John Libbey Eurotext

Science et changements planétaires / Sécheresse

Coupling technology with traditional knowledge and local institutions to deal with change in rural households: A focus on the semi-arid tropics Volume 24, issue 4, Octobre-Novembre-Décembre 2013

Figures

See all figures

sec.2013.0404

Auteur(s) : Victoria Reyes-García1 victoria.reyes@uab.cat, Matthieu Salpeteur2 matthieusalpeteur@hotmail.com, Laura Calvet-Mir2 lcalvetmir@gmail.com, Tarik Serrano-Tovar2 kirat@gmail.com, Erik Gómez-Baggethun2 Erik.Gomez@uab.cat

1 ICREA-Universitat Autonoma de Barcelona Institut de Ciència i Tecnologia Ambientals 08193 Bellatera Barcelona Spain

2 Universitat Autònoma de Barcelona Institut de Ciència i Tecnologia Ambientals 08193 Bellatera Barcelona Spain

Reprints: V. Reyes-García

In ecology, disturbances are typically defined as discrete events in time that disrupt ecosystem structure and change resources, substrate availability, or the physical environment (White and Pickett, 1985). Major ecological disturbances include fires, droughts, flooding events, wind storms, insect outbreaks, whereas anthropogenic disturbances include forest clearing, introduction of exotic species and political and economic crises. Ecological and anthropogenic disturbances are often coupled. Because disturbances are a constituent part of the dynamics of social-ecological systems, all societies have developed mechanisms to cope with them. Contemporary industrial societies typically rely on engineering and technological means to control any form of disturbance and variability threatening food production or other aspects of the resource base for survival (Holling and Meffe, 1996). But before the advent of industrial mechanization and energy-based technological innovations, societies primarily relied on other types of adaptation strategies. A common feature of those adaptation strategies is that they were based on a long-term understanding of the dynamic relations between human cultures and the environment (Colding et al., 2003b, Tengoe and Belfrage, 2004; Haque and Etkin, 2007). Round the world individuals have developed a myriad of innovations that – when spread by cultural transmission (imitation, teaching, emulation, and the like) (Mesoudi et al., 2012) – have been instrumental in helping societies to cope with environmental and other types of change. These innovations take place through a variety of mechanisms including chance (accident, copy errors), invention (trial-and-error, exploration), refinement of existing knowledge, and recombination or exaptation (applying existing knowledge to new functions) of knowledge.

The storyline of this paper is that, for long stretches of history, and in many places still today, the effectiveness of locally developed technologies has relied on coupling them with a) a deep knowledge of the local environment (traditional ecological knowledge); and b) a set of shared rules, norms and conventions on how to apply society's technology and knowledge (locally evolved institutions).

Traditional knowledge refers to the cumulative body of knowledge, practices and beliefs evolving by adaptive processes and handed down through generations by cultural transmission. When this knowledge relates to the relation of living beings (including humans) with one another and with their environment, it is referred to as traditional ecological knowledge (Berkes et al., 2000). Researchers agree that this type of knowledge evolves over time from long-term observation and learning from crises and mistakes (Olsson and Folke, 2001). Furthermore, because of its intrinsic dynamic nature (Gómez-Baggethun and Reyes-García, 2013), traditional ecological knowledge has had an important role in building resilience to disturbance among rural and indigenous communities that rely on ecosystem services or natural resoources as primary means for subsistence and/or sources of income (Berkes and Turner, 2006; Gómez-Baggethun et al., 2010).

Institutions are defined as humanly created formal and informal mechanisms that shape social and individual expectations, interactions, and behaviour (North, 1990; Ostrom, 1990). Institutions can be expressed through “expectations, stability and meaning” and are essential to coordination between individuals and between organizations (Vatn, 2005). In that sense, institutions are different from organizations. Organizations are better understood as actors while institutions regulate the interaction between these actors (Vatn, 2005). Local institutions can contribute to the resilience of social-ecological systems because they allow societies to store their collective memory (Berkes et al., 2003; Folke, 2004; Berkes and Turner, 2006; Barthel et al., 2010) and because they have the potential to glue the community together and promote social cohesion in the face of disturbance and crises (Gómez-Baggethun et al., 2012). This last aspect is of primary importance since anthropological research on small-scale societies shows that social relations and reciprocity systems within communities are likely to deteriorate during crises and famine, thus increasing the likelihood of robbery, murder, and revolts (Sahlins, 1972; Ember and Ember, 1992).

The core argument of the paper builds on Agrawal's (2008) framework on strategies of adaptation to climate change. Agrawal proposes a framework to analyze adaptive capacities, or the preconditions that enable actions and adjustments in response to change, and coping mechanisms, or the existing resources that are used to achieve goals during and immediately after disturbance. His framework is based on the idea that most strategies that help households to cope with environmental or other types of shocks do so primarily by spreading risks. In that sense, main strategies to cope with change or shocks include mobility, storage, diversification, and pooling and sharing. Mobility helps households spread risks across space; storage helps households spread risks across time; diversification is used to spread risks across asset classes; and pooling and sharing contribute to spread risks across households or communities.

We add two main aspects to Agrawal's framework. First, Agrawal analyzes those four strategies as basic mechanisms through which households address livelihoods’ riskiness in face of climatic change. Here we argue that those strategies have also been instrumental in guiding adaptation to other types of economic, social, and environmental disturbances, thus helping societies to deal with many types of change and shocks. Second, we argue that the effectiveness of the four strategies proposed is dependent, not only on specific technological developments, but also on a) the traditional ecological knowledge of the society (Gómez-Baggethun et al., 2012) and b) the correct functioning of the local institutions that regulate their implementation and use through periodical readjustment (Ilich, 1975).

We devote the next four sections of the paper to present one example of each of those strategies and devote the last section of the paper to discuss the role of traditional ecological knowledge and local institutions in dealing with change in the semi-arid tropics. Our examples draw from different contemporary indigenous and rural societies and highlight the role of traditional ecological knowledge and local institutions in managing the resource base of households and communities. We then discuss how traditional ecological knowledge and local institutions contribute to mobility among pastoralists in Gujarat (India). The next section analyzes the contribution of local institutions to the adequate functioning of water tanks, a traditional water storage technology in South India. We then analyze the links between crop diversification, a well known risk spreading strategy, and traditional ecological knowledge using a case study from a hunter-horticulturalist Amazonian society. The following section focuses in explaining how sharing knowledge about local landraces can function as a mechanism to preserve them in a context where markets dominate seed acquisition.

In the last section of the paper, we focus on the role of traditional ecological knowledge and local institutions in dealing with change in the semi-arid tropics. In the arid and semi-arid regions of the world, accounting for approximately 30% of the world total area and 20% of the total world population (Sivakumar et al., 2005), climate variability and unpredictable occurrence of water extremes, especially droughts, are major sources of ecological disturbances. As traditional agricultural systems are dependent on local ecological conditions, water extremes are also a principal source of fluctuations in food production. Thus, throughout history, extremes of droughts and floods, but also of heat and cold and other various forms of violent weather changes have generated fluctuations on the agricultural systems in these regions, generating shocks and variability in food production (Mirza et al., 2001). But semi-arid tropics, societies have also traditionally developed a set of mechanisms to deal with such disturbances. The cases of pastoralist mobility in Gujarat and water storage in South India are good examples of mechanisms that have traditionally helped societies in the semi-arid tropics to deal with ecological disturbances. We propose that attempts to increase the adaptive capacity of such social-ecological systems to deal with disturbances should make an effort to couple technological innovations with sociocultural elements.

Mobility: Nomadic and semi-nomadic pastoralists in Gujarat

Mobility is a strategy that helps households and communities to spread risks across space (Agrawal, 2008). Through movement, human groups extend their spatial range of action and take advantage of resources spread across large areas, including those that are out of reach for sedentary populations. For small-scale societies depending on natural resources for their livelihoods, mobility helps to cope with spatial and temporal variations of resources’ availability. Mobility is considered as the last coping mechanism for agricultural populations in face of environmental risks (McGregor, 1994), but it is also a well-known and widely used strategy for a wide range of societies across the globe living in different ecosystems and depending upon different food production strategies: hunting-gathering, slash-and-burn agriculture, and nomadic pastoralism (Ingold, 1986; Oteros-Rozas et al., 2013).

In the semi-arid tropics, mobility is primarily associated to nomadic and semi-nomadic pastoralism (Niamir, 1995; Niamir-Fuller, 1999). In such marginal areas, periodic moves allow groups to adapt their grazing pressure to the low carrying capacity of the environment (Dyson-Hudson and Dyson-Hudson, 1980; Bonte et al., 1996). Furthermore, nomadic societies are characterized by a high flexibility, which allows them fast adaptation to changing ecological and socio-political conditions (FAO, 2001; Casimir and Rao, 2003). Researchers have made attempts to assess the specific features of nomadic societies linked with the mobile way of life, highlighting that nomadic pastoral societies have developed traditional ecological knowledge and local institutions that allow them to adapt to the high variability of natural resources availability.

We illustrate this in a case study of the Raika-Rabari1 of North-western India. The Raika-Rabari represent the main group of transhumant pastoralists in Western India. They are mainly found in Gujarat and Rajasthan states. Initially specialized in camel herding, they progressively switched to sheep and goat keeping as this pack animal was progressively replaced by mechanized transports (Prévot, 2007). Over the last two decades some have shifted to sedentary cow and buffalo keeping for milk production and wage labour (Choksi and Dyer, 1996; Köhler-Rollefson, 1999). The sale of wool is the main income source for groups living in Rajasthan, whereas the sale of young males in the meat market is the main income source for groups living in Gujarat. Different specializations and varying degrees of agricultural dependence, nested in a broader multi-resource economic system, have been considered key features in the adaptability of the pastoralist way of life (Khazanov, 1984; Pratt et al., 1997).

The transhumant Raika-Rabari conduct seasonal migrations, moving during the dry season (October through May) and returning to their villages during the monsoon time (June through September). As such mobility implies unstable relations with people outside the community, the Raika-Rabari have developed specific institutions to deal with environmental as well as social variability and change. Seasonal migration is carried out mostly by men, although they are sometimes accompanied by their wives and older children. The Raika-Rabari cluster in nomadic groups, called dangs, of varying size but common political organization (Agrawal, 1999). Depending on the circumstances, households and their herds move alone or in clusters of up to 40 households. The creation of such groups depends on friendship and trust relations, and they include individuals from different kin groups and villages. Each group has a leader, called Nambardar in Rajasthan (Agrawal, 1999) or Patel in Gujarat (Swayam, 2001), chosen by the dang members for his experience in migration and animal keeping, knowledge of the areas to be visited, negotiation abilities, and trustfulness. This leader is in charge of the decision-making for matters affecting the whole group: he negotiates the sale price of animals and wool with middlemen; he makes arrangements for group stays and grazing opportunities with farmers and village councils; he deals with government officials to obtain grazing rights in forest reserves; and he is the legal responsible if problems arise. When the dang has to move to a new place, the leader will scout the potential grazing areas and gather information from group members and villagers to choose the next migration place (Agrawal, 1999; Prévot, 2007).

This local institution developed by the Raika-Rabari help them to adapt to change in two different ways. On the one side, the convey flexibility to the composition of migrating units (Khazanov, 1984). For example, grouping can help nomadic communities to obtain better prices through collective negotiations, thus increasing the economic returns of nomadic units while ensuring the required relations with people outside the community. Similarly, splitting can help them to reduce and spread their pressure on fragile environments during critical years and on high-density farming areas, and in doing so, limit conflicts with farmers. On the other side, the delegation system in decision-making, allows group members to optimize the advantages of migration, as they save the time normally dedicated to negotiate or otherwise relate with outsiders, and invest this time in herd management (Agrawal, 1993; Agrawal, 1998; Agrawal, 1999), although excessive centralization can also reduce resilience if the whole group depends on a single leader.

Migrations and animal husbandry require specific traditional knowledge to ensure the success of temporary moves and the sufficient reproduction of herds. Information regarding local ecosystem composition, cycles, and climatic conditions allows herders to know which grass or tree species will be available in each area at a particular time. For example, in the Kutch region of Gujarat, individuals generally make a distinction between “rainy season” and “dry season” grasses, names that indicate the temporal availability of particular species (figure 1). Pastoralists are aware of the effects of the different grasses on milk production and offspring bearing, as well as of the healing properties of numerous species (Köhler-Rollefson, 1997a; Mistry et al., 2003). As migration often involves passing across agricultural lands, and pastoralists partly feed their animals from the fodder remaining in the fields after the harvest, they also master information on agricultural cycles, crop quality, and farmers’ attitude, criteria that partly determine the choice of migrating routes. Animal keeping and breeding practices also constitute an important part of the traditional ecological knowledge of these populations. For example, pastoralists monitor birth rates to manage the size of herds to adapt them to local environmental conditions and to household's needs. Moreover, they keep and maintain specific breeds adapted to the local environments, able to support the harsh, often site specific, climatic conditions as well as to survive to the long migrations (Köhler-Rollefson, 1997b).

A good example of the adaptive capacity of mobile pastoralists is provided by the Dhebar Rabari, a subgroup of the Rabari located in the Kutch area of Gujarat. Before the partition of India and Pakistan, they used to migrate to the Sindh area in Pakistan, where large pastures were available. The closing of the border with the newly created Pakistan (1947) forced them to find new pastures. They started to migrate towards the east, moving across Gujarat and reaching other states of North and Centre India (Maharashtra, Madhya Pradesh, Chhattisgarh). Thus, the socio-political crisis help the Dhebar Rabari to discover new rich pasture lands, where they continue to migrate.

The fitness of nomadic pastoralist livelihood to marginal areas and the contributions it makes to local economies plays a key role in sustaining continued food production and security in the semi-arid tropics. If nomadic pastoralist societies are highly vulnerable and threatened in many parts of the world nowadays, they also have a great potential for adaptation and mitigation of climate change (Davies and Nori, 2008), and their efficiency in food production can be higher than “modern” food production systems (FAO, 2001). Thus the development of technologies aiming at improving the food security in the semi-arid tropics should take into account the specific knowledge and institutions these societies have built through times.

Storage: Water tanks in South India

Storage helps reducing risks across time and has been historically used to address food as well as water scarcities. Water tanks provide a vivid example of a storage technology that has traditionally allowed people in the semi-arid tropics to deal with the characteristic rainfall variability of the region.

Water tanks are shallow water reservoirs ranging from a few hectares to over a thousand hectares and formed by constructing earthen embankments that dam rainfall and seasonal runoff in situ and that extend across the natural drainage flow. Water tanks are found in almost every village in arid and semi-arid regions, where rainfall is low (350-800 mm), interannual variability is high, and the soil has low permeability reducing percolation (von Oppen and Subba Rao, 1980; Agarwal and Narain, 1997; Gunnell and Krishnamurthy, 2003). Those areas include the dry zone of Sri Lanka (Li and Gowing, 2005) and the semi-arid southern and central India (Gunnell and Krishnamurthy, 2003).

Water tanks in South India have been built for over 3,000 years, and the development of agriculture in the region seems to be linked to their expansion (Mosse, 2003). By impounding runoff water from the monsoon rains, tanks have been a critical technology supporting agriculture in the region (Jayatilaka et al., 2003). Besides their role in agriculture, water tanks have become a central element of local agro-ecosystems, providing a wide variety of other socio-economic uses (i.e., fresh water for domestic uses, fish, silt, grass) and ecological functions (i.e., contribute to flood control and runoff mitigation, provide protection of the biodiversity of the surrounding area, avoid erosion, recharge the water table) (Meinzen-Dick, 1984; Wade, 1987; Janakarajan, 1993; Palanisami and Meinzen-Dick, 2001; Ratnavel and Gomathinayagam, 2006; Prabhakar, 2008). Those multiple uses and functions benefit many different sectors of the society, including farmers and non-farmers, with marginal sectors (i.e., Scheduled Castes) using tank resources in more diverse ways than other sectors of the population (Reyes-García et al., 2011).

As with other common pool resources, water sharing and the management of water infrastructures calls for cooperation and a set of managing rules within and between villages (Ostrom, 1990). The management of water tanks implies high coordination not only to achieve the equitable distribution of water between head- and tail-reach farmers and other beneficiaries, but also to conduct tanks’ and canals’ regular maintenance, including desilting and cleaning the canals and the tank, preventing encroachment or water diversion, and strengthening the canal walls or deepening the canals beds. As many tanks are interconnected in complex systems through water canals, tank management involves many stakeholders and institutions across villages, which in South India has developed into a complex management system involving both local and state level institutions. Customary local institutions have typically managed tank resources (water, fish, grass, trees, etc), whereas the tank infrastructure has typically remained under state authority (Janakarajan, 1993; Vaidyanadhan, 2001) (figure 2).

Despite the fact that, for centuries, tanks have been the main source of surface irrigation in South India, their irrigated area has steadily fallen during the nineteenth and twentieth centuries (Janakarajan, 1993; Palanisami and Meinzen-Dick, 2001; Aubriot, 2008). The causes of this decline are multiple and complex, but many authors interpret the sharp decline in tank use in direct relation with the decline of local institutions for tank management (Agarwal and Narain, 1997; Palanisami and Balasubramanian, 1998; Aubriot and Prabhakar, 2011). For example, farmers in Tamil Nadu who invested in pumps have been freed from the constraints of surface water availability and collective irrigation (Aubriot and Prabhakar, 2011). As typically only the wealthiest owners have been able to invest in pumps and bores and as they were also those playing an important role in collective decisions for water management, their shift to other forms of irrigation has resulted in the neglect of institutions regulating water tanks’ management, and a consequent decline of water tanks (Janakarajan, 1993; Palanisami and Balasubramanian, 1998).

Over the last two decades, many researchers, policy-makers, and donor agencies have become increasingly disenchanted with groundwater and large-scale irrigation systems such as big dams (Moris and Thom, 1990; Hussain and Hanjra, 2004; Webb, 2006) and have shifted their focus to farmer-managed irrigation systems (Watson et al., 1998). The shift has occurred parallel to a trend to decentralize water management programs from the state to local users (Parker and Tsur, 1997). Water tanks have been seen as an opportunity, as they provided a millenarian-proved technology for water storage, while decentralising water management to local populations (Meinzen-Dick and Zwarteveen, 1998; Webb, 2006). In the 2000, the Tamil Nadu State government enacted a Farmer Irrigation System with the aim to transfer the management of tanks to farmers’ organizations. Under this Act, farmers in the command area of a tank, and only them, were compulsorily included in Water Users Associations (WUA), the new formal organizations that should assume the duty of managing the water tanks and its uses, including planning the water rotational system, solving conflicts, and removing encroachers. Although WUAs aimed at increase local participation, they were created without taking into account the existence of customary institutions for tank management. Those associations have been controlled by new elites linked to local powers, thus creating a myriad of new local conflicts (Aubriot and Prabhakar, 2011).

In sum, programs aiming at the ’revival’ of water tanks in South India have revolved around the idea of making technical interventions to restore water tanks to their original design, paying no attention to institutional issues other than creating incentives for farmers to manage the tanks themselves. Several authors have hypothesized that a successful approach for the ‘revival’ of water tanks should focus on institutional, as well as on technical issues, involved in the functioning of this water storage infrastructure (Meinzen-Dick and Zwarteveen, 1998; Webb, 2006). In that sense, water tanks constitute an example of how coupling technology and local institutions might be necessary to increase the adaptive capacity of social-ecological systems.

Diversification: Crop diversity among Tsimane’ hunter-horticulturalist in the Amazon

Diversification is a universal risk spreading strategy that can be adopted in relation to a wide variety of productive and non-productive assets, consumption strategies, or livelihood activities (Turner et al., 2003) and that can operate at several spatial and temporal scales. Diversification protects livelihoods by pooling risks across households’ and societies’ assets and resources. Diversification is a reliable risk management strategy against adverse environmental (Bentley, 1987; Zimmerer, 1996) and economic shocks (Perreault, 2005) to the extent that benefits generated by the different assets are subject to uncorrelated risks and to the extent that the returns given up by investing in diversification are lower than the security provided by diversification.

It is well known that in relatively remote rural settings, households diversify agricultural production in a variety of ways. For example, to protect food production against localized risks related to environmental and economic variability, households plant several crops or varieties of crops, scatter plots, stagger the planting season, and use intercropping (Altieri, 1989; MacDonald, 1998). Among those strategies, diversification of crops and varieties has received large attention by researchers (Brush, 1992; Brush et al., 1995; Zimmerer, 1996). Because crops and crop varieties vary in almost every conceivable trait, from seed size, height, and fruiting time, to response to heat, cold, drought and ability to resist specific diseases and pests, or to nutritional qualities and taste, farmers have traditionally relied on a diverse set of crops and crops varieties to enhance food security by enhancing response diversity to disturbance.

The study of crop and varieties diversification in agricultural systems has often focused on determining the number of different crops and crop species in farmers’ fields and analyzing the relation between crop diversity and food security. For example, in a study of a transition from traditional shifting agriculture to intensive horticultural production among the Yucatec Maya, Humphries (1993) found that the high crop diversity in subsistence plots buffered the effects of environmental uncertainties. Similarly, in a study of the Colombian Amazon, Hammond et al. (1995) found that crop diversity provided a buffer against environmental risks, such as rainfall fluctuations, attacks from pests, and plant diseases. In a study in the Andean region of Paucartambo, Brush et al. (1995) documented that the peasants’ choice of crops was shaped by their protective role. Similar conclusions were obtained by Gómez-Baggethun et al. (2012) regarding crop diversification strategies among farmers in the Doñana region, in south-western Spain. But less attention has been paid to the detailed knowledge that accompanies this diversification strategy and the dependence of crop diversification on traditional knowledge systems. This was the goal of a study we conducted among the Tsimane’, a semi-autarkic society in the Bolivian Amazon (Reyes-García et al., 2008).

The Tsimane’ are an indigenous population whose economic system is based on hunting, slash-and-burn farming, and – increasingly – wage labour. The Tsimane’ obtain their main staples (upland rice, plantains, maize, and manioc) from plots cleared from old or fallow forest. Rice is the main cash crop, typically grown in newly-cleared fields. After the rice harvest, fields may be partially replanted with maize, manioc, or plantains. Maize and plantains are used both for household consumption and for sale, whereas manioc is primarily sown for household consumption. Other crops planted in small patches include pineapples, peanuts, watermelons, squash, sweet potatoes, and sugar cane (Huanca, 1999; Vadez et al., 2004). In a study of Tsimane’ agriculture, Piland (1991) reported that most of the diversity of species was found in fallow plots and home gardens, and that higher variability was found within a crop species than among different crops species. In a nutritional study in two Tsimane’ villages, Byron (2003) found that 47% of food items consumed by Tsimane’ households came from farm production, making food security highly dependent on agricultural production.

Among the Tsimane’, crop diversification is higher among households who are more dependent on agricultural production for household consumption, than among households more dependant in other livelihood strategies. Crop diversification is positively associated with the production of consumption crops (i.e., manioc and maize), but not with the production of cash crops (i.e., rice), highlighting the important connection between crop diversification and household food security. Interestingly, our research also shows a significant and positive association between the level of traditional ecological knowledge of the male household head (the person mainly responsible for agricultural production among the Tsimane’) and the number of different crop species in household's fields, suggesting that traditional ecological knowledge is a key covariate of crop diversity. Since Tsimane’ rely on intercropping and polycropping for agricultural production, systems that mimic natural environments, it is possible that Tsimane’ who have a better knowledge of relations in natural ecosystems might be better able to apply some of this knowledge on their own agricultural fields, which would allow for higher productivity, as Altieri (1999) has observed in similar agroecosystems.

In sum, the association between individual levels of traditional ecological knowledge and crop diversity in farmers’ fields suggests that traditional knowledge systems can help individuals and societies to deal with change and shocks by enhancing the protective role of crop diversification. Researchers and policy makers have stressed the importance of conserving the world's agricultural diversity, mainly to retain the capacity to develop new crop varieties that might contribute to human adaptation to global environmental change (Vadez et al., 2012). However, crop diversity conservation programs that do not take into account the traditional knowledge systems that accompany crop diversity developed in situ might fail short as they might miss the cultural information that make of crop diversification a successful risk-insurance mechanism for rural households.

Pooling and sharing: Distributing landraces knowledge through a social network

Pooling refers to joint ownership of assets and resources whereas sharing refers to joint use of assets and resources (Agrawal, 2008). Pooling and sharing within a community have been mostly studied in relation to assets and resources, suggesting that this strategy helps households to spread risk from one to another in the face of small, idiosyncratic shocks that typically affect one or few households, such as illness or death. But pooling and sharing do not seem to help well against large or covariant shocks affecting all the households (Morduch, 1995; Townsend, 1995; Kurosaki and Fafchamps, 2002).

Societies not only share resources, they also share the knowledge on how to use those resources. Differently from assets and resources, knowledge is a public good, as it is non-excludable (the use by one individual does not exclude the use of other individuals) and non-rivalrous (one individual's use does not reduce availability to others). This is mostly the case for traditional knowledge systems, where knowledge is socially transmitted and widely shared among community members (Reyes-García et al., 2003; Richerson and Boyd, 2005). Although some small-scale societies have developed systems to protect part of their knowledge (Barth, 1990), a high degree of information sharing is required to maintain the cultural transmission processes (Richerson and Boyd, 2005), especially in oral societies. As we shall see in this section, sharing knowledge can be considered an institution that allows societies to store collective memory of the community of how to manage their natural resources, thereby helping to protect communities against covariant shocks.

Previous research suggests that the exchange of knowledge is crucial for the effective governance of natural resources (Bodin and Crona, 2009), an idea we wanted to test in an agricultural context. Researchers have shown that germplasm (in the form of seeds, seedlings, or other propagules) and its associated knowledge are often exchanged together (Vogl and Vogl-Lukasser, 2003; Acosta-Naranjo and Diaz-Diego, 2008), which implies that networks of seed exchange should play an important role in agrobiodiversity in situ conservation. To test this idea, we studied how the position of a person in a social network related to the number of landraces cultivated in his or her garden and to the level of landraces knowledge hold by the same person (Calvet-Mir et al., 2012). The study was conducted among home garden tenders in a rural area of the Catalan Pyrenees, north-eastern Spain.

Home gardens in the study area traditionally harboured a wide diversity of crops and crop varieties. Before the 1970s, shops were absent in the study villages and the access to the towns was difficult, especially in winter. Because of these constraints in access to markets, farmers typically kept seeds from one year to the next or obtained them from neighbours who had stored them. Although the practice of storing seeds drastically changed in the 1970s, when the accessibility to the market town improved, around 20% of the crops in the studied gardens still do not come from commercial propagules (Calvet-Mir et al., 2011). The practice of storing and exchanging seeds and propagules is mainly maintained in relation to local landraces. Nowadays, local landraces can also be acquired from a local seed bank established in the area in 2005 in an effort to strengthen in situ agrobiodiversity conservation. The seed bank provides gardeners with local landraces, and gardeners are expected to grow the crop and to return new seeds to the bank.

Results from our study suggest that the number of landraces a gardener grows and her/his level of knowledge on such landraces are highly dependent on the position of the person in the social network. People with larger relative importance in the network of seeds exchange, i.e., people with higher centrality, also have higher levels of knowledge about landraces and grow a larger number of landraces in her/his fields. An implication of this finding is that the pooling of knowledge through the network helps ensure landraces collective knowledge as well as the maintenance of the landraces themselves. The finding that centrality in the seed exchange network is associated with local landrace knowledge reinforces previous findings on the importance of seed exchanges in maintaining local agrobiodiversity (Thiele, 1999).

Interestingly, we also found that less than 10% of the gardeners in the study area were active collaborators with the local seed bank, the formal organization in charge of maintaining local landraces. Furthermore, our analysis suggests that the informal network of seed exchange played a more important role than formal institutions in the effective in situ maintenance of agrobiodiversity. The same findings have been reported in other contexts (Thiele, 1999; Bodin and Crona, 2009). In sum, in our case study informal networks to share knowledge on resource management were critical to maintain collective memory and promote in situ maintenance of agrodiversity. In a sense, our results help conceptualize social networks as human biologic corridors to facilitate the conservation of agrobiodivesity knowledge by sharing it among the network actors.

The role of traditional ecological knowledge and local institutions in dealing with disturbances in the semi-arid tropics

Global environmental change adds new challenges to traditional agricultural systems in semi-arid regions. While certain regions of the globe will be favoured by climate change as it might bring under production areas that were traditionally too cold, the arid and semi-arid tropics, where cropping conditions are already under environmental pressure, will be the most affected. Increasing surface temperature and decreasing rainfall trends have already been reported for several countries in tropical Asia (Sivakumar et al., 2005). Similarly, trends in African rainfall have changed substantially over the last 60 years and a number of theoretical, modelling, and empirical analyses have suggested that noticeable changes in the frequency and intensity of extreme events, including floods, may occur even with only small changes in climate (Sivakumar et al., 2005). Modelling work projects increasing warming in the semi-arid tropics of Asia and Africa, changes in the nature and characteristics of monsoon, a decrease in freshwater availability, and increased frequency of drought (IPCC, 2007). All those changes are expected to have serious impacts on food security in the semi-arid tropics of Asia and Africa (Sivakumar et al., 2005). Moreover, environmental stress, not only poses risk on agricultural production, but also on other provisioning ecosystem services (pastures, game, etc.) that often constitute important complements for the income and nutrition of local populations.

National and international policies have dictated a set of mitigation strategies with a market top-down management component to reduce the sources or enhance the sinks of greenhouse gases in an attempt to limit the effects of global climate change. There are also national and international policies oriented to help societies to adapt to global environmental change. For example adaptation strategies oriented to reduce their vulnerability to climate change in the semi-arid tropics include measures such as the use of improved (sometimes through Genetically Modified Organisms) crop varieties, the improved use of fertilizers, the alteration of the timing or location of cropping activities, changes in pest, disease and weed management practices, a better use of seasonal climate forecasts, or the construction of big infrastructures to deal with water scarcity or flooding (Harris, 2002; Colding et al., 2003a; Orr and Ritchie, 2004; Schlecht et al., 2006; Rockstrom et al., 2010; Vadez et al., 2012).

A common trend of those adaptation strategies is that they heavily rely on engineering and technological means that are often decoupled both from the local knowledge of the ecosystems where the technologies should be implemented and from the local institutions. As the case studies presented here suggest, and as resilience scholars have suggested (Adger et al., 2003; Colding et al., 2003a; Adger et al., 2005; Pelling and High, 2005), whether a technology might contribute to help societies to deal with change depends, not only on the technology itself, but also on the way the technology is adopted and implemented.

Conclusion

In conclusion, it is widely accepted that adaptation strategies should be prescriptive and dynamic, rather than descriptive and static (Sivakumar et al., 2005), but it is far less acknowledged that adaptation strategies might not be successful in creating social-ecological resilience to natural disaster unless technologies fit local knowledge systems and local institutions (Colding et al., 2003). Therefore, any attempt to implement adaptation strategies oriented to reduce the vulnerability of agricultural systems to projected climate change should not only draw on the implementation of new technologies, but also tap into social-ecological memories embedded in local institutions and traditional knowledge systems (Barthel et al., 2010). Traditional knowledge systems and local institutions could be important complements to science and technology in creating successful adaptive strategies to protect food security in the semi-arid tropics and elsewhere.

Acknowledgements

Research was funded by grants from the programs of Cultural Anthropology of the National Science Foundation, USA (BCS-0134225 and BCS-0322380), the 6th Framework Program, EU (CT-2006-036532), and the Ministerio de Educación y Ciencia, Spain (SEJ2007-60873/SOCI and CSD2010-00034). V. Reyes-García thanks a mobility grant to BRIC countries (Vicerectorat the relacions internacionals, Universitat Autònoma de Barcelona, Spain) for the writing stage, and Resilient Dry Land Systems, ICRISAT-Patancheru (India) for providing her with office facilities.