1 Assessing changes in the agricultural systems of selected sites in Cambodia, Laos and Vietnam: A Participatory GIS approach Hands and Minds connected to boost eco-efficiency in smallholder crop-livestock systems Louis Parker and Nguyen Thi Than April 2018 2 Acknowledgements We would like to thank Sabine Douxchamp (CIAT) for her valuable insights and guidance, and also Adrian Bollinger (CIAT) who facilitated the initial meetings with the partner organizations. We would like to thank the national coordinators and their respective organizations for their enthusiasm, perseverance and dedication, notably Dr Lyda Hok (Royal University of Agriculture, Cambodia), Dr Chau Thi Minh Long (WASI) and Dr Phonepaseuth Phengsavanh (The National Agriculture and Forestry Research Institute NAFRI). We would like to acknowledge the participation and inputs from Clément Bourgoin who provided integral support for the GIS analysis and also for the survey design on Commcare, and also Dharani Burra who worked tirelessly on the data storage, analysis and provided valuable inputs for the trainings. This work was carried out in the frame of the project “‘Hands and Minds connected to boost Eco- efficiency in Smallholder Systems” Funding was provided by the German Federal Ministry for Economic Cooperation and Development (BMZ). Citation Parker, L., Nguyen, T, T. (2018) Assessing changes in the agricultural systems of Cambodia, Laos and Vietnam through Participatory GIS. Technical Report. International Center for Tropical Agriculture. 3 Contents 1. Abstract ..................................................................................................................................... 5 2. Introduction ............................................................................................................................ 6 3. Objectives ................................................................................................................................. 6 4. Study sites................................................................................................................................. 7 5. Methodology and methods ................................................................................................... 15 5.1. Case study in Cambodia ................................................................................................ 22 5.1.1. Household information .......................................................................................... 22 5.1.2. Village – plot level crop data in Cambodia ........................................................... 25 5.2. Case study in Laos ........................................................................................................... 37 5.2.1. Household information .......................................................................................... 37 5.2.2. Village – plot level crop data in Laos ..................................................................... 39 5.3. Case study in Central Highland, Vietnam ..................................................................... 54 5.3.1. Household information .......................................................................................... 54 5.3.2. Village – plot level crop data in Vietnam .............................................................. 57 6. Discussions ............................................................................................................................. 69 6.1. Cambodia ........................................................................................................................ 69 6.1.1. Padal Village ........................................................................................................... 69 6.1.2. Prouk Village ........................................................................................................... 71 6.1.3. Luon Village ............................................................................................................. 72 6.2. Laos .................................................................................................................................. 73 6.2.1. Muon, Or Anh and Tar village ................................................................................ 73 6.2.2. Pa Khom village ...................................................................................................... 74 6.2.3. San Khing village .................................................................................................... 74 6.3. Central Highlands, Vietnam .......................................................................................... 75 6.4. Limitations ...................................................................................................................... 76 7. Conclusions and recommendations .................................................................................... 77 References ...................................................................................................................................... 79 4 Annexes 1 ....................................................................................................................................... 81 Annexes 2 ....................................................................................................................................... 86 Annexes 3 ....................................................................................................................................... 92 5 1. Abstract Background: The Mekong River delta is a region undergoing rapid socio-economic and environmental changes. As part of the BMZ funded project, ‘Hands and Minds connected to boost Eco-efficiency in Smallholder Systems’ we have conducted an assessment to identify the changes in cropping systems occurring in specific project sites in Vietnam (Central Highlands), Cambodia (Ratanakiri Province) and Laos (Xiangkhouang Province). We have implemented a participatory GIS methodology to compile spatially referenced information on the changes occurring at the plot level for the period 2007-2017 and recorded information on the climatic and biophysical stresses present. We have also collected information on the household characteristics in order to better understand the context in which the cropping system changes are occurring. Objective: i. To identify and describe the changes in the cropping systems from 2007 to 2017 occurring in the project sites ii. To collect descriptive data at the household and village level which can be used to better understand why the specific changes in cropping system are occurring and how these changes manifest themselves across the sites and surveyed farmers Method: We implemented the survey on the commcare application using hand held tablets and used GPS to map the field plots of the households (between 78-94 Households for each country). We used Geographic Information Systems (GIS) to store, analyse and present the data. Conclusions Overall we can see that the agricultural system in the Central Highlands, Vietnam, is largely orientated towards the market, with cash crops being produced since and notably before 2007. The changes in crop patterns is complex, but we can see an increase in the production of pepper and a reduction in coffee. Cambodia has seen a reduction in the forest and fallow plots with increases in the production of cash crops such as cassava. However, this is not uniform across all the villages, with Padal undergoing the most substantial increase in agricultural production, with forest and trees being replaced by cassava and cashew. In Laos we can see an increase in the production of maize, particularly in Pa Khom village. We also see an expansion in the area of the tea in Or Anh village, a cash crop largely grown for export. 6 2. Introduction The BMZ funded project entitled ‘Hands and Minds connected to boost Eco-efficiency in Smallholder Systems’ has been implemented by the international Center of Tropical Agriculture (CIAT) and local partners in order to better understand the changing nature of smallholder agriculture in Laos, Cambodia and Vietnam. In the following report we will focus on the component of the project that considers ‘how a shift from traditional semi-subsistence but often highly diverse and complex smallholder farming systems to increasingly specialised and intensified ones’ (project proposal) may be occurring in the project sites. In order to capture the changes in cropping systems we have adapted and implemented a Participatory GIS (PGIS) methodology (de Haan 2009) to capture information on the changes in cropping systems that have occurred from 2007 until 2017. As part of the study we also have captured detailed information on the households and also the village characteristics which we attempt to integrate within the findings to provide a more holistic picture of the transitions in cropping systems and the context in which these changes are occurring. 3. Objectives The objectives of this study are: i. To identify and describe the changes in the cropping systems from 2007 to 2017 occurring in the project sites ii. To collect descriptive data at the household and village level which can be used to better understand why the specific changes in cropping system are occurring and how these changes manifest themselves across the sites and surveyed farmers iii. To valorize and build on local knowledge for land use monitoring and planning iv. To generate local capacity to conduct applied research and guide land use planning In order to achieve these objectives we have implemented a PGIS methodology across the study sites. PGIS can be thought of as the integration between Participatory Learning and Action (PLA) methods with Geographic Information Systems (GIS). PGIS facilitates the representation of local people’s spatial knowledge through two- or three-dimensional maps and has been widely used amongst the geospatial community to capture local knowledge and identify specific trends (see Basupi et al 2016, Aynekulu et al 2006). The PGIS approach is useful in that it can produce map products that can facilitate decision-making processes, as well as support communication and community advocacy. PGIS practice is geared towards community empowerment through 7 tailored, demand-driven and user-friendly applications of these geospatial technologies. Good PGIS practice is flexible and adapts to different socio-cultural and biophysical environments. It often relies on the combination of ‘expert’ skills with local knowledge. Unlike traditional GIS applications, PGIS places control on access and use of culturally sensitive spatial data in the hands of those communities who generated it. This applies to both the data collection and ideally the development of maps and products which are useful to the communities. 4. Study sites The following section will outline the general socio economic context and the key agricultural commodities produced in each country. The study sites where the PGIS methodology was conducted will then be introduced, we will then document the number of households sampled in each of the respective villages and the local government partners responsible for conducting the survey: Cambodia: Cambodia is currently undergoing rapid economic, social and environmental changes. The Gross Domestic Product (GDP) of Cambodia tripled (UNDP 2017) between 2005 - 2015 and annual growth rates of 6.8% and 6.9% were recorded for 2017 and 2018 respectively (World Bank 2018a). The population has grown from 12 million in 2000 to roughly 16 million in 2018, and is projected to reach 22 million people by the year 2050 (World Population Review 2018). The agricultural sector has experienced profound changes over recent decades, with increased yields, productive use of labor and the expansion of farmland, contributing to one of the highest rates of agricultural growth (5.3%) recorded globally (World Bank 2018b). Hor et al (2014) have noted that agricultural expansion is one of the driving forces behind deforestation in Cambodia. Forests covered roughly 74% of land area in Cambodia in 1990 and by 2010 this decreased to 57%. Furthermore, Davis et al (2015) have noted that nearly 2 million hectares of land have been leased on long-term concessions to international and domestic investors, and that forest loss rates between 29%-105% higher than in the land areas outside of the concessions have been recorded. The context in Cambodia is incredibly complex and a combination of domestic and regional policy are driving the changing agricultural and forest landscape of Cambodia (World Bank 2018a). Rice has traditionally been the most important crop in Cambodia and is the most widely cultivated (Figure 1); other important crops include cassava, maize and soy bean. 8 Figure 1. A graph showing the harvested area of key crops in Cambodia for the period 2011- 2016. Source of data: FAOSTAT (2018) The PGIS survey has been conducted in Ratanakiri, a province located in the north east of Cambodia, bordering Vietnam. Agricultural practices in the province are changing and traditional practices such as shifting cultivation are being replaced by more intensive farming systems (World Bank 2015, World Bank 2018a). From Figure 1, we can see that landuse change has occurred south of Pruk, we can also notice that there is large agricultural or urbanization presence at the main town of Ban Lung Town. Interestingly we can see that there is land use change following the river towards Padal village. This is notable from the green surroundings opposed to the browner lighter grey shades of land around Padal. 9 Figure 2: The location of the study villages in Ratanakiri, Cambodia. The PGIS activity in Cambodia was conducted with the Royal University of Agriculture and the principal focal point was Dr Lyda Hok. In total 78 households were sampled across three villages: Luon (25 Households), Pruk (33 Households) and Padol (20 Households). The fieldwork was conducted in September 2016 over a three week period (including training and data collection). Laos: Over the past decade Laos has experienced rapid economic development (IFAD 2018) with an annual GDP growth rate of roughly 7% (2017-2015) recorded for the past 3 years (CIA, 2018). Although poverty rates have declined by 40% over the last 15 years, it is estimated that 26% of population continues to live below the poverty line (IFAD, 2018) and that prevalence of poverty is highest in the mountainous regions, particularly those in the north east bordering to Vietnam (IFAD 2018). Despite a reduction in the contributions of agriculture to national GDP (FAO, 2018), the sector remains important, contributing to 20% of national GDP (2017) and importantly providing the principal employment for 73% (2017) of the labour force (CIA, 2018). In regards to harvested area, the top 3 crops in Laos are rice, cassava and maize. 10 Figure 3. A graph showing the harvested area of key crops in Laos PDR for the period 2011-2016. Source of data: FAOSTAT (2018) The PGIS survey was conducted in the Xiangkhouang Province, north east of Laos, roughy 100km from the border of Vietnam. Phonsavan is the capital of the province and can be seen in Figure 4. The National Agriculture and Forestry Research Institute (NAFRI) was the lead partner responsible for conducting the research, with Dr Phonepaseuth Phengsavanh as the team leader and focal point. In total 94 households were sampled across five villages: Mouan (10 households), O Anh (14 Households), Pa Kkhom (29 Households), San Khing (22 Households) and Tar (18 Households). The fieldwork was conducted in March 2017 and took a total of 3 weeks to complete (including training and data collection). 11 Figure 4: The location of the study villages in Laos. Vietnam: Over the past 30 years Vietnam has moved from one of the poorest countries to a lower middle income nation (World Bank 2018c). The political and economic reforms of Đổi Mới, commenced in 1986 and have transformed the country (World Bank 2018c). Despite drastic changes the agricultural sector remains important both in regard to employment (40% of labour force emplyed in agriculture in 2017) and contribution to GDP (contributing to 15.3% of GDP in 2017). Rice remains the most widely grown crop in Vietnam, with maize, coffee, cassava and rubber also being widely grown (Figure 5). 12 Figure 5. A graph showing the harvested area of key crops in Vietnam for the period 2011-2016. Source of data: FAOSTAT (2018) The PGIS has been conducted in the The Central Highlands, a region of Vietnam, bordered by Quang Nam Province to the North, Quang Ngai, Binh Dinh, Phu Yen, Khanh Hoa, Ninh Thuan and Binh Thuan to the East, Dong Nai, Binh Phuoc, bordered by Attapeu (Laos) and Ratanakiri and Mondulkiri (Cambodia) provinces. While Kon Tum has its western border with both Laos and Cambodia, Gia Lai, Dak Lak and Dak Nong share the border with Cambodia. Lam Dong has no international border. The Central Highlands is equal to the total area of 5 provinces, the Central Highlands is 54,641 km2. In essence, the Central Highlands is not a unique plateau but a series of adjoining plateaus. Kon Tum Plateau is about 500m high, Kon Plong Plateau, Kon Ha Nung Plateau, Playku about 800m high, M'Drak High Plateau is about 500 m high, Buon Ma Thuot Plateau is about 500 m high, Mo The highland is about 800-1000 m high, Lam Vien plateau is about 1500 m high and Di Linh 13 plateau is about 900-1000 m high. All of these plateaus are bounded to the east by high mountain ranges and massif (Nam Truong Son). The Central Highlands can be divided into three terrain sub regions and three sub-climates, including North Central Highlands (corresponding to Kon Tum and Gia Lai provinces, formerly a province), Central Highlands (corresponding to Dak Lak and Dak Nong provinces), South Central Highlands (corresponding to Lam Dong province). Central Highlands have lower elevations and higher temperatures than the two northern and southern sub-regions. Located in the tropical savanna, the climate in the Central Highlands is divided into two seasons: the rainy season from May to October and the dry season from November to April, in which March and April are two hot and dry months. Best. Due to the influence of the altitude, while in the highlands of 400-500 m high, the climate is relatively cool and rainy, with the high plateau of over 1000 m, the climate is cool all year round, characterized by high mountain climate. Compared with other regions in the country, the socio-economic conditions of the Central Highlands are difficult, such as the lack of skilled labor, underdeveloped infrastructure, the commonality of many ethnic groups in a region. Small land and low living standards. However, the Central Highlands has the advantage of natural resources. The Central Highlands has 2 million hectares of fertile basalt soil, which accounts for 60% of the basalt land in the country, which is suitable for industrial crops such as coffee, cocoa, pepper, mulberry and tea. Coffee is the number one industrial plant in the Central Highlands. The area of coffee in the Central Highlands is now over 290 thousand hectares, accounting for 4/5 of the country's coffee area. Dak Lak is the province with the largest coffee area (170 thousand ha) and Buon Ma Thuot coffee is famous for its high quality. The Central Highlands is also the second largest rubber growing region after the Southeast, mainly in Gia Lai and Dak Lak. Forest resources and forest land in the Central Highlands are at risk of serious decline due to various reasons, such as the small area of unpolluted and non-migrant forests. Encroachment of forest for living and production (agricultural land of the whole region increases very rapidly) as well as illegal deforestation and forest product exploitation. Due to the depletion of forest resources, the volume of timber fell continuously from 600,000 to 700,000 cubic meters at the end of the 1980s and early 1990s to about 200-300 thousand m3 per year. At present, local authorities are experimenting with allocating land, leasing forest land to organizations, households and individuals for stable use and allocating forest and contracting forests for households and communities in village. 14 With regard to the diversity of cropping systems and natural conditions, we have selected Dak Lak and Dak Nong provinces as pGIS research areas.  Dak Lak Province is located in the central part of the Central Highlands, at the head of the Serepok River system and part of the Ba River, within the geographical coordinates of 107028'57 "- 108059'37" East and 1209'45 ". "- 13°25'06" North latitude. The average height of 400 - 800 meters above sea level.  Dak Nong is located at the southwestern gateway of the Central Highlands, in the coordinates 11 ° 45 to 12 ° 50 north latitude and 107 ° 12 to 108 ° 07 east longitude. The center of Dak Nong province is located 125 km from Buon Ma Thuot city along national highway 14, 250 km south of Ho Chi Minh City. The north and northeast of Dak Nong is bordered by Dak Lak province, to the east and south-east by Lam Dong province, to the south by Binh Phuoc province, and to the west by Dak Nong province to the Kingdom of Cambodia The border is about 120 km long. The two border crossings are Dak Per border gate in Dak Mil district and Bup'rang in Tuy Duc district. The government partner for conducting the PGIS was WASI, and the lead focal point was Ms Chau Thi Minh Long. The research was undertaken for 91 households divided between Dak Nong (52 Households) and Dak Lak (39 Households). 15 Figure 6: The location of study in Central Highlands, Vietnam 5. Methodology and methods The following section will provide information on (i) survey preparation and materials (ii) survey implementation (iii) data cleaning and finally (iv) data analysis. The PGIS survey can be conceptualized into two main components, one is compiling information on the household, which is undertaken on a singular occasion with each respective household. The second component is the field level data which is undertaken for each of the respective fields managed by the household. Before the training is conducted it is important to identify the respective household who will be included in the study. In our study we sampled a random subset of the households who had been included in the ROHMIS study (a previous survey conducted as part of the Hands and Minds project). Identifying these households was organized by the lead partner who was able to work with the village head to identify the respective households. Survey Preparation 16 As noted Participatory GIS (pGIS) is a tool to collect high-resolution spatial data based on local knowledge and perceptions. The methodology we use captures information through a survey (Annex 1) which was originally developed to assess the diversity of potatoes in the Andes at varying altitudes (de Haan 2009). The PGIS methodology was reviewed and refined as part of the Hands and Minds project. Notably, the survey was transferred from a paper based survey to Commcare (dimagi 2018), an open data kit (ODK) platform that we operated on huawei tablets (model: MediaPad t1 7.0). The initial phase focused on ensuring that the survey questions were entered correctly onto the commcare platform and that the survey could be completed and data collected and stored on the central database. Commcare provides the cloud storage for data collection which can be accessed through the commcare app (dimagi 2018). Data is accessed through an excel sheet which is devised as part of the survey formatting to ensure that the responses are clearly classified and eligible. The data can be extracted from the cloud to a pc and analysed offline. Beyond the environmental benefits of reduced paper wastage, the use of commcare enables in the field changes to the survey to be undertaken, although it is important to note that internet access is required to implement alterations to the survey structure and for the changes to be updated to the tablets. Nevertheless, through the ODK platform new questions can be included or redundant questions removed across all surveys and devices. This enables the survey to be more fluid and responsive to any changes that may occur during the research process. Undertaking the survey on tablets via commcare saves time in regards to data entry, as the information is automatically updated to the cloud, furthermore, near real time analysis of the data is made possible. The PGIS survey (Annex 1) can be divided into two sections, the first (Annex 1: Household datasheet) collects information on the household, specifically in regards to descriptive socio economic indicators. It is undertaken with the family heads of the household but captures information on both the female and male head of household and also family members. This section of the survey is short, equivalent to one page of questions and is intended to be undertaken in a conversational format and is implemented on a singular occasion for each sampled household. The second section (Annex 1: Field properties registration sheet) of the survey focuses on field level data and is undertaken with the head of the household for each of their respective fields. This section of the survey is intended to capture a broad plethora of indicators relating to both biophysical and socio economic stresses, the final page of the survey 17 relates to the crop history over the past ten years for the respective field and the planned crop cycle over the subsequent three years. It is important to note that before asking the field level questions (Annex 1: Field properties registration sheet) we use the GPS to map the farmer’s respective field (Annex 2). The farmer will be asked to walk the perimeter of their field, carefully tracking the boundary. The enumerator will take a GPS point at each corner of the field (Annex 2). The enumerators are trained to decide how many points are required, but the key is to accurately capture the shape of the field, as shown in Annex 2. The codes of the field are matched to the famer’s code. For instance, if farmer 001 has 5 fields, then we will map the first field of farmer 001 with the code 001-1 in the GPS device. The methodology process for a particular household can thus be summarized as the following: the enumerator conducts the household level questions on a singular occasion for the respective household, the enumerator then accompanies the household to their respective field, they then walk the perimeter of the field together taking the GPS points (Annex 2), subsequently, the enumerator asks the household the respective field level questions for the mapped field (Annex 1: Field properties registration sheet). This is repeated for each of the fields of the respective household. Materials The following list provides the necessary materials required to undertake the training of the enumerators and necessary equipment required for the survey of 100 households. GPS Units x8: In order to capture the location and structure (see survey implementation for further information) of the respective fields we use handheld GPS (Model: Garmin GPSMAP 64s). Annex 2 provides information on how the GPS surveys are operated. Each GPS requires two a4 batteries. Batteries a4: 48 Tablets: 8 (1 for each enumerator) Paper A2: 10 sheets (this is required for the training of the enumerators) Pen markers (multiple colours): 10 18 Projector: 1 Laptop: 1 Training materials (survey + GPS instructions): 8 (1 for each enumerator) Survey Implementation The implementation of the survey can be divided into two sections, notably (i) Training and shared objective setting with partners (ii) and the implementation of the survey  Shared objective setting with partners In each of the case study countries national government partners were included in the design, implementation and training of the PGIS methodology. It is integral that local government are involved. The first step is to sit with the respective government official responsible for leading the survey and to obtain consent and mutual understanding of how the information will be used. It is important for the survey questions to be reviewed to ensure that they do not infringe on culturally or politically sensitive information. Once consent has been given by the national government the main contact point is responsible for working with the local government partner for recruiting a total of 8 staff. We attempted to ensure that this team would consist of 2 national government staff, 2 local government staff and ideally 4 local farmers or students from the survey sites. However, due to various constraints regarding time and resources this was not always possible. Nonetheless, it is important that the team of enumerators possess individuals who understand the local context. Recruiting local farmers and students is a means to ensure ownership over the results and a way to involve local people in the decision making process. In all the study sites across the three countries, local government officials were part of the enumerator team. This was vital as they were able to converse in the local language and also provide helpful insights into the survey design and also question structure. Once the enumerator team has been established, the initiative to conduct pGIS is shared with village authorities, notably the village head and the local stakeholders, for instance the farmers union. In our case we used our main contact point in the government to oversee the sharing of this information and to garner the necessary permission required to undertake the study. Once permission to undertake the PGIS has been granted, CIAT scientists and the main government focal point worked together to plan the training and survey implementation. It is important to work closely with the partners to ensure that data collection is undertaken at a suitable time for both the partners, enumerators and farmers. Before commencing the project the 19 survey is translated from English into the local language and uploaded onto commcare and downloaded onto the tablet devices.  Training of local team of pGIS surveyors In each of the countries a single training over 2 to 3 days was given to ensure that all enumerators were comfortable with the methodology. For example, in the case of Laos the enumerators were recruited from government and therefore had significant understanding of undertaking research and also handling of tablets and GPS devices, this meant that the training could be undertaken in 2 days. Understanding the level of proficiency is required in order to set the length of the training and this is best deduced by working closely with the lead focal partner. The intention of the training is to provide detailed information on the survey structure and questions, data capture using tablets and operation of the GPS units in order to capture the field structure and location. In our experience, working closely with an identified lead focal person from the partner organization leading up to the project commencement helped to ensure that questions and queries could be more effectively addressed. The lead focal person from each country (see section: Study Sites) had already attended trainings at the CIAT office and were therefore aware of the expectations and complexities surrounding the survey design. The first day of the training focused on the PGIS survey (Annex 1), with the morning session dedicated to understanding the purpose of the project and reviewing the PGIS survey. The afternoon session focuses on the field level questions (Annex 2) and how to ask the questions regarding the stresses and the history of the field plot. As noted the intention is for the questions to be asked in a fluid manner, for example, there are questions relating to the climatic stresses and also the biophysical stresses (eg pest and disease, fertility) that affect the field. The enumerator asks these questions in an open format but can list the options if the household is unsure of the question, this will enable the household to answer a simple yes or no response. If the option is not available then the enumerator can record the climate or biophysical stress (Annex 2). During the training it is important to include opportunities for the enumerators and CIAT staff to discuss the relevance of the questions to the local context and to refine the survey if necessary. In the trainings completed the survey questions were deemed to be sufficient and understandable for the local contexts across the countries. Once the survey questions are fully understood information on how to complete the survey on the tablets is undertaken. The tablets have the survey loaded onto the commcare app. The purpose of the training from the CIAT scientists is for the enumerators to be able to input the responses and also to sync the data to the cloud. During the training, a CIAT scientist (Dr Dharani Dhar) was responsible for checking and altering the survey on commcare. 20 The second day of the training focuses on understanding and operating the GPS devices. Information on entering the respective code for the each farmer is given (annex 2). As noted, the enumerator is required to map the fields of the respective farmer. To ensure basic understanding of how GPS operate, a section of the morning is also allocated to describing triangulation and satellite location (Woodford 2018). The efforts are then focused towards understanding how to turn on the GPS, detect a satellite location, and to mark a point (Annex 2). The enumerators are then divided into pairs and take it in turns to practice the PGIS methodology. They will practice asking the questions in a fluid way and of also taking the GPS points and then asking the field level data questions. During this period the CIAT staff and the lead partner will divide the subset of farmers between the enumerators, ensuring that each enumerator has a similar number of farmers to survey. Local knowledge of the enumerators and the partners is used to plan how to sample all farmers in the respective villages. This requires some flexibility as it depends upon the location and accessibility of the villages. The entire duration of the PGIS survey was between 2-4 weeks for each country.  Implementation and data collection It is required that an initial meeting is set up with the village head and selected farmers who will participate in the survey; the lead enumerator and the CIAT staff will then describe the main objectives and outputs of the survey. This provides an opportunity for input from the farmers who are encouraged to add information to the survey, or raise any concerns. Additionally, this initial meeting can be used to mark communal spaces (e.g. pastures, forests), community boundaries and other landmarks if present onto high-resolution satellite image print outs. The exercise can be done with a group of village elders who draw the spaces with markers directly onto the map. The initial meeting is usually undertaken at the village hall, for example, this may be a temple or meeting house which the community use for such events. Based on the sample size (n = households) and the size of the survey team, each team member gets a fixed number of households assigned. Each team member coordinates visits to all individual fields of each of the assigned households. They can introduce themselves to the households during this initial meeting and arrange when to visit the respective fields. It is important that efforts are made to fit into the schedule of the household. Therefore, when possible it is recommended to visit the fields when the household plans to visit. The data collection process begins after the meeting has finished and continues until all households have been sampled. The enumerator team then move onto the next village where they repeat the aforementioned process. 21 During the data collection process Ms Than Thi Nguyen (CIAT) was responsible at the end of each day for compiling the GPS devices and copying over the GPX files onto her computer, recording the village, date of collection and enumerator responsible.  Data analysis The data can be categorized into three distinct groups (i) household (ii) Field properties and (iii) GPS. The household and field properties data are both stored in individual .xls and are opened and cleaned in Microsoft excel. The data cleaning requires ensuring that the Household codes are entered correctly, this is important as it is required to facilitate the join to the spatial data (GPS). The GPS data is stored as GPX files and is opened in QGIS 3.2 (see Annex 2: HOW TO EXPORT YOUR GPS DATA TO A GIS SOFTWARE). Each GPS device is attached to one enumerator who was responsible for the data collection. As mentioned, Than Thi Nguyen stored the GPX points at the end of each evening with the points classified by date and village. The file includes the GPX points collected by each enumerator. Each entry of GPX points is converted into a point .shp file which is saved (Annex 2). The point files are then converted into Polygons to represent the geometry of the respective fields for the villages in the respective country (Annex 3). This is done in ArcMap 10.6 using the ‘production points to line or polygon’ tool (ESRI 2016a). The accuracy of the conversion is manually reviewed using ArcMap. It is necessary to ensure that the conversion has functioned successfully. Errors may be present when the Field ID is not consistent amongst points. In this case the point can be manually edited to ensure that they possess a consistent Field ID using ‘modifying field properties’ ESRI (2016b). We can then ‘merge’ (ESRI 2016c) the polygons into one polygon shapefile which we can join to the field level data from our .xls file. In order to implement this we load the respective field level data .xls file into ArcMap 10.6. We then undertake a join function (ESRI 2016d) using the common ‘form.plot’id’ of the respective .xls file and the polygon.shp (eg the common field id) to make the join. We then have a spatially referenced polygon file with our respective fields with the information from the field level survey. We can then begin our descriptive statistics. In regard to household data we used excel to calculate the respective general statistics regarding the education, ethnicity and socio economic indicators of the villages for each country. This provides some general understanding of the overall context in each country. In regard to field level statistics, we used arcmap to calculate the area of all the plots, using the ‘calculate geometry tool’ in the fields attribute table of the polygon field shapefile for each country. This enabled us to produce general information on the number of plots of each crop and the area (ha) it occupies. We did this for each of our countries and then for each of the villages in the respective countries 22 (see results). We then opened this dbf file in excel to create graphs of the respective top 5 crops in each of the villages (for example see Figure 4 in results section). In order to understand the changes in cropping patters we used a land use change matrix function (Intersection tool) in ArcMap 10.6 for two distinct time periods: 2007 to 2012 and 2012 to 2017. A video tutorial is available (GIS for Planning 2017). We used our polygon .shp for 2007 (column: ‘c_wet_2007’ in the attribute table) and for 2012 (column: ‘c_wet_2012’ in the attribute table) as the input files to compare and we saved this as a new shp file, we then open the attribute table of the new feature to create a new column entitled change ‘2007_2012wet’, and we use ‘calculate geometry’ (see GIS for Planning 2017) to calculate change in area for each of the features. We then copy the data to an excel sheet and use a pivot table to select the area change (ha) for the ‘values’, we enter our land use 2007 as the ‘Columns’ of our table, and the Land use 2012 for the ‘Rows’ of our table (see GIS for Planning 2017 for example), we can then see the changes and save table (for example see results: figure 5). We repeat this for each of our villages. This process is done for the 2007-2012 period and then repeated for the 2012-2017 period. In conclusion, we create two transition matrix tables (one for 2007 – 2012 and 2012-2017) in excel (using pivot table) and which we include in our results section to reveal the changes in the area (ha) of the respective crops during this time interval periods. Results In the following section we present the household information for the case study sites, we then describe and discuss the main results for each of the villages in the respective case study sites. 5.1. Case study in Cambodia 5.1.1. Household information Among the 78 households surveyed, Tom Poun ethnic group is the majority compared to other ethnic groups, followed by the Cha ray and Khmer. Table 1: Distribution of ethnic groups in Cambodia Ethnic Frequency Percent Cha ray 19 24.4 Khmer 13 16.7 Other 1 1.3 Tom Poun 45 57.7 Total 78 100.0 23 The education level of the surveyed villagers is shown in the following figure: Figure 7. Distribution of Education groups in Cambodia Comparison of education levels among ethnic groups is shown in Table 3: Table 2. Comparison of education levels among ethnic groups community * education Cross tabulation education Total no primary primary incomplete secondary secondary incomplete superior community Cha ray 14 2 3 0 0 0 19 Khmer 1 3 6 0 2 1 13 other 0 0 1 0 0 0 1 Tom_poun 19 10 10 2 3 1 45 Total 34 15 20 2 5 2 78 24 Table below shows the percentage of people who originate from the village. Table 3. The original residents of Cambodia (percent) Original Frequency Percent no 7 9.0 yes 71 91.0 Total 78 100.0 Among 78 households, 7 households migrated from other localities in the past. Below we display whether family are originally from the village based on the different ethnic groups. Figure 8: Community * Family comes originally from the village Most households are not involved in off farm employment (61.5%). Table 4: The household also involved in off-farm employment Off-farm-employment Frequency Percent No 48 61.5 Yes 30 38.5 Total 78 100.0 25 5.1.2. Village – plot level crop data in Cambodia According to the data given by farmers during the pGIS survey, the distribution of the plots and land area for the sampled HH is presented in Appendix 1. The households with the highest number of plots are the 43rd household with 6 plots in Loun Village, 6th, 18th, 26th household with 5 plots in Prouk Village and in Padal Village are the 108th, 120th with 2 plots. In 3 villages surveyed with 108 plots and 87.69 ha are shown in the following graph: Figure 9: Number of plots and total cropping area/household for each village 5.1.2.1. Agriculture Production system We have identified the five major crops in regard to number of plots and area grown for each village. In the village of Loun: 5 main crops are: Cashew, Cassava, Rice, Soybean and mix of Cashew - Cassava, of which Cassava and Cashew occupy the largest area. Meanwhile, Prouk's main crops are Beans, Cashew, Beans – Cashew and Rice. Rice occupies the largest area then Beans. With Padal village, the main crops are Cashew, mixed of Cashew - Cassava, Rice and Cassava. Details of major crops are shown in the following chart: 26 Figure 10: Top 5 crops in villages in 2017 To better understand, we can see the distribution of the crops of the villages in Cambodia through the following map: 27 Figure 11: Spatial distribution of crops across the seasons of 5 major crops in 2017 In 2012: we can see the change of major crops according to the following chart: 28 Figure 12: Top 5 crops in villages in 2012 Spatial distribution of cropping system in the three villages in Cambodia for 2012 is shown in the following map: Figure 13: Spatial distribution of crops across the seasons of 5 major crops in 2012 In 2007: The major crops grown in 2007 are shown in the following graph below: 29 Figure 14: Top 5 crops in villages in Laos (2007) In 2007, the main crops in Prouk village were Rice, Cassava followed by Forest, Cashew, and Beans. At Loun Cashew, Cassava, Forest, Soybean and Fallow. While dominant crops dominated the villages, Padal village was largely deserted without cultivation. But in 2017, the area vacated has fallen significantly. This shows a significant shift in the area of cultivation. Spatial distribution of land use in three villages in Cambodia for 2007 is shown in the following map. 30 Figure 1: Spatial distribution of crops across the seasons of 5 major crops in 2007 5.1.2.2. Transition in Cropping system In Padal village: The following table shows the transition area between the crops in each period 2007 - 2012 and 2012 – 2017 Table 5: Transition matrix of agriculture production system in Padal (2007 to 2012) (ha) Crops system 2007 Crops system 2012 Cashew Cassava Fallow Forest Rice Trees Cashew 1.26 0 0 0 0 0 Cassava-Rice-Fallow 0 0.24 0 0 0 0 Fallow 0 0.56 4.38 0 1.70 0 Forest 0 0 0 3.83 0 0 Rice 0 0 0 0 0.32 0 Trees 0 0 1.44 0 0 0.93 31 In the period 2007 to 2012: Cashew area is unchanged, the area planted between Cassava - Rice and a part of vacant land is completely converted into Cassava. The rice area increased for conversion from bare land (1.7 ha). Some Trees move to Fallow (1.44 ha), either by harvest or due to weather conditions affecting crops. Table 6: Transition matrix of agriculture production system in Padal village (2012 to 2017) (ha) Crops system 2012 Crops system 2017 Cashew Cashew - Cassava Cassava Rice Cashew 1.26 0 0 0 Cassava 0 0 0.56 0.24 Fallow 0 1.36 1.22 3.24 Forest 1.96 0 0 1.87 Rice 0 0 0.73 1.29 Trees 0 0 0.93 0 In the period 2012 – 2017: Almost empty Fallow that no longer switch to crops Cashew - Cassava and Rice, The area of forest plantation also lost due to the shift to Cashew (1.96 ha) and paddy rice (1.87 ha). This shows that there is a significant shift in the crop when most of the Forest and Trees are cut off to other crops. In Prouk village: The following table shows the transition area between the crops in each period 2007 - 2012 and 2012 – 2017 Table 7: Transition matrix of agriculture production system in Prouk village (2007 to 2012) (ha) Crops system 2007 Crops system 2012 Beans Cashew Cassava Fallow Forest Others Rice Beans 6.53 1.45 0 0 0 0 0 Cashew 0 9.98 0 0 0 0 0 Cassava 10.18 0 12.02 0 0 0 0 Fallow 0.38 0 0 3.73 0.93 0.33 0 Forest 0 0.69 3.44 2.31 8.62 0.06 4.92 Others 0 0 0 0.65 0 6.58 0 Rice 3.89 0 0 1.34 0 0 15.75 In the period of 2007 - 2012: Cassava area decreased from 22.20 hectares to 15.46 hectares, in which the biggest decrease was the Forest area (from 20.04 hectares to 9.56 hectares), part of rice cultivation (4.92 hectares), Cassava (3.44 ha), Fallow (2.31 ha). The crop area increased 32 rapidly as Beans, the cause is a part of area of Cassava (10.18ha), Rice (3.89ha) shifted to make the area jump from 7.98 ha to 20.98 ha. Table 8: Transition matrix of agriculture production system in Prouk village (2012 to 2017) (ha) Crops system 2007 Crops system 2012 Beans Cashew Cassava Fallow Forest Others Rice Beans 20.91 0 0 0 0 0 0.99 Beans - Cashew 0 0 0 0 5.31 0 0 Beans - Potatoes 0 0 2.83 0 0 0 0 Beans - Rice 0 0 0 1.50 0 0 0 Cashew 0 12.12 0 0.56 0 0 0 Cashew - Cassava 0 0 1.44 0 0 0 0 Cassava 0 0 11.18 0.49 0 0 0.44 Cassava - Rice 0 0 0 0 0 0 2.17 Fallow 0.08 0 0 2.31 0 0 1.09 Forest 0 0 0 0 0.93 0 0 Others 0 0 0 0 0 6.97 0 Pepper 0 0 0 0.16 0 0 0 Rice 0 0 0 3.02 3.31 0 15.98 In the period 2012 - 2017 there is a clear shift between the crops when people switch to intercropping, in particular the intercropping between Beans - Cashew (5.31 ha from the plantation moved), Beans and Potatoes (2.83 ha from Cassava) etc. During this period the area of Beans dominate because Beans (20.91ha) are intercropped with other crops. In Loun village: In the period of 2007 - 2012, cashew area increased 2.22 hectares, planted area decreased by 1.63 hectares due to switching to Cashew - Cassava intercropping. Table 9: Transition matrix of agriculture production system in Loun village (2007 to 2012) (ha) Crops system 2007 Crops system 2012 Cashew Cashew, Cassava Cassava Cassava, Rice Fallow Forest Rice Rubber Soybean Cashew 10.76 0 0 0 0 0 0 0 0 Cashew, Cassava 0 2.00 0 0 0 0 0 0 0 Cassava 0 0 8.47 1.36 0 0 0 0 0 Fallow 0 0 3.40 0 1.17 0 0 1.78 0 Forest 0 1.63 0 0 0 3.18 0 0 0 Rice 0 0 0.65 0 0 0 0.68 0 0 Soybean 0 0 0.84 0 0 0 2.21 0 1.84 Soybean, Cashew 2.22 0 0 0 0 0 0 0 0 33 Soybean, Rice 0 0 0.78 0 0 0 0 0 0 In the period 2012 -2017, Cassava area increased, the area of cashew plant has little change. Most of main crops have little change, mainly with changes in Forest area and short-term crops such as Soybean, Rice, etc. Table 10: Transition matrix of agriculture production system in Loun village (2012 to 2017) (ha) Crops system 2012 Crops system 2017 Cashew Cashew, Cassava Cassava Maize Others Rice Soybean Cashew 12.53 0.45 0 0 0 0 0 Cashew, Cassava 0 3.63 0 0 0 0 0 Cassava 0 0.45 12.51 0.81 0 0 0.36 Cassava, Rice 0 0 1.36 0 0 0 0 Fallow 0 0 0.76 0 0.41 0 0 Forest 0 0 1.92 0 0 1.27 0 Rice 0 0 2.60 0 0 0.29 0 Rubber 0 0 0 0 1.78 0 0 Soybean 0 0 0 0 0 0 1.84 Most of the plots are irrigated by rain fed, with very few plots having an active irrigation system. We can see the distribution of the plots in space: Table 11: Irrigation of surveyed in 3 villages in Cambodia Irrigation Number of plots Total area (ha) Loun Rain fed 40 42.97 Padal Rain fed 18 14.66 Prouk Cannel 1 6.58 Pump 2 1.34 Rain fed 46 87.10 34 Figure 2: Irrigation map of the surveyed plots in Cambodia According to the space distribution of the plots only the village Prouk has some plots equipped with pumps for irrigation. Based on the survey, we developed climate risk maps for villages in Cambodia focusing on drought, erosion, soil. 35 Figure 3: Map of Affected by Drought to plots In addition to being affected by drought, the crops are also influenced by other factors such as Erosion, soil fertility. 36 Figure 4: Presence of Erosion Figure 5: Effects of Poor Fertility on plots 37 5.2. Case study in Laos 5.2.1. Household information Among the 90 households surveyed, 78% were Hmong ethnic group and 22% were Lao Loum. Table 12: Distribution of ethnic groups in Laos villages Ethnic Frequency Percent Hmong 70 78 Lao Loum 20 22 Total 90 100.0 The education level of the surveyed villagers is shown in the following figure and table: Figure 6 . Distribution of Education groups in Laos Comparison of education levels among ethnic groups is shown in Table below: Table 13: Comparison of education levels among ethnic groups Community education Total No Primary Primary (incomplete) Secondary Secondary (incomplete) Hmong 13 14 17 11 15 70 Lao Loum 0 4 2 5 9 20 Total 13 18 19 16 24 90 38 Table 14: The original residents of Laos (percent) Original Frequency Percent no 12 13.3 yes 78 86.7 Total 90 100.0 Among 93 households, 12 households migrated from other localities in the past. Based on that result we compared the number of ethnic migrants who are originally from the village as indicated in Figure below: Figure 7: Community * Family comes originally from the village. The majority of HH were not involved in off farm employment. Table 15: The household also involved in off-farm employment Off-farm-employment Frequency Percent no 57 70.3 yes 36 29.7 Total 93 100.0 39 We also looked at the Livestock holdings of the Households in Laos: Table 16: Livestock in Laos (heads/total households/all villages) Livestock Minimum Maximum Sum Mean Std. Deviation pigeon cows goats 0 9 36 .39 1.540 buffaloes 0 9 107 1.16 2.170 pig 0 12 100 1.09 2.197 chicken 3 120 2494 27.11 18.182 ducks 0 50 586 6.37 7.839 fish 0 200 1306 14.20 32.486 horses 0 0 0 0.00 0.000 5.2.2. Village – plot level crop data in Laos During the implementation of pGIS in Laos, we have selected and surveyed in five villages: Mouan, Or Anh, Pa Khom, San Khing and Tar, most of the crops here are mostly mono- cropping. Table 17: Monoculture of villages in Laos Villages Name Monocropping no yes Total Mouan 12 12 Or Anh 1 28 29 Pa Khom 7 41 48 San Khing 10 28 38 Tar 4 20 24 Total 22 129 151 According to the table above, despite the disparity in numbers between the plots of village but we still 22 plots where the plant cultivation system of the village are focused on growing crops mixed together. In total of 5 villages surveyed with 151 plots and 104.55 ha are shown in the following graph: 40 Figure 22: Number of plots and total cropping area/household for each village in Laos 5.2.2.1. Agriculture Production system With the current survey results in 2017, cropping systems in Laos are shown in the table below: 41 Table 18: Agriculture Production system in Laos in 2017 Crops system in 2017 Mouan Or_Anh Pa_Khom San_Khing Tar Total plots Total Area Plots Area Plots Area Plots Area Plots Area Plots Area Banana - Cassava 2 0.59 2 0.59 Banana - Fodder - Maize - Rice 2 1.98 2 1.98 Banana - Fodder - Sugarcane - Vegetables 1 0.17 1 0.17 Bell Pepper 2 0.10 2 0.10 Cassava 1 0.03 2 0.57 3 0.60 Cassava - Fodder 1 0.34 1 0.53 2 0.87 Cassava - Fodder - Fruits 1 1.82 1 1.82 Cassava - Fodder - Tea 1 1.03 1 1.03 Coffee - Fodder - Maize - Rice 1 0.85 1 0.85 Coffee - Fruits - Rice - Vegetables 1 0.99 1 0.99 Fallow 2 0.10 2 0.10 Fodder 2 0.58 7 1.66 5 1.16 5 2.52 4 1.99 23 7.91 Fodder - Maize 1 0.54 1 2.07 2 2.61 Fodder - Rice 4 3.95 4 3.95 Maize 24 24.75 1 0.09 1 0.08 26 24.92 Maize - Peanuts 1 1.04 1 1.04 Maize - Rice 1 3.36 1 0.10 2 3.47 Maize - Rice - Peanuts 1 1.29 1 1.29 Maize - Vegetables - Beans 1 0.23 1 0.23 Peanuts 1 0.21 1 0.21 Rice 9 6.86 7 4.55 12 7.51 18 11.95 10 5.92 56 36.79 Rice - Fallow 1 0.67 1 0.67 Rice - Vegetables 1 0.34 1 0.34 Tea 13 12.01 13 12.01 Vegetables 1 0.01 1 0.01 Grand Total 12 8.11 29 19.80 48 43.37 38 18.31 24 14.97 151 104.55 42 From Table above, we can see five major crops: Rice, Maize, Tea, Fodder and intercropping between Fodder - Rice. Most of the crops in the villages are Rice and Maize. The tea plantation is mainly in Or Anh village. Most of the other plots have interlocking between the crops. To better understand the process of crops change that has happened at the household level we conducted a survey on Agriculture Production system since 2007. This was done through farmer interviews. We selected the year 2012 and 2007 in a five-year cycle to facilitate the transition calculation of the following crops: 43 Table 19: Agriculture Production system in Laos in 2012 Crops system in 2012 Mouan Or Anh Pa Khom San Khing Tar Total Plots Total Area Plots Area Plots Area Plots Area Plots Area Plots Area Banana - Cassava 2 0.59 2 0.59 Banana - Fodder - Maize - Rice 2 1.98 2 1.98 Banana - Fodder - Sugarcane - Vegetables 1 0.17 1 0.17 Bell Pepper 1 0.07 1 0.07 Cassava 1 0.03 1 0.03 Cassava - Fodder 1 0.34 1 0.53 2 0.87 Cassava - Fodder - Fruits 1 1.82 1 1.82 Cassava - Fodder - Tea 1 1.03 1 1.03 Coffee - Fodder - Maize - Rice 1 0.85 1 0.85 Coffee - Fruits - Rice 1 0.99 1 0.99 Fallow 1 0.31 4 3.01 11 9.01 6 2.89 6 1.71 28 16.94 Fodder 1 0.27 5 1.10 1 0.21 3 1.71 2 1.14 12 4.42 Fodder - Maize 1 0.54 1 0.54 Fodder - Pepper 1 0.32 1 0.32 Fodder - Rice 4 3.95 4 3.95 Forest 3 0.96 1 0.82 4 1.79 Fruits - Rice - Vegetables 1 0.34 1 0.34 Maize 18 15.00 2 0.11 1 0.36 21 15.46 Maize - Peanuts 1 1.04 1 1.04 Maize - Rice 2 5.43 1 0.10 3 5.54 Maize - Rice - Peanuts 1 1.29 1 1.29 Maize - Vegetables - Beans 1 0.23 1 0.23 Rice 10 7.53 5 3.74 11 9.10 17 9.99 7 4.54 50 34.89 Tea 10 9.41 10 9.41 Grand Total 12 8.11 29 19.80 48 43.37 38 18.31 24 14.97 151 104.55 44 Table 20: Agriculture Production system in Laos in 2007 (ha) Crops system in 2007 Mouan Or Anh Pa Khom San Khing Tar Total Area Total Area Plots Area Plots Area Plots Area Plots Area Plots Area Banana - Cassava 2 0.59 2 0.59 Banana – Fodder - Maize - Rice 2 1.98 2 1.98 Banana - Fodder - Sugarcane - Vegetables 1 0.17 1 0.17 Bell Pepper 1 0.07 1 0.07 Cassava - Fodder 1 0.34 1 0.34 Cassava - Fodder - Fruits 1 1.82 1 1.82 Cassava – Fodder - Tea 1 1.03 1 1.03 Coffee - Fodder - Maize - Rice 1 0.85 1 0.85 Coffee - Fruits - Rice 1 0.99 1 0.99 Fallow 1 0.31 13 11.11 16 14.07 7 3.71 7 2.31 44 31.51 Fodder 1 0.27 3 0.42 2 0.53 2 1.15 1 0.98 9 3.35 Fodder - Maize 1 0.54 1 0.54 Fodder - Rice 3 2.37 3 2.37 Forest 5 2.64 4 3.57 1 0.03 1 0.82 11 7.06 Fruits - Rice - Vegetables 1 0.34 1 0.34 Maize 12 14.95 2 0.11 1 0.36 15 15.42 Maize - Rice 1 0.10 1 0.10 Maize - Rice - Peanuts 1 1.29 1 1.29 Maize - Vegetables - Beans 1 0.23 1 0.23 Rice 10 7.53 5 3.74 11 6.99 17 9.73 9 6.20 52 34.18 Tea 1 0.32 1 0.32 Grand Total 12 8.11 29 19.80 48 43.37 38 18.31 24 14.97 151 104.55 45 5.2.2.2. Transition in cropping systems So back in 2012 and 2007, the Laos study sites had 151 plots. The five main crops grown by farmers are Rice, Maize, Tea, Maize – Rice intercropping system in 2012 and Rice, Maize, Forest, Fodder in 2007. During these two years, Fallow occupied a sizeable area and hardly ever used to grow any crops. This shows a transition in the agricultural production system over the period from 2007 and 2012 to 2017. Therefore, in order to facilitate the assessment of agriculture production system change for the period 2005 - 2012, and 2012- 2017, we selected the same five major crops and system as of 2017. Table below shows the Matrix conversion area between crops of the five major crops intercropping system. Table 21: Transition matrix of agriculture production system in 2007 to 2012 (ha) Crops system 2012 Top major crops 2007 transition Fallow Fodder Forest Maize Rice Tea 7.42 0 1.68 0 0 Forest 0 0 1.79 0 0 Cassava - Fodder - Tea 0 0 0 0 0 Rice 3.83 0 0.85 0 30.21 Fodder 0.83 3.03 0 0 0.56 Fallow 12.68 0 0 3.13 1.13 Fodder - Maize 0 0 0 0 0 Coffee - Fruits - Rice 0 0 0 0 0 Fruits - Rice - Vegetables 0 0 0 0 0 Cassava 0 0 0.03 0 0 Maize - Vegetables - Beans 0 0 0 0 0 Coffee - Fodder - Maize - Rice 0 0 0 0 0 Cassava - Fodder 0.53 0 0 0 0 Banana - Fodder - Sugarcane - Vegetables 0 0 0 0 0 Maize 4.15 0 1.68 8.93 0.71 Banana - Cassava 0 0 0 0 0 Cassava - Fodder - Fruits 0 0 0 0 0 Maize - Rice 2.07 0 0 3.36 0 Bell pepper 0 0 0 0 0 Banana - Fodder - Maize - Rice 0 0 0 0 0 Maize - Rice - Peanuts 0 0 0 0 0 Fodder - Pepper 0 0.3168 0 0 0 Maize - Peanuts 0 0 1.04 0 0 Fodder - Rice 0 0 0 0 1.58 46 In period 2007 to 2012: According to 2007 table 20, the Fallow area is 31.51 hectares, but by 2012 it is only 16.94 hectares (table 19). This shows that the Fallow has been reduced by half and moved to other crops of which 7.42 hectares were planted to Tea, 4.15 (table 21) hectares were planted to Maize, 3.83 hectares were planted to Rice (table 21), and some were converted to other crops (table 21). Rice is the main crop of the area surveyed with an area of 34.18 hectares in 2007, rising to 34.89 hectares due to part of the area from Fallow (3.83 hectares) and Forest (0.85 hectares). The area of Maize is almost unchanged when it retains the area previously planted (15.42 - 15.46 ha). In addition, significant increase of Tea increased from 0.32 to 9.41 hectares (of which 7.42 hectares from Fallow, 1.68 hectares from Forest). Table 22: Transition matrix of agriculture production system in 2007 to 2012 (ha) Crops system 2017 Top major crops 2012 Fallow Maize Maize- Rice Rice Tea Tea 2.45 0 0 0 9.41 Rice 4.52 1.10 0 30.36 0 Cassava- Fodder - Tea 0 0 0 0 0 Fodder 2.25 0.10 0 0 0 Fodder - Maize 0 0 2.07 0 0 Coffee - Fruits - Rice - Vegetables 0 0 0 0 0 Rice - Vegetables 0 0 0 0 0 Vegetables 0.01 0 0 0 0 Cassava 0.57 0 0 0 0 Maize - Vegetables - Beans 0 0 0 0 0 Fallow 0.10 0 0 0 0 Coffee - Fodder – Maize - Rice 0 0 0 0 0 Cassava - Fodder 0 0 0 0 0 Banana - Fodder – Sugarcane - Vegetables 0 0 0 0 0 Maize 6.82 14.25 0 3.86 0 Bell pepper 0 0.02 0 0 0 Banana - Cassava 0 0 0 0 0 Cassava - Fodder - Fruits 0 0 0 0 0 Maize - Rice 0 0 3.47 0 0 Banana - Fodder - Maize - Rice 0 0 0 0 0 Maize - Rice - Peanuts 0 0 0 0 0 Maize - Peanuts 0 0 0 0 0 Peanuts 0.21 0 0 0 0 Rice - Fallow 0 0 0 0.67 0 Fodder - Rice 0 0 0 0 0 47 In period 2012 – 2017: The Rice area increased 36.79 hectares (of which 4.52 hectares from Fallow, 1.1 hectares from Maize), besides also partially Rice planting area (3.86 hectares) and Fallow (6.82 hectares) Maize cultivation switch makes Maize acreage increased rapidly (from 15.46 hectares to 24.92 hectares). Other crops in this period but also the shift that shift represents only a relatively small area. To better understand, we can look into more detail about the restructuring of plants at the village level. VILLAGE LEVEL ANALYSIS In Mouan Village: According to the table above, Mouan Village has 12 plots with 3 main crops: in 2017 with 9 plots Rice (6.86 hectares), 2 plots Fodder (0.58 ha) and 1 plot Rice - Fallow (0.67 ha). However, before that in 2012 and 2007 there are 10 plots were Rice (7.53ha), 1 plot was Fodder (0.27ha), 1 plot was Fallow (0.31ha). As a result, there was virtually no transplant in Mouan village between 2007 and 2012, until the transition period from 2012 to 2017, there is a slight change in the area of rice decreased by 1 plot, Fodder increased to 2 plots. For clarification, this translation can be shown in the following table: Table 23: Transition agriculture production system in Mouan Village 2007, 2012 to 2017 (ha) Crops system 2017 Crops system 2012, 2007 Rice Fodder Fallow Rice 6.86 0 0 Fodder 0 0.27 0.31 Rice - Fallow 0.67 0 0 According to the table on 0.67ha area from Rice to Rice - Fallow, the area planted Fodder increased due to the Fallow area. Now that we see no diversity of crops in the village Mouan village. In Or Anh Village: In 2017, there are 29 plots of which 13 plots are Tea (12.01 ha), 7 plots are Rice (4.55 ha), 7 plots are Fodder (1.66 ha), and intercropped Cassava - Fodder - Tea, Fodder - Tea. By 2012: 10 plots are Tea (9.41 ha), 5 plots are Rice and Fodder, Fallow 4 plots. In 2007, number of Fallow plots accounted for 13 plots the in total 29 plots surveyed in Or Anh village, followed by Rice and Forest. Detail of this transition can be shown in the following table: 48 Table 24 : Transition agriculture production system in Or Anh Village 2007 to 2012 (ha) Crops system in 2012 Crops system in 2007 Fallow Forest Cassava – Fodder - tea Rice Fodder Fodder - Maize Tea Tea 7.42 1.68 0 0 0 0 0.32 Forest 0 0.96 0 0 0 0 0 Cassava - Fodder - Tea 0 0 1.03 0 0 0 0 Rice 0 0 0 3.74 0 0 0 Fodder 0.67 0 0 0 0.42 0 0 Fallow 3.01 0 0 0 0 0 0 Fodder - Maize 0 0 0 0 0 0.54 0 During the period 2007 - 2012, there was a major change in Fallow and Forest, with the largest area being Fallow: 7.42 ha to Tea, 0.67 ha to Fodder. The forest area is reduced to 1.68 hectares due to the shift to Tea. The area of rice cultivation is almost unchanged. Table 25: Transition agriculture production system in Or Anh Village 2012 to 2017 (ha) Crops system in 2017 Crops system in 2012 Tea Forest Cassava-Fodder tea Rice Fodder Fallow Fodder- Maize Tea 9.41 0.15 0 0 0 2.45 0 Rice 0 0.82 0 3.74 0 0 0 Cassava- Fodder-Tea 0 0 1.03 0 0 0 0 Fodder 0 0 0 0 1.10 0.56 0 Fodder-Maize 0 0 0 0 0 0 0.54 In period 2012 – 2017 no more fallow area, this area has changed to Tea (2.45 ha) and Fodder (0.56 ha), the rice area also increased to 0.82 ha due to the change from Forest. In Pa Khom Village: - The number of plots surveyed in 2017 in the relatively large area accounted for 48 plots, of which 24 plots were Maize, 12 plots were Rice, 5 plots were Fodder, and the remaining 7 plots were planted alternately among crops. - In 2012, there were only 18 plots of Maize, 11 plots of rice, and 11 plots of fallow. 49 - Back in 2007, the number of Maize reduced to just 12 plots, with no change in the paddy fields, but the fallow increased to 16 plots. From there, we could saw that from 2007 to 2012 and 2012 to 2017 fallow has moved to most other crops. Detailed information on this crop conversion can be shown in the following table: Table 26: Transition agriculture production system in Pa Khom Village 2007 to 2012 (ha) Crops system in 2007 Crops system in 2012 Rice Fallow Maize Forest Banana-Fodder-Maize-Rice Maize-Rice- Peanuts Fodder Fallow 1.05 4.84 3.13 0 0 0 0 Rice 5.23 3.01 0 0.85 0 0 0 Maize 0.71 4.15 8.46 1.68 0 0 0 Banana-Fodder- Maize-Rice 0 0 0 0 1.98 0 0 Maize-Rice 0 2.07 3.36 0 0 0 0 Maize-Rice- Peanuts 0 0 0 0 0 1.29 0 Fodder-Pepper 0 0 0 0 0 0 0.32 Maize-Peanuts 0 0 0 1.04 0 0 0 Fodder 0 0 0 0 0 0 0.21 Table 27: Transition agriculture production system in Or Anh Village 2012 to 2017 (ha) Crops system in 2012 Crops system in 20017 Rice Fallow Maize Banana- Fodder-Maize- Rice Maize- Rice Maize-Rice- Peanuts Fodder- Pepper Maize- Peanuts Fodder Fodder 0 0.53 0.10 0 0 0 0.32 0 0.21 Rice 5.23 1.54 0.74 0 0 0 0 0 0 Maize 3.86 6.73 14.16 0 0 0 0 0 0 Banana-Fodder Maize-Rice 0 0 0 1.98 0 0 0 0 0 Fodder-Maize 0 0 0 0 2.07 0 0 0 0 Maize-Rice- Peanuts 0 0 0 0 0 1.29 0 0 0 Maize-Peanuts 0 0 0 0 0 0 0 1.04 0 Peanuts 0 0.22 0 0 0 0 0 0 0 Maize-Rice 0 0 0 0 3.36 0 0 0 0 Between 2007 and 2012: 3.01 ha of fallow and 0.85 ha of forest converted to rice. The area of maize cultivation also increased due to a portion of rice (0.71 ha), fallow (4.15 ha) and forest (1.68 ha). The forest area in 2007 has shifted completely to crops such as maize, rice and intercropping between Maize-Peanuts. In the next period from 2012 to 2017: The evolution of 50 the transition is similar to that of 2007 to 2012, with the majority of the area planted by people changing their fallow. In this period, the area of rice was reduced in part due to the over- conversion to maize (1.59 ha). In San Khing village: With 38 plots surveyed in 2007, the main crops in the village are Rice (17plots with 9.73ha), Fodder (2 plots with 1.15ha) and Fallow (7 plots with a large area of 3.71 ha. By 2012, the main crops are Rice (17 plots with an area of 9.99 ha in which 0.82 ha was from Fallow), Fodder (3 plots with an area of 1.71 hectares) and the number of plots of fallow reduced to 6 plots with an area of 2.89 hectares. The transition agriculture system is shown in the following table below: Table 28: Transition agriculture production system in San Khing village 2007 to 2012 (ha) Crop system in 2012 Crop system in 2007 Rice Fallow Coffee, Fruits, Rice Fruits, Rice, Vegetables Forest Maize, Vegetables, Beans Coffee, Fodder, Maize, Rice Cassava, Fodder Banana, Fodder, Sugarcane, Vegetables Maize Fodder Banana, Cassava Bell pepper Rice 9.17 0.82 0 0 0 0 0 0 0 0 0 0 0 Fallow 0 2.89 0 0 0 0 0 0 0 0 0 0 0 Coffee, Fruits, Rice 0 0 0.99 0 0 0 0 0 0 0 0 0 0 Fruits, Rice vegetables 0 0 0 0.34 0 0 0 0 0 0 0 0 0 Cassava 0 0 0 0 0.03 0 0 0 0 0 0 0 0 Maize, Vegetables, Beans 0 0 0 0 0 0.23 0 0 0 0 0 0 0 Coffee, fodder, maize, rice 0 0 0 0 0 0 0.85 0 0 0 0 0 0 Cassava, fodder 0 0 0 0 0 0 0 0.34 0 0 0 0 0 fodder 0.56 0 0 0 0 0 0 0 0 0 1.15 0 0 Banana, fodder, sugarcane, vegetables 0 0 0 0 0 0 0 0 0.17 0 0 0 0 Maize 0 0 0 0 0 0 0 0 0 0.11 0 0 0 Banana, cassava 0 0 0 0 0 0 0 0 0 0 0 0.59 0 Bell pepper 0 0 0 0 0 0 0 0 0 0 0 0 0.075 By 2017, the main crops of the village are Rice (18 plots with an area up to 11.95ha was converted from Fallow), Fodder (5 plots with 2.52ha), fallow is almost still very little, only 2 plots with area 0.1ha. It is also possible to see an increase in plant area due to the conversion of unused fallow. 51 Table 29: Transition agriculture production system in San Khing village 2012 to 2017 (ha) Crops system in 2017 Crops system in 2012 Rice Fallow Coffee, Fruit, Rice Fruits, Rice, Vegetables Cassava Maize, Vegetables, Beans Coffee, Fodder, Maize, Rice Cassava, Fodder Fodder Banana, Fodder, Sugarcane, Vegetables Maize Banana, Cassava Bell pepper Rice 9.99 1.96 0 0 0 0 0 0 0 0 0 0 0 Fodder 0 0.81 0 0 0 0 0 0 1.71 0 0 0 0 Coffee, Fruits, Rice, Vegetables 0 0 0.99 0 0 0 0 0 0 0 0 0 0 Rice, Vegetables 0 0 0 0.34 0 0 0 0 0 0 0 0 0 Vegetables 0 0.01 0 0 0 0 0 0 0 0 0 0 0 Cassava 0 0 0 0 0.03 0 0 0 0 0 0 0 0 Maize, Vegetables, Beans 0 0 0 0 0 0.23 0 0 0 0 0 0 0 Fallow 0 0.10 0 0 0 0 0 0 0 0 0 0 0 Coffee, Fodder, Maize, Rice 0 0 0 0 0 0 0.85 0 0 0 0 0 0 Cassava, Fodder 0 0 0 0 0 0 0 0.34 0 0 0 0 0 Banana, Fodder, Sugarcane, Vegetables 0 0 0 0 0 0 0 0 0 0.17 0 0 0 Maize 0 0 0 0 0 0 0 0 0 0 0.09 0 0 Bell pepper 0 0 0 0 0 0 0 0 0 0 0.02 0 0.07 Banana, Cassava 0 0 0 0 0 0 0 0 0 0 0 0.59 0 In Tar village: With total area of plots in village: 14.97ha (24 plots). The main crops are: Rice (9 plots-6.20ha), mix of Fodder – Rice (3 plots – 2.37ha) and Fallow (7 plots – 2.31ha) in 2007. In 2012: The area of rice cultivation decreased (7 plots - 4.54ha), the area of intercropping of Fodder-Rice increased (4 plots - 3.95ha), fallow area also decreased (6 plots - 1.71ha). Table 30: Transition agriculture production system in Tar village 2007 to 2012 (ha) Crops system in 2012 Crops system in 2007 Rice Fallow Forest Cassava, fodder, fruits Maize, Rice Maize Fodder, Rice Fodder Rice 4.54 0 0 0 0 0 0 0 Fallow 0.08 1.62 0 0 0 0 0 0 Forest 0 0 0.82 0 0 0 0 0 Cassava, Fodder 0 0.53 0 0 0 0 0 0 Cassava, Fodder Fruits 0 0 0 1.82 0 0 0 0 Maize, Rice 0 0 0 0 0.10 0 0 0 Maize 0 0 0 0 0 0.36 0 0 52 Fodder, Rice 1.58 0 0 0 0 0 2.37 0 Fodder 0 0.16 0 0 0 0 0 0.98 By 2017: The area of rice increased (10 plots – 5.92ha). However, compare with 2007, it was reduced a part to convert to other crops. This shows that the area of crops has increased and decreased in each period. Table 31: Transition agriculture production system in Tar village 2012 to 2017 (ha) Crops system in 2017 Crops system in 2012 Rice Fallow Forest Cassava, Fodder Cassava, Fodder, Fruits Maize, Rice Maize Fodder, Rice Fodder Rice 4.54 1.02 0 0 0 0 0.36 0 0 Cassava 0 0.57 0 0 0 0 0 0 0 Fodder 0 0.03 0.82 0 0 0 0 0 1.14 Cassava, Fodder 0 0 0 0.53 0 0 0 0 0 Cassava, Fodder, Fruits 0 0 0 0 1.82 0 0 0 0 Maize, Rice 0 0 0 0 0 0.10 0 0 0 Fodder, Rice 0 0 0 0 0 0 0 3.95 0 Maize 0 0.08 0 0 0 0 0 0 0 The majority of HH in Laos were rainfed. Table 32: Irrigation in Laos Total Irrigation ha % Cannel 10.61 10.15 Rain fed 93.94 89.85 Grand Total 104.55 100 The distribution of the Irrigation map of the surveyed plots in Laos was shown in the following map: 53 Figure 23: Map of irrigation in Laos (surveyed Villages) Most of the plots did not record climate stress as a substantial issue. Table 33: Affected by Climate conditions (ha) Climate sensitivity Total Area Number of plots Cold 3.77 4 Drought 9.24 15 Flooding 4.98 6 None 86.55 126 Grand Total 104.55 151 Affected by risks: Table 34: Affected by Risks (ha) 54 Risks Total area Number of plots None 58.41 93 Other risk 0.21 1 Pest 29.87 41 Pest Diseases 0.41 1 Pest diseases, Poor fertility 1.05 1 Pest, Poor fertility 4.10 4 Pest, sedimentation 3.73 2 Pest, soil erosion 0.96 2 Poor fertility 1.52 2 Sedimentation, poor fertility 2.37 2 Soil erosion 1.92 2 Grand Total 104.55 151 Of the 151 surveyed in the table above, the number of affected pests accounted for about 1/3 of the total. Although the number of affected plots is low, the affected area accounts for a high proportion (about 40% of the surveyed area). 5.3. Case study in Central Highland, Vietnam 5.3.1. Household information Regarding the distribution of ethnic groups in Vietnam, there are now 4 ethnic groups living in the sampled villages (case study: Dak Lak and Dak Nong). Among the 91 households surveyed, Kinh ethnic group is the majority compared to other ethnic groups, followed by the Mnong and Thai. Table 35: Distribution of ethnic groups in Central Highland villages Ethnic Frequency Percent Kinh 83 91.2 Mnong 5 5.5 Tay 1 1.1 Thai 2 2.2 Total 91 100.0 The education level of the surveyed villagers is shown in the following figure: 55 Figure 8: Distribution of Education groups in Central Highland Comparison of education levels among ethnic groups is shown in Table below: Table 36: Comparison of education levels among ethnic groups community * education Cross tabulation Community Education Total no Primary Primary incomplete Secondary Secondary incomplete superior Kinh 1 8 6 31 15 22 83 Tay 0 0 0 0 1 0 1 Thai 1 0 0 1 0 0 2 Total 4 9 7 32 17 22 91 56 Table 16 shows the percentage of people who originated Table 37: The original residents of Central Highland (percent) Original Frequency Percent No 83 91.2 Yes 8 8.8 Total 91 100.0 Among 91 households, there are 83 households migrated from other localities in the past. Based on that result we compared the number of ethnic migrants with the originals as indicated in Table and Figure below: Figure 9: Community * Family comes originally from the village The majority of sampled Households did not possess off farm employment (64%) as shown below. Table 38: The household also involved in off-farm employment Off-farm-employment Frequency Percent no 64 70.3 yes 27 29.7 Total 91 100.0 57 Table 39: Livestock in Central Highland (heads/total households/all villages): Livestock Minimum Maximum Sum Mean Std. Deviation Pigeon 0 100 153 1.96 11.819 Cows 0 11 152 1.95 2.412 Goats 0 10 28 .36 1.503 Buffaloes 0 2 3 .04 .252 Pig 0 100 373 4.78 17.638 Chicken 0 500 2734 35.05 63.694 Ducks 0 100 310 3.97 14.056 Fish 0 1300 11750 150.64 318.631 Horses 0 0 0 0.00 0.000 5.3.2. Village – plot level crop data in Vietnam Perennial trees including coffee, cashew, pepper, Sugarcane, and fruit trees and so fifth are strength of the Dak Lak and Dak Nong. According to the data given by farmers during the pGIS survey, the distribution of the land area in Central Highland is presented in Appendix…, the households with the highest number of pieces (bold and underlined) are the 200th and 30th household with 5 plots in Dak Nong, 170th, 95th, household with 6 plots in Dak Lak and the household with the highest is 2nd with 9 plots. In total of 2 villages surveyed with 203 plots and 125.77 ha are shown in the following graph 58 Figure 10: Number of plots and total cropping area/household for each village in Central Highland 5.3.2.1. Agriculture production system In 2017: From Table below, we can see five major crops: for each village intercropping system. This table shows the area statistics of the five major crops and Figure 18 shows the spatial distribution of the areas of 2017. Table 40: Top major crops grown by Central Highland farmers in 2017 Crops Dak Nong Number of Plots Area (ha) Coffee 36 16.02 Cashew-Coffee 10 7.77 Coffee-pepper 10 6.82 Cashew 10 5.89 Cashew-Coffee-Pepper 8 3.13 Dak Lak Sugarcane 18 32.49 Coffee 14 7.28 Pepper 11 5.85 Cassava 5 3.89 Rice 16 3.38 In Dak Nong: 5 main crops are: Coffee, Cashew, Cashew-Coffee, Coffee-Pepper, Cashew-Coffee- Pepper of which Coffee occupy the largest area. Meanwhile, In Dak Lak's main crops are Coffee, Cashew, Cassava, Pepper and Sugarcane. Sugarcane occupies the largest area then Coffee. Details of major crops are shown in the following chart: Figure 11: Top major crops in Central Highland in 2017 59 To better understand, we can see the distribution of the crops of the villages in Central Highland through the following map: 60 Figure 12: Spatial distribution of crops across the seasons of 5 major crops in 2017 In 2012: We conducted a survey on land-use change since 2012 to better understand the process of land use change that has happened at the household level. This was done through farmer interviews. Table 40: Top major crops grown by Central Highland farmers in 2012 Crops Dak Nong Number of Plots Total Area Coffee 44 20.91 Cashew coffee 14 9.28 Cashew 16 8.85 Fallow 4 2.19 Rice 8 1.87 Dak Lak Sugarcane 19 29.96 Coffee 28 15.69 Maize 10 6.27 Cassava 5 4.67 Cashew 6 3.54 61 Similar in 2017, top major crops grown in 2012 is shown in the following graph below: Figure 13: Top major crops grown in Central Highland The main crops in Dak Nong were Coffee, Cashew-Coffee followed by Cashew, Rice, Fallow. At Dak Lak is Sugarcane, Coffee, Maize, Cassava and Cashew. This shows a significant shift in the area of cultivation. Spatial distribution of land use in villages in Central Highland for 2012 is shown in the following map. 62 Figure 14: Spatial distribution of crops across the seasons of 5 major crops in 2012 In 2007: The top 5 crops did not change compared to 2012 in Dak Nong and Dak Lak provinces However, there is a change in area, which indicates that there is a shift of crops from one stage to another. Table 41: Top 5 major crops grown by Central Highland farmers in 2007 Crops Dak Nong Number of Plots Area (ha) Coffee 39 19.30 Cashew 23 13.51 Cashew coffee 8 4.47 Fallow 7 3.57 Rice 9 2.63 Dak Lak Sugarcane 17 27.69 Coffee 28 15.17 Maize 13 8.65 Cassava 5 4.67 Cashew 7 3.86 63 Spatial distribution of crops across the seasons of 5 major crops in 2007 can be shown following figure: Figure 30: Spatial distribution of crops across the seasons of 5 major crops in 2007 64 5.3.2.2. Transition To better understand the process of crops change that has happened at the household level we conducted a survey on Agriculture Production system since 2007. In Dak Nong: The following table shows the transition area between the crops in each period 2007 - 2012 and 2012 – 2017. Table 42: Transition matrix of agriculture production system in Dak Nong (2007 to 2012) Crops system 2012 Crops system 2007 Cashew Cashew-Coffee Coffee Fallow Rice Others Cashew-Coffee 4.78 3.89 0.62 0 0 0 Coffee 0.92 0 18.41 0.35 0.76 0.21 Cashew-Coffee-Pepper 0 0.58 0 0 0 0 Fallow 0 0 0 2.19 0 0 Cashew 7.81 0 0 1.04 0 0 Coffee-Pepper 0 0 0 0 0 1.07 Maize 0 0 0 0 0 1.23 Rice 0 0 0 0 1.87 0 Forest 0 0 0 0 0 0.32 Pepper 0 0 0 0 0 0.10 In the period 2007 - 2012: Coffee area increased while cashew area decreased, specifically: Cashew area (4.78ha) and coffee (0.62ha) between Cashew-Coffee (4.78ha). The area for growing coffee increased due to the area from Cashew (0.92ha), Rice (0.76ha), and Fallow (0.35ha). Table 43: Transition matrix of agriculture production system in Dak Nong (2012 to 2017) Crops system 2017 Crops system 2012 Cashew Cashew-Coffee Coffee Fallow Rice Others Cashew-Coffee-Pepper 1.09 1.46 0 0 0 0.58 Coffee 0.83 0 14.42 0 0 0.49 Coffee-Pepper 0 0 5.75 0 0 1.07 Cashew-Coffee 0 7.48 0.00 0.29 0 0 Pepper 0 0 0.46 0.93 0 0.10 Cashew 5.89 0 0 0 0 0 Cashew-Fruits 1.04 0 0 0 0 0 Maize 0 0 0 0 0 1.05 Rice 0 0 0 0 1.87 0 Fodder 0 0 0 0.96 0 0 65 In the period 2012 - 2017: Area of coffee plantation reduced to 14.42 ha as part of the transition to intercropping between Coffee-Pepper (5.75ha), Pepper (0.46 ha). The area of Cashew also decreased from 8.85 ha to 5.89 ha, due to switch to Coffee (0.83ha), Cashew-Fruits (1.04 ha). Table 44: Irrigated in Dak Nong Irrigation Total of Area Cannel 1.557101 Pump 38.859833 Rain fed 5.991494 Total 46.408428 We can see the distribution of the Irrigation map of the surveyed plots. Figure 31: Irrigation map of the surveyed plots. - Due to the tropical monsoon climate, it is divided into two distinct seasons: the rainy season from May to November and the dry season from November to April, leading to the 66 structure of irrigated areas in uneven seasons. , low usage area. We also investigated some of the risks of climatic conditions affecting the area of the crop, most of which are affected by drought and heat. This is also the reason for changing the crop area to select the crop that best suits on the plot. Table 45: Effecting of weather conditions on the plots in Dak Nong Climate sensitivity Total Area Number of plots Drought 6.78919 17 Drought - Heat 37.648894 72 Drought – Heat - Flooding 0.341976 2 Heat - Flooding 0.666605 2 None 0.961763 1 Grand Total 46.408428 94 - Pest problems also cause people to change crops accordingly. Table 46: Pest problems on the land in Dak Nong Risks Sum of Area Number of plots Diseases 0.259126 1 Diseases - Soil erosion 0.938499 2 None 0.961763 1 Pest 1.098534 3 Pest - Diseases 22.164833 42 Pest Diseases – Poor fertility 7.771259 18 Pest Diseases - Sedimentation - Poor fertility 0.246211 1 Pest Diseases - Soil erosion 1.73381 3 Pest Diseases - Soil erosion - Poor fertility 8.326118 18 Pest - Poor fertility 0.160462 1 Pest - Sedimentation 0.262713 1 Pest - Soil erosion 2.03593 2 Poor fertility 0.44917 1 Grand Total 46.408428 94 Table 47: Products from plots are used to Use production Sum of Area Number of plots None 1.410933 2 Home Consumption 1.869631 8 Sale 43.127864 84 Grand Total 46.408428 94 67 In Dak Lak: In period 2007 - 2012: Cashew area decreased 0.32 hectares, coffee area increased by conversion from Maize (0.47ha) and Sugarcane (0.57ha). Sugarcane area increased due to part of Maize (3.92ha) and other crops transferred. In this stage, we see a significant increase in the area of sugarcane and coffee. Table 48: Transition matrix of agriculture production system in Dak Lak (2007 to 2012) Crops 2012 Crops 2007 Others Cashew Cassava Coffee Maize Sugarcane Cashew 0 3.54 0 0 0 0 Cassava 0 0 4.67 0 0 0 Coffee 0 0 0 14.65 0.47 0.57 Maize 0 0 0 0 4.13 2.14 Others 18.25 0.32 0 0.53 0.12 0 Sugar cane 1.07 0 0 0 3.92 24.98 In period 2012 – 2017: During this period, Sugarcane area continued to increase from 29.96 hectares to 32.49 hectares. Coffee area decreased sharply by switching to other crops (5.05ha) and Pepper (3.87ha). Table 50; Transition matrix of agriculture production system in Dak Lak (2012 to 2017) Crops 2017 Crops 2012 Cashew Cassava Coffee Maize Others Sugarcane Others 3.16 1.95 5.05 2.74 10.41 3.16 Cassava 0.38 2.73 0 0 0.79 0 Coffee 0 0 6.78 0 0 0.50 Pepper 0 0 3.87 0 1.98 0 Rice 0 0 0.00 0.33 3.05 0 Sugarcane 0 0 0 3.19 2.99 26.30 - Effects of Irrigation: Of the 109 plots, most of the plots have pumping stations for irrigation and the rest depends on rain-fed. 68 Table 49: Irrigated in Dak Lak Irrigation Total Area Number of plots Pump 45.129317 71 Rain fed 34.232762 38 Grand Total 79.362079 109 - Impact of Climate Condition: most of the plots are affected by drought and heat: Table 50: Effecting of weather conditions on the plots in Dak Lak Climate sensitivity Total Area Number of plots Drought 20.287323 27 Drought - Flooding 0.150756 2 Drought - Heat 54.327449 71 Flooding 1.528542 3 Heat 2.113499 3 None 0.95451 3 Grand Total 79.362079 109 - Impacting by pests: Table 51: Pest problems on the land in Dak Lak Risks Total Area Number of plots Diseases 0.197667 1 Diseases - Poor fertility 0.500695 1 Diseases - Soil erosion 1.022061 1 None 10.690969 8 Pest 8.416327 10 Pest - Diseases 14.230204 23 Pest - Diseases - Poor fertility 25.489399 37 Pest - Diseases - Sedimentation 0.061509 1 Pest - Diseases - Sedimentation - Poor fertility 0.4009 1 Pest - Diseases - Soil erosion 1.336373 2 Pest - Diseases – Soil erosion - Poor fertility 4.171536 4 Pest - Poor fertility 6.374514 9 Pest sedimentation 0.089247 1 Pest - Soil erosion 3.377519 2 Poor fertility 1.708518 5 Sedimentation - Poor fertility 0.323337 1 Soil erosion 0.228504 1 Soil erosion - Poor fertility 0.7428 1 69 Grand Total 79.362079 109 - Using the production: Table 52: Products from plots are used to Use production Total Area Number of plots None 1.966681 3 Home consumption 3.88395 19 Sale 73.511448 87 Grand Total 79.362079 109 6. Discussions In the following section the changes in agricultural cropping systems will be described in general terms for each country, an in depth analysis for each village will then be provided. In order to generate greater understanding of the changes in cropping system the GIS shapefiles for each of the villages have been analysed. This involved identifying at what year the changes occurred, through visualizing the changes in ArcMap 10.6. The GIS shapefiles are available upon request. 6.1. Cambodia The findings from the pGIS survey reveal interesting trends in the transition of the agricultural systems in Ratanakiri, Cambodia. We can see a trend to less fallow and forest based systems to more market based crops. We also see that most plots suffer from Risk factors, such as soil erosion, poor fertility and pest, although the type of system present and intensity of the system influences the risk presence, nonetheless, understanding this distribution can help us plan the introduction of farming practices that minimize the environmental impact of the crop on each piece of land. The following section will describe and identify these transitions of the sampled villages. 6.1.1. Padal Village The villages is located on the banks of the river in the north eastern region of Ratanakiri, in close proximity to the Vietnam border (5 km). The village was established post 2001 (not 70 present in Google Earth Images from 2001) and therefore may be considered an example of a frontier village that is driving the expansion of agriculture in the more remote areas of Ratanakiri. Transition - Increase in cassava, rice and cashew plots - Decrease in forest and fallow plots There has been a shift to more intensive agricultural production with a reduction in the number of plots left fallow and also the conversion of all forest plots to crop production. For example, in 2007 a total of 8 fields were fallow in the wet season, 4 plots were classified as forest, 3 are classified as Trees, which indicates a degraded forest, and only 1 plot of cassava, cashew and rice respectively were recorded (Figure 15). By 2017 no plots contained forest or trees, just one plot was fallow, and 10 plots of rice, 4 plots of cashew and 5 plots of cassava are present (Figure 11). In the year 2013 - 2014 we can see a final push from forest to rice, with last 2 forest plots making the transition to cultivation. In regard to the spatial distribution of the crops, it can be seen that rice is generally grown on the flood plain roughly 1km from the river. This area has been cleared of forest and trees in previous years and possess a track which means that travel time from the house is low (<20 minutes walk). In total, 4 plots of Cassava (3) and cashew (1) are grown in close proximity (<400m) from the river bank and take longer to reach from the village (3 of the 4 plots have travel time from the house hold > 25 minutes). Notably, in 2017 the plot located on the opposite bank of the river makes a conversion from fallow to rice. A travel time of 42 minutes from the house hold is recorded. This may signal increased activity of agricultural expansion on the opposite side of the river, however, further research would need to be conducted to understand what is driving this expansion, and whether it may be an isolated case, or part of a general push to cultivate the land in this area. It is notable that of the 9 rice fields (2016), only 1 plot is for “sale”, 1 is for both sale and home consumption and all other remaining plots are for home consumption. This indicates that the village is growing rice to sustain its population. This may be because the village is relatively isolated and therefore may indicate that less rice is bought at the market. Alternatively, the 8 71 plots of cashew (4) and cassava (5) are reserved for “sale” or for “both” domestic consumption and sale, therefore indicating that these crops are more market orientated. Risk factors affecting surveyed plots Pest (11) and pest and disease (5) are the most important risks to crops in Padal. Worms, parrots, mouse, crab and pink bugs were some of the pests that were recorded. The worms affected plots near to the river. Impacts of soil erosion (combined with disease 1 plot) and poor fertility (1 plot) were not recorded in large numbers. The village is located on the river bank so is flat. Additionally, the village was established recently (post 2001) and therefore soil nutrients is likely to be relatively high. 6.1.2. Prouk Village Located approximately 18km from Ban Lung town, Prouk village is located in the central southern region of Ratanakiri (Figure 2). This area is on the northern edge of the heavily cultivated land areas which can be seen in the satellite imagery where the dark grey and brown meets the relative green north of Prouk. Transition: There is a reduction in Forest. In 2007 there is a notable cassava and bean rotations, with the presence of rice, cashew and Forest. By 2012 forest is cleared in further plots from village. In 2014 there is an increase in number of cassava fields replacing those fields which were forest-rice-cassava. For example, by 2012 of the 13 forest plots (recorded in 2007), 5 have been converted to rice, 2 for cassava, 1 for cashew, and by 2014, all forest plots have been converted, with cassava (4 plots) rice (4 plots) and beans (1 plots) and cashew (1plot) being the most popular crops. Beans are also introduced to these plots often replacing cassava for one year before resuming cassava (cassava bean rotation is maintained). Rice is generally for home consumption, whilst other crops are for sale. The village contains alternative commercial crops. For instance in 2011 Potato is introduced to one plot (3297m2) near the road (10mins walking) for commercial sale. The field remains potato until 2017 but no new potato fields are recorded. In 2016 there is a small (1555m2) pepper field (commercial sale) located close to the household and road (6 mins walking) pesticides are used and the field is irrigated. Additionally there is a very small plot of mango (598m2) which is developed in 2012 from forest, with the mangos for home consumption. 72 Risk factors affecting surveyed plots: In total 19 plots (of 49) have no Risk. Pest and/or poor fertility affect 18 plots (of 49). It would be interested to analyse whether the plots with the presence of risks are plots experiencing transitions in cropping systems. 6.1.3. Luon Village Luon Village is the closest to province capital Ban Lung Town, Ban Lung, roughly 3.5km distance. The village is market orientated and we can see that the cropping systems are responsive to change. Further research could explore whether it is market forces, climate stresses, or other risks which are driving the changes. Transition - There is a reduction in the number of plots which are forest. 2007 = 4 plots (48089m2) and in 2017= 0. In 2015 there is 1 plot of forest (15679m2). - There is relatively few plots dedicated to rice (2007- 3 plots / 3 – 2017) and these are dedicated to both “home consumption and sale”. - There has been an increase in the number of cassava fields and cashew which has largely replaced soybean (2007=4 plots / 2017=1 plot). - 1 field of rubber is planted in 2007 (17846). In 2017 there are rubber/ cassava fields (1) and in 2017 also 1 field with Rubber (dry season) and soya bean (wet season) Risk: In 2017, only 4 fields had no “Risk” factors. These fields are relatively small in size. 1 plot: cassava/fallow cycle of 15679, 1 plot: rice/other/fallow cycle of 12661, 1 plot: a previously forest plot until 2015 when it had 1 year fallow 2016 and became a cassava field in 2017 of 15679m2. The larger fields tend to suffer from soil erosion and poor fertility. These fields are located in the southern areas of the village (Appendix 1) Spatial Dynamic - Near to Ratanikiri town - Already marketed orientated in 2007 73 6.2. Laos The agricultural sector in Laos is beginning the transition to a more intensive and market orientated system, although large disparity regarding agricultural intensification exists and as can be seen in the PGIS dataset, it is still possible to find farmers practicing traditional agro technologies and practices such as shifting cultivation. Notably we identified that just 21 plots of 151 applied agrochemicals. Furthermore, of the 151 surveyed plots, over half (84 plots) were for home consumption, whilst the produce from 49 plots was allocated to both domestic consumption and sale, and 19 plots dedicated solely to the market. This reveals that there is a dynamic of subsistence in the surveyed villages in Laos with overall less market focus. In the following section we will focus on capturing and describing for the respective villages the transition in cropping patterns, the risks farmers face and the spatial dynamic which influences the manifestation of these variables. Summary - More Fodder crops- especially fodder and maize or rice combinations - More fields intercropped (15% of plots) - Less pesticide used - Increase in Maize - A decline in fallow fields 6.2.1. Muon, Or Anh and Tar village We see a shift in the northern areas of the village (Or Anh) from fallow to tea. This is an interesting transition in which the households are growing a cash crop. According to informal discussions with several Households, the majority of the tea trees are grown within agroforestry systems. The first plot switches from fallow to tea in 2010, the plots further north switch in the next years with the final recorded plot making the switch in 2015. In the south west of the area (Mouan village), we see that rice dominates and remains stable and all plots are for home consumption. Pest and poor fertility affects a number of these plots likely due to the continued production of monoculture rice. Tar has a large number of plots dedicated to fodder and/or fodder mixed cropping. 74 Risks: The results suggest that the area has much less incidence of Risks, with 49 of the 65 plots stating that there were no risks to the plots. Of the remaining plots, the biggest Risk was deemed to be pests (9 plots) with the majority of these plots growing rice (6 plots) and the remaining allocated to tea (3plots). 6.2.2. Pa Khom village In Pa Khom village the expansion of maize recorded in the area is attributable to the conversion of forest and fallow land. The transition usually follows a forest, fallow, maize cycle. With 3 plots undergoing the transition for forest to fallow in 2010. Roughly 9 plots switch from fallow to maize, although we also see some plots grow rice for several years between fallow and maize. Finally, in the dry season many of the maize plots remain fallow, allowing the fields to recover. It is notable that Maize (22 plots) is always either for sale (11 plots), and/or sale + home consumption (11), while the other most significand crop, rice (12 plots) is almost exclusively for home consumption (7plots) , home consumption and sale (3plots) and just 1 plot totally dedicated to sale. Risks: 22 of the 48 plots recorded no Risk. 19 of the plots recorded pests as a risk, with Rice (9plots) and Maize (7) being the crops grown. 6.2.3. San Khing village San Khing does not display a large change in cropping patterns. There is an increase though in the number of rice plots with 2 plots moving from fallow in wet season to rice, although 1 plot moves from rice to fodder. In 2016, there are 18 plots of rice and in the 2007 (wet season) there are 16 plots, so one can see rice remains stable. In 2007, we see 2 fodder plots, by 2017 this increases to 5 fodder plots. We notice that there are small plots (average size of 0.38ha) that are allocated to vegetables, such as bell peppers and also coffee, in mixed cropping systems (7 plots) with the produce for home consumption and sale. While the area of rice fields is larger, on average (0.66 ha). Risk: 22 plots had no risk (of 38 plots), while 13 plots recorded pest and just 2 soil erosion plots. Again, it is notable that there is less incidence of soil erosion, and poor fertility, furthermore, there is not a notable spatial cluster of the pests, and rather a spread of disease throughout the plots. 75 6.3. Central Highlands, Vietnam In Dak Nong: The main crops of Dak Nong are: Coffee, Cashew, of which Coffee occupies the largest area. In addition to coffee trees, the provinces of the Dak Nong also have long-life industrial crops of high economic value, such as pepper, rubber trees. However, agricultural development across the region in general, as well as the development of long-term industrial crops in particular, has not yet met the regions potential. Agriculture is still fragmented, with low quality and value, while major agricultural products such as coffee, pepper, rubber, and cashew nut are mainly raw exports, leading to a low added value and incomes for local farmers. In Dak Lak: The main crops is similar Dak Nong. In fact, the geographic location, topography has caused the resources of Dak Lak is greatly affected by the climate conditions. At the same time, the current use of land by the people is one of the causes for increasing the complexity of climate change for land resources. Massive development of coffee and pepper areas... not in line with the plan, together with the conversion of many areas of dipterocarp forests into rubber plantations has increased the erosion, soil degradation and ecological environment changes. coupled with the decline of water resources, has a negative impact on land resources and land use patterns. Apart from objective reasons, such as climate change which drastically increases abnormal drought and flooding that cause severely negative impacts on agricultural production, the subjective reasons are that policy mechanisms have not been properly adjusted on a par with production practices, especially the policies on land, investment and credit for agricultural development. In addition, the prevalent situation is that management and use of agricultural land is not under planning, especially the planning for growing coffee and pepper. Cause changes in plant area in Dak Nong: According to the report on land use in agricultural production of Dak Nong province, there are some reasons leading to low land use efficiency: - Because the topography of the province is strongly fragmented, it is mainly hilly and mountainous, leading to areas of agricultural production of small, decentralized, dispersed in distant. Based on that, the team analyzed some of the results from interviewing people about how much time they spend moving from home to their land and by what transportation. According to the results of the survey, most of the plots are far from the 76 area where people live and it is quite difficult to move there by walking, most of which are using vehicles such as: Motorbike and take about 30 minutes to get there. The detail can follow in Appendix… - The agricultural infrastructure works only for certain areas, especially irrigation works, and canal systems that do not lead water far away, mainly for the area around lakes and dams etc., irrigation works etc., resulting in low land use coefficient, causing difficulties in transportation, increasing production costs etc., so the production value per unit area is not high. According to the survey most of the area is mainly pumped for irrigation, but still part of the area using natural rainwater, which also affected part to the crops in that area can withstand the impact of conditions drought, and it also causes their partial conversion to other crops have the ability to cope with more water shortage conditions. - Literacy levels are still low, uneven; cultivation skills are low, working individually, fragmentation, etc., many producers inefficient, and income mainly from the activities of non-agricultural but still want to keep the land as a mechanism for hedging due to lack of formal social security in the countryside. 6.4. Limitations The study is an ambitious attempt to capture the changes in cropping systems in a broad spectrum of villages located in different geographies across the three countries. The survey was conducted with the same households which were included in a previous survey (Rohmis) that sought to identify a plethora of information on cropping practices, socio economic status and decision making at the household level. Despite the best attempts to ensure 100 households were sampled in each country it was not possible in some cases to identify the same households sampled as part of Rohmis. Nonetheless, in Laos (94 households), Cambodia (78 households) and Vietnam (91 households) a substantial sample was achieved, however, with greater resources and time it would have been beneficial to increase the sample size for each country. Furthermore, the reason for the lower sample size in Cambodia was due to one of the villages included in Rohmis not being included in the PGIS survey as resources were not available. Although it is possible to infer general trends in the cropping systems for the villages sampled, it is difficult to make assumptions of the regional trends that are occurring. Increasing the geographical distribution of the villages sampled would enable broader assumptions to be made. 77 7. Conclusions and recommendations The PGIS activity has produced a wealth of information that can shed light on the agricultural transitions occurring in the three study sites. It is notable that both differences between the three case study sites and also within the study sites exists. Cambodia has seen a reduction in the forest and fallow plots with increases in the production of cash crops such as cassava (see section 5.1.2.2). However, this is not uniform across all the villages, with Padal undergoing the most substantial increase in agricultural production, with forest and trees being replaced by cassava and cashew (Table 5 & 6). Like Pruk and Luon village, Rice in Padal is largely for domestic consumption, whilst the cassava and cashew are for the market. We also notice that in Pruk village we have the trialing of novel cash crops such as Potato on small plots near to the household. This suggests that households are entrepreneurial and willing to try less traditional crops. A spatial component to the types of Risks present in Cambodia exists, with the more recently established village of Padal having less issues with soil fertility, and erosion, and more incidence of pest and disease, whilst Pruk and Luon village which have been market orientated for longer recorded problems with both soil fertility and pest and disease. Laos displayed less market orientation with limited application of agrochemicals, and more plots allocated to domestic consumption (GIS database). Nonetheless, we can see an increase in the production of maize, particularly in Pa Khom village. We also see the increase of tea in Or Anh village, which is a cash crop that has expanded and is grown in agro-forestry systems. Overall we can see that the agricultural system in the Central Highlands, Vietnam, is largely orientated towards the market, with cash crops being produced since and notably before 2007. The changes in crop patterns is complex, but we can see an increase in the production of pepper and reduction in coffee (see section 5.3.2.2.). We have provided a description of the household information at each village and also the crop changes occurring at the plot level. Further analysis should attempt to capture the drivers behind the agricultural transitions which are occurring, in order to identify which households are driving the changes in cropping systems. A cluster analysis could be used to group the indicators at the household level of those farmers who are undertaking agricultural transitions. 78 This work could be linked to the Hands and Minds research on Risk taking farmers and draw on a previous survey called RHoMIS (a rapid baseline survey carried out with + 1300 households in the three sites) to further classify risk taking farmers and link that to the cropping changes. Comparing the rates of agricultural transition between the countries could also be undertaken. For example, are greater levels of crop system change found in Vietnam and is this positively correlated to indicators such as market orientation? Another topic for further consideration may be to compare the use of agrochemicals application, monocrop information, and also socio economic information on the Households, to explore how this influences the incidence of pest and diseases. 79 References Aynekulu, E., Wubneh, W., Birhane, E., Begashaw, N. (2006) Monitoring and Evaluating Land Use/ Land Cover Change Using Participatory Geographic Information System (PGIS) Tools: A Case Study of Begasheka Watershed, Tigray, Ethiopia. The Electronic Journal of information systems in developing countries. 25 (1) pg 1-10 Basupi, V, L., Quinn, H, C., Dougill, J, A. (2017) Using participatory mapping and a participatory geographic information system in pastoral land use investigation: Impacts of rangeland policy in Botswana. Land Use Policy 64, pg 363-373 CIA (2018) Laos [online] Available from: https://www.cia.gov/library/publications/the- world-factbook/geos/la.html [Accessed on 10th October 2018] Davis, F, K., Yu, K., Rulli, C, M., Pichdara L., D’Odorico, P. (2015) Accelerated deforestation driven by large-scale land acquisitions in Cambodia. Nature geoscience. [online] Available from: https://cloudfront.escholarship.org/dist/prd/content/qt8194188h/qt8194188h.pdf [Accessed on 10 October 2018] De Haan (2009) Potato Diversity at Height: multiple dimensions of farmer-driven in-situ conservation in the Andes. PHD Thesis Wageningen [online] Available from: http://www.farmersrights.org/pdf/PhD%20Thesis%20Stef%20de%20Haan.pdf [Accessed on 10 October 2018] Dimagi (2018) CommCare [Online] Available from: https://www.dimagi.com/commcare/ [Accessed on 15 October 2018] ESRI (2016a) Converting Points to Lines or Polygons [online] Available from: http://desktop.arcgis.com/en/arcmap/latest/extensions/production- mapping/converting-points-to-lines-or-polygons.htm [Accessed on 15 October 2018] ESRI (2016b) Modifying Field Properties [online] Available from: http://desktop.arcgis.com/en/arcmap/10.3/manage-data/geodatabases/modifying- field-properties.htm [Accessed on 15 October 2018] ESRI (2016c) Merge features into one feature [online] Available from: http://pro.arcgis.com/en/pro-app/help/editing/merge-features-into-one-feature.htm [Accessed on 2 October 2018] ESRI (2016d) How to: join an Excel spreadsheet (.xls) to a feature class in ArcMap [online] Available from: https://support.esri.com/en/technical-article/000008903 [Accessed on 2 October 2018] FAO (2018) Laos [online] http://www.fao.org/laos/fao-in-laos/laos-at-a-glance/en/ [Accessed on 10th October 2018] FAOSTAT (2018) CROPS. Area Harvested. Country [online] Available from: http://www.fao.org/faostat/en/#data/QC [Accessed on 10 October 2018] 80 GIS for Planning (2017) How to Create Land Use Change Matrix Table/ Membuat Matriks Perubahan Penggunaan Lahan. [Online ] Available from: https://www.youtube.com/watch?v=D2UTdli1KoA [Accessed on 15 October 2018] Hor, S., Saizen, L., Tsutsumida, N., Watanabe, T. (2014) The impact of agricultural expansion on forest cover in Ratanakiri Province, Cambodia. Journal of Agricultural. 6 (9) 46-59 IFAD (2018) Laos. [online] Available from: https://www.ifad.org/web/operations/country/id/laos [Accessed on 10th October 2018] UNDP (2017). United Nations Develpment Programme, Human Develpment Reports [online] Available from: http://hdr.undp.org/en/countries/profiles/KHM [Accessed on 10 October 2018] Woodford, C (2018) GIS [online] Available from: https://www.explainthatstuff.com/howgpsworks.html [Accessed on 15 October 2018] World Bank (2015) Cambodian Agriculture Transition: Opportunities and Risks. World Bank Group. Economic and Sector Work. [online] Available from: http://documents.worldbank.org/curated/en/805091467993504209/pdf/96308- ESW-KH-White-cover-P145838-PUBLIC-Cambodian-Agriculture-in-Transition.pdf [Accessed on 10th October 2018] World Bank (2018a) Cambodia data [online] Available from: https://data.worldbank.org/country/cambodia [Accessed on 12th March 2018] World Bank (2018b) Cambodia Agriculture [online] Available from: http://www.worldbank.org/en/country/cambodia/publication/cambodian- agriculture-in-transition-opportunities-and-risks [Accessed on 12th March 2018] World Bank (2018c) Vietnam [online] Available from: http://www.worldbank.org/en/country/vietnam/overview [Accessed on 12th March 2018] World Population Review (2018) 2018 World Population. [online] Available from: http://worldpopulationreview.com/ [Accessed on 2nd April 2018] 81 Annexes 1 PROTOCOL (participatory GIS for land use mapping) Context Climate change is anticipated to result in land use change, including a spatial and temporal reshuffling of crops. Changes can be monitored based on solid baseline data and key metrics from key localities that serve as benchmark sites for timeline comparison (i.e. the Climate Smart Villages). Participatory GIS (pGIS) is a tool that can help to collect high-resolution spatial data based on local knowledge and perceptions. Products derived from pGIS can aid local people in decision making and land use planning if local concerns are incorporated into survey from the onset. Objectives (i) To construct robust baseline landuse maps and indicators to systematically monitor future change (ii) To valorize and build on local knowledge for landuse monitoring and planning (iii) To generate local capacity to conduct applied research and guide landuse planning Method (in a nutshell) - Step 1: Shared objective setting with partners and prior informed consent The initiative to conduct pGIS is shared with village authorities and the local population. The main objectives and potential outputs are clearly explained. Local authorities and people are stimulated to add additional research questions, information needs or concerns to the exercise. - Step 2: Training of local team of pGIS surveyors A team of local surveyors, ideally youth, from the village(s) are trained in the methodology. Particularly, the use of GPS equipment, data collection and registration, and mechanisms to coordinate field visits with farmers. Trainings events take 3 to 4 days and are hands-on. - Step 3: Implementation and data collection HOUSEHOLD LEVEL. Based on the sample size (n = households) and the size of the survey team, each team member gets a fixed number of households assigned. Each team member coordinates visits to all individual fields of each of the assigned households. Basic household data are registered in the “household data sheet” (annex A), coordinates of each field in a “fields coordinates sheet” (annex B), and standard field characteristics in a “field properties sheet” (annex C). Data are registered based on farmer knowledge or perceptions, as well as surveyor observations. COLLECTIVE LEVEL. Communal spaces (e.g. pastures, forests), community boundaries and other landmarks are drawn on high-resolution satellite image print outs. The exercise can be done with a group of village elders or authorities who draw the spaces with markers on the map. - Step 4: Data analysis 82 Data is initially organized in databases that are searchable by code, household or field. Data analysis includes descriptive statistics and key indicator computing. The georeferenced data can be made visually explicit on thematic maps. - Step 5: Sharing and discussion of results A key step involves the sharing of results. Initially with the local authorities and villagers to generate discussion and give back information (including maps) that may serve or inform local landuse planning. Eventually results can also be made available in the public domain through the publication of baselines or journal articles. Samples size Normally 100 to 200 households that are randomly selected, depending on the extent of the area being mapped, the total human population size and resources available for the actual pGIS surveys. Key metrics that can be established (examples) Social data  Number of livestock heads per by species, household  Education level and distribution in the village (without, primary, secondary, superior)  Labor type involved in family farmer (family, contracted, reciprocal)  Others …. Household level crop data  Land property size by household (m²)  Land property dispersion (no field / household)  % + area (m²) per crop species by household  Number of crops species / varieties per household  Others …. Village-level landuse data  % + area (m²) of irrigation and rain-fed land  % + area (m²) decicated to annual and perennial crops  % + area (m²) dedicated per crop species / variety  % + area (m²) under external input use  % + area (m2) prone to different climate effects (drought, lodging, etc.)  Crop species area distribution by altitudinal belt  10 years rotation design and trends  Others …. Expected outputs  High resolution thematic maps of landuse practices  Robust baseline metrics of landuse issues at the household and community level  Quality maps showing landuse categories and issues 83 Annex 1: Household Data Sheet HOUSEHOLD DATA SHEET Household Code Basic data Community Sector Family name Family data Male head household (name) Age – male head Female head household (name) Age - female head Education family heads Male head Without formal education Primary complete Secondary complete Primary incomplete Secondary incomplete Superior Female head Without formal education Primary complete Secondary complete Primary incomplete Secondary incomplete Superior Number of adult household members (> 18 yrs) Number of non-adult household members (< 18 yrs) Data crops (present season) Number of plots with annual crops Number of plots with perennial crops Family Economy Family comes originally from the village? Yes No Head of livestock Number Number Number Horses Sheep Rabbits Cows Pigs Fish Goats Chicken Other _____ Buffaloes Ducks Other _____ Predominant distribution of agricultural labor Family Contracted ($) Mutual help (minga) Is the household also involved in off-farm employment? Yes No 84 Key Field Properties Registration Sheet FIELD PROPERTIES REGISTRATION SHEET Household code Field code Basic field data Crop(s) Annual / perennial Number of varieties Type of varieties (improved / local or landrace) Date of planting Harvest date(s) Name of sector Travel time from field to home Min. __________ / Transport ___________ Travel time from field to road Min. __________ Type of soil tillage Full tillage Minimal tillage No tillage Do you use agrochemical on this field (pesticides, fertilizers, etc.)? Yes No Property type (mark) Family Communal State Other _____ Irrigation (mark) Rain-fed Irrigated (cannel) Irrigated (pump) Slope (mark) Flat Modest slope Strong slope Farmers perceived climate sensitivity of field (mark *) Drought prone Heat exposed Other _____ Water lodging Wind exposed Other _____ Cold stress Other farmer perceived risks for this field (mark *) Pest ____________ Sedimentation Other ______________ Diseases ________ Salinity Other ______________ Soil erosion Poor fertility Other ______________ Source of planting material for this field 85 Own Exchanged Bought Subsidized Primary final use of produce from this field (mark) Home consumption Sales Both consumption - sales Additional uses for the produce from this field (mark) Planting material Processing Other ______ This year’s content (2015) Crops (or fallow / trees) Current crop C B A Past field content (2005-2014) ** Crops (or fallow / trees) Crops (or fallow / trees) 2014 C 2009 C B B A A 2013 C 2008 C B B A A 2012 C 2007 C B B A A 2011 C 2006 C B B A A 2010 C 2005 C B B A A Future field content (2016) Crops (or fallow / trees) 2016 A B C * = Only consider plot or field specific factors!! Not general factors that affect all crops or production zones ** = go back to the year that the farmer can still remember 86 Annexes 2 Participatory GIS for land use mapping A) HOW TO USE YOUR GPS DEVICE ? Installing battery: Turn the D-ring counter- clockwise, and pull up to remove the cover 1) Switch it on/off: Press the button 4 seconds and release it 2) Check the battery: Make sure you have enough battery (check the green bars) Press the button 3) Check the satellite signal: Make sure you have the five bars in green !! It may take 10 seconds to find the satellites and show 5 bars !! 87 B) HOW TO MAP YOUR FIELD PLOT ? Take as many GPS points as possible to get the perfect shape of your field plot. For example, at every corner like in the example below. Make sure to enter the same GPS name for every point. 4) Create a GPS point: By pressing the "MARK" button. Make sure to be in the right spot 5) Fill the information page: Use the down arrow and press the "ENTER" button to move in the page Name your GPS point by writing Household Id and field id . Ex: 001-1 (which means Household 1, field number 1) In the note section, write your initials. For example, Huong Pham = HP Press the "DONE" button when you are finished Field plot GPS point moving buttons 88 C) HOW TO MAINTAIN YOUR GPS DEVICE ? - Set to power save mode - Don’t keep it switched on over night - Reduce the luminosity 89 C) HOW TO EXPORT YOUR GPS DATA TO A GIS SOFTWARE ? 1) Install the QGIS software : https://www.qgis.org/en/site/forusers/download.html Select the QGIS Standalone Installer Version (32 or 64 bit depending on your computer). Then follow the installing procedure. 2) Connect your GPS to your computer via the USB cable Click on the device that appears / Click on the Garmin folder / Click on the GPX folder / Find your GPS points stored in a daily file (in .gpx format). 3) Open QGIS software. Drag and drop your gpx file directly into QGIS 90 Select to import only waypoints and tracks. 4) Save your GPS points Click right on the name of your waypoint file. Then click on the "save as" button. A box appears and you need to enter the pathway where you want to save the data. Also select the right SCR. Do not worry about the other options. Click on ok when you are done. 91 You can save the same way the track file. Pathway SCR 92 Annexes 3 Point to Polygon conversion. (visualized in QGIS 3.2)