2.3.1. Capacity building training workshops

In September 2022, 34 farmers were selected in Chieng Yen in Van Ho district, Son La province, to participate in a technical training workshop on designing home gardens, agroforestry, organic farming, intercropping, pest and disease management and composting (Figures 2 and 3). Participants were also trained on how to use sensors and apply their smartphones to monitor and report agricultural indices in their own fields.

2.3.2. Selecting households participating in the experiment

A total of 18 households with existing home gardens voluntarily registered for the project (see Appendix A). The small sample size was intentionally chosen to enable close facilitation, participatory learning and in-depth observation of CSA practices under real household conditions. As an exploratory pilot, the study aimed to test feasibility and refine approaches rather than generate statistically representative findings.

Household selection was purposive, based on the following criteria:

• Willingness to participate and commit to the project throughout its duration.
• Availability of home garden space suitable for establishing experimental plots.
• Representation of diverse socioeconomic conditions and farming experiences.
• Interest in adopting CSA practices and serving as peer educators within the community.

Participating households (Figures 4 and 5) were required to follow project guidelines, attend all training sessions, and contribute labour for garden management and monitoring.

2.3.3. Co-design CSA experiment

Households participated in a focus group discussion (FGD) to identify priorities and propose changes to their home gardens, including the selection of suitable fruit trees, annual crops and management practices (Figure 6). The discussion allowed farmers to share experiences and constraints, while researchers and extension staff introduced feasible CSA options adapted to local conditions.

Each household then developed a garden layout plan, mapping the placement of crops, composting sites, and irrigation systems with technical guidance from experts. The process combined farmers’ traditional knowledge with scientific advice on intercropping, soil fertility, and biological pest management.

This participatory co-design ensured that the experimental models reflected each household’s needs, labour capacity, and land characteristics. It also strengthened local ownership and learning, laying the foundation for sustained CSA adoption in the community.

2.3.4. Experimental design and citizen science co-implementation

Each household implemented a suite of CSA interventions adapted to local conditions, including:

• Agroforestry and intercropping within home-garden systems.
• Nutrient and pest management through bio-treatments using Metarhizium anisopliae and effective micro-organisms (EM).
• Livestock waste management through vermiculture composting to enhance soil fertility.
• Water management using low-cost irrigation technologies.

To evaluate the effectiveness of these practices under actual farming conditions, each household established two plots within the same home garden, an experimental plot and a control plot. Conducting both treatments within the same garden ensured comparable conditions, as the soils were loamy with a slightly acidic pH (5.5–6.5), moderate organic matter content (2.1–2.8%) and uniform texture and drainage. Plot sizes ranged from 200–400 m², depending on household land availability.

The experimental plots received biocontrol treatments (M. anisopliae and EM), which were incorporated into the soil before planting and reapplied every 14 days, while control plots received no treatments. From November 2022 to January 2023, households tested these treatments on Kohlrabi (Brassica oleracea var. gongylodes) and from March 2023 to May 2023 on cucumber (Cucumis sativus), to assess their effectiveness in pest and disease management through a citizen-science monitoring approach.

The project adopted a citizen-science model, where farmers acted as co-researchers in monitoring and documenting changes in their own gardens. Citizen science refers to the active participation of non-scientists in collecting, analysing and interpreting data, enhancing both scientific rigour and community learning. Farmers recorded observations on crop growth, pest and disease incidence and yields using the ESRI Survey123 mobile application (Esri, n.d.). Pre-loaded digital questionnaires enabled participants to capture quantitative data, qualitative notes and photographs directly from the field, which were automatically uploaded to a central database. These data were later aggregated and analysed to generate timely, evidence-based feedback, allowing households and researchers to jointly refine CSA practices.

To support implementation, each household received equipment and inputs, including a pumping machine, a drip-irrigation system (Figure 7), vegetable seeds and fruit-tree seedlings such as jujube (Ziziphus mauritiana), persimmon (Diospyros kaki) and guava (Psidium guajava), and vermiculture starter kits for composting. All inputs were selected and designed through participatory co-design discussions with farmers.

2.5.1. Quantitative data analysis

Quantitative data obtained from the household questionnaires were entered into Microsoft Excel for cleaning and analysis. Descriptive statistical tools were employed to summarise the data, including frequency, percentage, mean and range to describe household characteristics, garden sizes and income levels.

2.5.2. Qualitative data analysis

Qualitative data derived from survey responses were analysed using thematic analysis following the six-phase framework proposed by Braun and Clarke (2006). This approach was selected to systematically identify, organise and interpret patterns of meaning across participants’ narratives related to home garden management and CSA practices.

Researchers first familiarised themselves with the data through repeated reading of transcripts and detailed field notes to gain an overall understanding of participants’ perceptions and experiences. The qualitative data were then transcribed, reviewed and coded, with initial codes assigned to recurring ideas, words and expressions relevant to the study objectives. These codes were discussed collaboratively by the research team to ensure consistency and reliability in interpretation.

The codes were then grouped into themes representing practices applied in experience. Themes were iteratively reviewed, refined and aligned with the overall research purpose to ensure that they captured all relevant insights and nuances within the qualitative data. Illustrative quotes from participants were included in the results section to highlight individual perspectives and provide contextual depth to each theme. This analytical process facilitated the identification of both shared experiences and variations among households, thereby strengthening the interpretation of farmers’ perceptions and behavioural changes following project interventions (Braun & Clarke, 2006).

3.3.1. Knowledge and capacity building

Prior to the project, most households lacked technical knowledge of home-garden design and management. Trees were often planted haphazardly, received minimal care and produced poor-quality fruit that had little or no market value. Through participatory training and hands-on guidance, farmers developed practical skills in agroforestry layout, pruning, soil improvement and intercropping design. They learned to combine perennial fruit trees, such as jujube (Ziziphus mauritiana), crispy persimmon (Diospyros kaki) and guava (Psidium guajava), with annual crops and legumes to improve both productivity and soil health. The survival rate of newly planted fruit trees reached 90%, with several already bearing fruit and expected to generate additional income in the following season (Figures 9 and 10).

The training also transformed farmers’ perceptions of land use. Many participants shifted from subsistence-oriented gardening to a more diversified, market-oriented approach. As one farmer noted, “In the past, we only grew rice and peanuts for sale, but now we use the land around our home to produce many different crops for the market” (P18). This shift reflects an emerging entrepreneurial mindset supported by improved technical capacity.

Through the project’s training activities, farmers enhanced their knowledge and skills in integrated pest and soil management. Participants learned to apply biological and ecological control methods, such as intercropping aromatic plants (garlic and onion) and rotating crops across seasons to disrupt pest cycles. “Since we started rotating crops and mixing with garlic and onions, the pests have reduced significantly,” shared one participant (P1).

Farmers also gained practical experience using effective micro-organisms (EM) and Metarhizium anisopliae to improve soil health and reduce disease incidence. These new practices not only promoted healthier crops but also reduced dependence on chemical inputs. “Previously, we were unaware of soil-borne diseases and did not treat the soil before planting. Since learning to apply lime and biological treatments, our crops are much healthier and require less care,” noted a farmer (P14).

Capacity building also extended to soil fertility management through composting and vermiculture. Before the project, farmers typically applied dried manure directly to crops, leading to odours, nutrient loss and disease. After training, they began using covered composting pits, mixing livestock manure with EM and green materials to accelerate decomposition and produce nutrient-rich organic fertiliser. “Composting now takes only 20–30 days and the product is lighter, cleaner and more fertile,” explained a participant (P11).

Innovative local adaptations emerged as well, such as composting golden snails collected from rice fields.

“Previously, our fertiliser source was limited to cattle manure. However, after training, I ventured into composting golden snails in the fields. This alternative has proven to be an exceptional fertiliser for our crops and vegetables, particularly when augmented with EM products, effectively neutralising any odours. I am delighted to have this supplementary organic fertiliser for our crops, addressing the challenge of organic fertiliser scarcity. I tested the liquid fertiliser on vegetable and rice seedlings and it worked wonders. The plants thrived exceptionally well, displaying remarkable resilience even during cold spells. Materials to make these fertilisers are also very cheap.” (P6, Figure 9)

Vermiculture training further enabled farmers to recycle waste and produce organic feed, thereby reducing livestock costs and earning additional income from worm biomass sales (Figure 10).

3.3.2. Income generation and market potential

Integrating CSA practices into home gardens initially led to improvements in household income and productivity. Diversity of crops in home gardens contributed an estimated 10% rise in annual income (USD 100–200) through vegetable sales (Figure 11). Likewise, new fruit trees such as jujube (Ziziphus mauritiana), persimmon (Diospyros kaki) and guava (Psidium guajava) are expected to yield up to USD 500 per household from the third year (∼20% of annual income).

Households also adopted off-season cultivation with short-cycle crops such as H68 sticky corn, which added around USD 200 per season and improved year-round land use (Figure 12). As one farmer noted, “Planting corn between fruit trees gives us extra income and keeps the garden productive even in the dry season.” (P12)

The vermiculture model diversified income sources and reduced costs. Earthworm biomass sales generated USD 100–200 annually, while feeding worms to poultry lowered feed expenses by about USD 300 per 100 chickens. Farmers observed tangible benefits: “The neighbours’ chickens often get sick, but ours stay healthy because they eat worms,” explained one participant (P9).

Although total financial gains remain modest due to the small pilot scale, two-thirds of participating households began selling surplus vegetables locally and to tourists. As one woman shared, “Before, we grew only for our family. Now, we sell to visitors and nearby markets.” (P15)

3.3.3. Improved food availability and dietary diversity

The integration of organic vegetables into home gardens significantly enhanced household food availability and dietary diversity (Table 4). Before the CSA intervention, most gardens were dominated by perennial fruit trees and staple crops, with only 3–4 edible species, primarily wild vegetables such as amaranth (Amaranthus tricolour) and wild gooseberry (Physalis angulata), along with a few root or fruit species like taro and pumpkin.

The introduction of diverse Brassicaceae (Cruciferae) vegetables and root crops such as carrot (Daucus carota) and white radish (Raphanus sativus) increased both crop diversity and nutrient availability. These crops are rich in vitamins and minerals, easy to grow and mature quickly, contributing directly to improved household nutrition and resilience to seasonal food shortages.

As one participant shared, “Since we started growing vegetables, our meals have become much more diverse. We always have an abundance of vegetables at home, with a wide variety available throughout the seasons. Previously, we relied heavily on foraging in the forest, but now we no longer need to do so. The vegetables and fruits from our garden serve as food for our livestock if not consumed, reducing the need to purchase animal feed or spend time foraging in the forest for pig and chicken feed. There was even a time when I fell while foraging. Therefore, having a vegetable garden around the house is very convenient, saving time spent in the forest and allowing for immediate use of the produce.” (P8)

Improved irrigation and organic fertilisation also allowed households to expand vegetable cultivation to nearby fields during the second cropping season, optimising moisture and nutrient use while reducing labour demands. These results highlight that integrating fast-growing, nutrient-rich vegetables into diversified home gardens is an effective strategy for strengthening household food security in upland communities (Figures 13 and 14).

3.3.4. Improved food availability and dietary diversity

Households also reveal enhanced practices learned and applied in pest management, with most farmers reporting significant changes when using biocontrol for pest and disease management. These practices have led to healthier crops, reduced pest infestations and minimised environmental impact.

In terms of pest and disease control, we have implemented intercropping techniques with aromatic plants such as garlic and onions, along with crop rotation across different seasons on the same plot. These strategies have significantly diminished the occurrence of specific pests(P1).

Previously, we were unaware of the prevalence of soil-borne diseases and therefore neglected to treat the soil with lime and biological fungicides before planting. Since adopting this approach, our crops have experienced improved health. Additionally, employing micro-organism products has rendered our crops less susceptible to cracking and black spots, consequently reducing the workload associated with crop care and pest management
(P14, P9)

The results from the co-experiment with households via ESRI Survey123 (Figure 15) on the effectiveness of the application of effective micro-organisms (EMUNIC) and the Metarhizium anisopliae in pest and disease control also showed that the kohlrabi plants in the experiment plots grew better and were less susceptible to insects and diseases (Table 5).

3.3.5. Labour and time efficiency

Farmers reported substantial time savings and improved water management following the adoption of automated sprinkler irrigation systems. These systems prevented crop dehydration and reduced the need for manual watering, decreasing total water use by approximately 30%. “Before, we had to carry buckets of water to the field every day; now irrigation happens automatically, saving us hours of work,” noted one participant (P11).

The complementary use of straw mulch further enhanced soil moisture retention, reduced irrigation frequency and effectively suppressed weed growth (Figure 16). Together, these water-saving practices lowered labour intensity and improved crop health.

Moreover, by producing sufficient vegetables at home, households reduced the need to forage in nearby forests for livestock feed and wild vegetables. As one farmer explained, “With vegetables grown at home, we no longer spend one or two days a week in the forest collecting plants for pigs and chickens.” (P15). This shift not only saved labour and time, especially for women, but also helped relieve pressure on surrounding forest ecosystems.

3.3.6. Knowledge sharing, social cooperation and digital empowerment

The introduction of digital tools and participatory learning activities led to shifts in how farmers shared knowledge, cooperated and supported one another in adopting CSA practices. Before the project, most households relied on individual experience or informal discussions during community meetings to solve farming problems. Through the project’s CSA digital platform and the ESRI Survey123 mobile application, farmers began to engage more actively in exchanging information, monitoring their own gardens and communicating directly with experts.

Co-monitoring the experiment and documenting progress and results (Figure 17) shaped farmers’ habits of regularly reporting changes in their fields, allowing for timely adjustments, assessments and the selection of best practices tailored to their site conditions.

Beyond technical learning, the digital platform fostered stronger social cooperation and peer learning within the community. Farmers used the platform and face-to-face meetings to exchange crops for rotation, discuss planting schedules and share surplus produce. As one farmer noted, “We now talk to each other more about what to plant and when. If we all grow the same vegetables at the same time, it is harder to sell. So now we plan together to grow different ones.” (P11).

This cooperation not only reduced post-harvest waste but also diversified local diets and improved marketing opportunities. Some farmers described becoming “closer” to their neighbours and more united in developing their home gardens. Although the project directly engaged 18 households with established gardens, some participants noted that neighbouring non-participant households had shown interest in adopting similar practices. Informal exchanges of seedlings, composting techniques and pest management methods indicate an emerging spill-over effect within the community, highlighting the project’s potential to foster broader collective learning and engagement.

Institutional support ensured sustainability beyond the project’s duration. The Van Ho Commune People’s Committee officially approved the platform and assigned the Centre for Agricultural Services the responsibility of maintaining it (Figure 18). This long-term arrangement guarantees that farmers will continue to access information, receive updates on pests, weather and markets and market their products online.

The integration of digital tools and collaborative learning fostered a stronger sense of community ownership, mutual trust and empowerment. Farmers no longer see themselves as passive beneficiaries but as active contributors to collective knowledge and local innovation. This social transformation, linking digital literacy, cooperation and participatory learning, represents one of the significant outcomes of the CSA project in Van Ho commune.

3.3.7. Challenges and barriers to adoption

Although the CSA pilot brought clear benefits, adoption among households remained uneven due to labour, behavioural and technological constraints.

Labour shortages were mentioned as the most common barrier, particularly during peak farming seasons when family members prioritised rice or maise production. Women, who often managed xhome gardens, found it difficult to balance household duties with new CSA tasks. As one participant noted, “When the rice season comes, I have no time to look after the vegetables.” (P6).

Traditional practices also slowed behavioural change. Some farmers were initially reluctant to replace chemical pesticides or raw manure with biological treatments, believing they were less effective. However, as one farmer shared, “At first, I did not think composting with micro-organisms would work, but after seeing the vegetables grow better, I trust it now.” (P14).

Digital literacy and connectivity issues further limited participation in citizen science monitoring. Older participants often relied on younger relatives to upload photos or data due to weak internet signals and unfamiliarity with the app. “The network is sometimes too weak to send pictures, so I wait until the evening,” explained one farmer (P3).