Agroforestry Responses to 20 Years of Agricultural Expansion in the Central Highlands of Vietnam

Abstract

Vietnam’s Central Highlands, under the ‘boom’ of increased immigrant populations, forest loss, and agriculture expansion, have been facing unbalanced and unsustainable development. This study determines the existence of agroforestry in the region and how farmers responded to the expansion of agriculture over decades by secondary and survey data from 248 households of 20 communes in five provinces. After decades of mono-agriculture development, farmers faced a range of problems and consequently began applying agroforestry practices. About 65% of the surveyed households applied agroforestry and multi/inter-crop systems, but almost all the households were small farms. Agroforestry at both the farm and landscape levels existed and featured diversity. Plant and animal diversity in the agroforestry systems of the households was rich. We recommend that upgrading the agroforestry level can likely cope with the requirements of market-based production and also provide basic benefits. Thus, to help conserve the forests, improve the farmers’ livelihoods, and protect the ecosystem in the Central Highlands and the whole of Vietnam, a National Agroforestry Policy and technical guides for agroforestry at both the farm and landscape levels should be launched together to approach the market-based productions of cash crops and agricultural circular economic development in Vietnam and tropical regions.

1. Introduction

Vietnam’s Central Highlands, known as a wealthy and unique area, have become a sensitive and unstable region over the past 20 years. Even though it has always been considered as one of the most strategically important economic and political regions by the government, and has become a major agricultural commodity production region of the nation by producing key national products high in market demand and bringing in multi-billion-USD export revenue each year, a lot of issues have developed during this time. First, natural resources, including minerals, soil, water, and forests, have been exploited to the point of exhaustion. Natural forest destruction and devastation, both legally and illegally, have been very serious problems during this long period. For example, in 2015, competent agencies uncovered 6034 cases of violations of regulations on forest protection and development, up 463 cases from the figure recorded in 2014 (equivalent to 8.3% of cases) [1]. The government closed the forests in 2016, but forest devastation still happens even inside the state-owned organizations [2]. Second, the rapid increase of land conversion for agriculture, especially the rise of cash crops, has become one of the major reasons for forest loss. Agricultural economic development in the region has resulted in trade-offs among economic, environmental, and social objectives in the process of sustainable rural development. For instance, robusta coffee in the Central Highlands of Vietnam has become one of the most intensive and concentrated areas of coffee production in the world; however, it is cultivated by smallholders as an un-shaded and clean-weeded mono crop [3]. Mono cultivation always faces a lot of risks such as natural disasters, diseases, market fluctuations, and a high dependence on outside inputs for the production systems. Although agricultural production brings billions of dollars to Vietnam, the Central Highlands are still one of the poorest regions. In 2018, the average rate of poor households in the Central Highlands was 11.4% (excepting Lam Dong Province – only 2.85%), while the average rate of poor households in all of Vietnam was 5.3% [4]. In addition, the region has faced increasingly more frequent serious droughts, which are the result of climate change and which strongly impact both agriculture and the livelihood of farmers, especially ethnic minorities. Climate change, however, has been poorly understood by farmers and other rural stakeholders [3]. The population has boomed, particularly uncontrolled immigration which was the strongest during the decades when high coffee prices stimulated migration from Central and North Vietnam to the Central Highlands [5], and it is continuous growth that has become a social issue. Last, many types of conflicts have happened and continue to happen. Since 1990, many environmental problems have occurred that can be grouped into 16 types of conflicts that are particularly linked to natural resource use and management [6].

Therefore, under the ‘boom’ of immigration, forest loss, and agriculture expansion, the Central Highlands of Vietnam face unbalanced and unsustainable development. The explosion of mono-cultivated cash crops focusing on rubber, coffee, and black pepper that farmers have grown freely without the control of the government has not served to improve the development of the region wholly. After decades of growing these cash crops driven by market demands, most of the farmers have failed because they caught the tail of the market signals. Due to the fact that state planning and the markets for cash crops were broken, farmers in the Central Highlands were also bankrupt [7]. So, how can farmers be aware of and how can they respond to these problems?

Agroforestry is a possible solution and an adaptable choice for small farmers in the Central Highlands, which will be discussed further in the research results. Thus, this paper seeks to determine the current status of agroforestry as applied by households to understand whether agroforestry has responded to the expansion of agriculture in the Central Highlands over the past few decades and then discusses further recommendations for sustainable development at both the local and regional scales.

2. Study sites and methods

2.1 Study location

Vietnam’s Central Highlands include five provinces, namely Gia Lai, Kon Tum, Dak Lak, Dak Nong, and Lam Dong. Figure 1 shows the study sites distributed over the ecosystem zones of the region (e.g., Ngoc Linh Mountains; Sa Thay lower mountains; Play Ku and Con Ha Nung Highlands; An Khe lower mountains; Cheo Reo, Phu Bon, and Ea Sup Semi-plateau; Buon Me Thuat Highland; Man Drak mountains; Dak Nong, Dak Min Highlands; Chu Ang Sin mountains – Da Lat Armenian Highland; and Di Linh, Bao Loc Highlands [8]).

Figure 1.

Study sites in the central highlands, Vietnam (Source: Map based on information from the Department of Survey, mapping, and geographic information Vietnam).

2.2 Data and research methods

The series of major secondary data from 1999 to 2018 were collected from official legal documents. Of which, the data of the forest areas and coverages were from the Ministry of Agriculture and Rural Development (MARD), the data on the land areas were from the Ministry of Natural Resources and Environment (MNRE), and the others were from the Statistical Yearbooks of Vietnam. In addition, landsat images from Google Earth also were used in this research.

Primary data were collected from direct surveys given to 248 households in 89 villages of 20 communes belonging to ten districts of five provinces. The surveys were conducted at the end of 2018 and early 2019. Surveyed households were selected by the snowball sampling method. This approach involves two stages: (a) the identification of a sample of respondents with characteristic x at the zero-stage (s₀); and (b) the solicitation of referrals to other potentially eligible respondents believed to have characteristic x at snowball stages s₁ through s_k [9, 10, 11]. Applying this approach, the research conducted the zero-stage (s₀) at each of the 20 communes in the study locations; therefore, the number of villages and households surveyed was not equivalent.

Data were recorded, classified, and analyzed in Microsoft Excel (version 2010) and the R statistical software (version 3.5 in 2019) with the functions described by ggplots and Kruskal–Wallis test [12]. Of which, the cultivation systems were divided into five types, namely: AF1 – agroforestry types including fruit, industrial, and wood trees; AF2 – agroforestry types including fruit and wood trees/industrial and wood trees; AF-others – other agroforestry types (different from AF1 and AF2); Agr – mono cultivation (only fruit tree or only industrial tree); and Agr-multi – multi-cultivation (mixed or intercroped fruits or/and industrial trees). The ethnic groups were divided into three major groups, namely: I – indigenous minority groups; K – Kinh group; and Others – other minority groups (immigrant minorities).

3. Results and discussion

3.1 A glance of 20 years of land use changes for agriculture expansion in the central highlands

Under the impacts of the economic development policies of the Vietnamese Government, agriculture activities in the Central Highlands were rapidly extended during the period from 1999 to 2018. Agriculture extensions are connected closely with forest degradation. The data in Figure 2 present the changes of agricultural land areas, forest areas, and cover, and of the population over the time period. It can be seen that the increases in agriculture land (particularly perennial crop areas) and population are consistently together with the decreases in forest areas and cover, especially the decrease of natural forests. Specifically, in 1999, the total forest land was 2.99 million ha (including 2.93 million ha of natural forests) with a forest coverage of 56.8%, while agricultural land was 1.23 million ha (including 0.65 million ha of perennial crop areas); and in 2018, the total forest land was 2.56 million ha (a reduction of 14.0% compared to 1999), and the natural forest area was 2.21 million ha (a reduction of 24.7% compared to 1999) with the forest coverage of 46.0%, while agricultural land was 2.43 million ha (an increase of 96.8% compared to 1999) (of which, perennial crop areas were 1.42 million ha (an increase of 118.1% compared to 1999)). Additionally, hilly bare land was also considered to be linked to forest loss and reduced over the period from 0.88 million ha in 1999 to 0.18 million ha in 2018 (a reduction of 79.7%). Plantations slightly increased from 0.06 million ha in 1999 to 0.35 million ha in 2018 (an increase of 19.0%) but were not enough to replace the loss of natural forest areas in the region.

Figure 2.

Land areas, forest coverage, and population changes in the central highlands from 1999 to 2018. (Sources: Data from the Ministry of Agriculture and Rural Development (MARD), Ministry of Natural Resources and Environment (MNRE), and statistical yearbooks of Vietnam).

Before 2005, the major driver of land areas, forest coverage, and population changes in the Central Highlands was the economic development policies which targeted the building of new economic zones (NEZs) [5] with many state-owned agriculture enterprises (SAEs) and state-owned forest enterprises (SFEs). Until 2004, there was a total of 144 SAEs and SFEs that managed and used 2,012,721 ha of land area (of which, 59 SAEs managed and used 95,867 ha; and 85 SFEs managed and used 1,916,854 ha, including 1,046,047 ha of natural forests, 62,780 ha of plantation, and 214,938 ha of hilly bare land) [13]. Of note, before 2005, Dak Lak Province had 106 SAEs and SFEs that managed and used 1.65 million ha of land (occupying 86.13% of the total land of the province) with 20% of the total population (an average of 12.18 ha/person), while 80% of the population lived on the remaining land (an average of 0.53 ha/person) [14]. In regards to forestry, Dak Lak Province had 36 SFEs and 30 state-owned companies, which mainly harvested and sawed timber. They annually exploited about 400,000 m³ of round timber and planted 10,500 ha of plantation [15]. At that time, agricultural cultivation was plan-based, consisting of mono-crops, while agroforestry was only plantations (mostly pine or/and rubber plantations). The unequal average land area per person between SAEs and SFEs and local communities at that time became the reason for conflicts about landuse later on.

After 2005, the changes in landuse and forest coverage were driven by markets [16], particularly the markets for perennial crop products (e.g., rubber, coffee, black pepper, cashews, and fruits). The cash crop product markets strongly influenced every household in the Central Highlands, which drove farmers to change their landuse quickly and then be over-controlled by government policies; thus, all landuse plans of the government were bankrupt. Currently, coffee areas are 630,000 ha compared with 530,000 ha of government planning, and black pepper areas are 93,000 ha compared with 50,000 ha of government planning to 2025 [7]. In actuality, the government could not control the changes by any policies launched. Additionally, the population rose quickly, and within 20 years, the population increased by 44.58%, from 4.06 million persons in 1999 to 5.87 million persons in 2018 (Figure 2). This was combined with the expansion of agriculture (particularly perennial cash crops) following the market demands. Moreover, the market demands strongly pushed farmers to follow intensive farming practices with their cash crop cultivation, leading to higher inputs of chemical fertilizers, pesticides, and herbicides to increase crop yields. Farmers did not care about and apply agroforestry but more and more focused on mono cash crop fields. These things together resulted in negative pressures on natural forest areas. Consequently, a series of conflicts happened and continue to happen, including conflicts of landuse and management; of natural resource management (such as water and forests); of agricultural product prices; and of food safety [6].

Overall, state-owned stakeholders were leaders in destroying the natural forests with modern equipment following the assigned timber-harvesting target, but they planted fewer trees and followed unsustainable management practices for a long time. Then, both state-owned stakeholders and farmers exploited these forests, where almost all the big timber had been harvested, to turn them into farmland. Large areas of the remaining natural forests are poor and continue to lose both their area and wealth. A series of articles reported that many natural forest areas in Gia Lai Province have been illegally cut, and although they still look like a forest from the outside, they are empty inside [2].

Forest management by the government in the Central Highlands before 2015 was a big problem with the continuous loss of forests and decreases in forest coverage [17, 18], but after 2015, the forests have been better managed, and the government has kept the forest areas (about 2.56 million ha) and forest coverage (about 46.0%) stable (Figure 2). Being aware of the important roles of forests, MARD kicked off a big project aimed at achieving a forest coverage of 49.2% in the Central Highlands region by 2030 [19]. The better management by the government and the unstable changes of cash crop markets have influenced and changed the farmers’ awareness, leading to households shifting from mono cultivation to multi-cultivation and agroforestry, and thus, the agriculture landscape has been changing to a mosaic landscape like the agroforestry landscape as a higher approach level of ICRAF [20] and FAO [21].

3.2 The existence of agroforestry applied by the households in the central highlands

3.2.1 Agroforestry types of households

Based on the household survey data, five types of agricultural cultivation being applied by the farmers were determined, namely: (i) agroforestry mixed fruit plants, industrial plants, and woody plants (named AF1); (ii) agroforestry mixed only fruit plants or industrial plants with woody plants (named AF2); (iii) other agroforestry combinations (different from AF1 and AF2) (named AF-others); (iv) multi-cultivation mixed agricultural plants (without woody plants) (named Agr-multi); and (v) mono cultivation (only one plant type) (named Agr).

Figure 3a shows that the households applied AF1, AF2, AF-others, Agri-multi, and Agr at rates of 37.5, 20.6, 6.9, 30.6, and 4.4%, respectively. Of which, the households applying AF1 were mostly in the provinces Dak Nong, Dak Lak, Lam Dong, and Gia Lai, while households applying AF2 were mostly in the provinces Kon Tum and Gia Lai (Figure 3b). Analysis of the minority groups revealed that more Kinh people applied the AF1 and Agri-multi systems, while more indigenous minority people applied the AF2 system. Of note, the Agr system was applied only by the Kinh people (Figure 3c). Overall, at the farm level, farmers in the Central Highlands likely prefer agroforestry cultivation more than mono cash crop cultivation (about 65.0% of HHs applied agroforestry), which is a signal that farmers are approaching sustainable agriculture in the region.

Figure 3.

Agroforestry households (HHs) at the study sites (a – Percentage of HHs applying types of cultivations; b – Number of HHs by province; c – Number of HHs by ethnic group) (Source: Surveyed data in 2018–2019) Notes: AF1 – agroforestry types including fruit, industrial, and wood trees; AF2 – agroforestry types including fruit and wood trees/industrial and wood trees; AF-others – other agroforestry types (different from AF1 and AF2); Agr – mono cultivation (only fruit tree or only industrial tree); Agr-multi – multi-cultivation (mixed fruits or/and industrial trees); I – indigenous minority groups; K – Kinh people; others – other minority groups (immigrants).

3.2.2 Land areas of the households

Based on the household survey data, there are five types of land farmers use for cultivation, namely: (i) land for annual crops; (ii) perennial crops; (iii) woody crops; (iv) fish pond areas; and (v) forest land.

According to the survey data, the households had land for annual crops, perennial crops, woody crops, fish ponds, and forest land at rates of 19.4, 97.6, 6.0, 13.3, and 1.6%, respectively (Figure 4a). In this case, 100 could have all five types of land listed above. Almost all the households had perennial cropland, but only a few of the households had woody cropland (including plantations for wood and non-timber forest products) and forest land (mostly naturally protected forests, which households were allocated and paid to protect these forest areas). The households that had perennial cropland were distributed in all the provinces in the Central Highlands, and they preferred perennial cash crops over the others (Figure 3b). The minority ethnic households were twice as likely to use their land for perennial cash crops than annual crops (55/63 minority ethnic HHs had land for perennial cash crops, while only 26/63 HHs had land for annual crops). With the Kinh people, households were six times as likely to use their land for perennial cash crops than other crops (187/187 Kinh HHs had perennial cropland; 30/187 Kinh HHs had fish ponds; and 22/187 Kinh HHs had annual cropland). Households that had fish ponds were mostly Kinh people (30 Kinh HHs and 2 other minority HHs), while households that had woody land and forest land were mostly minority HHs (12 minority HHs and 3 Kinh HHs) (Figure 4c). Therefore, perennial cash crops appear to play an important role not only at the household level, but also at the local region level.

Figure 4.

Land areas of the households (HHs) at the study sites (a – Percentage of HHs with each type of land; b – Number of HHs by province; c – Number of HHs by minority groups with types of land; d – Average land area of the HHs by province; e – Average land area of the HHs by ethnic group; and f – Average land area of the HHs by types of cultivation) (Source: Surveyed data in 2018–2019) Notes: AF1 – agroforestry types including fruit, industrial, and wood trees; AF2 – agroforestry types including fruit and wood trees/industrial and wood trees; AF-others – other agroforestry types; Agr – mono cultivation; Agr-multi – multi-cultivation (mixed fruits or/and industrial trees). I – indigenous minority groups; K – Kinh people; others – other minority groups (immigrants).

The average total land areas belonging to the surveyed households were different among provinces. The average land of HHs in Dak Lak, Dak Nong, Gia Lai, Kon Tum, and Lam Dong Provinces were 1.28 (± 0.11) ha/HH, 2.84 (± 0.34) ha/HH, 1.98 (± 0.20) ha/HH, 2.39 (± 0.38) ha/HH, and 1.82 (± 0.16) ha/HH, respectively (Figure 4d). Data analysis showed that the average land of the households in Dak Nong Province was higher than households in Dak Lak and Kon Tum Provinces (X² = 24.259, p = 7.089e⁻⁵). However, there were no differences in the average land of HHs when accounting for the HH ethnic group. The average land of the indigenous minority HHs, Kinh HHs, and other minority HHs were 1.90 (± 0.19) ha/HH, 2.06 (± 0.14) ha/HH, and 2.91 (± 0.82) ha/HH, respectively (X² = 1.44, p = 0.486) (Figure 4e). Similarly, the average land areas of HHs applying each cultivation type were not different either. The HHs applying AF1, AF2, AF-others, Agr-multi, and Agr had land areas of 2.17 (± 0.20) ha/HH, 1.66 (± 0.20) ha/HH, 3.35 (± 0.75) ha/HH, 1.81 (± 0.12) ha/HH, and 2.49 (± 1.0) ha/HH, respectively (X² = 8.558, p = 0.073) (Figure 4f). Furthermore, Figure 4f also presents the households that preferred to cultivate the AF1, AF2, and Agri-multi systems.

Although there were no differences in average land area among ethnic HHs, the average land for perennial crops of ethnic HHs was significantly different. The average land for perennial crops of the indigenous minority HHs, Kinh HHs, and other minority HHs were 1.17 (± 0.15) ha/HH, 1.95 (± 0.13) ha/HH, and 2.60 (± 0.77) ha/HH, respectively (X² = 13.49, p = 0.001176) (Figure 4g). Recently, the indigenous minority HHs in the Central Highlands have had the lowest perennial crop land area on record. They had more land in the past, but many of them sold their good land to Kinh and immigrant minority people.

Overall, the average land area of the households in the Central Highlands is currently higher than before 2005, as noted in the above discussion, because the Vietnamese Government applied a policy to restructure the state-owned agriculture and forest stations and then reformed them to be state-owned enterprises. Furthermore, many of the state-owned enterprises must return a large section of land to local province committees, which is then allocated to the households. In addition, many farmers have invaded and shifted the forest lands into agricultural lands based on shifting cultivation over time.

3.2.3 Land areas for agroforestry of the households

The survey results also showed that the land areas the households used for agroforestry in the Central Highlands were significantly different among ethnic groups and types of cultivation. Land areas for agroforestry of the indigenous minority, Kinh, and other minority households were 1.24 (± 0.16) ha/HH, 1.65 (± 0.11) ha/HH, and 2.7 (± 0.83) ha/HH, respectively (Figure 5a) (X² = 8.915, p = 0.01159). The land areas of households cultivating AF1, AF2, AF-others, and Agr-multi were 1.79 (± 0.15) ha/HH, 1.13 (± 0.14) ha/HH, 2.36 (± 0.74) ha/HH, and 1.49 (± 0.10) ha/HH, respectively (Figure 5b) (X² = 13.93, p = 0.00752). On average, the agroforestry land areas of the households in the study sites were very small, but many farmers were utilizing agroforestry, which is a positive sign for the future. However, the small land areas of many of the households have been an obstacle to market-based agriculture production, especially in agroforestry. Only some of the farmers surveyed had over 3 ha per household of land for agroforestry (9 households seen as the 9 big dots in Figure 5b), which makes it possible for these HHs to approach the market requirement.

Figure 5.

Land areas for agroforestry of the households at the study sites (a – by ethnic group; b – by type of cultivation) (Source: Surveyed data in 2018–2019) Notes: AF1 – agroforestry types including fruit, industrial, and wood trees; AF2 – agroforestry types including fruit and wood trees/industrial and wood trees; AF-others – other agroforestry types (different from AF1 and AF2); Agr – mono cultivation (only fruit tree or only industrial tree); Agr-multi – multi-cultivation (mixed fruits or/and industrial trees). I – indigenous minority groups; K – Kinh people; others, other minority groups (immigrants).

3.3 Diversity of plants and animals in the households

3.3.1 Existence of plant and animal types in the households

The survey data showed that 96.0% of the households grew perennial industrial crops, 61.7% of the households grew fruit crops, 64.5% of the households grew woody and NTFP plants, 3.6% of the households grew medicinal plants, and 44.4% of the households raised animals and fish (Figure 6a).

Figure 6.

Households cultivating crops and animals in their farms at the study sites (a – HHs ratio; b – by province; c – by ethnic group; Kinh is a major Vietnamese group) (Source: Surveyed data in 2018–2019).

Surveyed households that planted perennial industrial crops were distributed in all five provinces of the Central Highlands (45/48 HHs of Gia Lai, 47/49 HHs of Kon Tum, 48/51 HHs of Dak Lak, 48/48 HHs of Dak Nong, and 50/51 HHs of Lam Dong), while three provinces had high rates of households that planted fruit crops (Lam Dong with 51/51 HHs, Dak Lak with 45/51 HHs, and Dak Nong with 42/48 HHs), and many households also planted woody and NTFP plants in all five provinces (37/48 HHs in Gia Lai, 32/49 HHs in Kon Tum, 28/51 HHs in Dak Lak, 37/48 HHs in Dak Nong, and 26/51 HHs in Lam Dong). A few households in the region raised animals and fish (29/48 HHs in Gia Lai, 29/49 HHs in Kon Tum, 17/51 HHs in Dak Lak, 19/48 HHs in Dak Nong, and 16/51 HHs in Lam Dong) (Figure 6b).

Analyzed by ethnic groups, the rates of indigenous minority households growing perennial industrial crops, fruit crops, woody and NTFP plants, and medicinal plants were 48/53 HHs, 14/53 HHs, 44/53 HHs, and 10/53 HHs, respectively, and 30/53 HHs raised animals and fish; while the rates of Kinh households growing perennial industrial crops, fruit crops, woody and NTFP plants, and medical plants were 182/187 HHs, 150/187 HHs, 110/187 HHs, and 5/187 HHs, respectively, and 75/187 HHs raised animals and fish (Figure 6c).

3.3.2 Plant and animal diversity in agroforestry systems of the households

The research determined that there were 69 types of plants the households applied in their agroforestry systems, of which coffee, black pepper, avocado, and durian were popular choices to grow (217/248 HHs grew coffee (equivalent to 87.5%), 113/248 HHs grew black pepper (45.6%), 125/248 HHs planted avocado (50.4%), and 108/248 HHs grew durian (43.5%)). In addition, a few households also planted other crops, such as L. glutinosa (56/248 HHs, equivalent to 22.6%), S. siamea (55/248 HHs, 22,2%), jack-fruit (28/248 HHs, 11,3%), and others (<10% HHs) (Figure 7).

Figure 7.

Types of plants in agroforestry systems of the households at the study sites (a – Numbers of HHs; b – Ratio of HHs with types of plants) (Source: Surveyed data in 2018–2019).

Figure 8 shows the quantity of plant types grown by the households in the Central Highlands. Of which, the average number of plant types of each household applying AF-others, AF1, AF2, Agr, and Agr-multi were 3.88 (± 0.61), 5.51 (± 0.17), 2.94 (± 0.14), 2.27 (± 0.19), and 4.36 (± 0.19), respectively (Figure 8a) (X² = 91.691, p = 2.2e⁻¹⁶). Therefore, AF1 had the most diversity in terms of types of plants used in agroforestry, allowing households to grow industrial, fruit, and wood plants together.

Figure 8.

Quantity of plants in agroforestry systems of households at the study sites (a – by types of cultivation; b – by provinces; c, d, and e – by ethnic groups) (Source: Surveyed data in 2018–2019) Notes: AF1 – agroforestry types including fruit, industrial, and wood trees; AF2 – agroforestry types including fruit and wood trees/industrial and wood trees; AF-others – other agroforestry types (different from AF1 and AF2); Agr – mono cultivation (only fruit tree or only industrial tree); Agr-multi – multi-cultivation (mixed fruits or/and industrial trees). I – indigenous minority groups; K – Kinh people; others – other minority groups (immigrants).

The plant diversity was different among households in the provinces. Households in Dak Lak, Dak Nong, Gia Lai, Kon Tum, and Lam Dong had 4.8 (± 0.26), 5.25 (± 0.25), 4.02 (± 0.31), 3.24 (± 0.2), and 4.52 (± 0.24) types of plants per household, respectively (Figure 8b) (X² = 39.018; p = 6.9e⁻⁸). Thus, households in Dak Nong Province had the most diversity of plants in their agroforestry systems.

By ethnic groups, the indigenous minority households had 3.57 (± 0.20) types of plants per household, Kinh households had 4.59 (± 0.14) types of plants per household, and other minority households had 4.62 (± 0.53) types of plants per household (Figure 8c) (X² = 11.956, p = 0.002534). Thus, Kinh households had the most diversity of plants in their cultivation systems.

Furthermore, the differences in plant diversity among households of ethnic groups were clearer when analyzing the types of perennial crops (including industrial, fruit, and forest plants). Figure 8d shows that the numbers of industrial and fruit plant types of the indigenous minority households, Kinh households, and other minority households were 2.17 (± 0.21), 3.45 (± 0.12), and 3.50 (± 0.46) types of plants per household, respectively (X² = 26.17; p = 2.07e⁻⁶). Figure 8e shows that the numbers of forest plant types of the indigenous minority households, Kinh households, and other minority households were 1.02 (± 0.02), 1.33 (± 0.06), and 1.0 (± 0.0) types of plants per household, respectively (X² = 13.837; p = 0.00098).

The research also determined that farmers in the Central Highlands have considered raising 11 major types of animals and fish, including cow, buffalo, goat, pig, chicken, swan, goose, duck, pigeon, sinkworm, and fish. Of which, the numbers of surveyed households raising chickens, cows, fish, and pigs were 56/248 HHs (equivalent to 22.6%), 45/248 HHs (18.1%), 33/248 HHs (13.3%), and 25/248 HHs (10.1%), respectively, while other animals were raised by fewer than 10% of the surveyed households (Figure 9a,c). Indigenous minority households preferred to raise cows (19/53 HHs, equivalent to 35.8%) and buffalo (7/53 HHs, equivalent to 13.2%), while Kinh households tended to raise chickens (51/187 HHs, equivalent to 28.0%), fish (30/187 HHs, equivalent to 16.0%), cows (22/187 HHs, equivalent to 11.8%), and pigs (17/187 HHs, equivalent to 9.1%) (Figure 9b). However, raising animals was not the main activity of most of the surveyed households, so the animal-raising farms were very small.

Figure 9.

Animals raised by the surveyed households at the study sites (a – number of households raising animals; b – number of ethnic households raising animals; c – Ratio of households raising animals) (Source: Surveyed data in 2018–2019) Notes: I – indigenous minority groups; K – Kinh people; others, other minority groups (immigrants).

3.4 Agroforestry perspective for future changes

Agroforestry can contribute to multi-benefits, but it is not a new cultivation or new technique with farmers in Vietnam. Agroforestry can improve soil health as it has the ability to enrich the soil’s organic carbon, improve soil nutrient levels, and enhance soil microbial dynamics [22, 23, 24]. It can help keep the high yields of cash crops with suitable shade and types of trees, such as rubber [25], coffee and cacao [26, 27, 28], other crops [29], and black pepper [30]. Agroforestry can stabilize household economics and help reduce poverty [31, 32, 33, 34] while also meeting the green market product value chain [34]. Using an agroforestry system can protect and conserve biodiversity [25, 35, 36], reduce the negative impacts of natural disasters, protect the environment [37, 38, 39], and serve other ecosystem services [35, 40, 41, 42, 43]. It can also help land adapt to and mitigate climate change [44, 45, 46] and help in sustainable agriculture development [47]. In addition, agroforestry can also conserve and use local or indigenous knowledge [28].

Overall, applying multiform cultivation systems involving agroforestry and a diversity of plants and animals will be a key to sustainable agriculture development for households in the Central Highlands. Importantly, the above results also indicate that farmers have been more and more aware of applying agroforestry practices by themselves with the high diversity of cultivation, plants, and animals, which are the keys to basic agroforestry [33].

Actually, there were dozens of agroforestry systems existing traditionally and modally in the Central Highlands, which were developed by farmers themselves or introduced by scientists [48]; however, there likely exist unclear obstacles to allowing agroforestry to be expanded. The total agroforestry area in Vietnam was estimated by ICRAF to be about 900,000 ha, and 10 million ha were identified as being suitable for agroforestry [49]. Moreover, further work by ICRAF continues to bring evidence to policy dialogs, including assessing the possible roles of agroforestry for Vietnam targets to international conventions, such as the Nationally Determined Contributions [45]. At the global level, many symposia have brought together agroforestry researchers, academicians, planners, policymakers, and administrators to discuss agroforestry research for sustainable development and to address future scenarios [46].

Traditionally, agroforestry was similar to circular agriculture and local organic and/or ecological agriculture at the household level by itself. However, under higher market demands, agroforestry in Vietnam must be upgraded to a higher level – a landscape agroforestry level, which ICRAF has been developing. Agroforestry in some places has already been moving to a landscape level (Figure 10). The landscape agroforestry level can not only cope with the requirement of market-based production, but can also provide basic roles and multi-benefits. Agroforestry can be a key aspect to develop an agricultural circular economy.

Figure 10.

An example of changing landscapes from mono agriculture expansion to agroforestry at one of the study sites (around Dak-tic’s hydropower reservoir located in Gia Nghia town, Dak Nong Province) (a) mono agricultural based landscape (Landsat image in 2006); (b) preparatory agroforestry based landscape (after Dak-tic’s hydropower reservoir was completed) (Landsat image in 2012); (c) agroforestry-based landscape (more diverse plots with trees) (Landsat image in 2018); (d) agroforestry landscape (zooming in from the yellow area in picture c) – Including trees and diverse plots of plants and cage fishery (Landsat image in 2019) (Source: Landsat images from Google earth free version (12^o00′14″N; 107^o38′24″ E)).

Therefore, to conserve forest areas and coverage, transform the lives of rural farming populations, protect ecosystems, and ensure food/fodder security through a sustainable basis in the Central Highlands and the whole of Vietnam, a National Agroforestry Policy will be a path-breaker as a legal instrument for making agroforestry a compulsory application. In addition, the technical guides for agroforestry at both the farm and landscape levels will also be launched together with the policy to approach the market-based production of cash crops.

4. Conclusion

Agriculture expansion in the Central Highlands of Vietnam during the period from 1999 to 2018 was closely connected with forest degradation. In 1999, the total forest land was 2.99 million ha, with forest coverage of 56.8%, while agricultural land was 1.23 million ha. In 2018, the total forest land was 2.56 million ha, with forest coverage of 46.0%, while agricultural land was 2.43 million ha.

The percentages of surveyed households applying AF1, AF2, AF-others, Agri-multi, and Agr were 37.5, 20.6, 6.9, 30.6, and 4.4%, respectively. The average land areas of HHs in Dak Lak, Dak Nong, Gia Lai, Kon Tum, and Lam Dong Provinces were 1.28 (± 0.11) ha/HH, 2.84 (± 0.34) ha/HH, 1.98 (± 0.20) ha/HH, 2.39 (± 0.38) ha/HH, and 1.82 (± 0.16) ha/HH), respectively. The average land areas of the indigenous minority HHs, Kinh HHs, and other minority HHs were 1.90 (± 0.19) ha/HH, 2.06 (± 0.14) ha/HH, and 2.91 (± 0.82) ha/HH, respectively. The average land areas of households applying AF1, AF2, AF-others, Agr-multi, and Agr were 2.17 (± 0.20) ha/HH, 1.66 (± 0.20) ha/HH, 3.35 (± 0.75) ha/HH, 1.81 (± 0.12) ha/HH, and 2.49 (± 1.0) ha/HH, respectively.

Agroforestry at both the farm and landscape levels existed in the region and featured diversity. There were 69 types of plants and 11 major types of animals and fish the households applied in their agroforestry systems, of which coffee, black pepper, avocado, and durian were popular choices to grow.

We recommended that a National Agroforestry Policy and technical guides for agroforestry at both the farm and landscape levels should be launched together to approach the market-based productions of cash crops, and circular agriculture development.

Acknowledgments

We would like to thank the households involved in the surveys and the staff of the Vietnam National University of Agriculture for their association with doing fieldwork.

Competing interests

The authors declare that they have no competing interests.