Миграция как адаптация к изменениям климата удалённых сообществ коренных народов: какие последствия мы можем ожидать?

Migration is well recognised as one adaptive strategy available to individuals and communities facing disruption to lifestyle and livelihoods as a result of climate change (Bailey 2011). Despite a rapidly growing international literature examining the impacts of climate change on human migration, there is no general model that provides an understanding of who might migrate when and from where to where. There are numerous studies examining migrations that have been linked to weather or climate in the past (most notably McLeman and Smit’s (2006) study of 1930s drought related migration from Oklahoma in the United States), and postulating migrations that might occur as a result of climate changes expected in the future (a recent example is the Asian Development Bank study (2012)). While few universal conclusions can be drawn from the research, it does reveal some ‘dimensions’ that should be explored when trying to model migration as an adaptation to climate change in specific cases. This paper examines these dimensions and considers their relevance to remote Indigenous communities in northern Australia.

The academic literature on migration as an adaptation to climate change among Indigenous communities is dominated by research involving James D. Ford and colleagues working in the north of Canada (e.g. Ford 2009), and Nicholas Tyler and colleagues working with Sami populations in northern Norway (e.g. Tyler et al. 2007). There is a much larger non-peer-reviewed literature that investigates migration (among other adaptations) in these locations and elsewhere (see, for example, Parker et al. (2006) looking at Indigenous populations of the Pacific Rim). Often, researchers examining climate change adaptations among Indigenous people provide only cursory attention to the is- 93

sue of human migration, perhaps because of an expectation that Indigenous people will not leave traditional lands unless absolutely necessary (as may be the case for communities in some island nations). This paper therefore addresses an important shortcoming in the existing literature by highlighting the role that migration might play as an adaptation to climate change by Indigenous people, with a particular focus on remote communities in the north of Australia.

In the remainder of the paper we provide a brief description of the context of our research (remote Indigenous communities in northern Australia) and the potential exposure of these communities to climate change. We then review the existing literature on migration as an adaptation to climate change, and specifically as an adaption for Indigenous populations, highlighting the various dimensions of migration (for example, whether it is temporary or permanent) that form a model within which migration as an adaptation can be investigated. We attempt to illustrate each dimension with evidence from ‘analogues’ (Ford et al. 2010) of remote Indigenous communities in northern Australia drawn from the literature concerning migration into and out of such communities. The paper concludes by discussing whether or not, in light of the multiple dimensions that are identified here, a general model of migration patterns expected to accompany climate change is feasible or even desirable.

Remote Indigenous Communities and Climate Change in Northern Australia

The Australian Government Office of Northern Australia considers its remit to encompass the region north of the Tropic of Capricorn. Under that definition, ‘northern Australia’ covers about one quarter of the Australian landmass, and includes substantial parts of the jurisdictions of Queensland, the Northern Territory and Western Australia. Based on the 2011 Census, it has an approximate Indigenous population of 160 000 people, which is about 13% of the total northern population and about 30% of the total Australian Indigenous population.

Around 300 000 northern Australians live in areas that are formally classified as either ‘remote’ or ‘very remote’ (see . The definitions of remoteness take into account the distance people live from major services and infrastructure and the sparsity of the population (Carson et al. 2011). Remote areas have poor access to labour markets, social services, technology and transport infrastructure (Carson 2011, Carson and Cleary 2010). Moreover, remote areas tend to have ‘split’ populations, with some residents experiencing socio-economic, health and other disadvantages, while others have high incomes and high access to education and employment opportunities (Taylor et al. 2011).

The population ‘split’ in northern Australia is partially explained by the differences in living conditions, demographic, and socio-economic characteristics between Indigenous and non-Indigenous Australians. Indigenous Australians in remote areas tend to be less engaged in mainstream economic, education and health systems than their counterparts in more urban parts of the country

(Taylor et al. 2011b). Many remote dwelling Indigenous people have English as a second, third, or subsequent language, making access to services even more difficult. Traditional cultural practices are still strong in many remote Indigenous communities, and this means people can be highly mobile (moving from place to place for ceremony, hunting, and social and cultural obligations) and have a high dependence on (cultural and food) resources extracted from the local environment (Prout 2008).

There are dramatic differences in very basic demographic characteristics between Indigenous and non-Indigenous Australians in northern Australia. Indigenous populations are much younger (for example, the Northern Territory Indigenous population had a median age of 22 years according to the 2006 Census of Population and Housing, compared with 34 years for the non-Indigenous population), have a higher proportion of females (98 males for every 100 females compared with 110 males for every 100 females in the non-Indigenous population), have lower life expectancies (in the order of 15-20 years) and higher fertility rates (almost double the non-Indigenous population). While Indigenous people are highly mobile, they are less likely to change their primary place of residence (about 15% do so every five years) than are non-Indigenous people in northern Australia (about 75% do so every five years). These demographic characteristics, combined with the fact that many (about two thirds) Indigenous people in northern Australia live in very small settlements (less than 2 000 people), means that local populations are subject to rapid and dramatic change (through births and deaths ‘clusters’ for example) and differ greatly from one another (Koch and Carson 2012). While generalisations about the nature of Indigenous populations are common, they must be treated with caution because of this demographic diversity.

Local Indigenous populations are frequently referred to as ‘Indigenous communities’, and this label has a number of colloquial and more formal meanings. ‘Community’ is colloquially used to refer to Indigenous people with a particular shared cultural heritage (as Taylor 2009 uses the term ‘peoples’) or to refer to the set of Indigenous people living in a particular geographic area (irrespective of the extent of their shared or separate heritage). More formally, however, there are identified ‘Indigenous communities’, which are discrete settlements with a large majority Indigenous population. These communities (we prefer to call them ‘settlements’) are often located on land that is subject to some form of Indigenous tenure (a complex set of arrangements which include freehold, leasehold, Crown reserve, exclusive or shared tenure etc.). Those lands are areas where Indigenous people have specific rights to use and residence. The system of tenure and rights also means that the lands and the people who live on them are subject to different rules and regulations than that which apply to other settlements (and to other Indigenous people). There may be restrictions on who can live in these settlements, who can access housing (and under what arrangements), how social security, royalties and other income is distributed, and, what access there is to alcohol, fast food and even literature.

The most explicit statement of these separate rules has occurred in the Northern Territory, where the Australian Government’s ‘Northern Territory Emergency Response’ in 2007 saw the declaration of over 80 ‘prescribed areas’ which are Indigenous communities subject to specific legislation and programs ostensibly designed to address child health and welfare issues (Taylor and Carson 2009). The other two jurisdictions have less universal regimes of separation although there is regular talk about extending the Northern Territory Emergency Response structures and some pilot programs to that effect. Nevertheless, northern Australian Indigenous communities are subject to a level of policy and legal scrutiny and controls over land and lives that far exceeds that applying to other northern settlements and to Indigenous people elsewhere in Australia.

One of the prominent issues around the control of Indigenous people is the extent to which, and the conditions under which, they move in and out of Indigenous settlements. Taylor and colleagues (2011a) discussed the ‘problem’ of Indigenous migration particularly out of remote settlements. Such migration has often been labelled ‘urban drift’ and is seen as a source of social disturbance in the places to which people migrate (Taylor and Carson 2009). In other cases, people (including researchers) are unwilling to discuss outmigration because of the implications a movement away from ‘country’ has for perceptions of the strength and persistence of culture (Taylor et al. 2011a). In recent times, a large outmigration of women and children from the remote Indigenous community of Yuendemu as a result of violence in the community sparked heated public debate about who (meaning “which government”) was responsible for these people once they had left the prescribed area (see, for example, ABC News 2010).

The discourse around Indigenous migration from (and to) remote settlements is concerned with complex issues such as the:

• •    extent to which government policy encourages or impedes migration;

• •    extent to which social, health, economic, environmental and hedonistic motives drive migration (and specifically what those motives are); and,

• •    relationships between the places where Indigenous people live and the places in which their culture is embedded (Taylor and Carson 2012).

It must be noted that Indigenous communities (as settlements) have emerged from a process of forced/fixed settlement of what were previously largely nomadic people. The extent to which the locations of fixed settlement match or allow access to traditional ‘spaces’ varies from community to community and can have a substantial impact on migration patterns and livelihoods (Prout 2008). The pattern of fixed settlement and persistence of traditional occupation has also led to networks of smaller communities (‘outstations’ and ‘homelands’) with varying strength of connection to the settlements from which their inhabitants originated (particularly during the ‘homelands’ movement in the 1980s). Outstations and homelands vary in size from a few to as many as 100 people, and have often been established as a way of separating rival families or helping people avoid difficult conditions in the ‘main’ com- ⁹⁶

munity (McDermott et al. 1998). Positioning climate motivated migration within this complexity is an enormous challenge for researchers, policy makers and Indigenous people themselves.

McKeon and colleagues (2009) summarised the contemporary expectations of the impacts of climate change in northern Australia. Moderate sea level rises are expected to affect some coastal communities, although it is not clear the extent to which this will occur, and which communities may be affected. The frequency of heat waves is expected to increase, and the intensity of (but not frequency of) cyclonic activity and associated flooding. There is also likely to be a loss of habitat for native animals and pastures for farmed animals as a result of wetland inundation and extended periods of drought affecting native and exotic flora and fauna in the arid areas. The timeframes in which such changes may occur are difficult to assess, although there are some arguments that changes are already being observed locally (Bird et al. 2013, Green et al. 2010, Petheram et al. 2010).

Existing literature on climate change induced migration of Indigenous Australians is limited. The report compiled by Green, Jackson and Morrison (2010) focused on exposure and vulnerability to climate change impacts more so than assessment of the adaptive strategies that may be available. The authors were firm in their conviction that Indigenous people would be able to adapt because of their long history of building environmental knowledge and managing their relationships with the land. Petheram et al. (2010), Howitt et al. (2011) and Bird et al. (2013) more directly considered issues around adaptation in the context of specific Indigenous communities in northern Australia.

Similar to other research, Petheram et al. (2010) noted adaptive strategies around engineering and infrastructure design and the redesign of transport systems. It was suggested that these adaptations in situ were the preference of Indigenous people in Arnhem Land in the Northern Territory where the research was conducted. However, there was also some consideration of migration as an adaptation, which will be covered in detail in the next section of this paper.

Howitt et al.’s (2011) research focussed on the community of Kiwirrkurra in Western Australia where the population was evacuated following flooding associated with Cyclone Abigail in 2001. The researchers argue that evacuation was used as a substitute for provision of better local infrastructure and the engagement of local Indigenous people in adaptation planning. The inherent adaptive capacity of Indigenous people was seen as nullified by the imposition of ‘adaptive policies’ by government agencies. As a result, it was not possible to assess the extent to which migration may have occurred without the government intervention.

Bird et al. (2013) documented multiple reasons for mobility among four northern Australian communities (Broome, Western Australia; Maningrida and Ngukurr, Northern Territory; WujalWujal, Queensland). The reasons for mobility included health andlivelihood and employment opportunities with a focus to be closer to essential services following the centralisation of many government services. Additionally, survey respondents discussed movement from out- 97

stations83 to these four somewhat larger centres during the wet season, to avoid being cut off from flooding. Discussions centred on how this temporary migration might exacerbate in response to climate change impacts (e.g. more frequent or severe extreme weather events such as cyclones, heat waves and floods or slow onset changes such as drought and loss of biodiversity).

In summary, remote Indigenous communities in northern Australia present a very complex context in which to study migration as an adaptation to climate change. Indigenous communities are diverse, subject to very high levels of surveillance and control, embedded in cultural notions of the environment and legal constructs of occupation of the land and distanced from resources that underpin adaptive strategies elsewhere. At the same time, however, they possess unique sets of knowledge and skills that present options for adaptation both by staying put and by moving. To begin to model what specific communities might do as a response to climate change demands an understanding of this context as it relates to the various dimensions of migration – the factors that determine how much migration occurs, from where to where, by whom, for how long, and with what consequences. The paper now moves to an articulation of these dimensions.

The Dimensions of Migration as Adaptation

The two specific examples of Indigenous adaptation to climate change in northern Australia (Petheram et al. 2010, and Howitt et al. 2011) immediately suggest at least four dimensions relevant to assessing the extent to which migration may be an adaptation. Howitt and colleagues are explicitly concerned with the extent to which migration is a voluntary, as opposed to mandated, adaptation. In their case example, migration away from the Kiwirrkurra community was mandated as part of government policy around emergency/disaster management.

Disaster Management Plans for remote Indigenous communities in the Northern Territory (particularly those in tropical areas susceptible to cyclones) also include provisions for mandatory evacuation. Petheramet al. (2010) and Bird et al. (2013), on the other hand, found that communities were considering or implementing temporary migration as a voluntary response. Clearly, it is easier to anticipate who will move under conditions of forced migration, and the use of mandatory evacuations has become more common in Australia generally in recent times following various bushfire (Haynes et al. 2010) and severe flood events in the eastern states (Taylor and Freeman 2010). In the broader community, people’s right to choose whether to ‘stay and defend’ or

‘leave early’ has been heavily debated, but one might expect that contexts in which government intervention in daily life is already high, such as in remote Indigenous communities in the north, the degree of individual choice will be low.

Petheram and colleague’s (2010) investigation of migration as an adaptation suggested that the preferred migration response would be to move to relatively nearby locations and within well established networks of locations (existing homelands and outstations, for example). In contrast, the Asian Development Bank (2012) report (as one example) talks about the need for some displaced people to move large distances to very unfamiliar environments. In general, the global climate change literature has a strong interest in international migration as a response (McLeman and Smit 2006), although many empirical studies suggest that moves within a single jurisdiction are likely to be more common given the additional resources and risks involved in international moves (McLeman and Hunter 2010).

Our understanding of current spatial patterns of migration by Indigenous people from northern Australia includes poorly tested assumptions that very little movement occurs across State and Territory borders. When such movement becomes apparent, concerns are raised about the jurisdictional responsibilities for both ‘servicing’ and policing Indigenous people (Taylor and Carson 2009). Similar concerns are raised around rural to urban migration even within a jurisdiction (Habibis et al. 2011).

It might be expected that moves over short distances following established migration (and other) networks would be easier to anticipate and measure than moves over longer distances to less predictable locations. However, even well established patterns of mobility among remote Indigenous populations in northern Australia can be difficult to model (Prout 2008). Furthermore, recent evidence suggests that the range of migration destinations that Indigenous people in remote northern Australia consider and pursue is increasing, and becoming more diverse, particularly for young people (Taylor 2012).

In addition to the spatial dimension of migration as an adaptation is the temporal dimension. Discussions between Indigenous people and Petheram et al. (2010) and Bird et al. (2013) specifically focused on temporary relocation. Howitt et al.’s (2011) case example also involved temporary relocation, but the mechanisms for facilitating return to the origin community were less well structured than the mechanisms for (forced) movement away. Clearly, whether migration is short- or long-term (or permanent) is in part a function of the nature of the impact of climate change. Communities (and even entire islands) that are inundated by rising sea levels, for example, will not be able to be repopulated. Ford and colleagues (2010) advocate permanent relocation of some specific communities in northern Canada, which face this level of threat.

Communities experiencing disruptions such as floods, fires, or cyclones, on the other hand, may be suitable for habitation very soon after the event, even if such events become more frequent and severe (although a ‘ratchet effect’ whereby continued exposure to even low risk results in dramatic responses 99

(Ford et al. 2006) may ultimately apply). However, the duration of a migration event is not simply a function of the habitability of the origin community, but of the attitudes of the individual migrant and the perceived utility of remaining absent for a particular period of time, moving on to new locations, or staying away permanently (Petrov 2007, for example, demonstrated this for Indigenous people moving from the north to the south of Canada). Utility may be measured in economic, social, cultural, health or other terms, and is likely to be a complex combination of these (Greenwood and Hunt 2003). It would be difficult to justify any assumption that all Indigenous people in all communities in northern Australia have the same specific perceptions of utility, even if there are cultural worldviews of utility.

The fourth dimension drawn from the publications specific to climate change and Indigenous communities in northern Australia relates to the extent to which migration is THE adaptation to climate change or is one of many options. Howitt and colleagues (2011) strongly emphasised the policies of forced migration as an adaptation to severe flooding in their case community, suggesting that government policy enforcing evacuation served both to limit government attention to other possible strategies and, more importantly, to deprive Indigenous people of the opportunity to use their traditional and local environmental knowledge to develop a range of alternatives. In contrast, Petheram et al. (2010), Bird et al. (2013) and Green et al. (2009) were able to document a number of adaptations proposed and pursued by Indigenous people and by external agencies – replacing or relocating physical infrastructure, better preparedness in terms of ensuring food, water and health supplies, and re-arranging work, social, and cultural schedules (see also, Ford et al. 2010).

In locations where these alternative adaptations can be successfully implemented, migration may be a less common response to either discrete weather events or slow-onset environmental changes. The great diversity of both exposure to climate change impacts and capacity to anticipate and respond observed among Indigenous communities around the world (Parker et al. 2006) suggests that general rules about the nature and extent of migration as adaptation are unlikely to emerge (Perch-Nielsen et al. 2008).

McLeman (2010) adds another perspective to this dimension by considering how population volatility in a community (in and out migration, ageing, changing fertility) might impact on migration as the preferred adaptation to climate change. McLeman provides the example of new in-migrants to a northern Canadian community from the urban south. These in-migrants have limited local knowledge of the climate and environment and are therefore poorly prepared to engage in adaptations beyond migration. Ageing populations may likewise lose the capacity to implement alternative strategies. Communities, which struggle to recruit and retain professional labour (in health services, for example), may also be more likely to consider migration as the adaptation.

Applying this to an Australian perspective, Bird et al. (2013) considered how population volatility might affect climate change induced mobility and migration. Bailey (2011) postulates that communities with excessive popula- tion volatility are more likely to experience dramatic disruption as a result of environmental changes. Moreover, these communities are more likely to use migration as a response to climate change, since they are acquainted to the movement of people into and out of the community, in addition to having social and familial networks, which are connected to outside communities and towns (Asian Development Bank, 2012).

McLeman (2010) suggests that communities with volatile populations can change more rapidly and substantively than those with non-volatile populations. Therefore, drawing from the literature, Bird et al. (2013) used demographic characteristics to determine the volatility of the four Indigenous communities across northern Australia. These characteristics included: the historical experiences of population growth and decline, the ‘churn’ of population brought about by temporary and long term migration to and from the community, mobility within the community and between communities, and ‘unbalanced’ age and sex profiles (Gloersen et al., 2009), and housing stress (Roiko et al., 2012).

The results of Bird et al.’s (2013) analysis shows that the extent of population change within a community as a result of a weather (or some other) event is not only likely to be related to its demographic path per se , but to its exact position on that path at the time of the event. Contributing to the extent of change the community may be subject is the vulnerability of familiar places to the same event. For example, extreme weather events such as floods, cyclones and heatwaves are likely to impact nearby and similarly remote communities to which the case study community is linked. As a result, migration may be pushed further afield to other, larger or less remote centres with more substantial hard and soft infrastructure. Similarly, migration into the case study communities may also occur if linked communities suffer greater impacts from the threat. The extent to which this may occur would be determined not only by the strength and direction of existing and past relationships between communities, but the relative infrastructure ‘wealth’ of the case study communities.

Demographic diversity and dynamism may not only affect who chooses migration or where they migrate to, but also how long they are away and whether they return. In northern Australia, for example, case studies of migration responses to Cyclone Tracy (Darwin, 1974) and Cyclone Les (Katherine, 1998) suggest diverse spatial and temporal characteristics not just between Indigenous and non-Indigenous populations, but also between long term and short term, rural and urban, young and old, and male and female residents (Haynes et al. 2011). That research also emphasises (along with Bailey 2011 and Bird et al. 2013) that in-migration to climate change affected communities is equally worthy of consideration as is out-migration. In the case of Katherine, there was an out-migration of existing residents to relatively distant places, but an in-migration of people who had been living in smaller (and presumably more vulnerable) nearby locations.

Much of the discussion here thus far has focused on what might be considered ‘true migration’, which is the movement of people away from their 101

‘usual’ places of activity. In the western literature, this is commonly taken to mean a move from one place of residence to another, whereas Indigenous people in places like Northern Australia may have multiple ‘residences’ and engage in regular movement between them (Taylor 2009).

In addition to climate change linking to migration events (the move to new places), it may be linked to changes in mobility (the established patterns of movement between places). Some of the Canadian (Ford et al. 2007) and European (Tyler et al. 2007) literature focuses on mobility changes – Indigenous people adapting by avoiding locations they would normally go to at particular times of the year, changing their patterns of movement between places where they sleep, work, and gather (Berkes and Jolly 2001). In particular, changes in the climate and environment may impact on patterns of mobility associated with ‘caring for country’, which can lead to personal and collective stress among Indigenous populations (Rigby et al. 2011). The extent to which such dislocation from traditional mobility practices might inspire ‘true migration’ has not been examined in the literature, perhaps because this distinction between mobility and migration has not previously been made explicit.

What has been explicitly discussed is the difficulty of separating climate driven migration from migration inspired by other factors (Adamo and de Sherbinin 2011). In our context, for example, Petheramet al. (2010) and Bird et al. (2013) found that climate change was a very minor issue in Indigenous people’s consideration of the future of their communities when compared to concerns about health behaviours, crime, poverty, government intervention and so on. In fact, some studies (including Petheram et al.) have found that they first needed to ‘educate’ Indigenous people about the significance of climate change before their research participants would make explicit commentary about perceived risks, vulnerability, impacts and responses. This aspect of the research process makes it difficult to assess the extent to which participants’ expressed views and intentions are likely to align with actual behaviour – a well known problem when trying to estimate migration intentions (Lu 1999).

Linking climate change to migration is additionally difficult using the existing literature because either it is assumed to have not happened yet (in the Australian literature) or the research is done only with those people who have not (yet) migrated (in the Canadian and European literature). In particular, several studies (Ford et al. 2008, McLeman 2010) lament that climate change threats to traditional modes of engagement with the environment are likely to further increase the distance between young Indigenous people (symbolically if not literally) and their communities – a gap that has been created over time through changes in economies, education and legal systems. There is a lack of research in the northern Australian context that considers the insights into relationships between climate and migration that might be gained from these young people, or from other people who have already migrated.

Such insights would contribute greatly to the debate about whether climate induced migration is likely to simply exaggerate existing characteristics of migration patterns (who migrates and where they go) or to inspire substan- tially new patterns. Bardsley and Hugo (2010) have suggested generally that, while the initial evidence is that new migration will follow established patterns, there will be ‘thresholds’ beyond which dramatically new patterns are likely to emerge. Ford, Berrang-Ford et al. (2010) suggest an increase in rural to urban migration and a dramatic change in the destinations for seasonal and circular migration. Fast and Berkes (1998) claim that Indigenous Canadian’s capacity to open up new migration paths has been diminished by the advent of fixed settlement and the move away from an explorer / nomad lifestyle, but this may be relevant only to short distance migration patterns.

The concern about the vulnerability of young Indigenous people to the impacts of climate change is also at least in part a concern about the vulnerability of Indigenous communities to the (continuing) loss of young people through climate induced migration. The majority of the literature we have consulted prefers to consider adaptations (including migration) as collective strategies embraced by Indigenous communities, or at least extended families (as with Petheram et al.’s (2010) discussion of voluntary relocation). Certainly, collective decision making is easier for external agencies (government departments, non-government organisations, emergency services etc.) to engage with and manage. It may also be easier for communities to argue their case for increased responsibilities for decision making where they can demonstrate a collective will. In an extensive analysis of migration patterns of Indigenous people in the Northern Territory, Taylor (Taylor 2012, Taylor and Carson 2009, Taylor and Carson 2012, Taylor et al. 2011a) recognises that migration often occurs in relatively large groups from within communities, but that it also often occurs by individuals, couples, and unrelated people from different communities. It is not only young people who are likely to migrate independently of ‘the family’ or ‘the community’, but there is certainly evidence that they are more likely to do so, and that this trend is likely to increase in the future (Taylor 2012).

Another sensitive dimension (also associated with, but not limited to, young people) is whether migration occurs in the search to continue practicing existing livelihoods or in the search to establish new livelihoods. The anthropological view (dominant in the Canadian research by Ford and others) is that the pursuit of traditional livelihoods drives selection of adaptive strategies. New circular and seasonal mobility patterns emerge to ensure access to traditional hunting grounds and species (Ford et al. 2010), new forms of transport are employed to ensure access to traditional lands (Cohen 1997) and so on. However, Ford and colleagues (2006) also note that the general pressures on even remote Indigenous communities are to move away from traditional livelihoods. Much of the pressure stems from a desire by governments to have ‘fixed settlement’ (Ford et al. 2006, Howitt et al. 2011, Bird et al. 2013). Fixed settlements reduce the capacity to use particularly short distance migration as adaptation and may therefore either lead to longer distance (and potentially longer term) migration or increased vulnerability in situ .

Muruiki et al. (2011) summarise the process in terms of ‘three migrations’ observed among Indigenous people in Kenya. The first migration was resource 103

driven and involved the search for food, water, and shelter (what we here have referred to as mobility). This migration was often culturalised into seasonal patterns associated with the climate and environment and intricately linked with patterns of trade, initiation rites, and marriage exchange. The second migration has been politically driven, a sentiment reflecting Howitt et al. (2011), through enforcement of fixed settlement and government policies which attempt to facilitate and restrict migration according to where and how governments would like to deliver services (Kainz et al. 2012). In many cases, the migration patterns required to survive the second migration were incompatible with those from the first migration. Some communities have adapted well (for example, by creating new ‘songlines’ linking locations in central and southern Australia (Tonkinson, 2011)), while others have suffered the consequences of dislocation from culture.

Climate change may contribute to the generation of a third migration which is both resource and politically driven. The political pressure to adapt comes with a set of preferred policies and strategies for what adaptation should look like (including migration), while the capacity to sustain the reshaped (following fixed settlement) and culturally vital relationship with the land will require further changes in mobility and migration. Which individuals, and which populations, desire or are required to seek new livelihoods under these circumstances is difficult to assess, but once again researching these issues with young people and those who have already engaged in the third migration will provide valuable insights.

There are limits in the northern Australian context, however, to what we can learn about future climate related migration from that which has occurred in the (recent) past. While analogues and case studies can be valuable tools (Ford et al. 2010), they are also problematic in terms of the assumptions that are made about homogeneity (the extent to which one Indigenous population resembles another) and temporality (the extent to which the past resembles the present and the future). Koch and Carson (2012) have acknowledged these problems in the context of attempting to develop models of demographic change in small populations in sparsely populated areas. Firstly, they describe the Modifiable Social Unit Problem (MSUP) as the potential to draw new conclusion based on relatively minor changes in how populations are constructed for comparison (changing definitions of ‘Indigenous’ or ‘resident’, for example). Secondly, the Modifiable Temporal Unit Problem (MTUP) is defined as the potential to draw new conclusions based on relatively minor changes to the timeframes used in analysis. This latter might be as simple as assessing demographic change in a location over a period of 6 years following a severe weather event compared with just 4 years (or 8 years).

Existing analogue research may be dealing with populations (social units) that are importantly different from the one we wish to learn about (following Taylor, 2009) while examinations of past and present migration patterns situate communities in a very different temporal context. In north- ern Australia, for example, existing climate related migration is likely to be largely proactive (i.e. anticipating further change) while climate related migration in 20, 40 or 60 years from now may be far more reactive.

The predecessor to MSUP and MTUP was the Modifiable Areal Unit Problem (MAUP) (Openshaw 1984). Openshaw argued that geographic scale matters, and that what is observed to happen at a regional level, for example, may not be what happens at a local level. Leyk et al. (2012) are particularly forceful in suggesting that this is the case when examining patterns of climate driven outmigration from rural South Africa. Regional ‘averages’ were not reflected in the characteristics of specific locations, and there was great diversity even within communities, with some neighbourhoods experiencing high outmigration and some experiencing low outmigration. Analysis of 2006 Census data underway by the authors of this paper relating to 104 Indigenous communities across northern Australia reveals great diversity in existing migration patterns. For example, average rates of outmigration of Indigenous men aged in their 20s from these communities between 2001 and 2006 was 13% (i.e. 13% of these men who were living in these communities in 2001 were not living in the same community in 2006), but the standard deviation was 17%, and some communities had lost as much as 60-70% of this population, while others had lost less than 5%. The diversity is not explained simply by the spatial distribution of communities, with substantial differences existing between neighbouring locations (for example, outmigration from various communities in the Torres Strait Islands for males in their 20s ranged from less than 4% to over 25%). It would be wise to commence investigation of migration as adaptation to climate change with an assumption that this sort of diversity is likely, and a careful consideration of what geographic scale presents an appropriate level of analysis (should central communities be considered with or separate from their homelands and outstations, for example?).

There may also be diversity between communities in terms of who chooses (or is forced) to migrate. Young people tend to be more mobile than older people, but there is considerable debate about whether young men are more likely to migrate out of remote Indigenous communities than are young women (Taylor 2011), and whether relatively wealthy people are more likely to migrate than those with fewer economic resources (Perch-Nielsen et al. 2008). Taylor’s (2011) work in particular demonstrated a diversity of age and sex specific migration patterns applying to remote Indigenous communities across the Northern Territory. It is not clear that Indigenous populations in that context are adopting the same migration trends (or other demographic trends such as reduced fertility (Johnstone 2011)). Research such as this emphasises the need for careful selection of analogues in particular.

Dimensions of Climate Induced Migration

Table 1

Summary of Dimensions of Climate Induced Migration

Dimension

Voluntary О Mandatory ________________________________________

Short distance О Long distance ______________________________________ Short term О Long term ________________________________________

Only adaptation О One of many adaptations _________________________ Outmigration O Inmigration _______________________________________

Mobility О Migration ____________________________________________

Climate as driver О Many drivers __________________________________

Exaggerate existing migration patterns О Inspire new migration patterns Individual decisions О Collective decisions ____________________________ Seeking similar livelihoods О Seeking new livelihoods ________________ Proactive О Reactive _____________________________________________ Regional О Local _____________________________________________

Well resourced О Poorly resourced

This paper has largely described the dimensions as independent of one another, but there may be relationships, which are more likely than not to occur. The research evidence for any such relationships is, however, very limited. It may be assumed that short distance migration is correlated with short term migration (as was the expressed wish of community members consulted by Petheram et al. 2010), but there have been no rigorous studies of such a relationship. It may be that less well resourced (economically) people move shorter distances, or find it more difficult to return (as suggested by Howitt et al. 2011), but again this is largely speculation at the moment.

What is clear is that climate change is just one of many migration pressures being exerted on northern Australian remote Indigenous communities. Government policy, socio-economic conditions, impacts of new technologies and cultural change appear to be considered as more immediate and dramatic issues than climate change. Not all of these pressures are acting in the same direction (i.e. encouraging the same patterns of in- or out- or nonmigration), so it is difficult to assess what the impact is or will be of adding (or emphasising) climate change to the mix.

Where migration does emerge as an adaptive strategy, it is likely to present challenges for the people who are migrating, for the places they migrate to, and the communities they migrate from. This is already the case with migration from remote to urban settlements in particular, but will apply equally where migration changes the patterns of occupation of outstations and homelands and affects the exchange of people between remote settlements. Nevertheless, migration is currently considered as part of the adaptation of remote communities (formalised in disaster management plans), and it is difficult to conceive of comprehensive adaptation strategies, which do not allow for some level of migration.

The challenge for communities and policy makers is to ensure that Indigenous people have appropriate migration options, and the skills and resources to exercise control over their migration decisions. For researchers, along with assessing the dimensions of migration as they apply to specific communities, there is a need to learn more from young people and those who have already migrated out of remote communities to better understand how individuals and communities (of origin and destination) can be better prepared to deal with the consequences. This task cannot be based on a general model of what migration is expected because each community will respond differently based on its unique circumstances.