システム農学 (J. JASS), 30(4) ： 131 ～ 141, 2014 131

Contributed paper

The Traditional Practice of Swidden Cultivation in the Djarai Ethnic Communities of Northeastern Cambodia

Sanara HOR*, Kei MIZUNO*, Shintaro KOBAYASHI** and Tsugihiro WATANABE* * Graduate School of Global Environmental Studies, Kyoto University, Japan ** Tottori University of Environmental Studies, Japan

(Received 23 April 2014; in final form 25 August 2014)

Summary

Djarai ethnolinguistic groups, who live in Ratanakiri Province of Cambodia, engage in the traditional practice of swidden cultivation. However, recent economic developments, such as the shifts toward large-scale plantations and cash crop production at the household level, have affected those swidden cultivators. In order to discuss the sustainability and adequacy of such new developments, it is important to clarify the traditional practice of swidden cultivation. This study examines traditional practices, farming activities, and land-use patterns related to Djarai swidden cultivation. It utilizes remote sensing and qualitative research methods and applies supervised classification and visual image analysis to ALOS AVNIR-2 and Worldview-1 data. The study also uses qualitative methods to collect and analyze in-situ information. Remote sensing and ground truthing techniques prove that Djarai swidden cultivation is interwoven with cash crop production. At the same time, qualitative methods reveal that Djarai swidden cultivation is a multiple cropping system that works in close relation with their traditional belief and land tenure practice. These traditional farming systems have been transformed due to development activities undertaken in the province, which have led Djarai villagers to change not only their farming system, but also the social system of their village. Key words: Djarai ethnic group, Land-use pattern, Northeastern Cambodia, Swidden cultivation, Traditional practice

1. Introduction that land uses and total tree covers remained stable in the period of 1953 to 1996 in many parts of Ratanakiri Many researchers have studied the traditional Province. practices of shifting cultivation and have reported Ratanakiri Province opened for development in the various problems caused by it, particularly the 1990s. Since then, swidden cultivators have been deforestation of some areas (Angelsen 1995, affected by recent economic development (Colm 1997). Goldammer 1988, Mittelman 2001). The Cambodian Large and small-scale land acquisitions have seriously northeastern province of Ratanakiri is the home of threatened indigenous livelihood (NGO Forum 2006a, ethnic minority groups that have practiced a traditional 2006b, Rosette and Michaud 2005). Recent economic farming system of shifting cultivation (Hean and Monie development has generated market pressure that 2002). Shifting cultivation, slush and burn agriculture, encouraged swidden cultivators to engage in new and and swidden cultivation are synonymous uses, but different forms of commercial agriculture, which resulted swidden cultivation describes the rice-based system in in major changes in land use (Fox 2002). Southeast Asian (Mertz et al, 2009). Fox (2002) finds The results of previous studies raise a series of questions: What is traditional swidden cultivation? In * Yoshida Hommachi, Sakyo-ku, Kyoto 606-8501, Japan practice, what kind of crop species are the swidden ** Wakabadai-kita, Tottori, 689-1111, Japan cultivators growing? What is the reality of swidden (Correspondence: [email protected]) cultivation? The goal of this study is to examine the 132 システム農学 (J. JASS), 30(4), 2014 traditional practice of swidden cultivation, farming Province had a population of 94,243, which increased to activities, and land-use patterns related to livelihood of 150,466 in 2008 (National Institute of Statistic 1998, local people who belong to the Djarai ethnolinguistic 2008). The annual growth rate was 4.6%. Ten ethnic group. minority groups comprised 62.6% of the provincial population. The province has a total area of 8,431.32 km2 2. The definition of shifting cultivation and borders the province of Mondol Kiri to the south and the province of Stung Treng to the west. Virachey Shifting cultivation has been prevalent since the National Park is located in the northern part of the Neolithic era, and it can be best understood by province. Lumphat Wildlife Sanctuary and Nsok examining its environmental, temporal, and cultural Protected Forest are located in the south. The Sesan and dimensions (Conklin 1961). The environmental Srepok Rivers cross the province flowing from western dimension refers to biotic, edaphic, and climatic Vietnam to the Sekong River, a tributary of the Mekong considerations; the temporal dimension refers to a series River. of swidden practices; and the cultural dimension includes technological, social, and ethnoecological considerations (Conklin 1961). In 1982, the Food and Agriculture Organization and the University of Ibadan defined shifting cultivation, which is also referred to as the swidden farming system, as “a system in which a short period of cultivation is followed by a long period of fallow” (Lanly 1985). Shifting cultivation was also found to be closely related to the natural vegetation-soil complex and cultivators (McGrath 1987). Furthermore, various groups have changed the meaning of shifting cultivation in order to support specific political interests (Jarosz 1993), and it is currently defined according to the characteristics of actual land-use practices (Upadhyay 1995). Therefore, the meaning of shifting cultivation conjoins environmental, chronological, cultural, political, land-use, and livelihood strategies in an area. Based on a series of questions and literature reviews presented in the previous section, we were provided a hypothesis, which mainly focuses on the origin of swidden cultivation including traditional belief and customary tenure. The hypothesis is “Djarai swidden cultivation is a multiple cropping system interweaving with traditional belief and tenure.” This study was conducted in Phi village, which is located in the Sesan commune of the Ouyadav district. 3. The study area The village is located 61 km east of Banlung, the provincial capital, and borders Vietnam. The village has 3.1 Introduction to study area a population of 632 people spread throughout 102 households (location illustrated in Fig. 1). There were Ratanakiri Province, lied about 600 km northeast of three reasons for selecting this village as the study area. Phnom Penh, is a mountainous area with elevation First, swidden cultivation fields were found around the ranging from about 70 m to 1,620 m. In 1998, Ratanakiri village, and villagers were maintaining the traditional Sanara Hor et al. ： Traditional Practice of Shifting Cultivation in Northeastern Cambodia 133 practice of swidden cultivation (see Fig. 2). Second, the villagers were maintaining the land tenure system in the Raw images DTM traditional manner. Third, external economic activities Georeference and orthorectification related to land development had been introduced by outsiders that challenged local people’s use of land

Pre-classification Calibrated images resources.

Designed classification scheme

Supervised Classification (AVNIR-2) Classification

New layer Smoothing

Accuracy Final result assessment Post-classification The field survey separated into three periods. The first 㻲㼕㼓㻚㻟㻌㻵㼙㼍㼓㼑㻌㼏㼘㼍㼟㼟㼕㼒㼕㼏㼍㼠㼕㼛㼚㻌㼜㼞㼛㼏㼑㼐㼡㼞㼑 period was between February 26 2012 and March 18 2012. The first-field survey was to gather baseline First, we performed image geometric rectification, information related to swidden cultivation. The second which included georeferencing and orthorectification. period started from August 20 2012 to September 20 The study performed image direct georeference on 2012 to collect training samples and interviewed Djarai ALOS PRISM and Worldview-1 data with 10 GCPs for villagers. The third-field survey took place between 3 polynomials. Then, we georeferenced the ALOS November 15 2012 and December 25 2012 in order to AVNIR-2 data based on the rectified ALOS PRISM verify the previous information. data. We generated a DTM using raster interpolation tools before we performed the image orthorectification. 4. Research methods We created the DTM based on ASTER GDEM and topographical map data. The topographical map data 4.1 Remote sensing and the GIS approach was produced by the Cambodian Ministry of Land This study used remote sensing and GIS approaches to Management, Urban Planning, and Construction and the outline the exact amount of land use and land cover Ministry of Public Works and Transport. All raster and (LULC) by utilizing three types of satellite images: vector data used in this study was projected to the UTM ALOS AVNIR-2, ALOS PRISM, and Worldview-1’s coordination system (UTM WGS1984 zone 48N panchromatic data. Land use (LU) refers to human meters). activities undertaken in a certain land cover (LC), which Second, we developed a classification scheme before refers to the observed biophysical or physical cover on we conducted supervised classification (Anderson et al. the earth’s surface (Gregorio and Jansen, 1998). ALOS 1976) with 273 training samples, which were AVNIR-2 data were used for deriving vegetation covers. categorized into 11 classes of land use. We randomized ALOS PRISM data were used for supporting ground 102 points for accuracy assessment and 171 points for truthing. Worldview-1 data were used for understanding signature development. Every band (band 1 through agricultural land uses. Both ALOS AVNIR-2 and PRISM band 3) of ALOS AVNIR-2 data was used for data were taken on February 16, 2011, while Worldview- supervised classification. The process of supervised 1 data were taken on February 13, 2012. classification was repeated several times until results The image interpretation was separated into three were consistent with the training samples. steps: (1) pre-classification, (2) classification, and (3) Third, a statistical filter was applied to eliminate post-classification (see in Fig. 3). clump areas smaller than 3u3 pixels, after which we 134 システム農学 (J. JASS), 30(4), 2014 performed an accuracy assessment. The results of this 5. Results and Discussions accuracy assessment are presented as an error matrix (Foody 2002). (see Table 1). In this study, we conducted supervised classification with the maximum likelihood classification method 4.2 Qualitative research because there were many types of land use and land cover We used the qualitative method to investigate human in the study area observed during the preliminary study. activities related to the agricultural system. We followed The image interpretation revealed the exact land use and Maxwell (2013) in designing our qualitative method and land cover quantitatively (Table 1). As the accounting of focused on research questions, goals, conceptual land uses is limited to direct human-induced usage of framework, and validity. land, the distinction between natural and human activities Individual interviews were conducted with 30 key cannot be identified using remote sensing data alone; as informants during field surveys through a semi- such, local information may be required in addition to the structured questionnaire in order to understand the reality of swidden cultivation (Rosenqvist et al. 2003). swidden cultivators’ activities, and group discussions Through the field survey, we could fully determine how were also conducted to clarify land-use patterns in detail. the lands are used (Maxwell 2013). Participatory rural appraisal methods, including participatory mapping, flow diagrams, and seasonal 5.1 Image classification results calendars, were partially adopted during interviews Defined through the results of the image classification (Chambers 1994, Narayasamy 2009). of ALOS AVNIR-2, the land areas of each land-use type We analyzed the qualitative data based on five steps: are identified in Table 1, and the LULC map is presented (1) initial coding and memo writing, (2) focused coding in Fig. 4. As a result of the accuracy assessment, a good and memo writing, (3) new data collection via theoretical signature reparability was achieved for each classified sampling, (4) continued coding, memo writing, and image. The overall accuracy value and overall Kappa theoretical sampling, and (5) shortening and integrating statistics were 89.2% and 0.8 respectively. memos (Charmaz, 2006).

㼀㼍㼎㼘㼑㻌㻝㻌㻱㼞㼞㼛㼞㻌㼙㼍㼠㼞㼕㼤㻘㻌㼍㼏㼏㼡㼞㼍㼏㼥㻌㼠㼛㼠㼍㼘㼟㻘㻌㻷㼍㼜㼜㼍㻌㼟㼠㼍㼠㼕㼟㼠㼕㼏㼟㻌㼍㼚㼐㻌㼕㼙㼍㼓㼑㻌㼏㼘㼍㼟㼟㼕㼒㼕㼏㼍㼠㼕㼛㼚㻌㼞㼑㼟㼡㼘㼠㼟 Classified PA Area FA PF SF DF EF MF SC SH MA SS WB RT Kappa Data (%) (ha) FA 61 1 0 0 1 0 0 0 0 0 0 63 92.4 0.9 3,327 PF 0 6 0 0 0 0 0 0 0 0 0 6 85.7 1 297 SF 4 0 5 0 0 1 0 1 0 0 0 11 100 0.4 2,685 DF 0 0 0 2 0 1 0 0 0 0 0 3 100 0.6 2,923 EF 1 0 0 0 5 0 0 0 0 0 0 6 83.3 0.8 7,200 MF 0 0 0 0 0 1 0 0 0 0 0 1 33.3 1 3,540 SC 0 0 0 0 0 0 1 0 0 0 0 1 100 1 3,847 SH 0 0 0 0 0 0 0 0 0 0 0 0 0 0 104 MA 0 0 0 0 0 0 0 0 1 0 0 1 100 1 17 SS 0 0 0 0 0 0 0 0 0 5 1 6 100 0.8 280 WB 0 0 0 0 0 0 0 0 0 0 4 4 80 1 467 CT 66 7 5 2 6 3 1 1 1 5 5 102 24,687 UA 96.8 100 45.4 66.7 83.3 100 100 0 100 83.3 100 The overall accuracy value: 89.2% and the overall Kappa statistics: 0.8

Note: FA = Farmlands, PF = Paddy fields, SF = Swidden/Fallow, DF = Deciduous forest, EF = Evergreen forest, MF = Mixed forest, SC = Secondary forest, SH = Shrub, MA = Marsh, SS = Sand and stone, WB = Water body, CT = Column total, RT = Row total, PA = Producer’s Accuracy and UA = User’s Accuracy Sanara Hor et al. ： Traditional Practice of Shifting Cultivation in Northeastern Cambodia 135

the gods offer them that land. After this, they return to the 5.2 Belief village or sleep next to the area to wait for the gods’ Table 2 presents the seasonal calendar related to response. The response can be provided via a dream or livelihood activities of Djarai swidden cultivators. The an animals’ cry. They could cultivate the forest land if Djarai’s swidden system is related to local belief. The they have a good dream of fishing, rowing a boat, or swidden cultivators perform rituals at each moment from crossing a desert. If they do not have a bad message selecting the forest to post-harvest activities. Before within seven days, they could develop the forestlands. clearing the forest, they have a meeting to find the new In contrast, they could not develop the forest land if area in the forest that they will use. They knock on the they have an inappropriate dream within 7 days, which ground and whisper, “We are going to use this plot” when might reference sexual affairs or particular animals such they find the proper plot. This statement is a request that as deer, tigers, or sacred birds that had been shot or died 136 システム農学 (J. JASS), 30(4), 2014

Table 2. Seasonal livelihood activities of Phi villagers

Months May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr Land preparation SF* SF* CF B Rice ceremonies 1st 2nd 3rd 3rd LE* LE* 1st Rice variety P M M M H 2nd Rice variety P M M M M H 3rd Rice variety P M M M M M M M H Vegetable species P M M H Cassava P M M M M M M M M H Legend: SF = Selecting forest land, CF = Clearing forest land, B = Burning, 1st through 3rd = Numbers of rice ceremonies, LE = Land expansion, * = Rituals related to Djarai land tenure practice, categorized by growing period, P = Start planting, M = Managing farmland, H = Harvesting  in that land. Villagers who experience a bad dream search chores and crop management, including weeding and for a different location. harvesting. They are also responsible for selecting In the cropping stage, they perform ceremonies that are varieties for the next season. Children assist their mothers divided into three rites known as the Rice Ceremony. The depending on their abilities. first ritual is conducted when the rice’s height is about 40 Weed competition is a fundamental problem for the to 50 cm, and the second ritual takes place during the farmers because traditional swidden cultivation only uses flowering stage. Those rituals call the gods to protect hand tools such as hatchets, blades, or hoes to manage their property. The third ceremony is performed after the weeds. The villagers recognize that one effective solution rice is harvested and before the rice is put in the would be to import some agricultural chemicals such as warehouse. That ritual thanks the gods and asks them to herbicide or pesticide. provide sufficient food in the following year. Even though they need cash to buy those agricultural To expand the land, they perform another ritual. The chemicals, the agricultural products that are harvested new land can be identified by a spiritual stick that divides from typical swidden plots are not bountiful enough to the old and new lands. This series of rites is outlined in sell. As a result, some farmers have introduced Table 2. production of cash crop from neighboring villages, in order to increase their family’s income. 5.3 Crop decision and farm management They sell cash crop products at local markets or export Swidden cultivators grow different crop species. them to outside markets in Vietnam. In turn, they import During the field survey, sixteen crop species were daily commodities. Their intuitive decision has formed a identified, including cucumber, sesame, gourd, pumpkin, mobility system (see Fig. 5) for trading agricultural mung bean, sweet potato, cassava, maize, pineapple, products. Chinese long bean, eggplant, banana, banana rough, chili, ridge gourd, and sponge gourd. At least five crop species 5.4 Djarai land tenure practice were found in every swidden plot (see cropping calendar The forest is one of the main sources of food. The in Table 2). This crop diversity creates the seasonal swidden cultivator collects non-timber forest products activities that help establish a more resilient livelihood. (NTFP), while wood products are commonly used for Such multiple cropping systems are parts of traditional housing. Those products stabilize the livelihoods of the land-use practices. swidden cultivators. The villagers carefully use forest In daily farm management, men are responsible for and other land resources according to their traditional collecting NTFP, fishing, hunting, and preparing zoning system, which consists of five categories: (1) swidden plots. Women are responsible for household housing area, (2) funeral forest, (3) spiritual forest, (4) Sanara Hor et al. ： Traditional Practice of Shifting Cultivation in Northeastern Cambodia 137

Market in Angdong Meas and Ouyadav

Phi villagers Nearby 䋾 Market in village Vietnam Phi villagers

Market in Banlung

Selling agricultural Accessing to hospital, products and buying meeting, or buying agricultural materials daily commodity (rarely) 䋽 Buying daily Trading daily commodity commodity within neighbour villagers

㻲㼕㼓㻚㻡㻌㻹㼛㼎㼕㼘㼕㼠㼥㻌㼐㼕㼍㼓㼞㼍㼙㻌㼛㼒㻌㼠㼔㼑㻌㻰㼖㼍㼞㼍㼕㻌㼢㼕㼘㼘㼍㼓㼑㼞㼟 a proposal to the owner. The village elders are usually the witness of such land ownership transformations. agricultural land, and (5) fallow land. Moreover, land without an owner can be shared among The housing area can be relocated if villagers become the villagers. seriously sick, houses are hit by lightning, or animals are a threat. The villagers also relocate when they start 5.5 Land-use patterns swidden cultivation in another area. The spiritual forest In addition to the ALOS AVNIR-2 data, a ground is the holy place of their animistic belief. The funeral truth survey and visual image classification of forest, located near the village, is utilized as a cemetery. Worldview-1 data were utilized in order to understand The spiritual and funeral forests can be accessed, but land-use patterns more concretely (Fig. 7). A mosaic farming and timber harvesting there are not allowed. The landscape of agricultural land in the study area agricultural and fallow areas are located around the displayed the land-use patterns of indigenous village and along the Sesan River (see Fig. 4). The fallow communities, migrant communities, and economic land lands are partly used by villagers for hunting. The land concession (ELC). boundary of each category of zoning is not clearly Two types of rice production systems were found in defined. this area—wet season and upland rice farming. Wet Traditionally, transferals of land ownership take place. season rice farming began in the Pol Pot Regime between The current cultivation and fallow plots are properties 1975 and 1979. The paddy fields were located in the held by individuals. For example, land can be passed western portion of the village (Fig. 8), where only seven from the older generation to the younger generation. Fig. villagers conducted this farming system from 1979 to 6 provides an overview of land expansion and ownership. 1997. After 1997, the areas were abandoned. In 2011, the young family (a couple who married recently In 2001, the Cambodian government re-enacted land and needed land resources to sustain their family) law. The law allowed the government to claim lands for inherited 0.82 ha of swidden plot from her father a year economic purposes. In 2006, an agricultural development after married. There was no land title. The witnesses were company signed an agreement with the Cambodian the village elders. Ministry of Agriculture, Forestry, and Fisheries to receive In addition, landed families can transfer a plot of land ELC for planting industrial crops, and it received 8,654 ha to landless families. To get land, landless families make of land to grow acacia or rubber. In the present, the 138 システム農学 (J. JASS), 30(4), 2014

until they can collect cashew fruit, and some villagers will try to expand their land without limitation. From 2006 to the present, Cham-people from different parts of Cambodia migrated to Tang Se village, which is located in the northern part of Phi village. According to the interview, the migrants came to this area for two main reasons. First, some of them, after returning from work in Malaysia and Thailand, sought private property for conducting agricultural activities. Second, some people who lived along the Mekong River (in the Kampong Cham Province) and whose livelihood depended on rice-based farming and fishing sought access to land resources when their aquatic resources became limited.

5.6 Transformation of swidden cultivation During the interviews, seven interviewees among thirty key informants responded that they had moved away from their traditional belief. One interviewee stopped performing rituals because he transformed his rice farm into a cash crops farm that grew cassava and cashew. Two families decided to stop performing ritual and introduce new agricultural technique because they utilize new land-use practices to increase agricultural products rather than spiritual belief. Five families became Christians. The Christian farmers stopped performing ritual, and they did not have a spirit forest or funeral forest. Land tenure practice has also changed since the introduction of cash crops in 2002. The transferal of company develops forest land for rubber plantations land ownership, which was previously handled among (Fig. 9). neighbor villagers and family members, has In 2002, the villagers found a new territory located disappeared due to land alienation. This situation north of Phi village that local people called the “Red encouraged a drastic influx of rural migrants into this Soil Area.” In the beginning, there were five families in area. Poffenberger (2009) argues that migrants moved the new area to develop the forest land into a cashew in large numbers into Oddar Meanchey Province, in farm. During the field observation, 34.9% of the village northwest Cambodia, and in the process increased the households practiced cash crop farming. Older cashew rural population and made land resources scarce. At the farms were found, and they were reportedly decreasing same time, Fox et al. (2009) find that indigenous people their production. In the first year, Phi villagers viewed land as a market commodity. The change in land developed forest land into cashew farms by integrating tenure practice affected land resources because Djarai rice and cashew in the same plot (Fig. 10). In the second villagers previously used land resources as common- year, swidden cultivators expanded the land, and they pool resources. In the theory of common-pool resources, were directly seeded with rice and cashews. Reportedly, the resource users have the right to sell or lease their some Djarai villagers will keep expanding their land collective properties (Schlager and Ostrom 1992). Sanara Hor et al. ： Traditional Practice of Shifting Cultivation in Northeastern Cambodia 139

Swidden cultivation is under transformation in the Acknowledgements Phi village. We found that Djarai swidden cultivation This work was supported by JSPS KAKENHI Grant practice has been transforming into mono-cropping Number 21405032. We appreciate Dr. Izuru Saizen and systems of agriculture. This transformation of their Dr. Narumasa Tsutsumida of Kyoto University for their farming system was not suitable for forest stewardship, encouragement. Additionally, we thank the Royal as observed elsewhere (Fox et al. 2000, Hariyadi and University of Agriculture (RUA) and the Provincial Ticktin 2012, Rerkasem et al. 2009). In addition to the Department of Agriculture of Ratanakiri Province that transformation of shifting cultivation, Cramb et al. facilitated and provided us vital information. We also (2009) argue that although cash crops might improve acknowledge the research assistance given by Mr. Thea the livelihoods of indigenous people, shifting SIVE (RUA’s staff) and Mr. Vinichhai VOERN’s cultivation can present a safety net in the face of market (RUA’s student) work as a community facilitator. fluctuations. Today, Djarai villagers are managing both traditional swidden cultivation and cash production. References Therefore, villager’s activities will increase pressure on Anderson, J.R., Hardy, E.E., Roach, J.T., and Witmer, land resources. R.E., 1976, A Land Use and Land Cover Classification System for Use with Remote Sensor 6. Conclusion Data. Geological Survey Professional, Vol. 964, pp. 1–41. Djarai swidden cultivation systems are multiple Angelsen, A., 1995, Shifting cultivation and cropping systems that intertwine traditional belief and “deforestation”: A study from Indonesia. World tenure. Recently, we found that it has been transformed Development. Vol. 23, No. 10, pp. 1713–1729. into mono-cropping systems of farming. A change in Chambers, R., 1994, The origins and practice of any element that composes the traditional farming participatory rural appraisal. World Development, system can cause other problems. There are at least two Vol. 22., No. 7, pp. 953–969. features of composing elements to be considered. Charmaz, K., 2006. Constructing grounded theory: a First, in farm management, Djarai swidden practical guide through qualitative analysis. SAGE, cultivation systems have the same temporal dimension London. in many areas. The swidden change farming systems Colm, S., 1997, Land rightsௗ: the challenge for’ integrate cashew and rice in the same field. The Ratanakiri's indigenous communities. Watershed, villagers expand their land every year in order to Vol. 3, pp. 1–9. produce rice. Some of them attempt to expand their land Conklin, H.C., 1961, The study of shifting cultivation. without limiting themselves. Current Anthropology, Vol. 2, No. 1, pp. 27–61. Second, in land tenure practice, Djarai villagers Cramb, R. A., Colfer, C.J.P., Dressler, W., transfer land ownership between neighboring villagers Laungaramsri, P., Le, Q.T., Mulyoutami, E., Peluso, or family members. Recently, land has been used as a N.L., and Wadley, R.L., 2009, Swidden commodity market. According to common-pool transformations and rural livelihoods in Southeast resource theory, villagers have the right to dispose of Asia. Human Ecology, Vol. 37, No. 3, pp. 323-346. their land, but land under open-access resources is Foody, G. M., 2002, Status of land cover classification vulnerable to forest degradation. accuracy assessment. Remote Sensing of Finally, the combination of the remote sensing Environment, Vol. 80, No. 1, pp. 185-201. technique and qualitative survey applied in this study Fox, J., 2002, Understanding a dynamic landscape: land proved to be a useful approach and will be increasingly use, land cover, and resources tenure in northeastern helpful in understanding swidden cultivation for the Cambodia. In Linking people, place, and policy: a designing area. The results of this study proved that the GISscience approach, edited by Walsh, J.S., and causes and mechanisms of change in land use and land Crews-Meyer, K.A., Springer, New York, pp. 113– cover should be in-depth analyzed. 130. 140 システム農学 (J. JASS), 30(4), 2014

Fox, J., Troung, D.M., Rambo, A.T., Tuyen, N.P., CUC, and importance. Tropical Forest Science. Vol. 13, No. L.T., and Leisz, S., 2000, Shifting Cultivation: A 4, pp. 671–690. New Old Paradigm for Managing Tropical Forests. Narayasamy, N., 2009, Participatory rural appraisal: BioScience, Vol. 50, No. 6, pp. 521–528. principles, methods and application. SAGE, New Fox, J., Vogler, J.B., and Poffenberger, M., 2009, Delhi, India. Understanding Changes in Land and Forest Resource National Institute of Statistic, 1998, General Management Systems: Ratanakiri, Cambodia*. Population Census. In Southeast Asian Studies. Vol. 47, No. 3, pp. 309–329. http://celade.cepal.org/redkhm/census/khm1998 Goldammer, J.G., 1988, Rural land-use and wildland National Institute of Statistic, 2008, General fires in the tropics. Agroforestry Systems. Vol. 6, No. Population Census. In 2, pp. 235–252. http://celade.cepal.org/redkhm/census/khm2008 Gregorio, A. Di, and Jansen, L.J.M., 1998, Land cover NGO Forum, 2006a, Land Alienation in Indigenous classification system (LCCS): Classification Minority Communities - Ratanakiri Province, concepts and user manual. FAO, Rome, Italy. Cambodia. Phnom Penh, Cambodia. Hariyadi, B., and Ticktin, T., 2012, From Shifting NGO Forum, 2006b, Indigenous people in Camboda. Cultivation to Cinnamon Agroforestry: Changing Phnom Penh, Cambodia. Agricultural Practices Among the Serampas in the Poffenberger, M., 2009, Cambodia’s forests and climate Kerinci Seblat National Park, Indonesia. Human change: Mitigating drivers of deforestation. Natural Ecology. Vol. 40, No. 2, pp. 315–325. Resources Forum, Vol. 33, No. 4, pp. 285–296. Hean, S., and Monie, T., 2002, Indigenous peoples: Rerkasem, K., Lawrence, D., Padoch, C., Schmidt-Vogt, ethnic minorities and poverty reduction, Cambodia. D., Ziegler, A.D., and Bruun, T.B., 2009, ADB, Manila, Philippines. Consequences of Swidden Transitions for Crop and Jarosz, L., 1993, Defining and explaining tropical Fallow Biodiversity in Southeast Asia. Human deforestation: shifting cultivation and population Ecology, Vol. 37, No. 3, pp. 347–360. growth in colonial Madagascar (1896-1940). Rosenqvist, Å., Milne, A., Lucas, R., Imhoff, M., and Economic Geography, Vol. 69, No. 4, pp. 366–379. Dobson, C., 2003, A review of remote sensing Lanly, J.P., 1985, Defining and measuring shifting technology in support of the Kyoto Protocol. cultivation. Unasylva, Vol. 37, pp. 17–21. Environmental Science & Policy, Vol. 6, No. 5, pp. Maxwell, J.A., 2013, Qualitative research design: an 441–455. interactive approach. SAGE, California. Michaud, A., and Rosette, J., 2005, Indigenous land Mertz, O., Padoch, C., Fox, J., Cramb, R. A., Leisz, S. J., crissis in Ratanakiri. NGO Forum on Cambodia, Lam, N. T., Vien, T. D., 2009. Swidden change in Cambodia. Southeast Asia: understanding causes and consequences. Schlager, E., and Ostrom, E., 1992, Property-rights Human Ecology, Vol. 37, No. 3, pp. 259-264. regimes and natural resources: a conceptual analysis. McGrath, D.G., 1987, The role of biomass in shifting Land Economics, Vol. 68, No. 3, pp. 249–262. cultivation. Human Ecology. Vol. 15, No. 2, pp. 221– Upadhyay, K.P., 1995, Shifting cultivation in Bhutan: a 242. gradual approach to modifying land use patterns. Mittelman, A., 2001, Secondary forests in the Lower FAO, Rome, Italy. Mekong Subregion: an overview of their extent, roles Sanara Hor et al. ： Traditional Practice of Shifting Cultivation in Northeastern Cambodia 141

◊✲ㄽᩥ  䜹䞁䝪䝆䜰໭ᮾ㒊䛻䛚䛡䜛ᑡᩘẸ᪘䝆䝱䝷䜲䛻䜘䜛ఏ⤫ⓗ↝⏿⪔స The Traditional Practice of Swidden Cultivation in the Djarai Ethnic Communities of Northeastern Cambodia  ி㒔኱Ꮫ኱Ꮫ㝔ᆅ⌫⎔ቃᏛ⯋㻖㻌 Sanara HOR ி㒔኱Ꮫ኱Ꮫ㝔ᆅ⌫⎔ቃᏛᇽ㻖㻌 Ỉ㔝㻌ၨ㻌 㫽ྲྀ⎔ቃ኱Ꮫ⎔ቃᏛ㒊㻖㻖㻌 ᑠᯘ឵ኴ㑻㻌 ி㒔኱Ꮫ኱Ꮫ㝔ᆅ⌫⎔ቃᏛᇽ㻖㻌 Ώ㑓⤂⿱㻌 㻌 せ᪨ 䜹䞁䝪䝆䜰䞉䝷䝍䝘䜻䝸ᕞ䛻ఫ䜐ᑡᩘẸ᪘䝆䝱䝷䜲䛾ே䜃䛸䛿ఏ⤫ⓗ䛺↝⏿㎰ᴗ䜢Ⴀ䜣䛷䛝䛯䚹䛧䛛䛧䚸㏆ᖺ䚸኱つ ᶍ䝥䝷䞁䝔䞊䝅䝵䞁㛤Ⓨ䜔㎰ᐙ䝺䝧䝹䛷䛾᥮㔠స≀⏕⏘䛾ᣑ኱䛺䛹䛾⤒῭㛤Ⓨ䛜㐍䜏䚸↝⏿㎰Ẹ䛻✀䚻䛾ᙳ㡪䜢 ୚䛘䛶䛝䛯䚹䛣䛾䛯䜑䚸䛣䛾䜘䛖䛺ᛴ⃭䛺㛤Ⓨ䛾ᣢ⥆ྍ⬟ᛶ䜔ጇᙜᛶ䜢᳨ウ䛩䜛ୖ䛷䚸䝆䝱䝷䜲㎰Ẹ䛾ఏ⤫ⓗ↝⏿㎰ ᴗ䛜䛹䛾䜘䛖䛺≧ἣ䛻䛒䜛䛛䜢▱䜛䛣䛸䛜㔜せ䛻䛺䛳䛶䛝䛶䛔䜛䚹ᮏ◊✲䛷䛿䚸⾨ᫍ⏬ീゎᯒᡭἲ䛸㉁ⓗ◊✲ᡭἲ䜢 ⏝䛔䛶䚸䝆䝱䝷䜲䛾ே䚻䛜⾜䛳䛶䛔䜛↝⏿㎰ᴗ䛻㛵䛩䜛ఏ⤫ⓗ⾜஦䜔㎰ἲ䚸ᅵᆅ฼⏝䝟䝍䞊䞁䛻䛴䛔䛶᫂䜙䛛䛻䛩 䜛䛣䛸䜢┠ⓗ䛸䛧䛯䚹⾨ᫍ⏬ീゎᯒ䛷䛿䚸ALOS AVNIR-2 䛚䜘䜃 Worldview-1 ⾨ᫍ⏬ീ䝕䞊䝍䛛䜙ᩍᖌ௜䛝ศ㢮ἲ 䛻䜘䜚ᅵᆅ฼⏝䞉ᅵᆅ⿕そ≧ἣ䜢ᢕᥱ䛧䚸䝆䝱䝷䜲䛾↝⏿㎰ᴗ䛿᥮㔠స≀㎰ᴗ䛸ΰᅾ䛩䜛ᙧ䛷ኚᐜ䛧䛴䛴䛒䜛䛣䛸䜢 ᫂䜙䛛䛻䛧䛯䚹䜎䛯䚸㉁ⓗ䝕䞊䝍䛾ศᯒ䛛䜙䛿䚸䝆䝱䝷䜲䛾↝⏿㎰ᴗ䛜ᙼ䜙䛾ఏ⤫⾜஦䛸័⩦ⓗᅵᆅᡤ᭷ไᗘ䛻ᐦ᥋ 䛻㛵㐃䛧䛯ከẟస㎰ᴗ䛷䛒䜛䛣䛸䛜᫂䜙䛛䛻䛺䛳䛯䚹㏆ᖺ䚸䝷䝍䝘䜻䝸ᕞ඲య䛷᥎䛧㐍䜑䜙䜜䛶䛔䜛⤒῭㛤Ⓨ䛿䚸䝆䝱 䝷䜲䛾↝⏿㎰ᴗ䚸䛭䛧䛶䝆䝱䝷䜲♫఍඲య䛻኱䛝䛺ᙳ㡪䜢୚䛘䛴䛴䛒䜛䚹 䜻䞊䝽䞊䝗䠖䜹䞁䝪䝆䜰໭ᮾ㒊䚸ᑡᩘẸ᪘䝆䝱䝷䜲䚸ఏ⤫ⓗ័⾜䚸ᅵᆅ฼⏝䝟䝍䞊䞁䚸↝⏿㎰ᴗ

 * 䛈606-8501 ி㒔ᗓி㒔ᕷᕥி༊ྜྷ⏣ᮏ⏫ ** 䛈689-1111 㫽ྲྀ┴㫽ྲྀᕷⱝⴥྎ໭ 1-1-1 (Correspondence: [email protected])