1.1 Climate change and its impacts

Climate change involves changes in mean values and variability patterns of meteorological parameters over a 30-year time scale at specified locations. The vulnerability of humans and ecosystems to climate change is to a considerable extent determined by changes in the frequency or severity of extreme events, including droughts and floods that directly affect human food supply and income (Cruz et al., 2007). Increased climate variability has been predicted for both East Africa and Southeast Asia, including Kenya and Vietnam. For Kenya, the reported 0.7–2.0°C increase in temperature during the last 40 years, together with irregular and unpredictable rainfall, has increased water-scarcity problems, alongside degradation of catchment areas and lakes (Mutimba, Mayieko, Olum, & Wanyama, 2010). The predicted changes in annual maximum and minimum temperatures in East Africa by the late twenty-first century are 1.8°C and 4.3°C, respectively (IPCC, 2007). Global climate change is expected to have significant effects on the Asian monsoon circulation with consequences for the mean and variability of regional climate in Southeast Asia, including Vietnam. Expectations are that climate warming will cause intensification of the monsoon, resulting in greater interannual and multi-decadal variability in the form of more frequent and more severe droughts and floods (Overpeck & Cole, 2007). In Vietnam, heavy rain, droughts and floods are becoming more frequent, particularly in central coastal provinces (Beckman, 2011; IPCC, 2007).

These changes are of direct economic importance to national governments. Degradation of water resources in Kenya costs an estimated KES 3.3 billion (about US$ 39 million) annually (Mogaka, Gichere, Davis, & Hirji, 2006), while in Vietnam the total damage caused by extreme flooding in 1996 and 2001 alone in the Red River Delta, Mekong Delta and Central Region was estimated at US$ 680 million (ADB, 2009). As a result, the Kenya Government launched the National Climate Change Response Strategy in 2010 and the Government of Vietnam adopted the National Target Program for Climate Change in 2008. However, operational aspects of these strategies are still a challenge. Research in Tanzania and Kenya has found that government tree-planting efforts that do not match local ambitions and preferences have a low chance of success and that indigenous knowledge could be integrated into formal adaptation planning (Kangalawe et al., 2011). Therefore, understanding the local context and the responses of households is critical in addressing the challenges of climate change. Formulation of appropriate policy options that can support effective local adaptation is also important.

1.2 Local response to climate change and diversification management

Maintaining a diversity of options within a portfolio is an important management tool that allows farming to survive climate variations. With portfolio diversity, many components that do not appear particularly productive for much of the time can suddenly assume key importance when others fail.

Some autonomous adaptation is already taking place in the areas vulnerable to climate-change-related disasters in Southeast Asia (Fransisco, 2008). This kind of adaptation is mainly made by farmers in order to cope with climate risks, and is defined by Malik, Qin, and Smith (2010) as ‘private adaptation’. In order to cope with climate risks, local farmers in central Vietnam have changed crop varieties, adjusted the seasonal calendar, built irrigation systems, dug pump wells, introduced alternative livestock breeds, found new feed and fodder sources, and participated in vaccination programmes organized by the government (Oxfam Vietnam and Kyoto University, 2008). Furthermore, paid work, trade in commodities, migrating to find work and sending money back to families are also common (Adger, 1999; Kelly & Adger, 2000). The farmers in the Mekong delta diversify local livelihoods through introducing agriculture and forestry activities along with rice farming (Vo, 2003). Adaptation strategies in Kenya mainly consist of crop diversification, mixed cropping patterns and tree planting for providing fodder and shade/shelter for crops, water-conservation measures and irrigation (Kabubo-Mariara and Karanja, 2007). Social networking and indigenous knowledge have been important in climate-change adaptations in Kenya, as they allow local communities to predict droughts and rainfall patterns and farmers to adapt by changing to drought-tolerant species or early-maturing crops (Ifejika, Kiteme, Ambenje, Wiesmann, & Makali, 2010; Lenhart, 2009).

Nguyen, Hoang, Öborn, and van Noordwijk (2013) showed how existing diversity-based approaches using trees on farms and in the landscape are relevant for current climate-change adaptation approaches in central Vietnam. The present study tested the approaches in a different landscape in Kenya and compared the results. It was assumed that farmers base their managed diversity portfolios on recalled performance in past extreme weather events and that analysis of past and existing portfolios would be an appropriate entry point for defining appropriate policy to help reduce human exposure to climate variability. Our argument is based on the observation that diversity-based approaches to agriculture have been used by farmers for thousands of years to minimize risk and to ensure at least some productivity in unfavourable years (Cooper et al., 2008; Lin, 2011). Therefore, policy support to promote what farmers like, and are familiar with, would bring the best result. This assumption is supported by the examples from Vietnam and Kenya mentioned above as well as from other places. For example, research in Ghana has shown that to confront future extreme climate events, local communities expect to use similar adaptation strategies to those found to be effective in the past, with some new additions (Codjoe & Owusu, 2011).

The overall objectives of this study were to identify local farming responses to climate variability and to examine the potential of policy support for on-farm diversification, for both climate-change adaptation and ensuring food security. Contrasting local ecological knowledge with current science and with public knowledge, policy discourses and perspectives can support more effective communication programmes (where key gaps are identified) and lead to policies with a good chance of implementation (where no major knowledge gaps exist; Clark et al., 2011).

2.1 Concepts and definitions

Adaptation is defined by the Intergovernmental Panel on Climate Change (IPCC, 2001) as ‘adjustment in natural or human systems in response to actual or expected climatic stimuli or their effects, which moderates harm or exploits beneficial opportunities’. This study investigated existing adaptation of farming and land-use systems to climate variability as a basis for adaptation to climate change, especially regarding changes in weather patterns and in the frequency and intensity of climate extremes. Adaptation can be treated as a state variable (‘being adapted’, ‘adaptedness’, ‘low vulnerability to existing stressors’) and as a process (‘increasing the degree of adaptedness’, reducing vulnerability to increase in stressors). Vulnerability to climate change can be considered to be dependent on the character, magnitude and rate of climate change and variation to which a system is exposed, its sensitivity and, when assessed over longer time frames, its adaptive capacity. Exposure to climate change can be interpreted as the changes in a location's climate variables (temperature, precipitation, wind speed and extreme weather events) beyond the impact-reducing effects of local buffers. Sensitivity describes the human–environmental conditions that exacerbate or ameliorate the hazard or trigger an impact. Thus exposure and sensitivity are intrinsically linked and mutually influence potential impacts. Adaptive capacity represents the human potential to implement adaptation measures in efforts to avert negative impacts (Gbetibouo & Ringler, 2009). Resilience, defined as ‘the amount of change a system can undergo without changing state’ (IPCC, 2001), is an important dimension of adaptation. According to Schipper (http://www.climate-adaptation.info/?page_id=51) ‘Resilience may either refer to the extent to which a system is able to absorb adverse effects of a hazard or it may refer to the recovery time for returning after a disturbance’. Thus highly resilient systems are characterized by an ability to endure despite high stress or to recover quickly and resilience allows individuals or systems to cope during an event without depleting all their resources or recovery capacity. However coping measures are generally short-term, while continuous or permanent threats require different survival strategies.

2.2 Study sites

The fieldwork was conducted at four locations: two villages in central Vietnam and two watershed-management divisions referred to as Focal Development Areas (FDAs) in Eastern Kenya.

The two villages (nos. 4 and 8) at the Vietnamese site are located in Cam My commune, Cam Xuyen district, Ha Tinh province (7°53′50″–18°45′40″N; 105°05′50″–106°30′20″E; 60–180 m a.s.l). The climate in the area is characterized by tropical monsoons, i.e. intense rainfall with regular flooding, during September and October (up to 2200 mm in October), followed by a long dry period from December to July, during which high temperatures in June and July and dry winds from the west exacerbate drought conditions. Village 4 has access to a water reservoir and forest, while Village 8 has better access to the main road and market. Village 4 cultivates more paddy rice, while Village 8 has more rain-fed crops and livestock (Nguyen et al., 2013).

The two FDAs in Kenya, Kithunguriri and Muthatari, are located in upstream and downstream sections of the Kapingazi River basin in Embu district, Eastern Province (0°22′–0°32′S; 37°27′–37°30′E; 1230–2100 m a.s.l). The main farming and economic activities in Kithunguriri are tea production and dairy cattle, while coffee, horticulture and off-farm work are more common in Muthatari. The Kapingazi River basin is located on the southern windward side of Mount Kenya and covers an area of 61 km². Mean annual rainfall varies from 1000 mm in the lower part of the catchment to 2000 mm in the upper part and mean annual temperature is 18–22°C. The highest temperatures occur in March and October (mean 26.3°C and 25.8°C, respectively) and the lowest in January and August (mean 12.1°C; Jaetzold, Schmidt, Hornetz, & Shisanya, 2006). The rainfall pattern is divided into ‘long rains’ in March−May and ‘short rains’ in October−December (Shisanya, Recha, & Anyamba, 2011), with light showers occurring during the remainder of the year (Ovuka & Lindqvist, 2000).

2.3 Climate patterns analysis

Trend analysis was conducted using 43 years (1965–2008) of rainfall and air temperature data for the Vietnamese site and 28 years (1980–2007) of air temperature data and 32 years (1976–2009) of rainfall data for the Kenyan site. Changes in monthly and annual air temperature and rainfall over time were statistically analysed using Minitab Software (version 14.0). Slopes with significant p-values were further analysed using pair-wise comparisons with dummy variables in Minitab. Detail about methods used and results of the statistical analysis for the Vietnam site can be found in Nguyen et al. (2013), while those for the Kenyan site will be reported in a future paper. Some main findings for the Kenyan site, cited in this paper, have been presented at Embu stakeholders' workshop.

2.4 Surveys on local perceptions of climate variability, its impacts and solutions

Participatory rural appraisal (PRA) tools, including village sketches, timelines, transect walks and brainstorming techniques, were used in combination with semi-structured interviews with individuals and groups, followed by a structured questionnaire. The semi-structured interviews were used to elicit local perceptions of climate variability, its impacts and solutions and the questionnaire to obtain data on local conditions (farm structure, access to land, labour composition, income sources and farming activities) and choice of tree species by different wealth groups. The total number of interviewees at all levels was 243 in Vietnam and 103 in Kenya (Table 1).

The field survey data were analysed and feedback was obtained through three workshops during 2011, namely two workshops for the researchers involved, held in Vietnam and in Sweden, and a one-day stakeholders' workshop in Embu district, Kenya. At the latter, survey results on climate-change analyses were presented, causes and solutions to water scarcity were discussed and technology and policy options for water conservation, including climate variability adaptation, were jointly explored. Twenty-five stakeholders attended the workshop, including farmers, researchers, administrators, government officers, extension officers and representatives from relevant projects in Embu district and local universities.

2.5 Comparative analysis

An extensive review of the literature on climate variability, local adaptation and relevant policy provisions in Southeast Asia and East Africa was conducted in order to examine the validity of our findings for the study sites in Vietnam and Kenya. The focus of this study is on three major aspects: (i) defining and comparing proof-of-climate variability from scientific and local perspectives; (ii) understanding local perceptions of climate-variability impacts on current farming systems and local responses and (iii) examining other approaches and supportive policies.

3.1 Climate variability, its impacts and local response

Local knowledge and statistical analysis at both sites indicated a trend of climate change towards lower rainfall and higher temperature, particularly in the driest months. The current intra-annual variation in temperature and rainfall and the incidence of extreme events appeared to be more pronounced at the Vietnamese sites, while at the Kenyan sites local people reported water scarcity, which was to a larger extent caused by increased water extraction (Table 2; Embu stakeholders' workshop). While interviewed Vietnamese farmers could recall their adaptation action in a particular year with extreme weathers, the interviewed Kenyan farmers tended to report their response to drought, in connection with the change in rain pattern (Table 2) and water scarcity in the Kapingazi basin (Table 3). In Vietnam, for example, due to the prolonged winter (extended by 38 days) and two flood events in 2008, approximately 70% of rice seedlings were killed, forcing farmers to replant rice varieties of short duration but lower yield, resulting in yield losses of up to 40%. Home gardens with tree-based systems and livestock were reported by the farmers interviewed to be one important ‘safety net’ when crops failed. The tea, acacia, eucalyptus, jackfruit and rattan grown in home gardens are resilient to extreme weather and can be sold to provide year-round income. Other crops such as banana, cassava and sweet potato supplement family food needs, while also providing animal fodder when rice crops fail. To cope with the increased climate variability, the crops listed above are being planted by farmers in village 4 in a new land-use form named ‘forest garden’ – a home garden type of illegal planting in forests near the village. We understood that farmers planted food crops in the forest garden as a ‘permanent’ strategy in anticipation of coming needs due to climate variability. Village 8 does not have forest land, but is using its good access to the market to raise more cattle for sale (Nguyen et al., 2013; Table 2).

At the Kenyan site, farmers' response to climate changes differed, with 50% of farmers (who rely for their livelihoods on tea and coffee production) interviewed admitting to ‘doing nothing’ when it gets too dry, but waiting until the weather improves. The other 50% of farmers interviewed reported a range of adaptation strategies, with diversification of the farm portfolio leading to diversified farm income and helping them to adapt to climate variability. Tea and coffee are the main crops, but these farmers also grow banana, cassava and beans and raise cattle. During the dry season, the farmers feed their animals with commercial feeds and with grasses which they ‘cut and carry’ from neighbouring areas (referred as zero grazing by farmers). Exotic tree species such as Grevillea robusta L. and Eucalyptus (90% of farmers with an adaptation strategy) are also commonly grown, particularly on sloping land in the upper FDA. The use of drought-resistant or early-maturing crop varieties, water tanks and drip-irrigation techniques, seasonal adjustment of farming practices, constructing food storage facilities, selling cattle and tree products, using savings and off-farm employment were all mentioned by farmers as their adaptation strategies to climate variability (Table 2). Embu town, which is only about 10 km away from the lower FDA, provides seasonal employment for farmers.

Despite the different impact levels of climate change and variability between the Vietnamese and Kenyan sites, there were similarities in terms of the patterns farmers were using to adapt. At the study sites in both countries, besides diversifying crops and varieties, planting trees and raising cattle, farmers also designed appropriate land use along landscape gradients as a strategy to buffer against increased weather uncertainty (Tables 2 and 3). For example, farmers in Vietnam moved to planting gardens in the forest during extreme events, whereas farmers in Kenya accessed the nearby town for seasonal employment. There is thus a need for development policies that incentivize diversification in agriculture and landscapes and that maintain existing diversity in order to create agricultural systems that are resilient to climate change and protect food production under future climate change.

3.2 Ways of enhancing local adaptive capacity

According to UNFCCC (2007), the most effective adaptation approaches to climate change for developing countries are those (i) simultaneously addressing a range of environmental stress and factors and (ii) synergizing with coordinated efforts aimed at poverty alleviation, enhancing food security and water availability, combating land degradation and reducing loss of biological diversity and ecosystem services, as well as improving human adaptive capacity. The main vulnerability factor at the Kenyan sites was identified as access to water, while access to forest land was the prime issue at the Vietnamese sites. Addressing the underlying cause of limited access to water and forest land is key to enhancing the adaptive capacity of local people at these sites.