50 60 70 80 9 100 January 2012 flood warning - hours 0 10 20 30 40 50 60 70 80 90 100 Cyclone Evan warning - hours Fraction sisptayed on Normal denay po shoe Figure 3. Early warning for each disaster by community. systems are reflected in the method of warning: 37% of communities in our study we first alerted to the pending flooding by storm clouds, high humidity, and risin, bai a contrast, 34% of respondents were alerted via television, radio, andlor fataraet sour “ (Brown, Daigneault, and Gawith forthcoming). With such short notice. ti ee flooding were generally limited to evacuation. a acer In December 2012, Tropical Cyclone Evan (a Category 4 storm) brought peak wi oe au ae hour. The cyclone destroyed more than 2,000 homes in Fi, oy : cates etween 11,000 and 14,000 people, and caused an estimated US$96.4 million (0 US$110.6 million in loss and damage nationwide (Simmons and Mele 2013; SPC and SOPAC 2013). Many of the phenomena that lead to tropical cyclones , lL understood, and recent developments in forecastin, js ‘predict th : ig allow meteorologists to predict the movement and speed of cyclones with increasing levels of precision. As a result compared with the January flooding, more than 90% of respondents received warnii via radio, television, and/or internet sources (Brown, Daigneault, and G; it forthcoming), on average, communities were aware of the trajectory of Cyclone ae more than 45 hours before its arrival (Figure 3). With such lengthy warnings, households were able to Move or secure household goods, cut branches and/or trees that ‘mi cht ca damage, reinforce roofing, shutter windows and doors, buy provisions, eraeaais id . most important — help their neighbours to do the same. : Ss 3. Methods The foundation for this study is an extensive socio-economic survey that i y ation r juantific So and indirect impacts of the January 2012 floods and Cyclone Evan Deane fl same year in the Ba and Penang river catchments, Respondents were drawn from villages (officially recognised entities that are exclusively iTaukei) and settlement (oosely organised clusters of houses that are largely Indo-Fijian) based on a probability— a xr eT — —_ —— ee oe 2 Journal of Environmental Planning and Management 2109 sample. In this way, 295 households from 14 rural villages (58% of those registered in the 2 catchment) and 14 rural settlements (representing approximately 32% of the Indo-Fijian & population) were surveyed in the Ba catchment. Similarly, 74 households from three 3/88 88999 ge geggss ges villages (60% of those in the catchment) and five settlements (50% of those in the z)es sss es 22 ssesce eee |, catchment) were surveyed in the Penang catchment. ‘3 e The household surveys were administered on tablet devices and consisted of questions = e on demographics, education, and health; cropping, livestock, fishing, and forestry; labour é income, remittances, durable goods, and housing; and time allocation. The survey also a $ included detailed questions on crop loss, direct damage to housing and assets, and z 5 indirect damage in the form of lost labour, money spent on cleaning supplies, medical Seo 5 . san 2 zi/BS 8888 Be ewevevsvoe we e costs, and money spent on packaged food during evacuation. 2/88 8888 88 SS88SR Bee ]2 In addition, the surveys contained a section on community resilience in which 2 Sesces Ss SSaaao 2 Sis z respondents were asked the extent to which they agreed with statements such as “The z 2 community has the skills and knowledge to limit the damage from natural disasters.” 5 P Answers to these questions were entered by respondents themselves, reducing the risk of g responses based on social desirability. Responses were entered on Likert scales with < ‘strongly disagree’ on one end of the scale, ‘strongly agree’ on the other, and 199 2 22 S g E gradients in between, allowing for greater flexibility and nuance than the usual 5-point 4 5 3 €3 = 3 Zz | Likert scale (Lozano, Garcia-Cueto, and Muiiz 2008). Scores above (below) zero £ : & 28 3F 3 Ae indicate agreement (disagreement), with higher scores indicating stronger agreement and £ 3 S 8 é z 8 & 5 2 lower scores indicating stronger disagreement. Responses to these questions were then 2 2 4 zg S ¢& ad = 2 used to construct a composite index of community resilience, described in Table 1. E eS 3 G2 8 = Be 2 |s Composite indices use mathematical formulae to produce a single metric based on a gs g85 2 22 $2 2 é eG | combination of theory, pragmatism, intuitive appeal, data availability, empirical analysis, =z g 3 2 g8 Ss ape 8 § & simplicity, reliability, validity, and comparability (Booysen 2002; Angeon and Bates s| |3¥ ede ¢ €2 82 234 S86 |g 2015). They are useful for understanding community resilience insofar as they integrate a ele ae gad B22 2 £ fe Beg 3 = number of individual indicators® into a single metric (Booysen 2002; Sullivan, Meigh, E S/22,52225 fg 28 g32 65 |é and Giacomello 2003; Sullivan and Meigh 2005). As such, composite indices are E/ S/S 2eea2 a : 23 aa Ff Ss (2 politically appealing, are commonly used to bridge gaps between social-scientific gls/F2ietess g2 22 282 82 [8 understanding and policy (Jing and Leduc 2010), and have been used to measure = g gaze 4 2 ge ef 5 Es FS = 32 8 | vulnerability, adaptability, and resilience to natural disasters (Mayunga 2007; Cutter, (4) ,8a28 BS Be ERPEEE; ,3 ee & Ash, and Emrich 2014; Angeon and Bates 2015). 2 Begeinse BE fggeee gs 22 4 Cutter, Ash, and Emrich (2014) argue that community resilience to natural disasters is 2 gee E a aos PEw2é : Bec RE £ derived from six forms of capital, although Aalbersberg (2011) has concluded that social ia 232 Seed 2 3 3 B33 4 geeee |= and human capital are the primary determinants of community resilience in the Pacific. 2 B28 2535 Ss £5 2522 232 Bas|é Hence, we focus on these two forms of capital. Specifically, the composite index of = gErrebss ghee igie eae & Sele community resilience used in this study comprises six general ‘domains’: four for social z 3 Ee Z F S22 222225 ga222222]2 capital and two for human capital. Social capital is measured by ‘cohesion,’ = Es gtiggac 5 E Bess ess escsls ‘cooperation,’ ‘social organisation,’ and ‘institutional support.’ Cohesion enhances s ges a= ees 38 se Bheeo Bess community resilience by enabling communities to draw together during times of stress 3 freaks ss $2R25 a3 S § & ae & (Aalbersberg 2011). Similarly, social organisation describes the extent to which - E gee & 8 g5 835 8 2. 28a 58/2 community members already work together on group decision-making (Berke ef al. 2 - BaRe BSE Sasol Fe 2008; Aalbersberg 2011; Maclean, Cuthill, and Ross 2013). Cooperation is important 3 : insofar as communities that are more inclined to work together to achieve objectives are g 5 x 2 generally more resilient (Davidson 2006; Berke ef al. 2008). Institutional support g 3 £ 3 provides a further indicator of community resilience as it describes the presence of inter- a ~ 5 2 2 2 organisational networks and support structures that can assist in developing the ale/2 se 2 § 2 2 functional workings of communities and offer support during recovery from disasters 22/8 ga 3 E q 3g Flald 8s é = i 3 (Norris et al. 2008; Maclean, Cuthill, and Ross 2013).— # ! | 2110 D. Gawith et al. Human capital is represented by the domains of ‘capability’ and ‘dynamism’ in the community resilience index. Capability is assessed through respondents’ views on the ability of their communities to navigate challenges and draw on traditional knowledge (Folke 2006; Mayunga 2007). The ability to learn, change systems, and adapt — referred to as ‘dynamism’ — is an additional aspect of capability (Barnett 2001) and is evaluated through respondents’ attitudes toward risk and views about the community’s attitude toward new ways of solving problems, Table 1 lists the survey questions that are reflected in each domain. Empirical values from the survey questions are combined to form composite index values. To ensure that composite indices reflect the relative importance of individual indicators, weightings may be assigned to individual measures during aggregation. However, there is considerable debate as to whether indicators should be weighted. For example, Booysen (2002) observes that any attempt at weighting indicators can be criticised, and Babbie (1995) and Atkinson (2003) argue that not applying weightings should be the standard approach. This appears to be the case, as Angeon and Bates (2015) point out in their review of composite vulnerability and resilience indices. Saisana, Saltelli, and Tarantola (2005) suggest that the sensitivity of index assumptions should be tested. We therefore constructed two composite indices — one with equal weightings applied to the individual indicators (which we term ‘unweighted’) and one with equal weightings applied to each domain (which we term ‘weighted’). In the first approach, different distributions of responses for each question means that indicators will have different impacts on the overall index score (Mayunga 2007). The second approach produces weights for each indicator, depending on the number of indicators in each domain. These implicit weights are likely to have minor effects compared with the explicit weighting of variables. Therefore, depending on whether the individual indicators or the domains in which they are placed are considered better representatives of community resilience, either approach could be seen to align with the suggestions of Babbie (1995) and Atkinson (2003). Despite the advantages described above, composite indices have come under wide criticism. Uncertainty stems from imperfect or incomplete data collection, subjective weighting, simplification of complex dimensions, and site specificity (Booysen 2002). In a broader sense, the selection of indicators remains arbitrary, as our understanding of the dimensions of resilience is incomplete (Alkire and Foster 2011). Darnhofer (2014, 466) suggests that “measuring the resilience of a system has proved to be like aiming at a moving target.” Indeed, Klein, Nicholls, and Thomalla (2003, 41) suggest that converting resilience from a concept to an operational tool is “a challenge that thirty years of academic debate does not seem to have resolved.” Given these limitations, we employ indices simply to provide an understanding of the relative social resilience of communities in the Ba and Penang river catchments; we make no claim to measure resilience in absolute terms. 4, Empirical model We use the survey data in a multiple regression framework to understand the relationship between community resilience and disaster impacts. Specifically, we regress loss and damage from natural disasters on the composite index of community resilience. Because personal assets are highly correlated with potential loss, losses are measured as a share of household wealth (for robustness, we also regress losses in levels on the composite index of community resilience). Journal of Environmental Planning and Management aul ho ee resilience is measured as the average social resilience score across all — olds : the community, as recommended by Adger (2000). Loss and damage were assessed at the communi it — ity level using data produced by household and community Because community resilience is c it i orrelated with economic resources ‘May 2007), we control for community wealth in our analysis. In addition, ae treatment from neighbours during emergencies), and the gender and age of the respondent (because different household members may have different access to inforation about damage caused by disasters), multiple regression approach also controls fo ‘cit It ; ap x ethnicity, because the collecti oe of iTc ‘aukei communities and the individualistic nature of | Indo-Fijian households 5 and thus their ability to endure hardship caused by natural disasters — are highly oe Cropping choices are also tied closely to ethnicity, with Indo-Fijian Cuseholds often specialising in sugarcane, and iTaukei households sometimes makin, multi-year investments in production of yagona, Models are estimated using tobit estimators to account for lower bounds equal to zero (both for loss and d eee jamage as a share of wealth and for total loss and damage), Y=x} B+, u~N (0,0?) ae J; is a latent variable equal to the observed damage measured at the household a daa when the latent variable exceeds 0. x is a vector of explanatory variables lescribed above, including community resilience, community wealth, ethnicity, household education, length of time in the i : community, gend u, is normally distributed. Thus, we have: Ms tol ages ie come te, _ fy if y>0 w= {g if yt <0 In contrast to the ordinary least s i ¢ uares with cer is consistent (Amen a ql sored data, the tobit estimator is ‘ ae on the hypothesis that community resilience affects damage that may ¢ influenced by human intervention (i.e., direct and indirect damage) more than loss for crop loss, direct damage to housing stock and ts indi isolate the potential ee Ig and assets, and indirect damage, to further hypothesis further, we compare the magnitudes of the different types of damage caused y Cyclone Evan (for which there was nearly two days’ waming, on average) with those caused by the January flooding (for which little waming was available), 5. Results —— 2 shows descriptive statistics for potential correlates of reported damage incurred y natural disasters, including community wealth, household education, length of time2112 D. Gawith et al. that household members have resided in the community, and the gender and age of the survey respondent. The mean household owns FJ$27,532 (US$15,625 in December 2012 exchange rates) in housing, durables, and liquid assets. Its highest educated members have 11.4 years of schooling, and its longest residing members have spent 46.1 years in the community, Nearly two-thirds of respondents are male, and the average age is 45.2. Indo-Fijian households, on average, hold over 50% more wealth than i7aukei households, statistically significant at the 1% level. This difference is consistent with the contrasting systems of social support available to each group, i.e., Indo-Fijians are not bom into the complex, overlapping social support groups to which iTaukei automatically belong and are thus individually responsible for insuring against risk. Indo-Fijian households are slightly more educated, with the mean highest household education 0.53 years higher than for iTawkei households, a difference that is statistically significant at the 5% level. The mean length of time that iZawkei households have spent in their communities is 3.87 years longer than Indo-Fijian households, a difference that is statistically significant at the 5% level. The sex and age of respondents are statistically indistinguishable between ethnicities in the survey. Summary statistics for flood and cyclone loss and damage are shown in Table 3. Loss and damage were greater during Cyclone Evan than during the January floods for both iTaukei and Indo-Fijian households. For both ethnicities during both events, crop losses made up more than 90% of total loss and damage, followed by direct damage; indirect damage was small. Crop losses for Indo-Fijian households are statistically indistinguishable from those for iTaukei households for the January flooding. However, the mean crop losses associated with Cyclone Evan for iTaukei are nearly FJ$2,000 (US$1,135) lower for Indo-Fijian households, which may be attributable to the timing of the cane harvest. Direct damage was more than FJ$100 (USS57) higher in Indo-Fijian households than in iTaukei households during the January floods and more than FJ$450 (US$255) higher in Indo-Fijian households during Cyclone Evan, differences that are statistically significant at the 10% and 1% levels, respectively. Indirect damage was higher among Indo-Fijian households than iTaukei households by more than FJ$40 (US$23) during the January floods and more than FJ$70 (US$40) during Cyclone Evan, likely due to higher participation in the wage economy. These differences are both statistically significant at the 1% level. Components of the community resilience index are described in Table 4 and Figures 4and_ 5. Specifically, Table 4 shows the mean and standard deviation for each of the six domains by ethnicity; it also indicates both the unweighted and weighted composite index scores. The full distribution for each measure of community resilience is shown via histograms in Figure 4. These distributions are summarised in Figure 5, which captures differences in mean scores across the six domains of resilience by ethnicity using an amoeba diagram. Overall, survey respondents report high levels of cohesion and dynamism, with mean scores of 54.4 and 49.6, respectively, on the —100 to +100 scale, Capability shows the lowest overall mean at 6.4.5 The mean unweighted community resilience index score is 31.5 while the mean weighted community resilience index score is 33.4. iTaukei communities score higher than Indo-Fijian communities in terms of cohesion, social organisation, dynamism, and institutional support. For example, the mean score for cohesion among iTaukei communities is 73.49, some 46 points higher than the mean score in Indo-Fijian communities, a difference that is statistically significant at the 1% level. The mean score for social organisation is similarly high in ‘Taukei communities vis-a-vis Indo-Fijian communities, also significant at the 1% level. Indeed, iTaukei Table 2._Summary statistics for covariates of loss and damage by ethnicity, based on survey data iTaukei Indo-Fijian All households Mean Std. dev. Mean Std. dev. Difference Std. dev. Mean Variable (15,669.59) 17,022.93 (6,089.98) 42,536.53 (12,649.10) 27,532.60 Average community wealth (FIS) Journal of Environmental Planning and Management (2.62) (19.51) 11.69 43.86 (1.99) (16.40) 11.16 (2.28) 47.73 (17.82) 38 46.13 Highest education in household (years) Length of time in this communi ty (years) Male respondent (dummy) Respondent age (years) (0.49) (13.62) 0.63 44.71 (0.48) (13.72) 0.66 45.54 (0.48) (13.67) 64 Ou 45.20 Note: Differences are tested using two-sided tests.“ p < 0.10," p < 0.05, and p <0.01 and indirect damage by ethnicity, based on survey data. Table 3. Summary statistics for crop loss, direct damage, iTaukei Indo-Fijian All households Difference Std. dev. Mean Std. dev. Mean Std. dev. Mean Variable (2,694.77) (839.62) 1,274.52 112.20 (65.44) (45.06) (5,215.84) 1,431.12 8.29 5.97 (4,353.60) (542.59) 1,366.61 Direct damage (FJ$) 5 Crop loss (FS) January 2012 flood (152.49) “ 46.88 (105.54) 22.82 Indirect damage (FJ$) (4,450.18) 3.99 513.42, ay (11,481.22) (9,295.13) 4,640.71 3,847.05 Crop loss (FI$) Cyclone Evan (1,323.37) (930.27) 56.08 (404.42) 244.47 Direct damage (FJS) Indirect damage (FS) (142.04) - 6.54 (26.35) 78.66 (99.75) 36.25 Note: Differences are tested using two-sided rtests.* p < 0.10,“ p < 0.05, and” p <0.01 2113— 2114 D. Gawith et al. Table 4. Summary statistics for the community resilience index by ethnicity, based on survey data. SSS All households iTaukei Indo- Variable Mean Std.dev. Mean Std.dev. Mean Std.dev. Difference ii =0.08 (33.87) 15.56 (42.98) Capability 036 (3861) -0. Cobssion. 441 (61.14) 73.49 (37.68) 27.16 (76.20) Cooperation 722 5.14) 26.15 (30.24) 28.76 «at : 1) 7 Social organisation 8.72 (46.41) 47.43 (34.28) 2.014 44) . Dynamism 9.57 (39.26) 58.45 (34.24) 36.89 (42.47) Institutional support 4.09 (51.02) 44.25 (43.63) (19.58 er. : iN Resilience indicator, 1.47 (24.78) 37.79 (19.70) 22.45. (28. unweighted a. Resilience indicator, 33.40 (26.67) 41.62 (20.12) 21.66 (30.32) weighted ests. *p < 0.10, p < 0. Note: Differences are tested using two-sided communities report statistically higher levels of arrose a oe _— - ili ighted or weighted. In contrast, well as overall resilience, whether unweig ted. Nr n for capability in Indo-Fijian communities exceeded that in iTaukei communities by Pe points, also significant at the 1% level. Scores for cooperation are statistically indistinguishable across ethnicities, Overall iTaukei Indo-Fijian 66S SSeS S8a8 ea pel to mitigate loss and damage arene that experience loss and E ier eee may eae anes e events, It i f ie g8/8o8 ¢ —. a2 4 se Se Sears acetbate deprivation in affected quinn Itis also possible = £2292 2s $8 3&2 Se08 85 $ 2014). Understandi re disasters (Warmer and tunities, reducing their 3 o[ae|(a2 ge & ae B5s§ g 225 3 loss and dam: anding the nature of the relationship betw der Geest 2013; Maru et al. 8 es sé B¢ : 3 |e e is ani et “ty pooh f a) |3 ey 22 68 £27 Gi 2 gale The importance of community resinee se ati aren | = e s8]e es mmunity resilience i : g 2) g8l298 Be og ez & oe s noted in the literature ience in making use S| |#/as/sSs S5 2 8: g8 ee ie a ed ae Sem ecbnetaradcd ri ee ere eae z| |& $22 55 28 25 S885 S8selt Gn ue Gea g 2 2 . FSF SF SH 1] damage through early w. importance of community resilien . Our findings provide % Bale eg 8 ¢ 9 af Saee F Be resilience cond ae aings. Furthermore, our findings sae reducing loss and g és s3 & g Ss Sea2 25 ge |v used over this timescal icteristic that is static over the event ti eee ne conten oe g i s és és s6 86 (| 85 systems that are able a As such, resilience is more than imescale to one that can be a) ee i intial es a 3 7 7 by ‘pstoms tat are able to maintain their core functions under siiieeo nacre e Ga, ff Gy £8 39 Referencing Adger (2006) to act pre-emptivey to reduce the imps Re eee - & 5 a2 33 ie . © aa Sumpter (2009, 365) tee et al. (2003), and cuem oan ces 6 ig] ae i a 5 alts and coupl at it “the most conspii : , Colten and a 3 § gio 222 BE —£ 3s gigs ipled human-envir »st conspicuous diffe 3s g BEG 2s ae «= €)/2e oe ‘onment system: ic rence between nonht a (2 £E g§58 fo Seb be @ Bley events and institute indivi ystems is the ability of h nonhuman a ijg G2 pk gee Bb og gS gy By fe fe individual and inst 5 umans t0 learn fi z 8 3 4 gle form of cot nt institutional adj i learn from extreme by 5 8 8 & 2 38 mmunity resilience identified in pear We consider the responsive ly provides a farther distinctionee ll 2118 D. Gawith et al. between the way non-human and human systems manage hazards. In addition to an ability to learn from the past, human systems can make use of community resilience through information about the future. If resilient communities have the ability to mitigate loss and damage when they receive early warning of an impending disaster, then during Hurricane Katrina New Orleans neighbourhoods with high community resilience would have suffered significantly lower impacts than neighbourhoods with low community resilience. Importantly, we would not expect community resilience to be a significant mitigator of loss and damage if New Orleans were instead struck by a disaster with little warning (e.g. an unanticipated earthquake), because more resilient communities would lack sufficient time to operationalise that resilience. 7. Conclusion The frequency and severity of natural disasters will increase under most climate-change projections (Preston ef al. 2006; Bates et al. 2008), and resilience is increasingly viewed as a critical component of mitigating loss and damage (e.g., Craeynest 2010). Considerable work has been undertaken to define resilience as it relates to human communities, and a number of attempts have been made to quantify resilience by identifying its social and economic traits and determinants. However, converting resilience from a concept into an operational tool has proved difficult. We construct an index of community resilience based on extensive survey data from 36 communities in Fiji. We then move beyond the quantification of community resilience to test whether it affects loss and damage incurred during disasters. Specifically, we regress loss and damage on community resilience for disastrous flooding and a cyclone that struck Fiji in 2012. To address possible endogeneity in assessing the impact of community resilience on damage, we exploit the fact that, on average, communities received warning only 5 hours in advance of the flooding, while they received warnings over 45 hours in advance of the cyclone. We separate loss and damage in order to isolate losses that were effectively unavoidable (i.¢., crop loss) from direct and indirect damage that might be ameliorated by communities working together, given sufficient warning. We find that community resilience had no effect on crop loss during the flooding or cyclone because this loss was effectively unavoidable. Similarly, community resilience had no effect on direct and indirect damage incurred during the January flooding, which we interpret to be a consequence of little advanced warning. Community resilience did, however, influence direct and indirect damage incurred during the cyclone. Specifically, each point of community resilience (measured on a 201-point scale) is associated with 0.28% lower direct damage as a share of wealth and 0.03% lower indirect damage as a share of wealth. These findings are robust to different measures of loss and damage and to different weightings being applied to the composite index of community resilience. Our results demonstrate that advanced warning and community resilience are both necessary conditions for reducing loss and damage, although neither is sufficient. This finding highlights the importance of robust early warning systems and demonstrates the influence of community resilience on loss and damage. Furthermore, our findings demonstrate that community resilience in Fijian communities is dynamic and responsive: indeed, community resilience is only an effective mitigator of climate-related disaster loss and damage when communities have sufficient time and ability to respond to pending threats. Journal of Environmental Planning and Management 2119 Acknowledgements This Raat ani was funded by the Climate and Development Knowledge Network under grant number Disclosure statement No potential conflict of interest was reported by the authors. Funding Climate and Development Knowledge Network [grant number RSGL-0024F] Notes 1. This definition refiects its Latin origin ~ resiliere, meaning ‘to jump back.’ 2 Without a wide consensus on the definition of ‘loss and damage’ (Al Faruque and Khan 2013; Calliari 2014; Mathew and Akter 2015), the term is often used to collectively indicate all impacts of disasters eg. United Nations Framework Convention on Climate Change 1995). In this manuscript, ‘loss’ refers primarily to crop loss; ‘damage’ refers to the replacement value of totally or partially destroyed physical assets and to money spent as a result of the disaster, pe definitions are consistent with the Damage and Loss Assessment definitions developed vy the US Economic Commission for Latin America and the Caribbean in 1972, subsequently ened the asain UNESCO, WHO, and others (see World Bank 2015). The survey cluded more than 2,. lata points for each household, many focused on loss and - seer to generate a reliable accounting, : — Individual indicators of community resilience are often hi, i nun ighly interdependent (Aalbersber: 2011), and thus of little use in inferential analysis. As such, Mayunga (2007) argues that resilience measures must move beyond focusing on particular indicators or dimensions, 4. Specifically: social capital; economic capital; housing and infrastructure capital; institutional capital; community capital; and environmental capital. 5. Capability was assessed using respondents” perceptions on their community’s ability to control ao = #: Healt loss and damage; the low scores may reflect the occurrence of 'iple weather-related disasters in the recent past coupled wit multiple past coupled with high levels of poverty in each 6, Sample sizes vary by specification. For both the Jan i a i ication. Fo juary flooding and Cyclone Evan, the sample a2 beitara ner ra in which at least one household was adversely affected by the aster, 1.€., to vulnerable communities. For crop damage, the samp ue »p damage imple was further restricted to References Aalbersberg, W. 2011. Pacific Adaptive Capacity Analysis Frame ; the Cope i of a ura zi Comm mites in th apache Islands steph mae ee Adger, W.N. “Soci: i ili i - i Ge oe ee eae Age ee a names een ee an preenen enn Legal and dintttonal Context in Bangladesh, Dhaka Bangladesh: Center fe Alkire, S., and J. Foster. 2011. “Counting and Multidi i" Public Bonomi 35 Cra) Satig and Multidimensional Poverty Measurement.” Journal of aTT 2120 D. Gawith et al. -LCA: Multi-Window Mechanism to i | Island States. 2008. Proposal to the AWG-LCA aaaeiross Lowy and Damage from limite Change Impacs,hip:fenfeosinaleakyots_prote ication/pdflaosisinsurance061208.pd? : ae PT Iga Regression Avalysis When the Dependent Variable is Trunceted Normal. A 6): 997-1016. i : : Re ce es 2015, “Reviewing Composite Vulnerability and Resilience Indexes: A tainable Approach : . 0): 140-162. Je Approach and Application.” World Development 72.(0): 1 Atkinson AB. 2005. "Mulsditensional Deprivation: Contsting Soval Welfre and Counting Se . . ” 51-65 hes." The Journal of Economic Inequality 1 (1): $1~65. Babbin E1995, The Practice of Social Research. Wadswort, Washington: Cengage Publishing Barnett, J. 2001. “Adapting to Climate Change in Pacific Island Countries: The Problem o} certainty.” World Development 29 (6): 977-993. baci fat Carapbell 3010. Climate Change and Small Island States: Power, Knowledge and "South Pacific. London: Earthscan. : a ee ooe, Colaba! Early Woming Systans for Natural Hazards: Systematic and People: Centred.” Philosophical Transactions of the Royal Society A: Mathematical, Physical an i iences 364 (1845): 2167-2182. : ates Lontonica, S Won and P, Pasko, eds. 2008. Climate Change and Water. Technical Paper of the Intergovernmental Panel on Climate Change. Geneva: polka GS 1964 Under the Ivy Tree: Society and Economic Growth in Rural Fiji. Berkeley: University of California Press. oo RR. Chuenpagdee, K. Juntarashote, and S. Chang. 2008. n Sead of Disaster Resiliency in Thailand: Social Capital and Aid Delivery.” Journal of i Planning and Management 51 (2): 303-317. oo Besken Fe and D. Jolly. 2002, “Adapting to Climate Change: Social-Ecological Resilience in a Canadian Western Arctic Community.” Conservation Ecology 5 Q): 18. Heel Booysen, F. 2002. “An Overview and Evaluation of Composite Indices of Development.” Soci icators Research 59 (2): 115—151. : a Boyd 2H Osbanrs Pi Egctaen, EL. Tompkins, M.C, Lemos and F, Mile. 2008,“Reiene Yand ‘Climatizing” Development: Examples and Policy Implications.” Development 51 (3): Serta « f Vulnerability and Adaptive |.W. Adger, and M.P. Kelly. 2005. “The Determinants of fulnes See ie National Level and the Implications for Adaptation.” Global Environmental 2): 151-163. Brown FA. Deharcault, end D. Gawith, Forthcoming, “Eeonomie Impacts of Climate Change on Soding in Fiji.” Climate and Development. : Bi eect Bway, SOIL “Agency. Capacity, and Resilience to Environmental Change: wVessone from Human Development, Well-Being, and Disasters.” Annual Review of ironment and Resources 36 (1): 321-342. - calle e014. “Loes and Damage: A Crieal Discouse Analysis: FBEM Working Paper No 084.2014; CMCC Research Paper No. RP0231. Milan, Italy: Fondazione Eni Enrico Mattei sin Principles of Ecosystem Stewardship: in Il, F.S,, G.P. Kofinas, and C. Folke, eds. 2009. Principles of Oaslancs Based Netaal: Resource Management ina Changing World. Dore : jence and Business Media, : Colne a aed A aris 3600, “Social Memory end Resilience in New Orleans.” Natural izard 2 355-364. ‘ Cracyaeat L610. “Lass and Damage ffom Climate Change: The Cost for Poor People in Developing Counties.” Actiondid International Discussion Paper. Johannesburg hp www. actionaid.org/sites/files/actionaid/loss_and_damage_-_discussion_paper_by_actionaid- Eas Ash, and C.T. Emrich. 2014, “The Geographies of Community Disaster ésilience.” Global Environmental Change 29 (0): 65-77. . Cotter SLL, Bames, M.Beny, C Burton, E Evans, E, Tale, and J. Webb, 2008." Pace-Based Model for Understanding Community Resilience to Natural Disasters.” Global Environmen Change 18 (4): 598-606. Journal of Environmental Planning and Management 2121 Daigneault, A., P. Brown, and D. Gawith. 2016, “Dredging Versus Hedging: Comparing Hard Infrastructure to Ecosystem-Based Adaptation to Flooding.” Ecological Economics 122: 25~35. http://dx.doi.org/10.1016/j.ecolecon, 2015.1 1.023 Damhofer, I. 2014. “Resilience and Why It Matters for Farm Management.” European Review of Agricultural Economics 41 (3): 461—484, Davidson, M. A. 2006. Designing for Disasters. Massachusetts: Massachusetts Coastal Hazards Commission, NOAA Coastal Services Center. De Vries, R.E. 2002. “Ethnic Tension in Paradise: Explaining Ethnic Supremacy Aspirations in Fiji” International Journal of Intercultural Relations 26 (3): 311-327. Douglas, C.H. 2006. “Small Island States and Territories: Sustainable Development Issues and Strategies: Challenges for Changing Islands in a Changing World.” Sustainable Development 15 (2): 75-80. Fiji Bureau of Statistics. 2012. “Population by Age, Race, and Province of Enumeration, Fiji: 2007 Census.” http:/www.statsfiji,gov. fi/index. php/2007-census-of-population Fiji Ministry of Information. 2012. “PM Meets Flood Hit.” Fiji Sun, January 29. Folke, C. 2006. “Resilience: The Emergence of a Perspective for Social—Ecological Systems Analyses.” Global Environmental Change 16 (3) 253-267. Gallopin, G.C. 2006, “Linkages Between Vulnerability, Resilience, and Adaptive Capacity.” Global Environmental Change 16 (3): 293-303. Godschalk, D. R. 2003. “Urban Hazard Mitigation: Creating Resilient Cities.” Natural Hazards Review 4: 136-143 Gordon, J. 1978. Structures. Harmondsworth: Penguin Books. Guha-Sapir, D., F. Vos, R. Below, and S. Ponserre. 2011. Annual Disaster Statistical Review: The Numbers and Trends. Brussels: CRED. Holling, C:S. 1973. “Resilience and Stability of Ecological Systems.” Annual Review of Ecology, Evolution, and Systematics 4: 1-21, IPCC. 2012. Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation. A Special Report of Working Groups I and Il of the Intergovernmental Panel on Climate Change, edited by C.B. Field, V. Barros, T.F. Stocker, D. Qin, D.J. Dokken, K-L. Ebi, MD. Mastrandrea, K.J. Mach, K.J. Platter, S.K. Allen, M. Tignor, and P.M. Midgley, Cambridge: IPCC. Jing, F. and B. Leduc. 2010. “Potential Threats from Climate Change to Human Wellbeing in the Eastem Himalayan Region: Climate Change Impact and Vulnerability in the Eastern Himalayas.” Technical Report 6. Kathmandu: ICIMOD. Kaly, U. C. Pratt, and J. Mitchell. 2004. “The Environmental Vulnerability Index (EVI) 2004.” Technical Report 384. Suva: SOPAC. Klein, R., R. Nicholls, and F. Thomalla. 2003, “Resilience to Natural Hazards: How Useful Is This Concept?” Environmental Hazards 5: 35—45, Kumar, &., and B. Prasad, 2004. “Politics of Race and Poverty in Fiji: A Case of Indo-Fijian Community.” international Journal of Social Economics 31 (5/6): 469-486, Leichenko, R. 2011. “Climate Change and Urban Resilience.” Current Opinion in Environmental Sustainability 3 (3): 164~168, Tozano, LM. E, Garcia-Cueto, and J, Muniz. 2008. “Effect of the Number of Response Categories on the Reliability and Validity of Rating Scales.” Methodology 4 (2): 73-79. Maclean, K.,M. Cuthill, and H. Ross. 2013. “Six Attributes of Social Resilience.” Journal of Environmental Planning and Management 57 (1): 144—156. Maru, Y.T., M. Stafford Smith, A. Sparrow, P.F. Pinho, and O.P. Dube. 2014. “A Linked {ulnerability and Resilience Framework for Adaptation Pathways in Remote Disadvantaged Communities.” Global Environmental Change 28: 337-350, Mathew, LM., and S. Akter. 2015. “Loss and Damage Associated with Climate Change Impacts.” In Handbook of Climate Change Mitigation and Adaptation, 1st ed., edited by W.-Y. Chen, J. Seiner, T. Suzuki, and M. Lackner, 1-23, New York: Springer. Mayunga, J. 2007. “Understanding and Applying the Concept of Community Disaster Resilience: A Capital-Based Approach.” Draft working paper prepared for the summer academy for social vulnerability and resilience building. College Station: Texas A&M University. s MoGree, S., $.W. Yeo, and S. Devi. 2010. Flooding in the Fiji Islands Between 1840 and 2009, Macquarie: RiskFrontiers.2122 D. Gawith et al. McMillen, H.L., T. Ticktin, A, Friedlander, $.D. Jupiter, R. Thaman, J. Campbell, J. Joeli Veitayaki, ef al. 2014. “Small Islands, Valuable Insights: Systems of Customary Resource Use and Resilience to Climate Change in the Pacific.” Ecology and Society 19 (4): 44. Monnereau, I. and S. Abraham. 2013, “Limits to Autonomous Adaptation in Response to Coastal Erosion in Kosrae, Micronesia.” International Journal of Global Warming 5 (4): 416—432. Munang, R., I. Thiaw, K. Alverson, J. Liu, and Z. Han, 2013. “The Role of Ecosystem Services in Climate Change Adaptation and Disaster Risk Reduction.” Current Opinion in Environmental Sustainability 5: 1-6. Narsey, W. 2008. The Quantitative Analysis of Poverty in Fiji, Suva: Vanuavou. Nelson, DR. 2011, “Adaptation and Resilience: Responding to a Changing Climate.” Wiley Interdisciplinary Reviews: Climate Change 2 (1): 113-120. Neumayer, E., T. Pliimper, and F. Barthel. 2014. “The Political Economy of Natural Disaster Damage.” Global Environmental Change 24: 819. Nicholls, R.J.,N. Marinova, J.A. Lowe, S. Brown, P. Vellinga, D. De Gusmao, J. Hinkel, and R.S.J. Tol, 2011, “Sea-Level Rise and Its Possible Impacts Given a ‘Beyond 4 C World’ in the Twenty-First Century.” Philosophical Transactions of the Royal Society of London A: Mathematical, Physical and Engineering Sciences 369 (1934): 161~181. Norris, F., S. Stevens, B. Pfefferbaum, K. Wyche, and R. Pfefferbaum. 2008. “Community Resilience as a Metaphor, Theory, Set of Capacities, and Strategy for Disaster Readiness.” ‘American Journal of Community Psychology 41 (1—2): 127~150. Nurse. L, and R. Moore. 2005. “Adaptation to Global Climate Change: An Urgent Requirement for Swaall Island Developing States.” Review of European Community and International Environmental Law 14 (2): 100~107. Parry, M.L., N. Amell, P. Berry, D. Dodman, 8. Fankhauser, C. Hope, 8. Kovats, et al, 2009. “issessing the Costs of Adaptation to Climate Change: A Review of the UNFCCC and Other Recent Estimates. London: IIED. Pearson, LJ., and C.J. Pearson, 2012. “Societal Collapse or Transformation, and Resilience.” ‘Proceedings of the National Academy of Sciences 109 (30): E2030—E2031. Pelling, M,. and D. Manuel-Navarrete. 2011. “From Resilience to Transformation: The Adaptive Cycle in Two Mexican Urban Centers.” Ecology and Society 16 (2): 11. Preston, BL., R. Suppiah, I. Macadam, and J. Bathols. 2006. Climate Change in the Asia/Pacific Region: A Consultancy Report Prepared for the Climate Change and Development Roundtable. ‘Aspendale: CSIRO. Rao, D. 2005. “Culture and Entrepreneurship: An Exploratory Essay.” Fijian Studies: A Journal of ‘Contemporary Fiji 3 (1): 57-86. Report of the Conference of the Parties on its First Session, Berlin from March 28 to April 7, 1995. http://unfece.int/resource/docs/Q37cop 1/07a01 pat Saisana, M., A. Saltelli, and S. Tarantola, 2005, “Uncertainty and Sensitivity Analysis Techniques as Tools for the Quality Assessment of Composite Indicators.” Journal of the Royal Statistical Society. Series A (Statistics in Society) 168 (2): 307323. simmons, M., and K. Mele. 2013.“Picking up the Pieces.” Suva: Mai Life, 18-21. SPC and SOPAC. 2013, Post-Disaster Needs Assessment: Tropical Cyclone Evan, 17th December 3013, Suva: Ministry of Strategic Planning, National Development and Statistics. Sullivan, C., and J. Meigh. 2005, “Targeting Attention on Local Vulnerabilities Using an Integrated Index Approach: The Example of the Climate Vulnerability Index.” Water Science and Technology 51 (5): 69-78. Sullivan, CA. JR. Meigh, and A.M. Giacomello. 2003. “The Water Poverty Index: Development and Application at the Community Scale.” Nanural Resources Forum 27 (3): 189-199. ‘Tompkins, E.L. 2005. “Planning for Climate Change in Small Islands: Insights from National Hurricane Preparedness in the Cayman Islands.” Global Environmental Change 15 (2): 139-149. Tompkins, E.L., and W. Adger. 2004. “Does Adaptive Management of Natural Resources Enhance Resilience to Climate Change?” Ecology and Society 9 (2): 10. Tompkins, ELL., W.N. Adger, E. Boyd, S. Nicholson-Cole, K. Weathethead, and N. Amell. 2010 “Observed ‘Adaptation to Climate Change: UK Evidence of Transition to a Well-Adapting Society.” Global Environmental Change 20 (4): 627-635, ‘Tumer, B.L.. RE. Kasperson, P.A, Matson, J.J. McCarthy, R.W. Corell, L. Christensen, N. Eckley et al, 2003. “A Framework for Vulnerability Analysis in Sustainability Science.” Proceedings ofthe National Academy Of Sciences 100 (14): 8074-8079. Journal of Environmental Planning and Management 2123 United Nations Framework Convention on Cli X imate Change, 1995, Report of the Conf Parties on its First Sess i i niece invesoure! ' SSeepitiorn” Berlin from March 28 to April 7, 1995. http:/unfece int/resource/ alker, B., CS. Holling, S., Carpenter, and A. Kinz ii ling, 8. and A. Kinzing, 2004. “R ar ttansfomabiliy in Scial-coloicl Systema.” Boolog and Sociay9 Gh. met Ky and K: van det Gees. 2013. “Loss and Damage ffom Climate Change: Loca-Level Evidence ‘ulnerable Countries.” International Journal of Global Warming 5 (4): Warner, K., K. van der Gest, and §. Kreft, 2013. “ Ke , and S. Kreft, 2013. “Pushed to the Limit: Evid i Change-Related Loss and Damage When P. i Linas Adenaten? 7 open inet age When eople Face Constraints and Limits to Adaptation.” ison, S., L.J. Pearson, Y. Kashima, D. Lusher, and C. Pearson. “Separati ‘ 7 , D. Lusher, and C. . 2013. i Maintenance Resilience) and Transformative Capacity of Social-Esologioal “Syston yi tea Scien 18 (1): 22 et ise, RLM., I. Fazey, M. Stafford Smith, S.E. Park, H.C. Eakin, E.R. | SE. .C. Eakin, E.R.M. Archer Van Gardei 3 Camptell 2014. “Reconcpuaising Adaptation to Climate Change a Part of Pater SE Change and Response.” Global Environmental Change: Human and Policy Dimensions 28 World Bank. 2015, “Damage and Loss Assessments.” http://go.worldbank.org/VSWHGVLN90