Agroforestry-The Most Resilient Farming System for the Hilly Northwest of Vietnam

Over 94% of the land of northwestern Vietnam is classified as sloping. Population growth has exerted pressure on local natural resources, with agricultural expansion on steep slopes resulting in forest degradation, landscape fragmentation and severe environmental consequences. Efforts to restore forest ecosystems have been made by the government, however, as the livelihoods of 80% of the population depend on agriculture, the reconversion of land to forest has proven to be an inappropriate solution. Agroforestry offers a potentially sustainable land-use solution, which could re-establish forests, restore ecosystem services, and stabilize local livelihoods. In this paper, we assess the potential of agroforestry development in the region based on the results of two interrelated surveys conducted in 21 representative villages in six districts of three northwestern provinces: 1) a farming system diagnosis implemented in 17 representative villages; and 2) an agroforestry adoption survey with 210 households practising agroforestry in 14 villages. The analysis was strengthened by four years’ experience in implementing an extensive agroforestry project in the region. The studies focussed on assessing key benefits and constraints of existing farming systems, including agroforestry practices, adopted by farmers and identifying the potential for agroforestry development in the region. The results showed that the dominant farming system in the north-western provinces was monoculture of staple crops on slopes, which provided relatively low economic returns. Soil erosion, land degradation, and water shortages resultant from intensive farming practices were the most significant environmental issues in the area. Tree-based farming systems were rare and mostly a result of spontaneous adoption by farmers. Given the mountainous landscape and the need for soil stabilization, agroforestry was seen by farmers as a viable approach. Improving the existing and adopting new, integrated agroforestry systems were identified as viable ways toward sustainable livelihoods in Northwestern Vietnam.


Over 94% of the land of northwestern
Vietnam is classified as sloping, of which 87% has slope above 25 o (Khanh, 2012).The region features six provinces with an area of 5.64 million ha and a total population of 3.5 million made up of 30 different ethnic groups with distinct languages, traditions and farming practices.An estimated 80% of the population depends heavily on agriculture for their livelihoods (IPCN Vietnam, 2016).Population growth, partly as a result of a government resettlement policy 1961-1998, exerts pressure on available natural resources.Deforestation and expansion of agriculture onto steep slopes using predominantly monocropping practices in the last few decades has resulted in forest loss, degradation of agro-ecosystems and landscape fragmentation that threatens environmental sustainability and food security (Hoang et al. 2010).
Efforts to restore forest ecosystems have been made by the Government of Vietnam through the Five Million Hectares Reforestation Program.As a result of this reforestation effort of the last few decades, forest area has increased from 7.8 million ha in 1981 to 13.1 million ha in 2008 (VNFOREST), however, the quality of forest remains poor as the underlying drivers of forest loss are still present (Stibig et al. 2007;Government of Vietnam 2007;CEM Vietnam).The primary reason for forest conversion is that local livelihoods still depend on small-scale agriculture; converting locally-valuable agricultural land to forest is not viable (Chaudhury, 2009).Indeed, farmers in the region have continued to clear forests (Long, 2009).
Along with deforestation, the frequency and intensity of extreme weather events (floods, landslides, frosts, droughts) have increased.The damages resulting from these extreme events increased by a factor of 10 from 2001-2005to 2006-2010(DONRE Son La 2014).
Globally, there has been a steadily increasing interest in the development of agroforestry systems to enhance resilience.FAO has also recently issued a guide for policy-makers, Advancing Agroforestry on the Policy Agenda, which aims to promote agroforestry in national policy frameworks and boost its impact (FAO, 2013).Agroforestry has been used widely to address soil erosion and rehabilitate soil fertility (Nair 1993;Pattanayak and Mercer 1996;Young 2002;Hasan and Ashraful 2006;Jama et al. 2006;Blanco-Canqui and Lal 2008;Rachel et al. 2012).Additionally, well-designed agroforestry interventions provide permanent soil cover, improve soil and water use-efficiency, restore tree cover and increase carbon stocks (Rao et al. 2007;Roshetko et al. 2007;Brenda, 2010;Charles et al. 2013;Roshetko et al. 2013).
Through soil fertility and, thus, productivity improve-ment, agroforestry contributes to food and nutritional security (Dawson et al. 2013;Jamnadass et al. 2013).Agroforestry is considered to be a best alternative for climate-smart agriculture (Kaczan et al. 2013); and provide an opportunity to combine the twin objectives of adaptation and mitigation (Verchot et al. 2007;Murthy et al. 2013;Mbow et al. 2014;Lasco et al. 2014).iii) Is there evidence that farming systems that integrate agroforestry practices are more resilient than systems that do not?
iv) What is the potential for expanding locally successful agroforestry systems?

Study location
The studies were conducted in 21 villages located between 300 and 1250 metres above sea level (masl) in 12 communes of six districts in the three northwestern provinces of Yen Bai, Son La and Dien Bien (Table 1).
These villages were purposively selected as they were part of the AFLI project.The villages were grouped into three elevation ranges with different predominant farming systems related to the elevation and local ethnic groups.The FSD was conducted in 17 of the total 21 villages and the agroforestry adoption survey in 14 villages (Figure 1 and Table 1 in the results section).

Farming systems diagnosis
The assessment methods included agroecological mapping, focus-group discussions (FGD), and structured interviews.

Agroforestry adoption survey
The main methods used were semi-structured interviews with 210 households in 14 villages who practised agroforestry and observations of the households' land during field visits.The information collected included i) basic information of the household, including their farming practices; ii) types of agroforestry systems farmers practised, with system components; criteria for species selection; area and year of establishment; iii) the reasons farmers adopted agroforestry practices; and iv) from where did they learn about agroforestry.
The interviews explored the spatial arrangements, species' selection criteria and farmers' evaluations of the benefits of the systems.For the purposes of this study, agroforestry was defined as a 'collective name for land-use systems and practices where woody perennials are deliberately integrated with crops and/or animals on the same land management unit.The integration can be either in spatially mixed or temporal sequence.There are normally both ecological and economic interactions between the woody and non-woody components in agroforestry' (ICRAF 1982).

Statistical method
The data collected in this study were analysed using descriptive statistics to calculate the frequency distribution and percentage of variables.The SPSS, version 16 (2007), program was used to generate the analysis.

General findings and issues
The transect maps of 17 villages showed similar characteristics.For example, the transect map of Nhop village (Figure 2) shows typical land uses associated with certain topographical, soil, water and erosion features.
The highest point of village transect maps was usually rocky mountains with natural forests, descending along steepto-moderate slopes with mono-cultivation of staple crops through comparatively flat areas with homegardens to paddy rice on the    Table 1).

The results of individual interviews
showed that agriculture was the main income source for over 95% of respondents, with the main cash crops being maize (82%), and paddy rice, coffee or tea (18%).
Landholdings in the areas averaged 1 ha/ household.The most important factors for the farmers in deciding which trees and/or crops to plant or which farming system to apply were the products' market availability, land and capital resources.According to farmers, the main factors in maintaining good yields were, first, soil erosion (mentioned by 85% interviewed), followed by water availability (65%) and soil fertility (52%).Only a few farmers mentioned inputs as the main factor for good yields.This clearly indicated that soil erosion and related issues -declining fertility and water availability -were the key problems in the region.

Economic analysis:
The profitability of each cropping system identified during the FGD in the 17 study villages is summarized in the last column of Table 2.The column also includes profit data of some agroforestry systems in the study areas, according to farmers' recollections for the previous year.
The detailed profit data of all dominant systems in each study village are provided in annex 2 Table A2.2.

Farming systems in three agroecological zones
Since cropping patterns and, hence, farming practices varied significantly between elevations, results are reported according to elevation and agro-ecological zone.
There were only slight variations within the range of elevations and from one location to another.Table 2 summarizes all dominant farming systems by elevation.

Farming systems below 400 masl
The dominant farming system at this elevation was monocultural hybrid maize with two harvests per year on sloping land.The average yield ranged 3-5 t grain/ha/crop.
The second most common crop was mono-  Perennial cash crops were dominated by bush tea in Yen Bai and coffee (Coffea arabica) in Son La and Dien Bien.Coffee plantations had expanded over the last few years as monoculture or, infrequently, in association with fruit or timber trees.Farmers preferred coffee to other traditional food crops because of its higher economic return.
However, only wealthier farmers could afford the investment to establish coffee plantations.Timber tree species -for example, mahogany (Chukrasia tabularis A. Juss) and canarium (Canarium nigrum) -or fruit trees were planted in limited numbers as shade in coffee plantations.Fruit tree species in homegardens -for example, longan, mango, plum and custard apple (Annona squamosa L.) -were grown scattered through the gardens and as shade trees for coffee.
Agroforestry systems: Coffee-based agroforestry systems were the most encountered intercropping system at this elevation.
Coffee was intercropped with fruit trees or with timber trees for shade (Table 3).Some systems with fruit trees and annual crops were also observed on a small scale.

High elevations: above 800 masl
The farming systems at high elevations were less diverse in terms of species than at lower elevations, however, Agroforestry systems: The types of agroforestry systems at this elevation were diverse.
The most common systems were coffee with shade trees, shan tea (Camellia sinensis O. Ktze) with short-term crops and amomum (Amomum sp.) or cardamom (Amomum aromaticum Roxb.) under canopy of natural or planted forest with son tra and pines.The systems differed from region to region, for example, in Son La the most common system was coffee, in Yen Bai shan tea, and in Dien Bien amomum or cardamom under natural or planted forest (Table 3).provides nuts and, eventually, high-quality timber.Some farmers planted Canarium as a shade tree for coffee.Macadamia was a newly introduced species that grew well in the Northwest and produced good yields (4 kg/tree in the 8 th year) (Mai Son district) and had potential for export to international markets.
At high elevations, farmers preferred temperate timber and fruit/nut species, including son tra, peach or shan tea.Walnut had shown good growth in Thuan Chau above 1100 masl, and could be a promising option.

General findings of the agroforestry adoption survey
The results of the study showed that farmers practised agroforestry for a number of reasons.The primary reason for 79% of farmers (N=210, multiple choice) was higher income and more effective land use; 28% practised agroforestry in order to provide shade for understory crops.Only 2% of respondents practised agroforestry for soil and water conservation (although farmers were aware of this benefit of tree cultivation) and 4% for more efficient fertilizer use.Thus, the majority of farmers primarily integrated trees on their farms to enhance income: 86% of respondents reported that their farming systems generated more income from the combined yields of tree and associated annual crops.Only a few respondents were not sure about the economic benefits of trees because they did not use the tree products or the benefits of trees were not clearly shown.
A few respondents reported negative effects because of tree and crop competition.Most farmers learned agroforestry techniques Observations revealed that spatial arrangements of the agroforestry systems were not optimal.In some cases, tree spacing was too narrow or too wide or the distance between trees and crops too close, increasing competition for light, water and nutrients.Issues also arose in species' selection and site matching.Tropical fruit tree species -for example, jackfruit (Artocarpus heterophyllus), mango, longan, lychee (Litchi chinensis), rambutan (Nephelium lappaceum)were planted at elevations above 800 masl outside of their recommended range.Temperate fruit tree species are advisable for these altitudes.In terms of system incomes, respondents reported that agroforestry systems brought higher incomes compared to monocultural practices.The systems with coffee and longan and coffee and canarium were good examples of the higher economic benefits of agroforestry (Table 2).This has been widely reported (Current and Scherr 1995;Craswell et al. 1997;Pye-Smith et al. 2008).

Discussion
The economic return of a farming system is one important criterion for farmers, especially poorer farmers, to adopt and maintain any farming technology (Scherr 1995;Dwivedi et al. 2007).The results of the study showed that the majority of farmers primarily integrated trees on their farms to enhance income (79%).Similar results have also been reported by Bugayong and Carandang (2003) in the Philippines, where agroforestry had been promoted as a suitable land-use option.The long-term adoption of agroforestry in the northwestern region of Vietnam will depend on the economic benefits.
The study results also indicated that agroforestry has high adoption potential in the region because most of the respondents (86%) reported greater income was associated with mixed tree and annual cropping compared to monocultures alone.This percentage could be higher if tree products were  2006, 2009;Dida et al, 2013;FAO 2014).
Homegarden enrichment with fruit and multipurpose tree species and vegetables diversifies household food supplies and generates additional income.Building on the survey results, and including farmer's species preferences, the agroforestry systems in

Conclusion
Given the prevailing hilly landscape in the region with predominantly monocultural the area.Within this region, the primary benefits of agroforestry would be to address chronic erosion (particularly on slopes above 25 o ), which presents a significant long-term threat to local livelihoods' systems.Recent trends away from swidden agriculture towards cash cropping have reduced the ecological complexity of local agro-ecosystems (Pham et al. 2015).The introduction of more complex agroforestry systems can increase the biodiversity of these systems and broaden the range of products produced; both of which will build resilience in local farming systems.The initial results from agroforestry trials initiated in 2012 by the Agroforestry for Livelihoods of Smallholder Farmers in north-western Vietnam (AFLI) project show that agroforestry systems with grass strips planted on contours significantly reduce soil erosion (by up to 74% during the 2016 rainy season) (La et al. 2016).Data on the economic returns are not yet available as the associated trees have not reached productive ages.The paper presents a combined result of two interrelated surveys: 1) a farming system diagnosis (FSD) that focused on identifying the dominant farming systems in the region, assessing the strengths, weaknesses and resilience levels of each system; and 2) an agroforestry adoption survey centring specifically on the agroforestry practices adopted by farmers in the region and exploring the potential for their wider adoption.The studies sought to answer the following questions.i) What are the dominant farming systems in the study sites; what are the strength and weakness of each system; and to what extent are they resilient within the smallholder livelihoods' systems?ii) What are the agroforestry practices adopted by farmers and what are the reasons for their adoption?

Figure 1 .
Figure 1.Map of the study location

Figure 2 .
Figure 2. Sample transect from a typical upland village, Nhop, in Thuan Chau district (1150 m) was becoming thinner over the years owing to intensive and continuous monocultural cultivation.This had resulted in significant loss of soil fertility and declining yields.Farming systems: Systems were characterized by fragmented fields on steep slopes that were remote and difficult to access.Mono-cropping of staple crops, especially maize or cassava, on steep slopes was common and intensive at all altitudes (Annex 2, TableA2.1),although above 800 masl these problems were less severe because the agricultural systems were mostly for subsistence.Resilience: The common mono-cropping systems with staple crops, especially maize, were characterized by low economic returns; high levels of vulnerability to pests, diseases, and extreme weather events; market price fluctuations; and exposure of the soil to erosion and degradation.Agroforestry: Homegardens with mixed systems were very common across the study area.However, away from the homesteads there was limited evidence of agroforestry systems.The percentage of households practising agroforestry was low, except in Long and Nhop villages where coffee-based systems were common and Ta Xua village where farmers planted shortrotation crops in shan tea plantations (see

(
monocultural staple crops (maize, cassava and upland rice) on sloping land continued to be a common practice.The crops grown by the H'mong people were mainly for subsistence.Compared to farmers at lower elevations, H'mong farmers primarily used local varieties of upland rice, maize and cassava with limited inputs, resulting in lower yields.The average maize yield varied 1.5-3 t/ha/year, cassava 10-15 t/ha/ year and upland rice mostly stable at 1 t/ha.Crop yields were said to be declining owing to germplasm degeneration.Some crop species were traditionally intercropped or planted in rotation during the rainy season.Paddy rice was mainly planted on terraces.Common tree species were pine (Pinus sp.), Vernicia montana, manglietia and Fokienia hodginsii.In homegardens, common fruit trees were pear (Pyrus sp.), plum Prunus salicina), peach (Prunus persica), persimmon (Diospyros decandra Lour) and son tra (Docynia indica).Some households planted son tra for income generation with the support of government reforestation projects.

(
USD 900-2300 ha/year), especially the systems with coffee planted in association with canarium and longan.The price of coffee fluctuated in the years of calculation, however, the price of canarium and longan remained stable keeping the system's profit high (USD 3000 and 10000 ha/year respectively)3.4.Tree-planting practices and preferred tree speciesMore than two-thirds (71%) of the interviewed farmers had trees on their farms for domestic use and income generation.Trees were planted mostly in homegardens or scattered around the farms, at the top or bottom of hills, or along the contours of slopes.The main challenges with managing trees that were mentioned by farmers were limited land resources (36%), limited options for species with high market demand (38%), lack of investment capital (17%), and pests and diseases (17%).Few farmers were familiar with the term agroforestry.Following a detailed explanation by the research team, 78% of the respondents reported they knew about the practice but few (10%) practised any form of agroforestry on their farms.All the interviewed farmers were well aware of the issues related to monocultural staple crops on slopes and were willing to plant more trees on their farms to reverse low soil fertility, increase soil protection and benefit from the long-term stability of tree-based systems.However, they were concerned about the availability of species of high economic benefit and the availability of markets.Species that the farmers preferred to plant on their farms are shown in Table 4. Below 400 masl, farmers preferred eucalyptus and acacia because of their suitability to the natural conditions and strong local demand for timber.Of the fruit tree species, farmers preferred late-fruiting longan and orange for their marketability.At 400-800 masl, farmers preferred eucalyptus and Canarium.Canarium is a multipurpose species with long rotation that Volume 5 Issue 1 June 2017 [ ]

4. 1 .
Resilience, economic benefits and the adoption potential of agroforestry systems Overall, the most vulnerable areas in all three agro-ecological zones were the monocultural staple crops on steep slopes, particularly, where farmers applied large amounts of fertilizer to maize (at 400-800masl).These practices were strongly associated with soil erosion, soil degradation, water shortages, pests and diseases, loss of biodiversity and other environmental consequences.Other research has reported similar negative impacts on soil quality and soil loss (Ijaz et al. 2007; Nguyen 2007; Eludoyin and Wokocha 2011) and biodiversity and livelihoods' resilience (Horrigan et al. 2002) when cultivating steep slopes with annual crops.Cultiva-tion on steep slopes is prone to drought and flooding (Bennett et al. 2012).Tea plantations, either with or without shade trees, were more resilient than monocultural staple crops.Integrating trees for shade over tea brought more economic benefits, diversified income, reduced soil erosion and increased biodiversity.Other tree-based systems -for example, fruit or timber trees with short-term crops or coffee with shade trees -provided the soil with more groundcover during the year, thus, reducing erosion and increasing carbon stock, water-holding capacity of the soil and biodiversity.Similar results have also been reported byTang et al. (2002) andSchroth et al. (2004).Farms with trees in the study area had a shorter recovery time after natural disasters, except for cold spells, demonstrating economic and environmental resilience, as reported bySimelton et   al. (2015).With regards to market vulnerability, the systems with trees in association with other short-term crops and/or bushes diversified incomes of households, thus, reducing market risks.As informed by farmers, the prices of maize and cassava and even coffee had fluctuated significantly in the last few years, affecting their livelihoods.Tree products could compensate for the losses caused by price fluctuations, reducing vulnerability and risk.
used for income generation, which would require more attention in species' selection, minimization of tree and annual crop competition through a sound system design, and improvement of farmers' marketing practices.Additional technical support in tree species' selection and system design could help farmers increase the benefits from their tree and crop systems.Most farmers (95%) learned agroforestry techniques from their neighbours or by themselves.This demonstrated that farmerto-farmer exchange was an effective way to disseminate new farming technologies.Similar results were also reported byKhaila et   al. (2015)  andMartini et al. (2016).The high percentage of farmers adopting agroforestry by themselves also indicated that they were unlikely to be fully aware of agroforestry options and would benefit from training.4.2.Potential agroforestry options for Northwestern VietnamGiven the issues related to monocultures of short-term crops on slopes, the advantages of tree-based system as discussed above, and that farmers are interested in tree-farming, agroforestry should be a priority land-use option and have great development potential in the northwestern region of Vietnam.Efforts in promoting agroforestry should focus, first, on improving existing agroforestry systems and gradually introducing new technologies and systems through a participatory approach.The improvement of the existing systems, as discussed above, could consist of species' selection and combinations, spatial arrangements of the components in the systems and system management.Moreover, agroforestry is a complex farming system that requires often higher and longer-term investment.The benefits from agroforestry are usually reaped in the middle and long terms, hence, given the poverty of the farmers in the region, promotion of agroforestry should be accompanied by strong government support.Simelton et al. (2015) and Buttoud (FAO 2013) came to the same conclusion.Highquality grafted seedlings should be used to generate income sooner for farmers; grafted seedlings of canarium, longan and others commence fruit production in the third year or earlier compared to the sixth or seventh year when grown from seeds.Many of the species mentioned by respondents have been identified as priorities for Vietnam (Gunasena and Roshetko 2000).Further research on those species and related agroforestry systems is appropriate.Agroforestry options for the region should be based on local contexts, including socio-economics, biophysics and market conditions (Coe et al. 2014).Farmers should have the option to adopt the whole, or part, of the recommended systems.Given the low income levels of farmers, agroforestry systems that combined trees and annual crops providing short-, middle-and long-term income have a high potential for adoption.Boundary planting and small woodlots in the agricultural landscape are also additional options for all elevations, which would provide additional income for farmers and in-crease landscape resilience (Manning et al. annual crops that lead to severe soil erosion, land degradation and low incomes, agroforestry should be the most suitable land-use option and thus has great development potential.Strengthening farmers' knowledge and understanding of agroforestry is an important priority.That could be achieved while improving existing agroforestry systems and gradually introducing new technologies and systems through a participatory research and technology development approach.The improvement of the existing systems should include species' selection, access to quality germplasm, system design and training in system management.Agroforestry requires often higher and longerterm investment than annual crops but the farmers in the region have low income levels restricting their ability to invest.Promotion of agroforestry should therefore be accompanied by strong support from government.A clear plan to promote agroforestry should be developed by the Ministry of Agriculture and Rural Development in conjunction with agroforestry research and development organizations.Agroforestry options for the region should be based on local contexts, including socio-economics, biophysics and market conditions.Agroforestry systems that combined trees and annual crops that provided short-, middle-and long-term income have high potential for adoption.Boundary planting and small woodlots in agricultural landscapes are also additional options for all elevations that would provide additional income for farmers and increase landscape resilience.Homegarden enrichment with fruit and multipurpose tree species and veg-The farmer-to-farmer extension approach in Malawi: a survey of lead farmers.Working Paper no. 189.World Agroforestry Centre, Nairobi, Kenya.http://dx.doi.org/10.5716/WP14200.PDF.Khanh Tinh Cong (2012) Project report on "Điều tra, đánh giá thực trạng đời sống đồng bào dân tộc thiểu số cư trú ở vùng cao núi đá, nơi thường xuyên xảy ra thiên tai (Assessment of current living conditions of ethnic minorities in the mountainous areas where natural disasters happen frequently)., ethnic policy department, Ethnic Committee of Vietnam.Assess on 5.05.2015, h t t p : / / a m b n .v n / p r o d u c t / 1 3 3 3 8 / di%E1%BB%81u-tra-danh-gia-thuctrang-doi-song-d%C3%B4ng-baodan-toc-thieu-so-nam-2012.htmlLa N, Do VH, Pham HT, Agustin M, Do TL, Hoang TL et al. (2016) Participatory Farmer Trials Results.Technical report no.19.Agroforestry for Livelihood of Smallholder farmers in Northwest Viet Nam project.World Agroforestry Centre, Hanoi, Viet Nam.pp.35-38 Lasco DR, Delfino RP, Espaldon MLO (2014) Agroforestry systems: helping smallholders adapt to climate risks while mitigating climate change.WIREs: Climate Change 5(6): 825-833.DOI: 10.1002/wcc.301.Long Vu (2009) Về quyền sử dụng đất làm nương rẫy của đồng bào dân tộc thiểu số ở miền núi.Land-use rights for upland farming of mountainous ethnic groups in Vietnam.Vietnamese Academy of Forest Sciences, Hanoi, Viet Nam.http://vafs.gov.vn/vn/2009/03/ve-quyen-su-dung-dat-lam-nuongray-cua-dong-bao-dan-toc-thieu-so-o-Climate Change 4-1: 131.DOI:10.4172/2157-7617.1000131.Nair PKR (1993) An introduction to agroforestry.Springer Science & Business Media

Table 1 .
List of study villages including percentage of households, adopted agroforestry cultural, high-yield cassava (Manihot esculenta Craz.), which was cultivated where soil fertility had declined and could no longer support maize.The average yield ranged 18-22 t/ha/crop.In general, cassava cultivation brought low economic returns owing to low and unstable prices (see Table2).In areas with available irrigation, monocultural rice (Oryza sativa) was planted with two crops a year.The average yield ranged 3-5 t/ha/crop, depending on local conditions.Bush tea (Camellia sinensis) was the leading cash crop.Fast-growing timber species, such as acacia (Acacia mangium, Acacia auriculiformis or acacia hybrid), eucalyptus hybrid (Eucalyptus urophylla x Eucalyptus camaldulensis), Manglietia conifera and Styrax tonkinensis were mostly planted on heavily-degraded soils unsuitable for annual crops.agroforestrysystems were bush tea with timber trees for shade (manglietia, acacia, melia (Melia azedarach), eucalyptus hybrid); fruit trees with annual crops -for

Table 2 .
Farming systems in the study villages, by elevation including system profit

Table 3 :
Agroforestry systems at three ranges of elevation in the study sites and corresponding number of households practising the systems

Table 4 .
Species farmers preferred to plant, by elevation range

Table 5 :
Potential agroforestry systems for Northwestern Vietnam