Lokale ecologische kennis van bodem- en waterfuncties bij boeren in Sumberjaya, Sumatra, Indonesië. (Farmers’ local ecological knowledge of soil and watershed functions in Sumberjaya, Sumatra, Indonesia) (by/door Wim Schalenbourg)

Who could understand it better than the farmer himself? He, who lives day in day out in close contact with his land, seeing the rain, feeling the earth, growing his plants. The endless cycles of rains and droughts and of life and death cause tremendous accumulation of experience and complex interconnectedness of knowledge, completely entrenched within the complete framework of life, generated by and continuously renewing what is termed ‘culture’; a world vision from the inside out. This research project has been carried out for the farmers, by the farmers and with the farmers; therefore a first word of thanks should be addressed to the many farmers of Sumberjaya, for their enthusiasm and interest, for their wise words and their hard work. At the same time, I would like to express my gratitude to all farming people in the whole world, for providing our society with our daily food, for maintaining the link between man and nature that has become so fragile in today’s scienticised and modernised globalising world and for keeping alive the agriculture that remains at the core of all cultures.

How on earth could a young scientist, educated in a clean environment of books and computers, dreaming of the real world somewhere out there, possibly ever think to know more than those people out there? How come so many western researchers still disdain the profound knowledge and rich experience of peoples all over the world? Fortunately, I was brought up with an open mind in a stimulating environment; hence a following thank-you should be directed towards my parents, to whom I owe endless gratitude for their incessant support and caring to my sometimes incomprehensible and casuistic dreams and desires. Mama, papa, all this would have been impossible without you. An immediate attention here is given to the people of my ‘second home’, the ‘wereldwinkel’ of Hasselt, who showed me at the right moment that another world is possible and that it is worth working for with all energy.

As the farmers, their living conditions and their environment are the focus of our work – and of thousands of other research and development projects – it is self-evident that those farmers are involved to the maximum extent. However, many projects fail because insufficient attention is paid to the experiences and aspirations of local people and the cultural embeddedness of their whole life environment. Therefore, I appreciate very much the work the World Agroforestry Centre (ICRAF) is doing in this respect, and I am very grateful towards the ICRAF team in Bogor for offering me this grand opportunity. Thank you, Pak Laxman, for you clear analyses and your useful comments throughout the whole field research. Thank you, Pak Bruno, for hosting me in your house in Bogor, for the long conversations and the enthusiast working spirit. Simultaneously, I value very much the fact that I was allowed to use this research for my final thesis at the Faculty of Agricultural and Applied Biological Sciences of the K.U. Leuven. After all, it is not obvious to write a thesis with a strong input of social sciences for the title of agricultural engineer. Thank you, Erik and Bart, for accepting my request and guiding me through the finalisation of my work.

The fieldwork in Sumberjaya was accomplished successfully thanks to Endy, thesis student from the University of Lampung, and the local ICRAF staff. Endy, tanpa kamu saya tidak mampu kemarin wawancara petani sebanyak, saya suka sekali meliti bersama kamu. Terima kasih, saya rindu dengan kamu dan orang lain di Sumber. Terima kasih Mas Rudy, Mas Eri, Mas Nugi, Mbak Vita, Maria dan Trudy. Working in such a nice atmosphere was really wonderful.

You deserve to be mentioned in capitals, HELEEN, for ceaselessly showing me the beauty of small things in life, confronting me with my own ideas and proving that even the pots in the kitchen can have deeper senses. Equal thanks are given to those who support me day in day out in all the challenges we face, thank you Andreja, Ellen, JoDi, Marga, Tore, the people at CoCoYam and all those others that I forget here and now but that I will remember forever.

Finally, this work is dedicated to all the people that struggle for a sustainable world full of respect and solidarity, against the oppression by the individualist and selfish system lead by consumption and economic growth. Together, we believe that a better world is possible!

Abstract

The Indonesian island of Sumatra consists of a chain of volcanoes and mountains running along its length and a vast lowland peneplain. One of the isolated valleys in the mountains is Sumberjaya. Sumberjaya remained relatively inaccessible by road and scarcely populated till the middle of the 20th century. Massive programmed and spontaneous immigration from neighbouring overpopulated areas triggered waves of deforestation and conversion to coffee gardens, which was the immediate cause of violent expulsion by governmental army forces. Currently, negotiation among all stakeholders should result in mutually beneficial solutions for these land use disputes. Therefore, it is interesting to investigate how farmers perceive the consequences of deforestation and the disintegration of the watershed functions in their area. This research has as objectives to construct a systematised knowledge base system using appropriate software programmes. Farmer knowledge on water and soil functions is obtained through in-depth interviews and questionnaires. Local interpretations of soil erosion, water quality and effects on the landscape level are aspects of attention. The collected knowledge is analysed thoroughly and compared to both conventional scientific knowledge and knowledge of other stakeholders in the area, such as non-farming villagers, agricultural extension officers and forestry service agents.

The findings of the research are that local farmers in Sumberjaya display a profound and detailed understanding of soil and water functions, reflecting a high degree of knowledge homogeneity between local subcommunities of farmers. Their local ecological knowledge is found consistent with and comparable to scientific hypotheses, and should be considered a complex of hybrid origin rather than a purely indigenous system. In spite of adequate farmer knowledge, little is done to prevent or inverse negative tendencies resulting from deforestation. Farmers clearly lack the incentives to implement their rich knowledge in practice. Reasons might be the harsh socio-economic conditions, the land tenure uncertainty and the difficulties of collective action to protect common resources.

List of used abbreviations

ICRAF: International Centre for Research in Agroforestry (now: World Agroforestry Centre)

1. Introduction

The subdistrict Sumberjaya (104º19’ - 104º34’ E, 4º55’ - 5º10’ S) with its 28 villages is located around the mountain Bukit Rigis (1623 m a.s.l.)[2] in the district of West Lampung and approximately equals the catchment area of the Way Besai, one of the main contributors to the Tulang Bawang river that flows east through the lowland peneplain. The subdistrict comprises 54,194 ha, of which 31,571 ha or 58.3% is officially State Forest land. In 1989, the total population was 87,390, consisting of Sundanese, Javanese, Semendonese, Balinese and some original Lampungese ethnic groups (Agus et al., 2002).

Sumberjaya is located in Lampung, the southern-most province of the island Sumatra, sometimes described as ‘North Java’, indicating its transitional nature between the densely populated island of Java and the rest of Sumatra. In the previous century there has been a lot of spontaneous and forced transmigration towards this province. As a result, only a minority originates from Lampung itself. Lampung province now has the highest population density of whole Sumatra and the lowest percentage of forest cover. Governmental programs for transmigration to Lampung province have stopped, but it is impossible to control spontaneous migration (Verbist, 2001).

The Sumberjaya area has a long history of human use, but population densities remained low until a few decades ago. Cultivation by the first Semendonese migrants in the first half of the 20^th century was a form of extensive shifting cultivation. The isolated position of the watershed prevented the region from being exploited on a larger scale (Ultee, 1949; Verbist et al., 2002).

In the early 1950s, Sundanese independence war veterans settled in Sumberjaya as the result of governmental transmigration programs. Spontaneous migrants from Java and Bali followed, attracted by the good soils (Charras and Pain, 1993; Benoit, 1999). These migrants brought along the techniques of irrigated rice cultivation and installed paddy fields in the valleys. Construction of new roads in the fifties caused transport costs to plummet, triggering more immigration and massive deforestation of the slopes, which in turn led to the actual conflict between the state forestry department and the farmers (Suyanto, 2000; Verbist, 2001).

Land use changed drastically since the 1970s, while immigration and population growth continued, especially in the 1980s. The rapid conversion of forest to coffee gardens after 1978 triggered commotion among foresters. There is an old widespread perception that local people would be unable to manage forests in a sustainable way (Junius, 1933; Huitema, 1935; Elmhirst, 1996; Kusworo, 2000).

Since the mid-1970s, some areas of cleared land have been reforested or intended to be reforested and since the late 1980s the Provincial government started to evict settlers, destroying their coffee gardens and houses located in the protection forest zone. Thousands of households were evicted; a part of them was transmigrated to North Lampung, while the rest had to leave on their own (Verbist et al., 2002).

In mid 1994, the construction of the World Bank funded Way Besai hydroelectric power plant (HEPP or PLTA) Project started, and to support this project the Forestry Department would enforce the disputed boundaries of State Forest Land together with a reforestation throughout the entire water catchment area (Kusworo, 2000). This led to a second wave of evictions and demolition in the villages located in the protection forest zone. Because of the large protest and objections of the villagers asking the status of their land to be officially clarified, the destruction of coffee farms and houses has been halted since 1996. The measures did not achieve their objectives as all coffee gardens destroyed and reforested are being reopened massively since mid 1998. Due to the ‘Reformasi’ movement following the economic and monetary crisis, the weakening of the centralised power diminished the enforcement capacity of the Forestry officials. This, together with high coffee prices, caused the natural forest clearing to reach a new peak from 1998 to 2000 (Kusworo, 2000).

The Forestry Department starts realising that their hard approach does not work and participates now in the negotiation support system for Community Forestry Management Schemes. Villagers hope that some part of their territory can be reclassified as private land and local communities hope to get formal recognition to manage their coffee farms in forest land through HKm (Hutan Kemasyarakatan: Community Forestry) certificates. Contracts and agreements between the Forestry Department and farmer groups could be used as tools to control how farmers use forest land. One of the constraints is that there are no clear criteria or guidelines on how farmers should manage this forest land, which creates a very confusing situation for the local community. Some farmers organised themselves in pre-HKm groups to fulfil the prerequisites (such as maps, management plan and group legalisation), and to co-ordinate the procedure of application in order to get the formal right to manage state forest land (Kusworo, 2000). By now (November 2002), temporary permits have already been given to 2 farmer groups in Gunung Terang village. Twelve other groups have applied and are in negotiation.

Today, coffee gardens (kebun kopi) are the dominant land use in the region, ranging from monoculture sun coffee to mixed agroforestry.[3] Quite some families have both irrigated rice fields (sawah) and coffee gardens (kebun kopi), which provide them the ideal combination of cash and food commodities (Kusworo, 2000). Since coffee prices plummeted the latest years, more and more farmers convert their coffee gardens to vegetable fields.[4]

In some villages (such as Sukapura, Tribudi Sukur, Simpang Sari and Cipta Waras), the local community protects intact natural forests according to schemes developed a long time ago, in spite of clashes with illegal loggers and farmers from other villages. They have so-called forbidden forests (hutan tutupan), village forests (hutan desa) and reserved forest that may be cleared (hutan cadangan). The local community’s reason to protect the forest is to conserve the springs supplying water for domestic use, rice fields and fish ponds and to avoid landslides. Other villages (like Suka Jaya, Fajar Bulan and Tri Mulyo) could not develop or enforce such protection schemes and there almost all forest adjacent to the villages has been cleared (Kusworo, 2000).

Deforestation still continues. Only on the crests, ridges and very steep slopes there is still forest left, where it would hardly interfere with watershed functions. Paradoxically, it seems that there are more trees in the coffee gardens on the private land than in the State Forest Land (Verbist, 2001).

The following is based on information out of internal evaluation and planning reports of ICRAF-SEA (ICRAF 1998, 2000, 2001).

In 1998, ICRAF and partners started a research project in Sumberjaya to have a closer look at watershed degradation and rehabilitation in foothills and mountain zones. ‘Watershed protection’ is the raison d’être for the ‘protection forests’ of the Indonesian state forest classification, with strong restriction on local use of such forest lands. In this approach, the main objective of the Forestry Department is to control erosion by forest protection. The conversion of forest to various other land use systems, such as coffee, is perceived to increase erosion and to have a large negative impact on downstream beneficiaries.

The Forestry department uses three criteria to define critical areas: rainfall, soil type and slope. However, these criteria are interpreted in such a way that not less than 80 % of Indonesia should be protected. It seems that the Forestry Department uses these apparent objective criteria to control an area as large as possible, rather than to provide a realistic framework for the management of critical zones.

One of the underlying principles for current Indonesian legislation to protect watersheds by classifying large areas as protection forests is the empirical USLE (Universal Soil Loss Equation). On the watershed level, however, this equation does not seem to take into account the spatial distribution of various land use types and thus the effect of filter elements as the natural forest cover, a paddy rice field or an agroforest. On the watershed level, it is not so important how many filter elements there are in the landscape, but far more where they are spatially located. Roads and footpaths on the other hand have the opposite effect of filters and work as channels of erosion, which also have to be taken into account.

Forest conversion in much of Southeast Asia is not a black-or-white deforestation process, but a gradual loss of forest functions in changing agroforestry landscape mosaics. Existing institutions and policies are largely based on a forest-agriculture land use dichotomy and this may lead to an unnecessary sense of conflict. The issue is of particular relevance where supposed watershed protection functions have been the basis for regulations of access to land, as in Sumberjaya. The perception that only natural forests are able to maintain these functions is an oversimplification.

Hence, the key hypothesis in ICRAF’s current research is that some farmer developed agroforestry mosaics are as effective in watershed protection functions as the original forest cover. As a result, conflicts between state forest managers and local population can be resolved to mutual benefit. Therefore, an iterative stakeholder analysis is in progress to allow articulation of the objectives of the stakeholders and to establish a negotiation support system in order to put stakeholders on a more equal footing and help them in negotiating an agreement over future resource use and access rights.

Assessment of the types of local land use that could meet the aims of watershed protection can be a much more efficient and cost effective way of achieving the government’s objectives. The idea is to collect several case studies in order to demonstrate that a lack of insight makes current regulations feeding conflicts, as in Sumberjaya. Through new biophysical insights in the landscape level processes of soil and water conservation, these conflicts can be avoided.

The objective of this research is to assess farmers’ perceptions of soil and watershed functions in Sumberjaya, Lampung province, Sumatra. This farmer knowledge of the interactions between trees, soils, crops and other elements of the agroforest ecosystem can complement and contrast the scientific ecological knowledge. Documenting and analysing this local knowledge, which is the base for current management decisions, helps to support the farmer as natural resource manager and plays an important role in finding a solution to mutual benefit of all stakeholders in the official negotiations concerning sustainable land uses that restore the watershed functions.

ICRAF already investigated farmers’ perceptions of erosion problems and conservation needs on the plot level, but not yet on the watershed level. The objective of this research is to focus more on the watershed level, and integrate and compare the farmers’ perceptions of the problems on different scales.

More specifically, our research focuses on assessing the local ecological knowledge about land suitability concerning erosion and watershed functions, for land uses such as coffee cultivation systems, protected forest, paddy rice fields, and others. Attention will be given both to the functioning of local structures to implement this knowledge and to the consequences for water quality in the watershed. As a result, our research has four main objectives.

What exactly do farmers know about the erosion problem, its causes and consequences for the watershed? Is there any local knowledge about landscape filter elements such as natural forest, agroforest and paddy rice fields? How do villagers select the best locations for these land uses and how do farmers perceive their roles in conservation of watershed functions?

Attention is given to analyse the method of cultivation and the dependence of the practiced farming system on the location in the catchment, the site topography and erosion hazard. The land tenure situation, ethnic background and tradition, time of conversion to coffee, experience in coffee farming in Sumberjaya, different social status and economic welfare of the farming family and importance of other crops on the farm, could be other determinant factors for farmers to choose for a specific farming system or land use.

If we can obtain a clear analysis of which factors dominate in their influence on the farmer’s choice for a particular cultivation method or land use, we will have a tool to help directing this choice into a sustainable direction, acceptable for all stakeholders, through extension, research or policies.

Another interesting point to focus on is water quality and farmers’ ideas about water problems, the causes and the possible solutions for these problems. How do farmers obtain the necessary water for household use, for irrigating the rice fields and for the fishponds? What are the alternative water sources the community disposes of? Do they notice an aggravating situation the last years and do they relate this to the deforestation and erosion problems? Is the protection of some forest zones near sources and rivers related to protection of water quality? Do farmers see any connection between landscape filter elements, such as irrigated rice fields and multistrata coffee, and water quality?

A third objective derives from the local community’s request to assist them in participatory village land use and land suitability mapping. There is a need for a land status map depicting state land, public land, community based state land management areas, etc, with commonly accepted boundaries, essential to be used in the negotiation process.

A first step in this direction might be to come up with a list of criteria that villagers and farmers use to distinguish, evaluate and classify different land units for different uses. In this way we might construct some kind of legend with guidelines on how to proceed with this participatory mapping process. An assessment of local knowledge about different soil types, their properties and suitability for different land uses is of primordial importance in this matter.

The Local Ecological Knowledge of farmers in Sumberjaya is compared to conventional scientific knowledge, as held by ICRAF and other scientists. In the end, it might also be interesting for the negotiation support model to parallel the local farmer knowledge with the perceptions of the other involved stakeholders, as there are non-farming villagers, extension officers of the local government, directors of the PLTA dam and officials of the Forestry Department. On which points do their views and ideas differ mostly from the farmers’ vision?

2. Local Ecological Knowledge: a review of the literature

Several concepts are used to characterise the knowledge held by local people all over the world. Now, it is commonly required by policies that this local, traditional or indigenous ecological knowledge (LEK, TEK or TK, IK) is incorporated into sustainable development, environmental assessment and resource management programs. However, the definitions of these knowledge concepts are inconsistent and unclear (Stirrat, 1998; Usher, 2000) and generally do not include the living and dynamic character of the knowledge (Ellen, 1998). Blaikie et al. (1997, p. 218) define ‘knowledge’ in general as:

“the way people understand the world, the ways in which they interpret and apply meaning to their experiences. Knowledge is not about the discovery of some final objective ‘truth’ but about the grasping of subjective, culturally conditioned products emerging from complex and ongoing processes involving selection, rejection, creation, development and transformation of information. This is inextricably linked to the social environmental and institutional contexts.”

The word ‘indigenous’, appearing as IK in publications since 1979 (Warren, 1998), refers to specific groups of people, defined by the criteria of ancestral territory, collective cultural configuration, and historical location in relation to the expansion of Europe (Purcell, 1998). The use of IK rests on recognition of the role of knowledge in the power relations constituted by the expansion of Europe, a perspective based on a humanistic unease with the effect of westernisation on indigenous peoples, and hence it constitutes a critique of aspects of Western knowledge (Purcell, 1998; Warren, 1998). The UN defines ‘indigenous peoples’ in a similar way as “having a historical continuity with pre-invasion and pre-colonial societies”. In the early 1990s however, the need was felt to expand the UN definition to histories of non-Western colonisation as it is rather the relationship to the nation state in its present role and to colonial and postcolonial processes that is important to consider (Muehlebach, 2001). Yet, many farmers in the world are non-indigenous peasants who do not benefit from a rich cultural tradition and draw upon the experience of only a limited number of generations (Ryder, 2003). Therefore, ‘indigenous’ might not be the most appropriate term to name the ecological knowledge held by rural people.

Another frequently used term is ‘traditional knowledge’, but one of the greatest qualities of this knowledge is said to be its ability to respond to feedbacks from the environment, to incorporate both each new generation’s experiences and reinterpreted aspects of knowledge of other origin, evolving continually and thereby remaining current and vital (Sillitoe, 1998; Berkes et al., 2000; Brodt, 2001; Davis and Wagner, 2003; Pandey, 2003). Many authors indicate farmers’ capacity to innovate, experiment and creatively search for improvements of their system. Thus their knowledge does not need to be traditional as it can also consist of recent innovations by local people (Pretty and Shah, 1997; Warren, 1998; Critchley, 1999; Critchley, 2000; Brodt, 2001; Critchley and Mutunga, 2003). Hence, it seems better to use the broader sense of the concept of ‘Local Ecological Knowledge’, as it does not make any necessary links with past traditions or community history. Broader and more acceptable definitions are as follows:

Much of the knowledge is implicit in the practical experiences and people are not familiar with expressing it in words; knowledge is indeed as much skill as concept (Sillitoe, 1998). This kind of ecological knowledge held by local people also exists in more developed parts of the world (Ellen and Harris, 2000; Olsson and Folke, 2001; WinklerPrins and Sandor, 2003).

LEK is much more than a collection of individual experiences and perceptions, as it expresses the richness and depth of human community and its relations with and understandings of the local environment and ecology. It is a socially and culturally rooted ‘knowledge system’, a more or less coherent structure that assembles a complexity of individual elements that jointly constitute ‘culture’ (Brodt, 2001; Davis and Wagner, 2003).

An important characteristic of local knowledge, as indicated by the definitions above, is its embeddedness in a cultural framework, as it is born through generations of intimate contact with the land (Sandor and Furbee, 1996; Stirrat, 1998; Kimmerer, 2002; and many others). Therefore, conservation of this local knowledge is inextricably connected to the ecological resilience of the environment (Berkes et al., 2000; Muehlebach, 2001; Pandey, 2003).

Many cultures are losing their LEK due to economic and social change and assimilation into the globalising, scientising and therefore homogenising world at a rate that may not allow us even to know what value such systems had (Sandor and Furbee, 1996; Brodt, 2001; Pandey, 2003; WinklerPrins and Sandor, 2003). Nevertheless, the often-marginalised local knowledge is argued to be able to survive despite of substantial disruptions to local lifestyles (Ross and Pickering, 2002). The diversity of local ecosystem management practices sustains because of the above-mentioned continuous adaptation to changing circumstances and because of social mechanisms such as local institutions, leadership, regulatory rules and norms, and cultural internalisation of traditional practices (Berkes et al., 2000).[5] However, farmers are not always able to adjust their knowledge to rapid changes in farming systems and increasing intensity of land use (Sillitoe, 1998; Ellis-Jones and Tengberg, 2000). Nevertheless, complete replacement and loss of local knowledge is uncommon as some parts of the knowledge system are likely to persist, depending on several attributes: the generality of the knowledge, the adaptability of the user, the hierarchy in the knowledge system, the interconnectedness of subsystems and the scale of application. Only that knowledge that seems useful to the users themselves – in spiritual, technical or other terms – is likely to survive (Brodt, 2001).

Top-down scientific approaches such as the green revolution, involving researchers and extension agents in the transfer of a technology paradigm largely fail the small-scale farmers, causing strategies to be reconsidered (Critchley et al., 1994; Sillitoe, 1998; Zuberi, 1998; Critchley, 2000). Often, farmers are still viewed as mismanagers of soil and water and have been advised, paid and forced to adopt new technologies procedures that often appear alien and sometimes deprive them of control over their own activities. Most efforts have been remarkably unsuccessful, undermining the credibility of technology transfer and modern science, wasting huge sums of money and causing widespread environmental degradation (Pretty and Shah, 1997; Sillitoe, 1998; Ellis-Jones and Tengberg, 2000; Kimmerer, 2002). The local people find themselves ill-adapted to their own habitat, as their indigenous knowledge has been decreed to be no longer useful. Lack of respect for others’ ways leads to offensive interference in their lives (Sillitoe, 1998; Zuberi, 1998). Agricultural-development research has focused mostly on cash crops and fertiliser trials and recommendations, whereas socio-economic problems make such commercial developments difficult and inappropriate as subsistence farmers lack the resources to take advantage of them. It would make more sense to assess indigenous practices regarding soils and fertility management (Sillitoe, 1998).

In academia, study of local knowledge occurred over the past four or five decades in the anthropological areas of ethno-science and human ecology (Sillitoe, 1998). Crucially depending on a change in paradigms that structure conceptions of development over the last two decades, the importance of LEK for a development-focused perspective has increasingly been recognised, causing a boom in LEK related research and applications (Critchley et al., 1994; Sandor and Furbee, 1996; Pretty and Shah, 1997; Norton et al., 1998; Sillitoe, 1998; Critchley, 1999; Critchley, 2000; Thomson, 2000; Hall and Bigler-Cole, 2001; Muehlebach, 2001; Huntington et al., 2002; Kimmerer, 2002; Ross and Pickering, 2002; Davis and Wagner, 2003; Grossman, 2003; Niemeijer and Mazzucato, 2003; Pandey, 2003; Robbins, 2003; WinklerPrins and Sandor, 2003; and many others). Farmers are now considered the potential solution rather than the problem. Applications of LEK are enormous in a wide range of disciplines such as agricultural development, soil science and land suitability assessment, technology development, soil and water conservation, ecology, pharmaceutical botany, ecosystem management, conservation biology, ecological restoration, participatory development, development assistance, medicine, land use and management, GIS analysis, agricultural extension, forestry, fish and wildlife sciences and many more.

Acknowledgement of LEK is presumed as critical to fostering the development of respect for what people know. LEK is human-centred, providing peoples and communities with a greater capacity to self-direct and self-manage, thereby empowering them through provision of control over core factors in their lives and livelihoods (Sillitoe, 1998; Davis and Wagner, 2003). LEK has value not only for the wealth of ecological information it contains but also for the cultural framework of respect, reciprocity and responsibility in which it is embedded (Kimmerer, 2002). Others also may have something to teach us: the cross-cultural study of their knowledge may advance our scientific understanding of natural processes by challenging our concepts and ideas (Sillitoe, 1998).

A growing number of multilateral development agencies (e.g., the World Bank and United Nations agencies such as ILO, FAO, UNEP, UNDP, WIPO and UNESCO) and bilateral agencies are recognising that involvement of existing knowledge systems can make projects more cost-effective and sustainable (Warren, 1998). Principle 22 of the Declaration on Environment and Development (on the Convention on Biological Diversity in Rio, 1992) states that “indigenous people and their communities have a vital role in environmental management and development because of their knowledge and traditional practices.” The Convention promotes the wider application of such knowledge with the approval and involvement of communities and encourages the equitable sharing of benefits. (Muehlebach, 2001; Kimmerer, 2002; Pandey, 2003). The United Nations even have a Working Group on Indigenous Populations (WGIP) where indigenous delegates themselves articulate notions of indigenous culture, politics and activism (Muehlebach, 2001).

Knowledge held by local people and knowledge of scientists and researchers are generally considered to differ, although they are judged complementary and can exchange useful information and views. Hence, experts call for more communication, combination, collaboration and integration with LEK systems, and give several examples in the literature (Sillitoe, 1998; Gobin et al., 2000b; Mackinson, 2001; Olsson and Folke, 2001; Zurayk et al., 2001; Huntington et al., 2002; Davis and Wagner, 2003; Pandey, 2003).

According to Kimmerer (2002), LEK and scientific knowledge have their source in common as both knowledge systems yield detailed empirical information of natural phenomena and relationships among ecosystem components and have certain predictive power. However, LEK observations tend to be qualitative, holistic and set within an ecosystem framework, whereas scientific observations tend to be quantitative, categorised and compartmentalised. Western science is conducted in an academic culture in which nature is viewed strictly objectively. The scientific community prides itself on data that are value free, while LEK is laden with associated values. It is much more than the empirical information concerning ecological relationships; it is woven into and inseparable from the culture, including an ethic of respect and obligations between humans and their environment. In indigenous science, nature is subject, not object. Such holistic ways of understanding the environment offer alternatives to the dominant consumptive values of Western societies (Kimmerer, 2002; Ross and Pickering, 2002).

Thus, several authors describe local knowledge systems as being an epistemological system separate and unique from all others and particularly Western science (Usher, 2000; Cools et al., 2003; Davis and Wagner, 2003; Robbins, 2003). Many authors classify knowledge systems into two broad and fundamentally different categories: western formal scientific knowledge and local indigenous informal knowledge (Blaikie et al., 1997; Sillitoe, 1998).

Ferradás (1998) argues that indigenous knowledge is a contested concept as “Indigenous knowledge” is considered the knowledge of an ‘other’ who becomes defined in opposition to an authoritative ‘we’, vaguely presented as scientists from the West, privileged enlightened revealers of truth. Anthropology and other social sciences have been addressing for some time the problems entailed in thinking of the world in terms of antinomies such as scientific western knowledge versus indigenous knowledge, for sure a well-entrenched dichotomy. Each pole is assumed as a totality with an internal logic and independent of the other (Ferradás, 1998).

Comparative studies do, however, indicate far more uniformity among the local and scientific systems than ever had been anticipated: there is little point in dichotomising them. Local knowledge is increasingly acknowledged to be scientific (Agrawal, 1995; Warren, 1998; Kimmerer, 2002; Niemeijer and Mazzucato, 2003; Robbins, 2003). Differences are not epistemological but more a matter of communication, politics, priorities, and a whole range of other institutional and cultural issues (Martin, 2003; Robbins, 2003; Wilson, 2003). Boundaries between scientific and local knowledge systems are falling apart once one realises the social, cultural and political character of all knowledge and worldviews. All scientific practice is filled with efforts at consensus, negotiation and struggle. Each side has its own filter through which people perceive and make sense of the real situation before them. Exclusive truth claims are not now tenable on the part of either science or local knowledge systems (Agrawal, 1995; Cleveland, 1998; Forsyth, 1998; Norton et al., 1998; Stirrat, 1998; Robbins, 2000; Barth, 2002; Kimmerer, 2002; Pandey, 2003; Robbins, 2003).

Many knowledge systems are in reality neither indigenous nor exogenous but rather hybrid in character, combining knowledge from different origins (Sillitoe, 1998; Brodt, 2001; Dove, 2002). Knowledge is in reality diversified, dynamic and heterogeneous, as clearly shown for migrant peasants all over the world. Knowledge production is seen as a process of social negotiation involving multiple actors and complex power relations (Nygren, 1999).

Study of LEK sits at the intersection of the natural and social sciences and the humanities. Therefore, hybrid methods and epistemologies are required and methods are a continual challenge (WinklerPrins and Sandor, 2003). Several authors mention the lack of detailed and efficient methodologies for documenting, presenting and operationalising LEK. However, other authors such as Ross and Pickering (2002) do not agree with this. Still, many researchers do not provide detailed descriptions of the methods they use (Sillitoe, 1998; Thomson, 2000; Usher, 2000; Davis and Wagner, 2003). This absence of discussion impedes further development of appropriate methodologies (Davis and Wagner, 2003). Frequently encountered problems arise from the fact that LEK often only has localised and cultural relevance, and that most information is presented in anecdotal form, difficult to classify and analyse (Sillitoe, 1998; Thomson, 2000). Interpretation and assessment of LEK alongside scientific criteria is contentious and difficult as much of the knowledge is symbolic and pragmatic, while we constrain understanding in reducing everything to words (Ellen, 1998; Sillitoe, 1998). Notably absent to date are accounts of how LEK has been employed in real, as opposed to hypothetical and theoretical, resource management settings (Davis and Wagner, 2003). The complex contexts in which development workers operate are often such that operationalising the insights gained from imaginative research programmes is an extremely difficult process. The result is that the products of successful research programmes are rarely used in the practice of development (Stirrat, 1998).

Some advocate a romanticised vision of LEK systems as traditional and ecologically sound, projecting onto them their own critiques of modernity and hence promoting the conservation of peoples and their lores as they imagine they should be. We need to guard against these perils (Ferradás, 1998; Sillitoe, 1998; Pandey, 2003). Valuable knowledge must be distinguished from myths and the science behind the traditions must be identified (Pandey, 2003). In practice, scientific or technical knowledge is used as a means to differentiate between ‘useful’ or ‘correct’ LEK and ‘useless’ or ‘incorrect’ LEK (Stirrat, 1998).

Several authors lament the widely used technique of extracting small parts of the useful LEK out of cultural context, translating it into positivist language, treating it as independent technical facts and incorporating it in the more familiar context of Western science as part of a paradigm that remains essentially top-down, although considered as participatory. Western scientific models are imposed uncritically, which results in distorted understanding and ill-informed, decontextualised knowledge results. Many so-called indigenous knowledge reports radically disembody particular bits of proclaimed useful knowledge from the rest of culture in a way that does a profound disservice to its potential importance (Ellen, 1998; Ferradás, 1998; Sillitoe, 1998; Thomson, 2000; Ross and Pickering, 2002; Kimmerer, 2002; Campbell and Vainio-Mattila, 2003). Scienticising and rewrapping LEK precisely causes the lack of theoretical and methodological coherence in its research (Ellen, 1998).

Still, in-depth understanding of the local culture is required to assess the impacts of interventions and to understand how projects can possibly improve local peoples’ livelihoods (Ellen, 1998; Zuberi, 1998). The considerable problems in trying to understand something about others’ sociocultural traditions are not to be glossed over; this is indisputably no easy or short-term task and misrepresenting them will lead to disillusionment. The time scale of ethnographic research is considerable, and this presents problems for development projects with their short-term orientation and politically driven requirement of quick returns. It is not just a question of the time it takes to learn the language and the sociocultural basics, but also a matter of the investment needed to win the trust and confidence of the people. Understanding another culture is extremely difficult as knowledge is transferred using idioms alien to science; translating another culture is inevitably distorting (Sillitoe, 1998). Researchers consider a few meetings with local farmers for “listening to community concerns” a way of understanding the local knowledge system of farming. Good intention is, however, insufficient for bridging the gap separating a different cultural history and place (Norton et al., 1998; Stone, 1998). Many Western scientists manifest a lack of respect for others’ knowledge traditions, assuming technological superiority as the answer for all problems (Sillitoe, 1998; Ross and Pickering, 2002).

However, Ellen (1998) opinions that many scientists – though not all – are willing to learn but have little understanding of LEK research and can find few anthropological expertise accessible to scientists. Anthropologists from their side indeed need to avoid producing esoteric and inaccessible ethnographic accounts. Their work is often sterile and undynamic from a developmental perspective (Sillitoe, 1998). Nevertheless, LEK is largely ignored in science curricula so that acknowledgement or understanding of LEK is rare in the scientific community (Kimmerer, 2002). The bottom line is perhaps that all researchers should ideally have some awareness of anthropological issues to promote an awareness of alterity and its implications (Sillitoe, 1998). Consideration and examination of alternative interpretations from a cross-cultural perspective trains scientists to think critically rather than passively accept a familiar paradigm and reveals cultural assumptions underlying Western science and technology. The culture of science is perceived as unwelcoming, exclusionary and hostile to traditional ways of knowing and a large division exists between training in science and humanities, causing very capable students to abandon their science education because of the perception that science prohibits the expression of cultural concerns and a personal connection to nature (Kimmerer, 2002).

Sillitoe (1998) argues that rapid rural appraisal (RRA) and participatory rural appraisal (PRA) cannot replace anthropological enquiries. It can take several years, not months or weeks, for someone unacquainted with a region to achieve sufficient anthropological insight into local knowledge and practices to illuminate developmental problems. Brokensha (1998) however replies that RRA is in most cases better than nothing, which is so often the only alternative. Providing that the anthropologist is familiar both with the region and the topic, he or she can make a meaningful contribution.

To bridge the gap between science and LEK, facilitatory methods that combine anthropological skills with science need to be promoted. Interdisciplinary work as it already takes place, combining the empathy of social scientists with the technical know-how of natural scientists implies a willingness to learn from one another (Sillitoe, 1998; Stirrat, 1998). New methodologies for dialogue with local knowledge holders will, however, not emerge until indigenous peoples have political and economic parity with development forces. The discourse on traditional knowledge has to be weaved into political actions that ensure the rights of indigenous peoples to define themselves, their knowledge and our access to it (Posey, 1998).

What passes as an interest in indigenous knowledge can at times be little more than a means by which commercial interests gain control over what were previously free resources (Stirrat, 1998). For many, indigenous knowledge has become a valuable commodity that can be patented and copyrighted (Ferradás, 1998; Muehlebach, 2001). Protection of LEK from exploitation has often been framed in terms of intellectual property rights, which are intended to ensure equitable benefits from the use of LEK. However, indigenous rights control over LEK is essential to cultural survival for the people who have generated and maintained this knowledge. Misappropriation of traditional ecological knowledge can lead to adverse consequences, such as resource exploitation and misuse of knowledge (Kimmerer, 2002; WinklerPrins and Sandor, 2003). An interest in indigenous knowledge can thus have a disempowering impact on the poor (Stirrat, 1998). Many debates are being waged on the significance and consequences that the commercialisation of indigenous knowledge might entail (Muehlebach, 2001). International indigenous activism is a transnational cultural political movement and has always aimed at what has gone wrong in the world today as many relationships are marked by exploitation, oppression and short-sightedness. Indigenous delegates attack the core of all civilising rationality, namely the denial of man that he is part of nature. They insist on the inseparability of two seemingly separable realms – ecology and ethnicity (Muehlebach, 2001). Indigenous concepts and beliefs challenge the hegemony of the Western construct of ‘power dominance and progress’ over the untamed ‘wilderness’.

3. Materials and Methods

As discussed in the literature review, there is a lack of attention to consistent methodology for documenting local ecological knowledge (Davis and Wagner, 2003). Nevertheless, the followed methodology corresponds to most descriptions and indications found in LEK research literature.

During the fieldwork, I co-operated closely with Endy Stefanus D., student of the University of Lampung (UNILA), both focussing on different but related aspects of the above mentioned objectives.

The design of a detailed knowledge elicitation strategy has been undertaken during a period of introduction and establishment with the source community of farmers in Sumberjaya. In this way key parameters that might account for differences in knowledge could be determined and used as variables to stratify the local informants.

Our sampling strategy was based upon previous data and reports such as the work of Kusworo (2000) who has investigated communal protection and farmers’ situation in a number of villages, and the work of Chapman (2001) and Risdiyanto on local farmer knowledge on erosion and soil conservation at plot level in three villages. As the soil and watershed functions and their integration on the landscape level were to be the core of our attention, we chose to select farmers from two subcatchments of the Way Besai watershed. After discussion with the local ICRAF field staff, we decided to concentrate on the Way Petai and Way Ringkih subcatchments. For the validation test of the elicited knowledge we selected additional farmers in the same two subcatchments and in an extra third test area, the Way Kumpai subcatchment.

The field research was initiated with explorative walks in the subcatchments to get acquainted with the area and the cultivation methods and associated problems, and to identify the factors of variation that were potentially determining for farmer knowledge differences.

With the help of aerial orthophotomaps and GPS, as accurate maps for the area were not available, it was possible to have a good view of the different catchments after only a short time of exploration. On the hand of visual landform characterisation, the subcatchments were divided in three areas, denominated upstream, midstream and downstream zone. The upstream area consists of a bunch of small streams coming out of the primary forest located on the top of the Bukit Rigis mountain. The most recently cleared land is located in this zone and paddy rice fields have been constructed in flatter parts of the valleys. The midstream zone starts where the small streams join to a single or a few rivers of larger size, but the valleys are still small and the slopes so steep that the paddy rice field presence is not possible over the full length of the river. The downstream area is characterised by a broad, meandering and slow streaming river, a wide and flat valley serving as large paddy rice field area, surrounded by hills with moderately steep slopes cultivated as coffee gardens. The downstream area ends at the tributary’s intersection with the Way Besai river. The downstream and midstream farmers mostly have their houses in or near the village, whereas some upstream farmers have permanent houses near their fields.

In both subcatchments, a limited but representative number of farmers was to be interviewed, in order to get an image of the content and coherence of the local knowledge. The interviewed farmer sample has been more or less equally stratified according to position in the watershed (upstream, midstream or downstream) and ethnic group (Semendonese, Sundanese or Javanese), in order to ensure coverage of the knowledge since different classes of people might have different views as suggested by earlier research (Chapman, 2001). For every farmer additional parameters that could possibly cause knowledge differences were collected, such as field type (coffee garden, paddy rice field, vegetable garden or a combination of these), age and type of coffee garden, age of farmer, length of stay in the area and ownership status of the land.

An essential issue in LEK research is the way local knowledge experts are identified and selected, since not all persons within local settings are equally knowledgeable (Blaikie et al., 1997; Brokensha, 1998; Warren, 1998; Davis and Wagner, 2003). Davis and Wagner (2003) suggest to base the selection on responses of large samples of resource users asked to identify those considered especially knowledgeable, in this way systematically gathering peer recommendations. However, a complex process involving social and political factors and relations affects the designation of local knowledge expertise. Hence, one can question whether a broadly consultative process truly leads to a consensus. Brodt (2001) advises a less systematic ‘snowballing’ technique, whereby villagers are asked for the names of most knowledgeable farmers who are subsequently interviewed and asked for more names of most knowledgeable farmers, and so on. A disadvantage of both techniques might be that villagers often point the village chief, the local doctor or the teacher of the school as most knowledgeable person, whereas those are obviously less entrenched in the frequent use and regular management of natural resources. An alternative approach can be to use the suggestions of local notables such as community leaders and representatives, or the referrals from local associations, initial key informants and early contact persons (Brodt, 2001; Davis and Wagner, 2003). Brodt (2001) also suggests that the observations of notable home gardens or field plantings can lead to the identification of knowledgeable interview candidates.

In practice, a combination of these suggestions was used, though not in a very systematic way. Leaders of the local farmer groups and the local ICRAF staff indicated most knowledgeable farmers who were assumed to show a reasonable willingness to participate in the research. Other candidates were selected during walks across the field plantings.

As the knowledge systems developed over generations of experiences and observations within very specific settings, deep knowledge is commonly ascribed to persons of advanced age and profound experience, although wisdom is not always associated by the local community with those who have accumulated most experience, i.e. the elder (Davis and Wagner, 2003). Nevertheless, it is important to focus on people that have a deep and rich historical relation with the environment and ecosystem of that specific locality (Brodt, 2001). Hence, we opted to focus on those farmers who already live and farm in this area for a long time, which enables them to see certain trends in changing watershed functions over time. Therefore, farmers who came into the area and started farming less than five years ago have not been relied upon for knowledge elicitation.

Four preliminary conversations with farmers were undertaken in order to gain familiarity with the farmer’s terminology and to get an idea of the basic and universal knowledge held by the farmers. After this, problem specification could be refined and a detailed topic list[6] for interviews could be set up, clearly delineating the objectives and providing a useful base to start the knowledge elicitation. Davis and Wagner (2003) confirm the importance of specifying the knowledge domain and determining which among the wide variety of experiences and observations is of particular importance and interest.

Knowledge elicitation is the process whereby selected informants are encouraged to articulate their knowledge (Dixon et al., 2001). As mentioned before, this is done through repeated interviews with farmers, but it can also be abstracted from written material, such as records from previous analyses in research involving farmers from the same villages. Techniques to gather information include semi-structured interviews, questionnaires, telephone surveys and participant observation (Davis and Wagner, 2003). Ethnographic techniques of knowledge elicitation as used by anthropologists have been recognised as useful in the development of expert systems because they capture insider knowledge, or knowledge described and explained from the informant’s point of view (Benfer and Furbee, 1990). We followed an ‘emic’ approach, in which knowledge is to be gathered in a way that reflects the structure of the indigenous knowledge system and allows the representation of that knowledge system to be intuitively correct to the informant (Werner and Schoepfle, 1987).

Starting from semi-structured topic lists was considered most appropriate to conduct domain-centred in-depth face-to-face interviewing with the goal of achieving demonstrable information saturation on a set of specified core questions (Gobin et al., 2000; Davis and Wagner, 2003; Visser et al., 2003). These open-ended interviews allow people to describe their knowledge using their own words. Particular words and expressions can be noted down verbatim in order to preserve their emic nature (Brodt, 2001). The interviews were largely informal as the conversations were allowed to flow in directions the farmer felt were important (Grossman, 2003). Because much knowledge is tacit and has been learnt through observation and experience, it is important to attempt to reduce social and intellectual barriers when interviewing; as researcher it is advisable to assume the role of student desirous to learn, and not of scientist or teacher. Own influence has to be minimised and judgement during the elicitation has to be suspended; the information as contributed by the farmer has to be considered as being the truth, as we are only trying to structure and understand their way of thinking (Dixon et al., 2001). Indeed, one must accept the risk that the results may not confirm preferences and presumptions, let alone produce desired outcomes (Davis and Wagner, 2003).

The interviews were conducted in Indonesian language (bahasa Indonesia), a language commonly known by all peasants in Sumberjaya. Sometimes however, the farmers used Sundanese and Javanese concepts and expressions, which rendered comprehension more difficult. It was an enormous advantage not having to work with translators, which mostly slows down the conversations and makes the barriers between researcher and farmer harder to bridge (Grossman, 2003). The use of appropriate language and concepts, comprehensible for the farmers and being part of their own vocabulary, is critical in knowledge acquisition (Benfer and Furbee, 1993). The interviews took place in the farmers’ houses or in the fields and lasted from half an hour to one hour. The interviews were completely registered using a tape recorder to be able to fully concentrate on the discussion during the interview and to ensure no important points were missed while writing statements afterwards. Tape recording of interviews indeed is a basic knowledge elicitation technique (Thomson, 2000).

The analysis took place immediately after the interviews, extracting all useful information from the recorded material in the form of natural language statements in Indonesian with the vocabulary as used by the farmers. Listening repeatedly to the recorded material has proven to be necessary to capture the background and context of the information articulated by the farmer, essential to be able to understand what exactly he or she is trying to formulate. New information had to be confirmed by other informants and was added to the topic list. Eventual doubts or inconsistencies could be clarified by going back to the same informants.

In this way 17 interviews[7] were accomplished in the Way Petai subcatchment, followed by a thorough analysis of the gathered knowledge using the appropriate software (see below). Consequently the research progress was evaluated, the information was compared with the research objectives and eventual gaps in the collected knowledge were identified so a revised topic list[8] could be set up, focusing more on those topics on which the information was still unclear and incomplete. Further on, nine farmers in the Way Ringkih zone were interviewed and the knowledge elicitation process was finalised when the new information brought in by subsequent interviews had declined to a negligible amount. Davis and Wagner (2003) opinion that a systematic methodological approach when identifying LEK experts is necessary to ensure the quality and accuracy of the gathered information. They suggest a minimum of three observations to be gathered respecting each particular ecological, environmental or resource behavioural knowledge claim. Not all topics are covered in all interviews but each topic should be covered by a large enough sample (Brodt, 2001). Hence, the ideal research design would require the identification of a pool of at least five LEK experts within each community area. Nevertheless, this approach does not guarantee the completeness or quality of documented information (Davis at Wagner, 2003). An alternative strategy as employed here consists of expanding the research process to new interview subjects until a ‘saturation point’ is reached at which little or no new information is being reported. The concept of reaching a ‘saturation point’ is considered to be sound and should be an integral part of LEK methodology, but according to Davis and Wagner (2003) researchers generally have neither the time nor the funds needed to continue the interview process indefinitely. On the other hand, when the research objectives are clearly delineated, one can question whether an elaborate systematic selection of informants doesn’t require more efforts than the less strict approach of continuing until saturation is reached.

Comparable LEK studies have been conducting more interviews, such as Grossman (2003), interviewing 31 farmers in 3 communities on soil fertility management on organic coffee farms in highland and lake regions of Chiapas in Mexico, and Visser et al. (2003), questioning 60 farmers in 3 villages on wind and water erosion processes and control measures in Burkina Faso. However, theses researchers do no mention any form of knowledge validation as accomplished here; taking this posterior expansion into account, the sample size of Sumberjaya farmers equally mounts to about 60 farmers in 3 communities.

Because agroforestry practises and their interference with watershed functions are rather complex, effective decision making in research and extension depends upon making effective use of all available knowledge. The value of augmenting scientific and professional understanding with knowledge held by local people is increasingly recognised. However, this knowledge is often qualitative and descriptive, incomplete or contentious, and of different complementary but incomparable sources (Dixon et al., 2001). Making indigenous knowledge accessible to other scientists and relevant to their research raises considerable methodological problems. Improved accessibility has to be built on approaches that range from robust commonsense schemes to those that draw on expert-systems computer technology, such as the one used here. The explosion in database technology may facilitate the recording and recalling of ethnographic information and its cross-referencing to relevant development fields, making it readily available to specialists. The methodology cannot be static or uniform and is subject to continual negotiation among stakeholders. The dynamism of indigenous knowledge not only increases the difficulties faced in attempting to grasp it but also compromises the attempts to represent it (Sillitoe, 1998). Therefore, efforts now are directed towards documenting and disseminating that knowledge in a systematic way. Computers indeed offer a range of tools to facilitate this, either by formalising descriptive local knowledge into a knowledge base, or transforming existing databases in order to better fit the world models of local inhabitants (Gonzalez, 1995).

Thus effective mechanisms are needed for assessing, recording, evaluating and synthesising knowledge on specific topics from these sources, and to achieve this we made use of the Agroecological Knowledge Toolkit for Windows, AKT 5, as developed by the School of Agricultural and Forest Sciences of the University of Wales, Bangor, UK. The AKT software provides an environment for knowledge acquisition in order to create knowledge bases from a range of sources. Database functions and a graphical user interface allow flexible exploration, retrieval and evaluation of the knowledge, also by using automated reasoning techniques. Creating a knowledge base involves four distinct stages; knowledge elicitation from the appropriate sources, converting the elicited knowledge into simple unambiguous statements, inputting those statements into AKT using formal representation and specifying or defining the formal terms used (Dixon et al., 2001).

While creating unitary statements, the original farmer terminology in Indonesian had to be interpreted correctly in order to identify and define the appropriate equivalent scientific terms in English. Unitary statements are created by extracting knowledge from the interview material, and breaking it down into simple statements each containing one ‘unit’ of knowledge. These ‘unitary statements’ form the intermediate stage between knowledge articulation and representation. Formal representation is the process of coding knowledge for input into a computer using a restricted syntax as defined by a formal grammar developed for the purpose. Formal representation results in statements with which you can reason automatically using computer software. Formal Term (keyword) specification is the process of identifying and organising key components of knowledge. Formal terms in AKT are either objects, processes, actions, attributes, values or user defined links. The core content of a knowledge base created within AKT is a set of unitary statements. Unitary statements represent knowledge that is perceived to be true by the source of the knowledge, even if not scientifically verified. Unitary statements are the smallest useful unit of knowledge, in that they contain knowledge that is useful without reference to other statements; they cannot be broken down any further into useful units of knowledge (Dixon et al., 2001).

A knowledge base is developed in AKT in order to create a synthesised report of the current state of knowledge on the defined topic. The knowledge may then be used for a range of reasoning tasks. In this way, the knowledge base can define gaps in understanding that constrain the productivity, stability and sustainability of the agro-ecosystem and consequently can be used as resource for further research and extension planning and prioritisation. Research can be directed to defined discrepancies between local knowledge and scientific information. Correlating the scientific and local knowledge can also broaden the range of applicability of the research results. This systematic knowledge base approach is not intended to produce definitive and, therefore, testable recommendations. Nevertheless, trial applications of the approach to date have demonstrated that it can have a real and significant impact on agroforestry based research and development programmes (Dixon et al., 2001).

The question remains whether the information stored within the constructed knowledge base can be considered a substantial part of the local knowledge system. How widely must statements, experiences and descriptions be shared within a community in order to be considered attributes of the local knowledge system? Researchers agree that highest reliability should be assigned to information that has been verified by several local experts and uncorroborated information should be discounted or left out. As mentioned before, Davis and Wagner (2003) recommend at least three different sources needed to validate a knowledge claim. It is clear that knowledge is not homogenous within a local population but varies according to the gender, class, age, occupation and social status of the respondent, comparable to the variations in our own society concerning knowledge about music, electricity or gardening (Blaikie et al., 1997; Brokensha, 1998). Warren (1998) distinguishes three types of knowledges for any domain: basic core knowledge possessed by virtually all members of a community, providing the basis for communication on a given topic; shared knowledge that expands on the core knowledge and allows persons engaged in related occupational niches to communicate in more specialised ways; and specialised knowledge within an occupational niche that most others in the community do not require. Irrespective of the contentious issue of being validated as part of the knowledge system, every claim can be tested for its representation and distribution amongst people within the community. Testing the knowledge base also leads to its augmentation by details not recorded in the elicitation process (Dixon et al., 2001).

After thorough analysis of the created knowledge base and all its subtopics, key statements covering all important aspects were selected to be tested. We opted for a questionnaire consisting of a list of 68 statements[9] derived from the knowledge base and of which half were inverted giving the contrary of what the informants actually said. Some controversial inconsistencies and remarkable statements have been included in the test in order to obtain more information on points that remained unclear so far. In general, if 75% or more farmers agree with a statement, then that statement can be regarded as common knowledge or core knowledge (Dixon et al., 2001). As appropriate stratification of the source community for testing the representation of the knowledge base can reveal differences regarding the knowledge held by different groups of community members, it was decided to use a similar stratification as for the knowledge elicitation process. 28 farmers[10] have been questioned; ten in the Way Petai, ten in the Way Ringkih, and eight in the Way Kumpai subcatchment, stratified according to their position in the catchment (upstream, midstream and downstream), while other parameters as ethnic group and type of field have been recorded as well.

It appeared to be useful to glance as well at the opinions and knowledge of the other stakeholders about these ecological landscape issues in order to have an indication of possible differences in perception to be clarified by future research. In this way, a very limited number of non-farming villagers, local government extension officers, direction of the PLTA Hydro electric power plant and officials of the Forestry Department, have been interviewed about the same topics, using adapted topic lists[11] as guidelines for the interviews. While analysing the knowledge base, the different knowledge claims have also been compared to and contrasted with conventional scientific conceptions.

Date	Activity
End of July, 2002	Language course
Beginning of August, 2002	Introduction and field work preparation in Bogor, at the ICRAF main office for South East Asia
August 13 – 21, 2002	Exploration and preliminary interviews in Way Petai subcatchment
August 22 – September 5, 2002	Interviews with farmers in Way Petai subcatchment and subsequent analysis
September 5 – 10, 2002	Evaluation of collected knowledge and adjusting topic list
September 11 – 20, 2002	Interviews with farmers in Way Ringkih subcatchment and subsequent analysis
September 21 – 27, 2002	Thorough analysis of knowledge base and preparation of knowledge test
September 28 – October 7, 2002	Questioning farmers in Way Petai, Way Ringkih and Way Kumpai subcatchments
October 8 – 12, 2002	Test analysis, collecting soil samples and interviewing other stakeholders
October 13 – 16, 2002	ACIAR workshop of ICRAF: presentation of results
October 17 – 22, 2002	Farmer group meeting: presentation and discussion
October 23 – November 15, 2002	Report writing in Bogor

4. Results

In total, we have accomplished 4 preliminary conversations and 26 recorded discussions with farmers of which 17 in the Way Petai and 9 in the Way Ringkih subcatchment. These 25 hours of recorded discussions were followed by 120 hours of analysis to extract natural language statements and 100 hours to enter formal statements in the Agroforestry Knowledge Toolkit AKT5. In total, 843 formal statements have been entered in AKT5, of which 105 attribute statements, 718 causal statements, 16 comparison statements and 4 link statements. 401 formal terms were created, of which 31 actions, 27 processes, 117 objects, 82 attributes and 138 values. Preliminary analysis of the created knowledge base (KB) in AKT5 prior to set up the test questionnaire took one complete week. To get reliable results, 22 hours of questioning 28 farmers – 10 in Way Petai, 10 in Way Ringkih and 8 in Way Kumpai – were needed.

Local knowledge as a whole can never be completely explained by the farmers with all its links and conditions, simply because great parts of the knowledge are implicit since farmers are not used to communicate their experiences in words. Hence, one of the greatest benefits resulting out of the highly systematised construction of a formal knowledge base, is that one can explicitly collect and represent the complex interlinked knowledge system, hereby gaining an insight in the views and understandings of the local communities.

The following description and interpretation of the Local Ecological Knowledge system held by Sumberjaya farmers result from thorough analysis of the complete knowledge base, containing the information of the first 30 semi-structured in-depth interviews complemented with the outcome of the 28 conversations using the test questionnaire.

The systematised knowledge base as created with the AKT5 software comprises several distinct but closely interlinked topics forming one coherent system. The different formal statements have been sorted according to these topics, and served as base for the description of the Local Ecological Knowledge system of Sumberjaya farmers. In Annex VII, you can find these statements in the same order as the description of the associated knowledge. Although it appeared difficult, or even impossible, to classify the knowledge in a hierarchical system as if they were independent subsystems, we opted to represent and describe the knowledge system in a logical order, starting from the forest, its functions and the major large-scale effects of forest clearance, floods and water quality. Subsequently, a focus is set on soil erosion, an essential and localised process exerting a crucial role in the phenomena leading to watershed degradation. The attention is also drawn to the impact of soil cultivation techniques on localised soil erosion processes and to the influence of different landscape elements and their position on the resulting watershed degradation. Eventually, local land suitability criteria and ethnopedology are glanced at. The gathered local knowledge is described, discussed and compared with scientific data and knowledge of other stakeholders.

Since the seventies, massive deforestation has been taking place in Sumberjaya, and the clearing of land for agriculture still continues up to now. These land use changes from original dense primary forests to open agricultural systems such as coffee gardens and irrigated rice fields are seen as the main cause for a wide range of landscape related problems.

Erosion and landslides occur at every rain event, causing the fertility of upland fields to decrease, the river water to become turbid and increasing its content of agrochemical residues, which makes it unsuitable for consumption. The river has lost its water discharge constancy, implying drought and lack of water in the dry season and regular flooding in the rainy season. These floods have a lot of negative effects, such as destruction of irrigated paddy rice fields, erosion of riverbanks, increased turbidity of river water and destruction of bridges and dams. However, also positive aspects as transfer of high fertility from the water to the inundated land have been noted.

Although farmers are aware of the wide range of negative impacts of forest clearing, they still continue opening the forest because the availability of good land is very limited, and often clearing land is the only way for them to be sure of an income. This implies that farmers also open land on very steep slopes, although they know that the only sustainable destination of this kind of slopes is to remain under forest cover or to be reforested.

Farmers are well aware of the high fertility of recently cleared black forest soil, because of its high organic matter content and the high availability of nutrients caused by burning the felled trees and bushes, making the application of fertiliser redundant.

Land near the forest is said to be more fertile, in the first place because it has been cleared more recently, and not because of the presence of the forest nearby, although some farmers think this also has a positive influence. Increased organic matter content and fertility is perceived to have a positive influence on crop growth, especially on coffee, the most important crop in the region.

The better structure, porosity, permeability and water holding capacity of soils under forest is well documented by scientists, together with the decline of these properties on forest clearance (Lal et al., 1986). Ultee (1949) writes that on those fields that have been covered with primary forest before, coffee production is going very well for many years, because these soils are very rich in organic matter. Van der Veen (1935) mentions that forest soil is ideal for coffee, because it is loose, aerated and has an optimal structure. [12]

On the other hand, farmers also realise that this natural soil fertility only lasts for a few years of cultivation, especially when the field slopes are steep. This decreased fertility evidently has a negative effect on the coffee plants, causing yellow leaves and dying twigs, indicating nutrient deficiency. Farmers say that the coffee crop growth decreases, and so do the coffee bean size and the eventual yield. Eventually the soil degrades, its colour fades from darkish black to more reddish and yellow, and the land has to be abandoned if no fertiliser is applied. This knowledge was the base of the old Semendonese ‘ladang’ shifting cultivation system, practised until recently in this region.[13]

Farmers identified the absence of a dense forest cover as primal cause of the deterioration of watershed functions, and own a profound knowledge of forest properties relating to soil and water processes. Forest is defined as dense vegetation with big trees that account for the forest’s main regulative functions.

Through the presence of trees, the forest cover protects the soil from direct impact of raindrops during rain events because the drops first fall on the tree and its leaves and later on the leaf litter covering the soil. Raindrops falling directly to bare soil with high energy make the soil more prone to erosion and landslides, enhance rainwater runoff and cause the runoff water to be more turbid.

The dense network of roots and the large amount of leaf litter covering the soil, as typical for all dense trees vegetation patches, mainly determine the forest’s ability to regulate water and soil conserving processes. The presence of leaf litter on the soil ensures that the soil remains covered, which does not only contribute to a decreased impact of raindrops on the soil, but also functions in maintaining the soil humidity by preventing the soil from drying out and cracking. Thus, together with tree roots, the soil cover enhances water retention in forest soils, and hence preserves soil humidity and water availability in the dry season.

Scientists note two established means for increasing soil water availability: reduced evaporation and increased infiltration (Young, 1997). Evaporation can be as much as five times higher in clearings than under a forest canopy (Lundgren and Lundgren, 1979). And a layer of leaf litter or a mulch of prunings substantially reduces evaporation, possibly having as much effect as the canopy (Young, 1997).

As outlined before, farmers perceive the forest soils as being fertile, and the leaf litter being decomposed to organic matter is thought to be responsible for this. As the leaf litter also causes increased water retention, which results in higher soil humidity, organic matter decomposition is enhanced. Scientists consider the effects of trees in maintaining soil organic matter levels through the supply of litter and root residues as a major cause of soil fertility maintenance and improvement. Organic matter affects soil physical properties, nutrient availability and biological activity, which all influence fertility and plant growth (Young, 1997).

According to the farmers, tree roots and leaf litter limit soil erosion by enhancing infiltration of rainwater and retention of both water and soil. Prevention of rainwater runoff and enhancement of sedimentation of any disrupted soil particles both contribute to the absence of high water turbidity, even in rainy periods. Scientists also realise that the dense surface-root system under natural forests serves both to improve infiltration, which is the main responsible factor in runoff reduction, and to hold the soil in place (Van der Veen, 1935; Hamilton and Pearce, 1987; Agus et al., 2002).

Farmers are aware that forests, by reducing both the quantity and the velocity of runoff water – predominantly through abundance of roots and leaf litter – play an important role in preventing river flooding and landslide occurrence during the rainy season.[14]

Extending these capacities to the landscape level, it is clear for the farmers that the dense forest cover functions as a sponge in retaining water and buffering the river flow, maintaining a steady water discharge the whole year true, with rare occurrence of flooding and drought. Scientists also widely found that forested watersheds produce less runoff than those from which forest has been cleared. Maintenance of forest cover in upper catchment areas is considered a fundamental rule of conservation, leading to reduction in storm runoff and increase in river base flow through seepage from groundwater (Young, 1997; Agus et al., 2002). Forests need to be conserved for the supply of water to the rivers (Ankersmit, 1940). Farmers know very well that a bold mountain cannot retain any water, causing the rainwater to flush down all at once during rain showers, transporting a high amount of soil and organic matter towards the river. The small rivers cannot possibly conduct this huge amount of water within a short time so floods are the logical consequence. When the forests were still large, floods were not only very rare, but it also took longer after the rain event before the river started flooding and the duration of the flooding itself was a lot longer. Nowadays, with bold hills almost completely deprived of their forest cover, rainfall frequently leads to river flooding, starting almost simultaneously with the rainfall and ending as soon as the rain stops. The bare mountain soils, not having enough trees or roots to enhance infiltration and to retain and store a significant amount of water during the rainy season, cannot maintain a steady river water flow as soon as the rains stop in the dry season. Consequently, the latest years, rivers start drying and severe water shortage problems occur, aggravating towards the end of the dry season, while only a few years ago, people could not have imagined lack of water. The region’s name, Sumberjaya, literally means “Source of wealth”, but at the same time this name reflects the area’s wealth of water sources and springs, according to the older villagers. One of the most essential forest functions is thought to be the protection of springs, by preventing them from drying out in dry periods and by preserving their water clarity and quality.[15]

In general, farmers understand that the hydroelectric power plant (PLTA) needs a steady flow of clear water to be able to function optimally and they know that this is the reason why the government wants the watershed to be reforested.

Thus, farmers in Sumberjaya are conscious of the importance of the forest in preserving the wealth of natural resources in their region, although little importance seems to be ascribed to biodiversity conservation.[16] They understand that sustainable land use has to include forest protection and reduction of the mountains’ boldness, either by reforestation or by an integration of big trees in their agricultural systems.

The villagers are aware that the forest retains and stores water through the presence of big trees and hence functions as buffer to ensure water availability in the dry season and to prevent flooding in the rainy season. They also noticed the strong influence of the coffee prices on the clearance activities, as confirmed by land use studies (Verbist et al., 2002).

The local coordinator for agricultural extension, Pak Yedi Rohyadi, is aware of the forest function in regulating the river flow stability. The conversion of most of the forest to agricultural fields causes drying of the paddy fields and fishponds during the dry season and river flooding during the wet season, because the water retaining trees are not present anymore.

The local director of the Forestry Department (Dinas Kehutanan), Pak Hartawan, is convinced of the need of a forest cover in maintaining river water quality and stability. Clearing the tree vegetation decreases the capacity of the soil to retain and store water, reducing water availability and causing massive soil erosion during rainfall. Although the dry season is short, the river gets very small and water shortages occur. During the wet season on the other hand, the river level rises and floods occur, taking along great quantities of mud, which makes the water very turbid. Forest clearance also affects natural springs in a negative way. Robbins (2003) states that it is often supposed that foresters see forests quite differently than the farmers who live at the forest edge and the indigenous people who dwell within. The modern state, represented by the foresters, has established such a relation with modern science and technology that it is now considered the major source of attack of all non-modern systems of knowledge. However, as indicated by these results, there is no meaningful epistemological division between state and local knowing, but it is rather the daily struggle over resources in local political economy that gives rise to contending accounts of nature and environmental change.

The Director of Operation of the PLTA hydroelectric power plant, Pak Sugeng, is aware of the function of the forest in retaining and storing the rainwater and hence ensuring a steady flow of water in the river, not too much flooding in the wet season and not too little water in the dry season, which is also important for the operation of the PLTA dam. During the dry season, the dam cannot operate at full capacity. Clearing the forest and replacing it with monoculture coffee gardens causes soil erosion to take place on large scale, which increases the sediment content and turbidity of the river water.

As illustrated in the chapter about forests, farmers know that flooding is caused by intense rainfall in the absence of mountain cover, enhancing a direct and fast runoff of rainwater into the small rivers that are incapable of transporting this abundance of water in a short time. The latest years, the decreasing forest cover has caused the river to flood at every heavy rain shower. Floods have a longer duration downstream than upstream because of the larger water captioning area, but decreasing forest cover diminishes the duration of the flooding events as well as the time lag between the onset of the heavy rains and the start of the flooding. The river water is always turbid when flooding, resulting from erosion of coffee gardens, erosion of riverbanks and destruction of paddy rice fields. The balance between the amount of rainwater infiltration and the amount of rainwater runoff, together with the speed of this runoff water are crucial factors leading to floods.[17]

Because of the heavy rains and the enormous flow of water during floods, a lot of soil particles, sand grains and stones are disrupted by the powerful currents and carried along with the turbulent water, causing high water turbidity. Farmers observed that together with those soil elements, also organic matter ends up in the water, increasing the fertility of the river water and hence its suitability for irrigation of paddy rice fields. Everything that is not firmly anchored is swept away, even the human dirt and litter in the village, likewise contributing to a higher water turbidity. The huge violent water flow in the enlarged river skews and scrapes both the riverbanks and the riverbed, increasing respectively the width and the depth of the stream, except where the bed consists of big stones too heavy to be displaced by the flood. While inundating the riversides, the destructive nature of flooding can cause landslides. These landslides can block parts of the river and consequently cause the river path to change.[18]

River floods can cause the river water to enter the paddy rice fields. This has various effects according to the farmers, depending on the force of the water flow, the duration of inundation and the stage of rice growth. Both first factors strongly depend on the paddy field position in relation to the river and the irrigation channels.

The high speed and power of the unregulated water flow is known by all farmers to have a negative impact on the paddy rice fields. The water can move and displace the rice plants, which is disadvantageous for their growth and hence for the eventual yield. Especially when the rice is still young, it is very sensitive for being displaced. Sometimes the rice plants are flushed away completely with the inundating water. Unregulated entrance of water in the paddy rice field leads to a faster flow and a shorter period of stay of the water in the field, which prevents sedimentation and enhances erosion of the soil. In extreme cases of heavy flooding, whole parts of the paddy field erode away with destruction of the field as consequence. Enormous quantities of paddy soil can dissolve in the river, which of course contributes to higher turbidity of the floodwater.

When the flood is less destructive with water flowing more slowly, no erosion but sedimentation and deposition of soil, organic matter and garbage takes place, fertilising the soil. Also when floods subside, they are commonly thought to deposit earth and leave behind a thin layer of fertile mud. By regular flooding and sedimentation processes, the soil fertility is continuously renewed, which makes application of fertiliser unnecessary. As a result, even for a paddy field with young rice plants, flooding can be advantageous as it causes sedimentation and hence higher fertility of the soil, which stimulates the rice growth. This is only possible when the water flow is very slow, for example in the area that is regularly inundated by the closing of the PLTA dam. On the other hand, excessive sedimentation can cause the rice plants to be buried by sediments and die. When the sediment is somewhat sandy and contains organic matter, it is perceived to be extremely beneficial for the paddy field, making its soil more loose and fertile, increasing the rice yields. Farmers note that the rice grains are rarely empty, as can happen on soils with low fertility, probably caused by nutrient deficiency. Besides improving the paddy soil, the sandy sediment can also be used as building material.

The floods also bring garbage and household waste to the rice fields, especially in those areas downstream of a village. This also increases the fertility of the soil and makes fertiliser application superfluous. Right before harvest, presence of trash and dirt is experienced to have a negative impact on the yield. Generally, flooding of the field is also perceived to lower the yields if the rice is getting yellow and almost ready to be harvested.

Mostly paddy rice fields do not remain flooded for a long time so the plants are not harmed and the fertility increases, but if the rice is under water for too long, for several days, it rots and dies.

As conclusion it can be stated that whereas flooding of rice fields brings some advantages, it always entails some risks for the farmers, which makes them preferring a steady flow the whole year through rather than this unpredictable alternating regime of floods and droughts.[19]

Villagers know that the decreasing forest cover is causing more and more river flooding during the rainy season, because the landscape is deprived of big trees that hold the water and prevent it from flowing downhill all at once. The floods cause landslides and bank erosion and can also destroy paddy fields. Heavy or long floods can kill the rice plants in the paddy fields. Floodwater is usually quite turbid due to erosion taking place in the sloping coffee gardens.

The local coordinator for agricultural extension, Pak Yedi Rohyadi, also believes that the decreasing forest cover leads to more frequent river flooding. Flooding of the paddy fields can destroy both the rice plants and the paddy field itself.

The local director of the Forestry Department (Dinas Kehutanan), Pak Hartawan, is aware of the increased occurrence of flooding when a forest cover is absent. He also mentions the high water turbidity of the river during floods.

The Director of Operation of the PLTA hydroelectric power plant, Pak Sugeng, realises that a decreased forest cover is responsible for frequent river floods during the rainy season. Floods can cause landslides to occur, which can damage all kinds of infrastructure.

The fast land use change from predominant forest to agricultural systems has not only a profound impact on the quantity of water in the rivers by floods and droughts, but also on water quality and its use for consumption and irrigation.

The most widely recognised water quality problem is high turbidity. Water is turbid if it contains suspended soil particles. Farmers say it contains soil, mud or organic matter. Water becomes turbid during and after rain events because rainfall causes erosion to take place, causing the runoff water to be turbid. Hence all processes and actions influencing erosion also have an impact on runoff water turbidity. A decreased soil cover, the absence of sufficient roots and leaf litter – characteristic for a land use transition from forest to coffee gardens – enlarges the impact of raindrops on the soil, causes easier dissolving of soil particles and increases the amount and velocity of water runoff if the field slope is steep, all leading to a higher water turbidity. On the other hand, the effect of erosion on water turbidity is diminished by soil sedimentation and filtering of runoff water. Runoff on soils with a hard consistency gives less turbid water than runoff on loose structured soils, although the amount of runoff water on the former is higher because of impeded infiltration. This is confirmed by experimental data from erosion plots in the area.[20]

The conversion from forest to agricultural fields and the location of landscape elements are perceived to have a large impact on the turbidity of the river water. Especially coffee monocultures are believed to cause higher water turbidity than other coffee cultivation systems. The absence of paddy rice fields seems to lead to higher river water turbidity when it rains, which means that paddy rice fields may function as some kind of filter elements. The riverside vegetation appears to be of utmost importance to preserve river water clarity; forest, strips of trees, bamboo, shrubs as well as grass are perceived to be filtering the runoff water. Every land use type as forest, coffee gardens, paddy rice fields, grassy vegetation patches, shrubs and bushes, bamboo strips, etc., influences the two determining processes that either increase or diminish water turbidity: soil erosion and soil retention. Farmer knowledge on soil erosion processes is elaborately described in the next part. Soil retention takes place through sedimentation and filtering processes that cause a reduction of water turbidity. The part on landscape elements deals with how and where these processes are favoured.

Flooding causes a higher turbidity of river water, because of increased erosion from riverbanks and inundated paddy rice fields. Garbage and dirt from the village, washed away and carried along by the rivers during heavy rainstorms contributes to higher water turbidity as well.

The river interchanges water with the paddy rice fields and fishponds located on the riversides. This means that if the water in the paddy fields or fishponds is turbid, the river water turbidity will increase and vice-versa. As described in the previous part about river flooding, sedimentation from turbid water can increase fertility of paddy soils. Preparing the paddy fields by hoeing the soil and planting rice is believed to be a source of turbidity that cannot be neglected.[21]

The duration of the river water turbidity depends on the location; upstream, the river water becomes clear again as soon as the rain stops, but downstream, it takes much longer because there is a lot of turbid water coming in from above. Spring water is believed to remain clear, even during the rainy season.

The farmers in this region considered three factors as being important in determining water suitability for consumption, namely water turbidity, water dirt and waste content, and water agrochemicals content.[22]

Turbidity lowers the water quality and makes it unsuitable for consumption, mainly for drinking and cooking, and sometimes even for washing. Some farmers say that when the water is only yellow, not black, you can still cook it. Others say it is better not to use the turbid water for cooking and drinking, but sometimes people have no choice as they have no access to any other water. However, they assert that they already got used to it and became immune. To reduce the turbidity of the water, people store it for a while and wait for sedimentation to take place. Others try to filter the water. In the upstream areas, the turbidity mostly doesn’t last long, only during and just after a rain event, so water turbidity is not really an issue there. Upstream people mostly use the clear spring water because it has the most reliable quality, but also river water can still be used because of its limited turbidity and pollution. Downstream people mostly use well or spring water, being less turbid than river water, certainly in rainy periods.

Trash and litter from upstream households end up in the river and make the water dangerous to drink and unsuitable to use for washing, especially downstream of villages as Simpang Sari. All the village garbage ends up on and alongside the streets where it is washed away by the next rain event, bringing it to the river. The use of the river as rubbish dump is still very common.

Residues from pesticides sprayed in the coffee gardens and the paddy rice fields can be taken along with the rain and end up in the river as well. Several farmers mention these pesticide residues as water quality problem constraining water consumption and leading to fish death, and more than half of the tested farmers confirmed this information which makes it a serious indication that agrochemicals are a growing threat to water quality. Some farmers say only insecticides and other pesticides to eradicate plagues are dangerous for humans and animals, others believe also herbicides can harm our health. Some farmers indicate that it depends on how long it takes after spraying the pesticides and before the rain sets in whether great quantities of pesticides are washed away towards the river, but opinions on this matter vary greatly depending on the type of pesticide used.

The villagers see conversion of the forest to coffee gardens, mostly without shade trees, as the main reason for increased erosion and hence turbid river water. Hence, this water is considered unsuitable for household consumption as drinking and washing. Turbid water, entering the paddy fields through irrigation or runoff, will cause soil particles to sediment in the field. This sedimentation causes an increased fertility of the paddy rice field soil, but it can also bury the rice plants if too abundant.

The local coordinator for agricultural extension, Pak Yedi Rohyadi, has the same opinion and states that the absence of trees is responsible for increased soil erosion, which causes the river water to be turbid during rain events. Turbid water is less suitable for drinking, so most people obtain their water from wells. Runoff water from the coffee gardens that enters the paddy fields leaves sediment over there, fertilising the paddy field as it contains organic matter.

The Director of Operation of the PLTA hydroelectric power plant, Pak Sugeng, is aware of water turbidity problems caused by increased erosion in coffee gardens without trees and by increased flood occurrence. Also organic litter such as rice straw and wood and household waste coming from upstream areas constitute a problem for proper functioning of the dam. The river water has to be filtered, and the filter has to be cleaned very often.

Erosion is the local process whereby rainwater removes soil and organic matter from the land. Farmers are aware of the ability of runoff to wash soil downslope, as compatible with reports of farmers’ perceptions of soil erosion in the whole world (Ryder, 2003).

Rainfall causes soil erosion to take place where land slopes are steep. The steeper the slope the bigger the erosion hazard. The ability of rainwater to disrupt and transport fine soil particles is strongly determined by the vegetation and the land cover. Little erosion happens under dense forest vegetation or vegetation of shrubs and bushes. The presence of big trees diminishes the erosion hazard on steep sloped terrain, and also grasses can anchor and retain the soil during rain events. The land cover determines to which extent several water and soil processes with a direct impact on soil erosion can take place. The vegetation type plays an important role in breaking the fall of raindrops and hence diminishes the amounts of dissolving soil colloids. The soil cover can retain water, enhance infiltration and diminish runoff, causing the water to flow below the soil surface. As a result, soil retention is enhanced and erosion limited. A lot of the beneficial properties of erosion preventing vegetation are attributed to the fact that the soil is covered by living plants or leaf litter and anchored by a dense network of roots. Young (1997) also emphasises the primordial importance of the soil cover for soil and water conservation, and he reasons that this is the primal mechanism through which tree vegetation can reduce erosion.

Erosion experiments in Sumberjaya revealed that clearance of forest for the installation of coffee based systems reduced infiltration through a decline in organic matter content and deteriorating soil physical properties, causing the cumulative runoff to increase tremendously during the first three years after clearance, whereas the situation tended to improve when the coffee reached its maximum stage after 7 years (Widianto and Suprayogo, 2002).

According to the farmers, decreasing soil cover doesn’t only have a direct influence on erosion by allowing the rainfall to reach the soil and enhancing the water runoff, but also by causing the soil to be hard and to crack during dry spells. During heavy rains these soil cracks can produce small landslides as bigger parts of the soil are swept away at once. Bare soils are perceived to permit more infiltration when they are dry than in humid circumstances, but on the other hand a lot of soil particles are easily carried away, resulting in high rates of soil erosion.

Farmers also notice a difference in erodibility between loose soils and hard soils. Water easily infiltrates in loose soils, causing the amount and speed of runoff water to be low, but their loosely structured aggregates are easily lifted an carried away by this little amount of runoff water, with high erosion and high water turbidity as a consequence. Sandy soils are considered to be quite loose, making sandy slopes less suitable for coffee cultivation because of their high erosion hazard. Hard soils are ascribed contrary properties as their hard and compact consistency impedes water to infiltrate and forces it to flow away over the surface, hence the runoff amount and speed is high. Nevertheless, the soil is hard and this ensures that the soil surface is neither disturbed by the impact of the raindrops nor by the runoff water flow, resulting in less erosion and lower water turbidity.[23] Scientific knowledge is not completely consistent with farmers’ insights in soil erodibility. Scientists relate higher experimental soil loss values on less porous, hard soil with lower infiltration rates on these soils (Widianto et al., 2002). For more details about farmer knowledge on soil characteristics and erodibility, see Annex IX on local soil types and properties in Sumberjaya.

The direct effects of erosion are an increase of runoff water turbidity and a decrease of fertility of the eroded soil. The easily erodible soil of steep coffee and vegetable gardens causes the runoff water to get more turbid. Only the topsoil is affected by erosion, but as this surface layer contains the bulk of the organic matter of the land, a lot of it disappears together with the runoff water. Hence, the farmers say that runoff water is fertile. Runoff water eventually ends up in the river, the paddy fields or the fishponds, causing the water there to be turbid, yellow and sandy.

As erosion removes the top layer of the soil, which contains almost all soil organic matter, the soil fertility decreases and so will the crop growth if no fertiliser is applied. Moreover, recently applied fertiliser is also washed away by the runoff water, as well as leaf litter and weed residues, which would have contributed to the soil organic matter content and soil fertility after decomposition. Farmers conclude that soil erosion is quite negative for the soil fertility of their steep fields. Young (1997) confirms that organic matter, which is concentrated in the topsoil in most tropical soils, is particularly exposed to erosion. Because of selective removal of fine clay-humus particles, eroded sediment is richer in organic matter than the topsoil from which it was derived. Farmers acknowledge that coffee needs a high soil fertility and high soil organic matter content to grow optimally. If the soil fertility is too low, the coffee leaves turn yellow, the twigs die and the coffee growth rate decreases as consequence of nutrient deficiencies. If no fertiliser is applied, the coffee beans will be small and the yields will be low.

After forest clearing for agricultural purposes, erosion causes the fertility of steep lands to decrease a lot faster than the fertility of flat lands. Excessive erosion also causes soils to degrade. Fertile soils with a high organic matter content have a dark, almost black surface layer. Farmers say that when the topsoil erodes away, red or yellow soil comes to the surface or the black soil colour “changes” to red, yellow or even white. Likewise, Colombian mountain farmers apply three names – black soil, mixed soil and red soil – to the same soil at different stages of soil erosion. They are aware of the erosive process that gives origin to these three soils (Ryder, 2003). Soil erosion goes together with a serious decline in soil fertility according to the farmers. Eventually, crops won’t grow anymore, cultivation has to be halted because the soil is degraded, and a vegetation of grasses known as alang-alang (Imperata cylindrica) occupies the land. Scientists similarly define land degradation as the temporary or permanent lowering of the productive capacity of land, which can include soil erosion, soil fertility decline, forest clearance and degradation, loss of biodiversity and others (Young, 1997). Farmers say that these degraded soils can be reclaimed through hoeing of the soil and removal of the grasses, terracing of the land and planting of trees. Scientists know that agroforestry is suited to soils of low fertility and to degraded lands, owing first, to the potential of many trees to grow on poor soils and, secondly, to their soil regenerative capacity (Cooper et al., 1996). In this way, reclamation agroforestry has proven to be successful in restoring degraded lands (Young, 1997).

Although on steep slopes a lot of soil erosion takes place, farmers testify that the eroded soil material, organic matter and fertiliser are deposited as sediments in the fields at the foot of the hills and enrich these foot slope soils that consequently gain fertility. Farmers that own these downhill fields are considered to be advantaged compared to those farmers with steep fields higher on the slope. Niemeijer and Mazzucato (2003) indicate from similar research in Burkina Faso that farmers recognise the impact of erosion and sedimentation as this is immediately visible in some locations and directly affects the productivity of their fields for better or worse. This widely recognised plot level phenomenon also exists on the large scale of the subcatchment. Farmers state that the soil material and organic matter eroded in upstream areas are deposited as sediments downstream and enrich the downstream soils near the river.[24]

Although sometimes farmers use the word for landslide (longsor) when talking about normal erosion, they are also aware of big parts of soil mass moving and sliding down at once from steep slopes. These landslides are caused when the soil stability is low and the impact of heavy rainfall, excessive runoff water or massive floods makes the soil move, especially when it is not anchored by vegetation. Also the huge amount and force of the river water during floods can cause whole parts of land on the riversides to slide into the river. Landslide occurrence on the riversides can block the river flow and hence cause the riverbed position to change.

If the soil cover is absent, long dry periods can produce cracking of hard soils, which enhances the occurrence of landslides during subsequent rain showers. The vegetation indeed plays a very important role in reducing the landslide hazard, not only by reducing impact of rainfall and runoff water, but also by withholding the soil from being carried away. The roots anchor the topsoil and its cover tightly to the substratum, essential for landslide prevention on steep slopes. In this way riverside vegetation of trees, bamboo, shrubs and bushes or grass can reduce the occurrence of landslides when the river floods.[25]

The villagers see that a decreased forest cover and a lack of big shade trees in the coffee gardens cause soil erosion to take place. This soil erosion makes the river water turbid and decreases the fertility of the steep coffee gardens, although the downslope (paddy) fields gain fertility through deposition. They say that landslides take place when the river floods.

The local coordinator for agricultural extension, Pak Yedi Rohyadi, believes that the absence of trees is responsible for an increase of soil erosion, which leads to low soil fertility and high water turbidity.

The local director of the Forestry Department (Dinas Kehutanan), Pak Hartawan, says high turbidity of the river water results from erosion in the coffee gardens that do not contain enough trees to retain soil and rainwater, from the paddy fields that are being hoed and planted, and from the destructive nature of floods.

The Director of Operation of the PLTA hydroelectric power plant, Pak Sugeng, is aware of the erosion problem that results from forest clearance and abundance of monoculture coffee gardens. This soil erosion causes the river water to be turbid.

Farmers’ understanding of erosion is compared to the scientific simulation model WaNuLCAS in Annex XV.

Farmers developed a wide range of cultivation methods in order to reduce loss of soil and soil fertility by erosion and thus to maintain a high production. Influencing the factors that directly or indirectly affect processes leading to soil erosion can prevent the occurrence of soil erosion through rainfall. According to the farmers, impact of raindrops on the soil, rainwater infiltration, the balance between subsoil water flow and surface runoff, the dissolution of soil particles and the retention of water and soil are the determining factors.

Farmers realise that keeping the soil bare is not a recommendable thing to do, certainly if the land slope is steep. The best option one has in that case to achieve reduction of soil erosion is the planting of trees or bushes and shrubs, but also the planting of coffee helps. Scientific research confirms farmer’s knowledge that also coffee diminishes soil erosion. 90% of the coffee roots are located in the upper 30 cm of the soil and the abundant fine hair roots are capable of retaining the soil although there is some runoff (Hartobudoyo, 1979). Other experiments in Sumberjaya confirm that coffee can reduce erosion through its intercepting crown and leaf litter, but only from more than three years after coffee establishment on (Widianto et al., 2002). As mentioned before, bare soil can crack during dry spells causing the amount of soil erosion to increase if the land is sloping, which indicates once more that steep soils should not be left uncovered.

Erosion prevention is closely related to cultivation methods in the coffee gardens and other fields on steep slopes. Farmers distinguish terracing, planting shade trees, constructing furrows, ridges and composting holes, and adapted weeding activity as main factors to prevent the rainwater from flushing away soil and organic matter.[26] Planting of cover crops such as Arachis pintoi is no common practice in Sumberjaya, though some field experiments are ongoing.[27] Scientific research in Columbia showed that for a steep coffee garden, the construction of terraces, infiltration pits and the presence of shade trees could reduce the runoff and especially the erosion to virtually zero (Arsyad, 1977). Other investigations of Indigenous Soil and Water Conservation (ISWC) techniques clearly showed the benefits and efficacy of these measures (Critchley et al., 1994; Tengberg et al., 1999; Ellis-Jones and Tengberg, 2000; Wakindiki and Ben-Hur, 2002).

Constructing terraces is one of the most quoted methods to prevent erosion and all tested farmers are convinced that steep sloping lands under cultivation should be terraced, whether they are managed as coffee gardens, as vegetable fields or as paddy rice fields. Terraces exist in many forms, from almost imperceptible undulations in slope surface to well-defined sculpted banks and horizontal platforms. Sometimes they are derived from converted paddy field bench terraces (teras bangku). On weed strip terraces (teras rumput) the steep sides are planted with weeds to enhance the longevity of the construction and hence diminish the maintenance labour requirements. Sometimes low ridges along the contour lines (gulud buntu) and systems of furrows and drains are also seen as terraces by the farmers, although we opted to describe the latter separately because of its slightly different functioning.

While terracing the land, the farmers say that it is important to make sure that the black soil remains at the surface because the black topsoil is the soil layer containing most organic matter. When terraces are not flat but ridge shaped the terrace function of retaining water, soil and organic matter gains importance. An alternative and less labour intensive method to make terraces is by placing wood such as tree stems and trunks horizontally on the soil, perpendicular to the slope direction. As a result, soil will be retained by these obstructions and terraces will be formed automatically. However, this method is conditioned by the requirement of an abundance of fallen trees or other available wood.

Farmers perceive that terraces prevent the rainwater from directly flowing downhill as the water flow is halted on every terrace. Thus both the amount and the velocity of runoff water are reduced, leading to declined soil erosion and reduced landslide hazard. The water, being slowed down and halted on the terraces, gets more time to infiltrate, causing the water to flow underground, which, together with diminished soil erosion, results in lower water turbidity. On-field experiments in Jember, East Java confirm that terraced coffee gardens with or without shade trees can drastically reduce soil losses compared to gardens without terraces, especially when the coffee is still young (Agus et al., 2002). Farmers are convinced that the higher amount of water retained by the terraces compared to non-terraced steep slopes ensures higher water availability during the dry season.

For terraced fields, less soil particles are dissolved and more soil is retained, which prevents the coffee roots from being uncovered and made visible at the soil surface. By retaining the soil, also leaf litter, organic matter and applied fertiliser are prevented from being washed away, leading to a maintenance of the soil organic matter content and soil fertility, hence the land’s suitability for coffee cultivation is preserved by the construction of terraces. Farmers testified that terracing has a pronounced effect on cultivation. Terraces increase the coffee growth and ensure the coffee leaves to remain healthy and green. Terraces also make walking, hoeing and harvesting easier and safer, noted as beneficial side effect.

However, the construction and maintenance of terraces is a hard job and requires quite some resources of capital, labour and also time, which often makes it difficult for the farmer to manage his fields as he knows he should.[28]

Already in 1863, Holle wrote that all steep slopes should be terraced in order to reshape the sloping terrain into a flat terrain and in that way to prevent earth from being washed away and to facilitate decent cultivation of the soil. He also warns for unfertile subsoil that might end up at the surface, and he advises natural weed strips on the terrace sides and the possibility of making furrows every four or five terraces to deviate excessive water. Ultee (1949) calls terracing one of the best ways to prevent erosion, varying from slight undulations to continuous high terraces, which are actually necessary in steep gardens. He also mentions the high costs to construct terraces, as a lot of soil has to be removed on a careful manner, because one should avoid the unfertile subsoil to surface. The best shape for a terrace would be lowest on the inner side of the hill, with shallow furrows to localise the rainwater, and with cover crops planted on the higher outer border. Garrity (1996) writes that terracing tends to reduce the slope length (as primary effect) and slope angle (as secondary effect). This reduces the energy of the water to move soil particles downslope.

Two other soil conservation techniques known by the farmers consist of constructing a network of furrows and drains (siring or parid) and digging regularly spaced squared holes, named infiltration pits or composting holes (lubang angin or rorak).

The rainwater and the runoff water enter into the furrows instead of flowing directly off hill. This reduces the speed and amount of the runoff water, making the currents less erosive and limiting the landslide hazard. In this way furrows on steep slopes can prevent the fertile topsoil from being flushed away during rainstorms. Sedimentation and retention of soil takes places as the turbid runoff water is slowed down and halted when entering the furrow. This reduces the turbidity of the water deviated by the furrows and drains towards the river, although during heavy rainstorms the furrow water can still be quite turbid.

More organic matter is retained, because of less runoff taking place and the runoff water being halted in the furrows. Coffee is generally regarded to grow better in soils rich in organic matter, leading to the perception of better coffee growth on steep fields with furrows. Entering weed residues and other plant material into the furrows to enhance their decomposition can enlarge this fertilising effect, similar to composting holes.

The steeper the slope, the more furrows are thought to be necessary, especially in the rainy season when the erosion risk is high, whereas during the dry season little runoff and erosion take place so farmers often leave the furrows closed. During the rainy season, the furrows are rapidly filled by eroded topsoil material, which is used to fertilise the land and to stimulate the coffee growth by heaping the fertile earth around the coffee stems. The deposited material can also be left in the furrow when the latter is not deeper than 50 cm, in order for the coffee roots to be able to reach the released nutrients.

In contrast to terraces and ridges, furrows do not only stop the runoff water, but also guide the excess of runoff water though a network of drains towards the river, in this way preventing destructive flooding and erosion of the coffee gardens. When testing the knowledge base, it became clear that the furrows still transport a large amount of water during heavy rainstorms, reducing hereby the sheet runoff in the field. The water ending up in the furrows still contains a lot of mud and topsoil, but part of this is deposited in the furrow and can be recovered. Ankersmit (1940) confirms that furrows are needed in gardens with steep slopes to lead the excess of rainwater to the river so that it won’t damage the plantation.[29]

Another soil conservation technique used by Sumberjaya farmers is the construction of composting holes or infiltration pits, known by the farmers as lubang (angin) or rorak. These square or rectangular holes of 20 to 50 cm deep have two main widely recognised functions; the first one is increasing soil fertility, the second one preventing soil loss through erosion. When of a more rectangular form, the name ‘dead end trench’ (saluran buntu) is sometimes used in scientific literature. The local name lubang angin literally means ‘wind hole’. Farmers say it improves the ‘winds of the roots’, which reflects the beneficial effect of these constructions on soil aeration, which is found to be very important for coffee (Agus et al., 2002).

As the name says, composting holes are used to gather and store organic matter and plant material from different kinds in order to enhance and accelerate the decomposition process. Weed residues and leaf litter from coffee plants and shade trees are collected in the holes, which also retain water, making the conditions more favourable for decomposition processes. The decomposition of this plant material leads to the formation of organic matter and humus, contributing to the fertility of the soil. Similar to furrows, the rich content of the holes can be spread on the field by heaping it to the coffee stems. These processes all lead to a high organic matter content of the soil, which is desirable for optimal health of the coffee plants.

While digging the holes, contact with the coffee plants causes rejuvenation of the hair roots, which stimulates their growth and activity. Agus et al. (2002) confirm that farmers know that, while making the composting holes, a part of the coffee roots are cut off, which stimulates the roots to grow and causes an increase in production.

Thus, the construction of composting holes is perceived to lead to higher soil fertility and hence higher coffee yields, and this is the reason why composting holes are also beneficial on flat lands, not prone to erosion. Agus et al. (2002) confirm that composting holes are also dug in multistrata gardens where their effect on soil erosion reduction would be minimal, indicating the farmers’ perception of their beneficial effects on soil fertility and coffee productivity.

In sloping fields, however, farmers construct infiltration pits with the additional objective of soil loss prevention. Ankersmit (1940) and Agus et al. (2002) confirm that infiltration pits are effective means for erosion prevention, by increasing infiltration and reducing water runoff. On slopes that are too steep, however, farmers opinion that it is not advisable to dig composting holes if the field is not terraced, as this would increase the landslide hazard.

When turbid runoff water enters the holes, infiltration and sedimentation takes place, gradually filling up the holes. This implies that the pits have to be dug regularly, especially during rainy periods. Some farmers remove the content of the old pits to fertilise the garden’s soil, while others dig the new holes on different places, exerting an alternating system of pit construction in order to have refertilised the whole field after some cycles of construction. Construction and maintenance of composting holes, especially in the rainy season, requires quite some labour force, time and money, which is perceived as being a major constraint for farmer’s initiative. Ankersmit (1940) and Agus et al. (2002) both mention the need for labour or capital for construction and maintenance as a disadvantage. Scientific measurements for erosion experiments in Sumberjaya reveal that soil loss from coffee gardens is very low more than three years after coffee establishment and indicate that conservation techniques like infiltration pits, ridges or weed strips did not significantly affect soil loss on fields with coffee trees of more than three years old. On the other hand, when looking at nutrient loss, runoff sediment from the plots with infiltration pits contained less organic C, evidencing the utility of infiltration pits in soil fertility conservation (Agus, 2002).[30]

A very important component of the coffee gardens is the shade tree, perceived to have a wide range of beneficial effects, far more than just providing shade as its name indicates. As can be seen in Table 2, farmers use a whole range of terms to point shade trees, reflecting their different functions and properties.

Having a leafy crown, producing litter that covers the surface and being firmly anchored in the soil by a root network, shade trees dispose of the same properties as forest trees.[31] Therefore, a multistrata coffee garden is thought to have similar capacities as the forest cover in regulating water and soil processes through similar abundance of roots and leaf litter. The literature describes coffee agroforests as holistic agricultural systems that structurally resemble coffee’s natural forest habitat; these systems can provide small farmers with an array of additional crops and services (Dicum and Luttinger, 1999). The Smithsonian Migratory Bird Council (1997) says that such an agroforestry structure results in a fairly stable production system, providing protection from soil erosion, favourable local temperatures and humidity regimes, constant replenishment of the soil organic matter via leaf litter production, and home to an array of beneficial insects that can act to control potential economic pests.

Both farmers and scientists reason that multistrata systems combine litter cover with dense root systems, thus both holding runoff when it first reaches the surface and promoting its subsequent infiltration (Young, 1997). If runoff is checked, erosion will also be controlled, as reflected in Kiepe’s study (1995), ‘No Runoff, No Soil Loss’. Erosion experiments in Columbia showed that steep coffee gardens with shade trees and without other soil conservation measures can drastically reduce runoff and erosion (Arsyad, 1977).

Thus all farmers see shade trees as an important and necessary means to reduce erosion on steep land. If no shade trees are present on the hill slopes, erosion is very likely to happen, certainly when no other soil conservation measures as terracing are applied. On the landscape level, farmers believe that shade trees are able to play an important role in regulating the river flow. If the mountains are covered with mixed shade coffee gardens, runoff of rainwater is limited and hence, flooding of the rivers is less likely to occur.

Farmers believe water is retained and stored by the shade trees, increasingly when more different species of trees are present. In this way, high water availability is ensured throughout the year, even in the dry season. Nevertheless, this is not perceived to be the case when the tree root system is superficial as superficial roots compete with the coffee roots for water, reducing the quantity of water eventually available for the coffee plants.

Shade trees also contribute to the maintenance of the organic matter content and fertility of the soil and to the recycling of nutrients deeper in the soil, testified both by Sumberjaya farmers (Chapman, 2001) and scientific studies (Ultee, 1949; Belsky et al., 1993; Campbell et al., 1994; Young, 1997).

Shade tree products such as fruits and wood are valued and appreciated by the farmers as contribution to the diversification of their produce. This range of products serves to diversify the output from farms, giving a broader economic base and greater food security (FAO, 1989).[32]

Shade trees can have a lot of direct and indirect beneficial effects on the coffee crops. One of the most vital of these shade tree services as recognised by all farmers is protecting the coffee from the heat of the sun, as its name indicates. Too much sunshine during the dry season has a negative effect on the coffee crop, predominantly because it causes a rise in temperature. Too high temperatures and too much sunshine retard the coffee growth, cause irreversible dehydration of the coffee cherries and eventually the death of the coffee plant. On the other hand, when the coffee is flowering more sunshine is needed; hence the shade trees need to be pruned. This is confirmed by Ankersmit (1940), who writes that on high, wet lands with a lot of shade, flowering may even not occur at all. When too many shade trees are planted in the coffee garden, farmers believe that the coffee plants do not receive enough sunlight and coffee yields decline.[33]

The shade trees also have a profound effect on the coffee garden soils, ensuring that even in the dry season the soil near the surface remains humid. When no trees are present, all farmers know that the soil dries out and hardens, which can cause the coffee fruits to fall down and the leaves to turn yellow or red. Eventually, the hard, dry and compact soil can cause the coffee plants to die because coffee prefers a loose and soft soil. Snoep (1932) writes that shade trees diminish the direct evaporation from the soil, but as through transpiration, shade trees extract a lot of water from the soil, which can end up loosing more moisture than without trees. However, experiments of that time confirm that this moisture decline is only true for greater depths, whereas the soil near the surface is always more moist under shadow cover, and it is just this surface layer that is of great importance for the coffee crop. The coffee needs the humid soil more than the shade itself, but when too much sunshine dries out the soil, the coffee suffers a lot (Holle, 1863). Ultee (1949) mentions that clayey soils can become so dry that they crack in the dry season when no shade trees are present.

Farmers say that a lack of shade also has a negative effect on the formation of coffee fruits, eventually leading to declined yields. If more shade trees are present, the coffee cherries are less likely to fall from the trees; both during the dry season, as consequence of the tempered drought, and during the rainy season, as effect of reduced rainfall impact.

However, shade tree presence retards the ripening of the coffee fruits, causing the harvest of the coffee to be delayed. Too many shade trees in the coffee garden is not a desirable situation either because it can make the coffee flowers, fruits and leaves fall down during the rainy season, which leads to a decreased yield. Scientists also say that too much shade is not desirable in coffee production; branches and leaves loose their strength, hence not enough flowering takes place and fruits don’t ripen easily (Holle, 1863; Ruinard, 1951). The optimal planting distance for shade trees according to the farmers varies from 5 to 12 m, depending on the species and the size of the shade tree and the slope of the field. Farmers have an extended knowledge on the characteristics of the different shade tree species, as can be reviewed in Annex XII.

Shade trees also reduce the growth and the amount of weeds; hence, shade coffee needs less weeding. Agus et al. confirm this (2002); shade trees reduce the labour requirements of weeding as weed growth is limited by the presence of the shade trees.

Various authors mention the necessity of shade trees in coffee gardens. Alberts (1915) writes that the annual yields, the productivity of the trees, the resistance against diseases, plagues and wind, and many other factors are to a great extent dependent on shadow. Productivity declines rapidly when no shade trees are planted.[34]

Weeds and techniques to remove weeds have a considerable effect on soil and water processes, besides their direct effect on coffee production. Weeds are especially exuberant during the rainy season, whereas in the dry season their growth is limited by drought. Farmers also noticed that an increased presence of shade trees or the presence of a cover crop as Arachis pintoi[35] results in a lower presence and slower growth of weeds. Faster weed growth results in more weeds, which leads to a higher need for weeding. Weeding is a laborious activity that costs time and hence money.

Farmers know that weeds have two different and conflicting effects; on the one hand they have a negative influence on coffee growth by competition for nutrients, but on the other hand they diminish erosion and retain the soil on steep slopes during rain events. Hence weeding and weeding frequency can have a direct effect on erosion. Weeds are firmly anchored in the soil by a dense network of roots, retaining the soil particles and preventing them from being dissolved and flushed away with the rainwater. Scientific measurements of runoff and soil loss in coffee gardens with lush natural weed vegetation and clean weeded coffee gardens, clearly demonstrate the capacities of weeds in limiting both runoff and soil loss to a minimum, resulting in a preservation of the organic matter content and related physical and chemical soil properties. Nevertheless, coffee growth declined significantly due to weed presence (Sriyani et al., 1997; Afandi et al., 2002).

Farmers also experience that weeds have a negative effect on soil fertility as they cause the coffee growth to cease, the amount of cherries to decrease, the leaves to turn from healthy green to yellow or red and eventually the coffee plant itself to die, all because the weeds take away the nutrients in the topsoil. The farmers literally say, “there is no food left for the coffee plants”. Farmers are convinced that weeding increases the soil fertility, not only because the weeds take away a lot of nutrients, but also because the soil is turned and aerated, the coffee roots are stimulated to grow and the weed residues decompose to increase the organic matter content of the soil. Thus weeding has a positive effect on coffee growth and the eventual yields.[36]

The ability of weeds to prevent erosion can be used in soil conservation techniques, but the weeds’ negative influence on the crops has to be minimised. The farmers know two distinct systems: weed strips and ring weeding.

Some farmers believe that simple contour strips of 20 to 30 cm between the coffee rows that are left unweeded can diminish soil erosion. They are called weed strips (strip rumput) or natural vegetation strips (NVS; strip tumbuhan alami). Attention must be paid not to leave the weeds too close to the coffee stems, which would result in increased competition for water and nutrients. This efficient way to reduce erosion on slopes that are not too steep does not need any capital investments and even decreases the labour need of weeding (Agus et al., 2002). Likewise, farmers know that the steep borders of terraces can be left unweeded, which increases the longevity of the so-called weed strip terraces (teras rumput). In contemporary scientific literature, these kind of strips, hedgerows and alleys of grasses, natural vegetation, shrubs or trees are described as very effective in soil loss prevention, enhancement of infiltration (Gintings, 1982; Agus, 1993; Kiepe, 1995; Young, 1997), retention and storage of water leading to increased availability during dry spells (Garrity, 1996), and enhancement of fast terrace formation by soil redistribution (Basri et al., 1990; Fujisaka, 1992; Sajjapongse, 1992). Nevertheless, farmer knowledge about the use of planted contour hedgerows on steep slopes seems to be quite limited.

Another cited way to reduce erosion is ring weeding (bokor), a method inherited from the pepper gardens of the lower areas to the north of Sumberjaya. The area around the coffee plant, a circle with a radius of approximately half a metre, is weeded while in the rest of the field weeds are allowed to grow. In this way the garden soil remains protected by a fair weed cover while the coffee plants are not too much disturbed by the weeds.

Previous studies in Sumberjaya showed indeed that some farmers are aware that selected weeding regimes such as leaving weed strips or ring weeding can limit erosion without too much effect on coffee yields (Agus et al., 2002).

Farmers are convinced of the negative impact of weeds on the coffee plants and that is why they are fond of clean coffee gardens. When the garden is clean-weeded, the fertility of the soil is believed to be high and the coffee leaves are healthy and green. Nevertheless, most farmers know that a higher frequency of weeding also implies a higher soil loss rate through erosion and hence a faster decline of soil fertility. Ankersmit (1940) writes that replacing clean-weeding with selected weeding regimes improves the water retention of the soil and drastically reduces soil degradation through runoff and erosion. Hence, for steep lands, farmers have to choose for the compromise between clean weeding and letting the weeds grow in order to achieve optimal results. Farmers consider three to six times a year as optimal weeding frequency, depending on the slope of the terrain, the application of soil conservation techniques as terraces and shade trees and the applied weeding technique. Ankersmit (1940) suggests maximum four or five times weeding per year.

Removing weeds can happen through a whole range of different techniques such as cutting off the weeds with a long bladed knife or sickle, hoeing with a small hoe (cangkul) and spraying contact or systemic herbicides. All these techniques can be combined with forking of the soil. These different methods have different effects on the presence of weeds; on soil structure, consistency and humidity; and on the coffee roots and the coffee crop as a whole. A discussion of these techniques and their effects can be read in Annex XIII. Farmers also use weed residues as source of organic matter, as summarised in Annex XIV.[37]

Villagers seem to be convinced of the need of planting big shade trees in the coffee gardens to diminish runoff and erosion processes, especially because these trees can provide the farmers with extra produce and because they contribute to the maintenance of good soil fertility. The interviewed villagers have not mentioned other techniques besides agroforestry.

The local coordinator for agricultural extension, Pak Yedi Rohyadi, defends constructing terraces and furrows as being the most important way to prevent erosion in coffee gardens. These earth constructions stop the water flow and prevent soil and organic matter from being flushed away, hereby preserving soil fertility. Also shade trees have a very important function as they contribute to a rehabilitation of the forest functions and fulfil a key function in the HKm (Hutan Kemasyarakat: community forestry) program of the government. They also protect the coffee plants and the garden soil during the dry season, their rapidly decomposing leaves fertilise the garden soil, the trees reduce the impact of rainfall on the soil by covering the soil surface and their deep roots can retain a lot of water. Shade trees can also diversify the production of the coffee farmers and hence be an important economic element. Monoculture sun coffee on steep slopes causes a huge amount of erosion to take place, with negative consequences for soil fertility and river water quality. The primary function of composting holes, mostly dug on flat or slightly sloping fields, is to enhance decomposition of collected leaf litter, weed residues and other plant material, which serves to fertilise the soil. He does not attribute any erosion preventive properties to these composting holes, in contrast to the farmers. Also weeds can retain the soil, but it is clear that they disturb the coffee growth so they have to be removed, even on steep slopes if these are terraced. Removing the weeds is preferably done through hoeing since this also loosens the soil and hence improves coffee growth. If farmers spray herbicides, they should only do this when they are too busy to hoe, and they should alternate spraying and hoeing because otherwise the soil because hard and the coffee hair roots cannot grow.

The director of the Forest Department (Dinas Kehutanan) believes that planting shade and fruit trees in the coffee gardens are the most important measure to limit soil erosion. Trees also have a positive influence on the garden’s soil fertility and they constitute a potential alternative source of income for the farmers through the production of fruits. Planting Arachis can prevent erosion in coffee gardens, although this practice is only in a try-out phase. Also natural weeds can retain the soil, but they are bad for the coffee growth so they have to be removed, there is no alternative solution yet. He doesn’t mention other techniques in his answer on the question how erosion in steep coffee gardens could be limited.

The Director of Operation of the PLTA hydroelectric power plant, Pak Sugeng, admits that he does not have any agricultural background and thus he says not to know a lot about agricultural problems or solutions from that side. Nevertheless, he realises that planting trees in the coffee gardens and practising agroforestry could help in diminishing soil erosion phenomena, as the coffee plants themselves are not capable of retaining water and soil. For the farmers, planting of trees can have an economic value through fruit production.

Farmers realise that reforestation of the land would rehabilitate the watershed functions. The questions now are on which places this forest is most efficient in maintaining the watershed functions and to which extent other land use and vegetation types can fulfil this forest function. Land use types and their cover characteristics mainly interact with watershed functions through their influence on water and soil processes as raindrop impact, infiltration, runoff, filtering, and water and soil retention.

All vegetation types are preferred above bare soils. Soil cover and preferably trees are needed to preserve the river qualities of water constancy and clarity. Also planting of coffee on bare soils can reduce soil erosion. Trees diminish erosion, and when they are planted below a coffee garden, they filter the runoff water. Tall grasses have the same ability to filter runoff water. A vegetation of shrubs and bushes (belukar) is believed to be quite efficient in replacing the forest cover. This vegetation type is characterised by a dense network of roots and a considerable layer of leaf litter, hereby reflecting all kinds of beneficial properties as described earlier for trees in the forest. The most important disadvantage of shrubs and bushes is that this kind of vegetation does not have any value for the farmers as no usable commodities are produced.

Scientific measurements of sediment yield in erosion experiments in Sumberjaya clearly show a drastic reduction of sediment through a grass strip of 1.5m and even more sediment was entrapped by a shrub strip of 3m, confirming the efficiency of these vegetation types as filter in the landscape (Afandi et al., 2002b). Scientists formerly thought that hedgerows acted by means of a ‘sieve’ effect, trapping the soil while the water continued to flow through, but experimental results showed the system also reduces runoff through an increase in infiltration (Young, 1997). Contour strips of natural vegetation have been observed to reduce soil loss as effectively as planted hedgerows or grass strips and in requiring less labour to establish, they would be more acceptable to farmers (Garrity, 1993).

The extent to which coffee gardens can replace the forest in maintaining watershed functions appears to be determined by the applied cultivation methods.

Shade trees play an important role, as they turn the monoculture coffee garden into a mixed multistrata agroforest. Also farmers realise this, as they say that flooding will decrease if coffee gardens are planted “as forests”. Shade coffee is assumed to reduce the occurrence of river floods and decrease river water turbidity, compared to monoculture coffee. Also earth constructions such as terraces, furrows and composting holes can help in replacing forest functions. While testing the database, almost all farmers argued that coffee gardens on steep slopes without shade trees and terraces cause a lot more and more turbid runoff water than shaded and terraced gardens. Weeding doesn’t only increase erosion, but weeds can also be used to reduce erosion as illustrated in the weed strips and ring weeding applications.

This shows that all measures taken to reduce erosion and conserve soil fertility contribute to replace the forest functions. From farmers’ knowledge can be concluded that multistrata coffee gardens with mixed shade and a variety of soil conservation methods applied are fairly appropriate in trying to recover and preserve the watershed functions, in comparison to coffee monocultures or bare soils. Agus et al. (2002) write that when the coffee and shade tree plants are more than about 8 years old, they by themselves become so efficient in preventing erosion that the application of other soil conservation measures is unnecessary, unless for soil fertility purposes. However, when the coffee garden is still young, these other techniques have proven to be needed to minimise erosion.

Riverside vegetation is thought to be of uttermost importance in preserving watershed functions and determines to which extent flooding effects such as landslides, bank erosion and river path changes take place.

Forest covering the complete riparian area can preserve the watershed functions almost to the same extent as forest covering the whole landscape. The river discharge constancy can be maintained, with rare flooding and occurrence of landslides. Erosion does not take place in the vicinity of the rivers, and the buffer forest makes sure the water ending up in the river is clear and clean. Hence all functions ascribed to a complete mountain forest cover are fulfilled by riparian forests, except of course for localised erosion and fertility aspects of those part of the land that are cleared, although this has no effect on the river. One farmer says that this is an old Semendonese tradition, and he suggests strips of 500m on both sides of the river. An exception on this rule is swampland, which may be cultivated as paddy rice field or coffee garden even when it is situated within this 500m range. Though all farmers seem to be aware of the benefits of a riparian forest, strips of about 50 meters would suffice.[38]

Trees planted on the riversides, not as a forest, but just a strip of a few meters, can fulfil some important functions. Water is retained and enhanced to infiltrate, which causes the main water flow to the river to take place below the surface. This, together with the filtering effect of the runoff water flowing over the surface and sedimentation and retention of soil particles, ensures that the water that ends up in the river is rather clear. Trees stabilise the riverbanks, making flooding effects as bank erosion and landslides less likely to occur. Shrubs and bushes on the riversides display similar properties in soil and water retention. Farmers often mention Calliandra as characteristic element of efficient riparian shrub vegetation. Also a vegetation of tall, strong grasses as Imperata (alang-alang) and cane (gelagah) is appreciated by the farmers as protective riparian vegetation. Planting bamboo on the riverbanks is a more valuable but also very effective alternative since bamboo has a high capacity of retaining water and soil due to its large amount of hair roots. [39]

Although farmers are convinced of the need of appropriate riparian vegetation, analysis of high-resolution remote sensing data has revealed that within the period of 1993 to 2000, 10% of the riverside land cover has been changed, dominantly to coffee gardens and rice fields (Wulandari, 2002).

One of the scientific hypotheses of ICRAF is that the paddy rice fields next to the river can work as buffer zone or filter elements where the turbid runoff water is halted and filtered.

The parts about flooding and water quality already described that turbid water can contribute to paddy field fertility or can affect the rice plants negatively by burying them when sedimentation is excessive or by displacing them when the water flow is too forceful. Now we will take a closer look to what happens with the turbidity of the water coming from the coffee gardens as runoff water or coming from the river as irrigation water, and to which extent the paddy field indeed functions as a filter according to the farmers.

When turbid river water ends up in the paddy field by means of irrigation inputs, sedimentation of soil is thought to take place. The more turbid the river water is, the more turbid the irrigation water will be, and hence, the more sedimentation is expected to take place. When the river water is sandy, sandy sediment can be found deposited in the paddy fields. Some farmers say clear water is better than turbid water, others say there is no difference and yet others think turbid water is more advantageous for the paddy field. A positive correlation between water turbidity and water force does exist; maybe this has influenced the farmers in their opinion about the suitability of turbid water for paddy rice fields? The overall effect depends on the presence of the destructive water flow; when the flow is smooth, the effects caused by the turbidity predominate. This interchange of water between the river and the system of paddy fields also implies that the degree of turbidity of the water in the paddy fields has an impact on the final degree of river water turbidity.

Nevertheless, the ongoing processes are more complicated and a lot seems to depend on the extent to which farmers can control the supply of irrigation water. If farmers do not regulate the irrigation inputs, too much water enters the paddy field during the rainy season and during river floods, which can cause the paddy soil to erode and to be swept away by strong water forces, hereby increasing once more the turbidity of the river through this extra input of soil material. At the same time, rice plants can be displaced and flushed away. When the currents are not erosive but the water is turbid, as is likely on flooded flat fields, the rice plants can die because of being buried by an excess of sediment. On the other hand, poorly controlled irrigation supply can cause shortages of water in the dry season, also having an adverse effect on rice cultivation.

By regulating both the inflow and outflow of water, farmers can not only control the amount of water in the paddy field, but also the flow speed and the length of stay of the water in the paddy field. In this way farmers can prevent the flow speed of the water from being too high, as can happen in uncontrolled irrigation regimes. Regulation involves an increased stay of the water in the paddy field and this is thought to enhance the deposition of soil particles. It is common farmer knowledge that this sedimentation does not only renews soil fertility, but at the same time lowers the water turbidity.

The conclusion is that paddy fields are capable of filtering irrigation water, but only when possible negative effects are diminished by strict regulation and control of the flow in the irrigated system.

Farmers observe that when rain causes turbid water to run down from coffee gardens and to enter the paddy fields, sedimentation takes place, which fertilises the paddy soil and diminishes the water turbidity, so that the water entering the river is not as turbid as it would be if it directly flew from the coffee gardens to the river. This positive effect on soil fertility through sedimentation makes turbid runoff water suitable for paddy fields, especially when there are no rice plants at that moment.

Nevertheless, apart from this fertilising effect, entrance of turbid runoff water also has less appreciated consequences. As the water coming from steep slopes is suddenly halted on the flat paddy field, too much soil material may be deposited at once, burying the rice plants under a layer of sediment and causing the plants to die. This is why most farmers say that it is better that turbid runoff water does not enter the paddy fields when rice plants are present at that moment. Moreover, depending on the land surrounding the paddy field and the intensity of the rain event, a big amount of runoff water can start moving the rice plants, which unavoidably leads to yield losses. Sometimes the flow of runoff water can be so overwhelming and devastating that it even starts eroding the paddy soil. When this happens, the bench terraces can collapse and the paddy fields can be destroyed, causing an immense input of water turbidity in the river. Thus, also for runoff water, regulation is the keyword both to sustain the farmers’ rice yields and to increase the efficiency of the paddy as filter.

This regulation can happen in two associated ways. The first and widely known method is constructing a furrow in between coffee gardens and paddy fields.[40] This furrow leads the excess of runoff water coming from the coffee gardens above directly to the river so that it doesn’t end up in the paddy field and potentially damaging effects are avoided. Of course, the need for making this kind of furrow depends a lot on the location and the situation of the above lying slopes. Farmers see, for example, that if the coffee gardens are terraced, the risk of large amounts of fast flowing runoff water coming down to the paddy fields is noticeably reduced. When the turbid runoff water, speeding downhill from the coffee gardens, is slowed down and halted in the furrow, quite some sedimentation takes place. This sediment can be collected by the farmers and spread on their rice fields to fertilise the paddy soil, so the fertilising advantages of runoff water entering the paddy field is not completely lost by constructing this furrow. One farmer says to believe more in the use of furrows that run through the paddy field towards the river, but the principles are the same. Regulating the inflow and the outflow of water in the paddy field is a second known mechanism to enhance beneficial aspects and reduce negative factors related to the entrance of runoff water, similar to regulation of irrigation water. Likewise, water is caused to flow slowly and stay longer in the paddy field, which causes sedimentation together with its earlier mentioned beneficial effects of higher fertility and lower water turbidity. Thus, from the farmers’ knowledge about runoff water entering the paddy fields can be concluded that regulation is needed to enjoy its beneficial effects and avoid its bad influences on rice cultivation and river water quality.[41]

According to the farmers, cultivation practices in paddy fields can have a considerable impact on the river water turbidity. Especially the installation of the paddy field, the construction of the bench terraces, the hoeing of the paddy soil and the planting of the young rice plants cause an input of turbidity that should not be underestimated, particularly when water continuously runs through the paddy field because of rainfall or irrigation.

The direct effect of rainfall on paddy water turbidity by impact of the raindrops is hard to separate from the turbidity inputs through runoff water caused by the same rain. Nevertheless, quite some farmers think that when there are rice plants on the field and these plants are not too small, rainfall cannot cause any extra input of turbidity. On the other hand, when the paddy soil is uncovered, recently hoed or planted, rainfall does increase the turbidity of the water in such a way that water from the paddy field entering the river is clearly more turbid than the river water itself, although also in this case one must consider the cumulative effect of turbid water entering the paddy field by runoff and irrigation and turbid water originating in the field itself by impact of the rain.[42]

Conversely, during longer dry spells, the amount of water in the paddy fields declines and consequently the turbid water inputs during rain events move faster through the system of paddy fields when appropriate and timed regulation measures are not taken. This faster turnover time leads to a less prominent sedimentation and a lower decrease of turbidity.

Considering the different aspects of the common farmer knowledge concerning paddy fields and their interaction with water and sediment flows in the landscape, some conclusions can be made about its efficiency to function as a filter element in the coffee agroforest landscape. According to farmers’ insights, the paddy field surely has the capacities to behave as a filter element, as turbid water from irrigation and runoff is slowed down and the consequent soil sedimentation processes imply reduced water turbidity. But one has to emphasise once more the importance of regulation in this perspective, as unregulated water inputs can have the opposite effect through erosion and destruction of the paddy fields.

On the other hand, necessary activities during the preparation of the fields such as hoeing and planting unavoidably lead to an immense input of turbidity in the river, which can be limited to a certain extent by cutting off the irrigation for a while and by not starting these activities when a lot of rain is to be expected. This implies that it might be better not to use the concept of filter element for the case of paddy fields, but rather to speak of temporary buffers, as quite some of the soil material that is halted in the paddy field will eventually end up in the river anyway, although too little is known about relative quantities of storage and release of sediment to draw a clear conclusion. The construction of a furrow in between coffee gardens and paddy fields as needed in many cases, also decreases the functionality of the paddy filter, as part of the turbid runoff water is deviated directly to the river.

The villagers are aware of the benefits of a forest cover, and are convinced that more trees in the landscape can improve the watershed functions. Especially the coffee gardens should be planted with more and bigger trees. Riverside vegetation and the effectiveness of shrubs, bamboo or grasses as filter strips have not been emphasised by the interviewed persons. The villagers do seem to be aware of the filter function of the riparian paddy fields, as they realise that turbid water entering the paddy field through irrigation or runoff looses turbidity because of sedimentation processes.

The local coordinator for agricultural extension, Pak Yedi Rohyadi, also believes that well-managed coffee gardens can improve the watershed functions through presence of trees but also through soil conservation techniques as terraces and furrows. The riversides should be planted with big trees, like a forest, up to 100 meters on both sides, for river water to maintain its clarity and stability, and to prevent landslides on the riverbanks. A cover of shrubs and bushes can limit erosion, but is not productive at all. Pak Rohyadi does not believe in the ability of shrubs as riparian vegetation to stop or filter runoff water. Although he is aware that turbid runoff water causes deposition of soil in the paddy fields, he says that the water leaving the paddy fields is still quite turbid and he doesn’t seem to be very convinced of its capacity to function as a filter.

The local director of the Forestry Department (Dinas Kehutanan), Pak Hartawan, is strongly convinced of the need of protection of the remnant forest and natural vegetation. He believes in the HKm project and hopes that by planting trees in the coffee gardens the forest functions can be partially recovered, although he emphasises that no new land may be opened on State Forest Land. Riverside vegetation like trees and bamboo is important to prevent riverside erosion and to filter turbid water from above, but it is not enough, there must also be forest on top of the mountains. The potential filter function of paddy fields is unknown to Pak Hartawan, on the other hand, he states that paddy fields that are being hoed or planted are responsible for a major input of turbidity into the river.

The Director of Operation of the PLTA hydroelectric power plant, Pak Sugeng, believes that the forest directly influences river water quality; hence it is of utmost importance that the remaining forest is protected. Presence of trees in coffee gardens as agroforestry system could also have a positive influence in the preservation of watershed functions. The importance of riparian vegetation is also clear to Pak Sugeng as big trees can minimise the occurrence of destructive landslides. For the same reason, the sides of the roads should not be cultivated, but planted with trees. He had no clue about influence of paddy rice cultivation in the riparian zone.

Farmer groups have asked to be assisted by local NGOs and ICRAF to conduct participatory land suitability mapping. Bearing this in mind, common perceptions of the most important criteria in determining land suitability have been tried to assess. Local soil knowledge is one of the fundamental aspects of participatory land suitability studies and is reviewed here. The overall conclusions for land suitability criteria can be found in Annex XVI.

Local soil knowledge or ethnopedology comprises the roots of modern scientific soil classifications (WinklerPrins and Sandor, 2003). Vernacular soil names have been used throughout history and helped to provide the basis of scientific classification. Dokuchaev studied and popularised folk names such as chernozem, solonetz and gley, but he also caused polemics with supporters of the ethnopedologic approach in the investigations of the soil cover (Krasilnikov and Tabor, 2003). The modern interest in ethnopedology is regarded as a return, at a higher level, to the methods that were widely employed in Russia in the 19th century (Krasil'nikov, 1999).

Niemeijer (1995) believes that working with indigenous soil classifications offers several important benefits. When a detailed inventory of soil resources is required, indigenous classifications are often much faster and cheaper than conventional soil survey techniques. In addition, the use of local soil terms can considerably facilitate communication between farmers, extension workers and researchers. A third advantage is that local soil taxonomy can offer important insights into the land-use considerations of farmers and their perceptions of soil/plant interactions. Local soil knowledge, as part of LEK, is attuned to local soil conditions; hence this knowledge can indeed offer guidance for realistic land management. Nevertheless, studies should not be limited to local taxonomies, but be broader and include knowledge of physical and ecological processes (Winklerprins, 1999).

Collecting and understanding local soil knowledge is much more difficult than simply asking some farmers what local soil names there are and how they can be grouped. The variety of research approaches such as interviews, the use of ethno-science tools and field identifications of soil types, have led to an equal variety of taxonomies, sometimes for the same ethnic group in the same region. It is extremely difficult to gain an emic or insider perspective by means of interviewing techniques alone. A more anthropologic approach based upon well-defined ethno-scientific procedures helps to bring out the indigenous semantic organisation of the soil terms (Niemeijer, 1995).

In many cases, correlation between the local and the scientific taxonomies is fundamentally impossible due to differences in conceptual basis of the classifications. Local soil classifications tend to be utilarian, based on visible properties, and divide soil into natural units, whereas scientific classifications are often based on soil genesis and additional laboratory-determined criteria. Scientific classifications tend to focus on the deeper soil horizons, which represent the more fixed and invariant characteristics of a soil. Moreover, they strive for universal applicability, whereas local soil knowledge is usually site and application specific, and quite heterogeneous, even on village scale. Local classifications emphasise the characteristics of the topsoil making their understanding of soils much more dynamic (Niemeijer, 1995; Niemeijer and Mazzucato, 2003). With its emphasis on practical agricultural management, indigenous classifications are more related to use oriented classifications than to pedology based systems. Indeed, nearly all components of local soil classification reflect its focus and intent: they are guidelines for evaluating productivity and making crop and soil management decisions (Sandor and Furbee, 1996; Gobin et al., 2000).

One has to realise that a thorough assessment of local soil knowledge, including taxonomy and knowledge about soil processes and pedogenesis, requires elaborate research work focussing specifically on these issues and involving several ethnographic techniques. Ethnopedology was not the primordial focus of this research, and consequently, the soil knowledge results are limited to basic information from interviews, tested and confirmed later on by a larger group of farmers.

Key soil morphological properties such as colour and texture seem commonly recognised by most cultures. Additionally, many cultural groups recognise landscape and soil formation processes and soil dynamic properties such as geomorphic processes and organic matter decomposition processes (WinklerPrins and Sandor, 2003). Several authors confirm the predominance of texture and colour in local soil classifications (Sandor and Furbee, 1996; Gobin et al., 2000; Ryder, 2003). Regarding consistency, soils are commonly described by their hardness or softness (Sandor and Furbee, 1996).

The main soil properties Sumberjaya farmers use to distinguish soil types in terms of suitability and productivity indeed appear to be soil colour and texture. The main distinguished colours are black, red, yellow and white. Texture is a concept that has to be interpreted in a broad way, as farmers seem to mix terms of texture and consistency. Farmers often mention the following properties: loose (gembur), sandy (berpasir), soft (lembut, lemah), hard (keras), sticky (lengket), clayey (liat), compact or massive (padat). During the interviews and the testing, farmers’ perceptions about these properties and their relation with fertility, suitability for cultivation and erodibility have been investigated. The results can be found in Annex IX: local soil types and properties in Sumberjaya.

Sumberjaya farmers are often found to use the concept of cold and hot soils (tanah dingin dan panas). The soil is called “cold” when the soil is fertile and humid because it can retain a lot of water. These are favourable growth conditions for seedlings and young plants, so that farmers say that soils have to be cold when planting if you want to have successful results. Humid soils have a more stabile temperature, whereas sunshine causes dry soil to heat very rapidly and to a great depth, having a negative effect on plant growth; maybe that is where these local concepts are derived from. Farmers are also aware of the different properties of topsoil and subsoil, as the surface layer of the soil is perceived to be a lot more fertile than underlying horizons. This increases the importance of topsoil conservation against erosion.

5. Analysis

When trying to develop a local knowledge base, one wonders where this knowledge originates from, especially in an area subject to migration and only recently cleared for agricultural purposes. Indeed, ethnographic case studies among migrant peasants all over the world testify the complex articulation of heterogeneous knowledges (Nygren, 1999). Every farmer brings along his traditional knowledge, as heritage from his ancestors, with clear differences between the different ethnic groups and places of origin. This traditional knowledge is mixed as the different tribal groups live closely together in the same area where daily contact, acculturation and interchange of knowledge and practices take place, still maintaining some differences in cultivation methodology and cultural traditions, but certainly unifying and enriching farmers’ knowledge.

It does not end there, since the local farmer communities also receive information and views from outside, especially in this conflict area where the government wants to protect the forest in order to preserve the watershed functions. This makes the farmers think and discuss about these issues and hence, this dimension is integrated in their LEK. LEK is not always unique to the local setting but sometimes arises as a consequence of ongoing regional and global exchanges of ideas and people (Davis and Wagner, 2003).

Further on, the government obliges the farmers to group, organise themselves in co-operatives in order to get the permission from the Forestry Department to cultivate certain parts of the State Forest Land. Although it is not clear yet what exactly is required to obtain such a permission, it is obvious that agroforestry practises and soil and water conservation measurements play an important role, to such an extent that local extension and local NGOs surely also have an impact through organised farmer meetings and demonstrations. This scientific and experimental knowledge and even terminology is intermingled with indigenous knowledge concepts, homogenising the final perception even more. Grossman (2003) similarly acknowledges that organic coffee farmers in Chiapas (Mexico) have a dual knowledge system about their soil, made up of experiences and phenomena that they can visualise and information retained from organic training workshops. Also Ryder (2003) indicates that some farmers attribute knowledge of soil erosion to contact with extension agents and with representatives of soil conservation projects. In his study of smallholder rubber cultivation in South-East Asia, Dove (2002) concludes that the resulting system is, like many other systems, neither indigenous nor exogenous but rather hybrid in character.

Nevertheless, it remains interesting to see how scientific knowledge is perceived, interpreted and integrated with local visions and experiences, to have experimental confirmation of research hypotheses when these are crosschecked by the farmers in the field, or to discover discrepancies or incompleteness in the scientific assumptions as input for future research subject refinement.

After completion of the series of in-depth interviews with farmers, the constructed knowledge base was analysed and discussed. The next step was to set up a questionnaire[43] dealing with the most crucial and controversial aspects of the knowledge system, in order to verify the previously elicited knowledge and to augment our understanding of it. At the same time, the representation and distribution of the knowledge across the community as a whole could be verified. Hence, a random sample statistically representative of the community was required for conducting this validation test. In general we opted to follow the contentious rule that when 75% or more farmers agree with a particular statement, it can be regarded as part of local core knowledge.

75% or more of the 28 farmers were found to agree with 62 of these 68 key statements that were included in the test, which means that the high figure of 91.18% of the tested statements can be considered as being part of core knowledge of Sumberjaya farmers[44]. The six non-validated statements partially originated from remarkable and controversial statements from the constructed knowledge base, and therefore augment our understanding of farmers’ reasoning. In Annex XVII, the hypothetical reasons why these six statements have not been validated are discussed. Many farmers also gave comments on the statements and conditions, revealing information that augmented our knowledge base.

The outcome of the test is that all farmers seem to have a fairly detailed and accurate knowledge of the covered topics, as reflected by the high number of validated statements. The test results also indicate that this knowledge is basically the same for all groups of farmers. This high degree of knowledge homogeneity may be called a surprise, as previous research and most literature indicated that local knowledge is likely to be quite heterogeneous and variable within a local population (Blaikie et al., 1997; Brokensha, 1998; Warren, 1998; Davis and Wagner, 2003), especially between the different ethnic groups resulting from their particular traditions and techniques (Chapman, 2001).

In contrast to our expectations, no indications for differences could be discovered during the series of in-depth interviews and the subsequent analysis of the collected knowledge. This made the analysis of the test results interesting to indicate possible differences in knowledge as held by different groups on farmers, distinguished by location (subcatchment), position (downstream, midstream or upstream), ethnic group (Javanese, Semendonese or Sundanese) and type of fields (with or without paddy field). Among all factors determined during the interview sessions, these had been distinguished as having highest probability to cause knowledge differences. Statistical analysis of the test results revealed no indications of knowledge differences between ethnic groups, between farmers from different subcatchments and between farmers with or without paddy rice fields. Only the position of the farmers’ fields seemed to have an influence, not because of any indication of different perceptions but rather concerning the extent of knowledge acquisition, as upstream farmers systematically proved to be a bit more knowledgeable than their colleagues from more downstream parts of the catchments.

Annex XVIII contains the scatter plots resulting from the correspondence analysis and a subsequent discussion of revealed disparities.

Our results clearly demonstrate that farmers are aware of the deforestation, erosion and water problems and that their knowledge about these phenomena is detailed and widespread. They also know about the processes, the reasons behind these problems and possible solutions and ways to overcome them. Still, if one wanders around in the area, if one looks at the coffee gardens of those farmers, one will notice that they rarely apply the soil and water conservation measures they know. Likewise, most of the riparian area is converted to paddy field or coffee garden, whereas farmers stressed the importance of vegetation buffers on the riversides. Of course, there are multistrata gardens, and there are farmers who apply and experiment with all kind of techniques, but the majority of the gardens still confirm this lack of implementation of ecological knowledge almost all have shown to possess. Automatically one starts asking oneself, when the farmers are really convinced of the use of conservation techniques and defend and explain them with enthusiasm, why then do they not apply them in practice?

If you ask farmers why most of them do not implement their knowledge despite all known advantages, the answer most often is because they are lazy, because they lack capital, labour force and time. Indeed it is true, constructing terraces and maintaining them, obtaining and planting seedlings, taking care of and pruning shade trees, digging composting holes and furrows; it all takes a lot of time without any short-term benefit. Other LEK studies confirm that lack of labour is one of the constraints farmers face when applying widely known techniques that reduce soil erosion (Visser et al., 2003).

The uncertain land tenure situation is an important factor that withholds farmers from investing in soil conservation techniques, as their main interests are short term profits, resulting in the simple monoculture sun coffee systems (Agus et al., 2002).

Another element to keep in mind is the monetary and economic crisis Indonesia has dealt with in the end of the nineties. Prices of all goods have skyrocketed, while average incomes stagnated. Consequently, an impoverishment of the Indonesian society has taken place and its effects still predominate today. The farmer’s main concerns are how to feed the family this week, how to send the children to school and how to collect the necessary money to survive. Although the coffee farmers in Sumberjaya have been able to save some money from the golden years of high coffee prices in the nineties, their situation now is aggravated once more by the coffee price that has plummeted from over 15,000 Rp. in 1997 to a mere 3,000 Rp. per kilo. This harsh economic situation forces the farmer’s attention to be focused on short-term profit in order to survive, rather than emphasising the long-term sustainability of his agricultural practice. The low coffee price doesn’t motivate the farmer to work long days in his coffee garden to ensure his future yields by implementing soil conservation measures. Thus the reason why farmers call themselves “lazy” is because they do not have any economic incentives to implement their rich knowledge. Other LEK research similarly reveals that farmers with a thorough knowledge of environmental requirements may cultivate inappropriate land for overriding socio-personal or economic reasons. Continued widespread application of erosive practices is not necessarily a consequence of farmer ignorance or inability to recognise which land is suitable for agriculture. Many farmers just do not have the privilege of choosing between suitable and unsuitable sites for their crops. They possess production-enhancing knowledge and skills but decide not to apply them because they perceive no economic benefit in doing so (Ryder, 2003). Unfavourable socio-political and economic circumstances are seen as the main burden for sustainable land use decisions and for application of local conservation techniques (Sillitoe, 1998; Brodt, 2001; Cools et al., 2003; Davis and Wagner, 2003; Niemeijer and Mazzucato, 2003).

A recent phenomenon is the conversion from coffee gardens to more profitable vegetable fields of chilli peppers, tomatoes and beans; it confirms the difficult situation the farmers face. Owners of multistrata coffee gardens cut down their whole field to plant vegetables on the cleared soil. They also know that these fields are very sensitive to erosion because of the low soil coverage of the small vegetable crops, but they admit they have no other choice than going for short-term profit when coffee prices are so devastating low. Nevertheless, most economic studies have shown agroforestry in a favourable light (Young, 1997). Comparisons of technified monoculture coffee and traditional, shade or multistrata systems have shown that the latter resulted in significantly higher net revenues per acre – and the increased benefit is even greater when externalities such as the wide-ranging social and environmental costs of modern intensive monoculture systems are incorporated (Dicum and Luttinger, 1999). Profitability assessment based on the macroeconomic parameters of year 2000, reveals that shade coffee system offers reasonable higher return to land and higher returns to labour than sun coffee does. In times of low coffee prices like now, the multistrata system would even be the only economically viable one (Budidarsono et al., 2000; Budidarsono, personal communication, 2002).

On the other hand, if the coffee prices will go up again and reach high levels, one questions himself whether the farmers will not be pushed to plant more monoculture coffee plantations and even convert some of the multistrata gardens because of high profitability of coffee plants. More research surely needs to be done to have more insight in how economics influence land use, but at least it can be said that economic factors have a major impact on all farmers’ practices.

On a last meeting with leaders of farmer groups, the research results were presented and the farmers were asked directly for their opinion on how implementation can be enhanced. A very important element would be a good organisation of the farmers in groups. According to them, the farmer groups play an important role in the preservation of their common environment. They have to work together, replant the riversides, protect the forest, spread information and assist each other in the application of soil conservation techniques and other cultivation practices. Only by bundling their forces, by taking the challenges together, they will be able to change the situation. The same was emphasised by the local coordinator for agricultural extension, Pak Yedi Rohyadi; farmers have to form groups to protect the remaining forests and to tackle the erosion and water problems together. It is clear that individual efforts will not solve problems concerning watershed functions, collective action is needed and that is why it is of utmost importance that the farmers are united in farmer groups. A reason why it is difficult for the farmer groups to exist and to implement soil and water protection at the watershed level lies within the history of the area. Recent migration, land clearing and struggle for the best lands have fed jealousy and conflicts between individuals living close to each other. The co-operation and trust needed for proper functioning of the farmer groups and for communal protection of the region is often lacking, and competition between farmer groups is not making the situation easier.

Probably only when the adverse consequences of their land management practices and the higher yields of the sustainable land use become more obvious and more pertinent than their personal quarrels, farmers will take the initiative and struggle for the preservation of their region. Hopefully, by then, it is not too late yet.

As shown throughout the presentation of the research results, farmers’ understanding and scientific knowledge are quite similar, so the mutual understanding between these scientists and farmers would be very good if they could manage to use the same terminology. A very limited number of interviews with other stakeholders has been conducted, merely to have an idea to what extent the farmers’ knowledge is shared by other groups. Consequently, no hard conclusions can be made out of the obtained information; the results only serve as indications or starting points for further investigation. The perceptions of the following groups have been found interesting to explore: non-farming villagers (Pak Osih, Ibu Osih and Pak Margani), the local head of agricultural extension (Pak Yedi Rohyadi), the local director of the Forestry Department (Dinas Kehutanan: Pak Hartawan) and the director of operation of the PLTA dam (Pak Sugeng).

Non-farming villagers seem to have quite a good idea of all topics, very comparable to the farmers’ knowledge. Their close contact with the farmer population that constitutes the vast majority of inhabitants of the villages and their roots in farming families explains this similarity of perceptions.

Except for the use of riparian vegetation and the paddy field filter function, also the perceptions of the extension officers seem to be in conformity with the farmers’ knowledge, although they seem to use both local and more scientific terminology.

The two other stakeholders, officers of the Forestry Department and even more the Direction of the PLTA, appeared to know less about agricultural issues and solutions, although they have a good understanding of the forest functions, the consequences of clearance involving erosion and water problems, and the necessity of natural riparian vegetation. The Head of the Forestry Department has some ideas about weeding and even heard about the Arachis cover crop, but his knowledge about conservation techniques seems to be quite limited. The Director of Operation of the PLTA admits that forestry and agriculture are not his domains, and his knowledge is limited to the basic processes involving forest presence and the conditions of the river water.

In the end, no radically opposed perceptions have been revealed during the limited number of discussions with the different stakeholders. This is a windfall for future cooperation for the management of local natural resources. Nevertheless, it is clear that the knowledge of some stakeholders is quite limited for a few topics. Providing them with correct information might broaden their vision and help to increase their willingness to co-operate.

6. Conclusions

Investigation of the local knowledge of Sumberjaya farmers concerning ecological processes and soil and water functions in their watershed resulted in the construction of an elaborate knowledge database, interlinking various aspects to a coherent complexity. Elicited information obtained from in-depth interviews and test questionnaires has been collected carefully in a systematised and computerised manner to gain insights in the local knowledge system as a whole, allowing for thorough analysis and interpretation of the knowledge compared to conventional scientific thinking and perceptions of other local stakeholders.

The final conclusions of this research project can be summarised as follows:

1. Sumberjaya farmers have a detailed and profound understanding of the watershed functions and how these can be preserved, quite similar to scientific knowledge;

2. The farmer knowledge is fairly homogeneous as no significant differences in knowledge have been revealed;

3. In spite of a thorough knowledge of the consequences of deforestation, erosion and water quality problems, a lot of farmers miss the incentives to implement this knowledge;

4. The perceptions of different stakeholders are not radically opposed to each other, and mainly differ in the extent to which they possess detailed knowledge on the different topics.

For future research, it might be a good idea to use the same questionnaire to test to which extent this knowledge is shared with new immigrants and with farmers in neighbouring watersheds with similar conditions. More detailed questionnaires could be designed to verify whether the indicated knowledge homogeneity truly prevails. The ecological knowledge of other stakeholders, of which in this research only some indications have been collected, should be elaborated intensively to be aware of all differing perceptions as an aid in negotiating future common land use management.

Another interesting and useful research topic is ethnopedology and local land suitability criteria, which has only been touched briefly in this research and surely deserves more attention.

Nevertheless, the most important conclusion from this research should be the third one, as farmers indeed exhibit a rich and detailed knowledge but are prevented from making full use of it, because of a variety of impediments of which the identity and relative importance still are to be investigated. In the end, whether farmers know or don’t know how to preserve the watershed will not be of great importance, what matters is whether they will act to preserve.

A participatory appraisal of the decisive factors that make farmers move towards a certain type of land use would be very useful in order to have an idea of what needs to be done to promote sustainable agriculture that restores and preserves the watershed functions. A certain technique is only effective as a means of conservation when local people adopt it (Young, 1997). Decision-making goes beyond empirical evidence and is influenced by a range of institutional and cultural issues (Martin, 2003). A lot of factors influence the farmers’ choices, such as market situation; capital, labour and time requirements; short-term profitability; land tenure situation; farmers’ preferences; and many more, but we do not have an idea of which of these factors really predominate. Immediate economic benefit usually comes high among decision-making criteria by farmers, and in this respect there should be more research into the use of cash crops as perennials (Young, 1997). Strikingly, this is exactly what some farmers meant when asking for more research and information about how to integrate valuable fruit trees in their coffee gardens.

References

AFFANDI, T., ROSADI, B., MANIK, T.K., SENGE, M., OKI, Y. and ADACHI, T. (1999). The dynamics of soil water pressure under coffee tree with different weed management in a hilly area of Lampung, Indonesia. In: Proceedings International Seminar Toward Sustainable Agriculture in Humid Tropics Facing 21^st Century. Bandar Lampung, Sept. 27-28, 1999. University of Lampung. 387-394.

AFFANDI, T. et al. (2002a). Erosion under Coffee Trees with Different Weed Managements in Humid Tropical Hilly Area of Lampung, South Sumatra, Indonesia. Journal of the Japanese Society of Soil Physics, 91, in press.

AFFANDI, T. et al. (2002b). Sediment Yield from Various Uses Practices in a Hilly Tropical Area of Lampung Region, South Sumatra, Indonesia. Journal of the Japanese Society of Soil Physics, 91: 25-38, in press.

AGRAWAL, A. (1995). Dismantling the divide between indigenous and scientific knowledge. Development and Change, 26: 413-439.

AGUS, F. (1993). Soil processes and crop production under contour hedgerow systems on sloping Oxisols. PhD dissertation, North Carolina State University, Raleigh, NC, USA, 141 p.

AGUS, F. (2002). Facilitating the Development of Conservation Farming and Agroforestry Systems. Compiled and summarised from the August 2002 report of the ASB Phase 3 team composed of researchers from Brawijaya University, University of Lampung and the Center for Soil and Agroclimate Research and Development. In: ICRAF – backgrounds for ACIAR project planning meeting.

AGUS, F., GINTINGS, A.N. and VAN NOORDWIJK, M. (2002). Pilihan teknologi agroforestri/konservasi tanah untuk areal pertanian berbasis kopi di Sumberjaya, Lampung Barat. World Agroforestry Centre, ICRAF Southeast Asia, 60 pp.

ANKERSMIT, N. (1940). De Robusta-Koffie-Cultuur in Nederlandsch-Indie. Uitgeverij W. van Hoeve Deventer. 108 p.

ARMITAGE, D.R. (2003). Traditional agroecological knowledge, adaptive management and the socio-politics of conservation in Central Sulawesi, Indonesia. Environmental Conservation, 30 (1): 79-90.

BARTH, F. (2002). An anthropology of knowledge. Current Anthropology, 43 (1): 1-18.

BASRI, I.H., MERCADO JR., A.R. and GARRITY, D.P. (1990). Upland Rice Cultivation Using Leguminous Tree Hedgerows on Strongly Acid Soils. IRRI Saturday Seminar Paper, March 31. Agronomy, Plant Physiology, and Agroecology Division, International Rice Research Institute, Los Banos, Philippines, 32 p.

BELSKY, A.J., MWONGA, S.M. and DUXBURY, J.M. (1993). Effects of widely spaced trees and livestock grazing on understory environments in tropical savannas. Agroforestry Systems 24, 1-20.

BENFER, R.A. and FURBEE, L. (1990). Knowledge acquisition: lessons from anthropology. AI Applications in Resource Management, 4 (3): 19-26.

BENOIT, D. (1999). Migration and Structure of Population. Transmigration and Spontaneous Migration in Indonesia: Propinsi Lampung. M. Pain, D. Benoit, P. Levang and O. Sevin (eds.). Bonty, ORSTOM.

BERKES, F. (1993). Traditional ecological knowledge in perspective. In: Traditional Ecological Knowledge: Concepts and Cases. Inglis, T.J. (ed.). Ottawa: Canadian Museum of Nature and International Development Research Centre. 1-9.

BERKES, F., COLDING, J. and FOLKE, C. (2000). Rediscovery of traditional ecological knowledge as adaptive management. Ecological Applications, 10 (5): 1251-1262.

BLAIKIE, P., BROWN, K., STOCKING, M., TANG, L., DIXON, P., SILLITOE, P. (1997). Knowledge in action: Local knowledge as a development resource and barriers to its incorporation in natural resource research and development. Agricultural Systems, 55 (2): 217-237.

BRODT, S.B. (2001). A systems perspective on the conservation and erosion of indigenous agricultural knowledge in central India. Human Ecology, 29 (1): 99-120.

BROKENSHA, D. (1998). Comments on Sillitoe (1998). Current Anthropology, 39 (2): 236-237.

BUDIDARSONO, S., ADI KUNCORO, S. and TOMICH, T.P. (2000). A profitability assessment of Robusta coffee systems in Sumberjaya watershed, Lampung, Sumatra, Indonesia. Bogor, ICRAF-SEA.

CAMPBELL, B.M., FORST, P. and KING, J.A. (1994). The influence of trees on soil fertility on two contrasting semi-arid soil types at Matopos, Zimbabwe. Agroforestry Systems 28, 159-172.

CAMPBELL, L.M. and VAINIO-MATTILA, A. (2003). Participatory development and community-based conservation: Opportunities missed for lessons learned? Human Ecology, 31 (3): 417-437.

CHAPMAN, M. (2001). Local Ecological Knowledge of Soil and Water Conservation Practices in Coffee based systems in the Way Besai watershed of Sumberjaya. Preliminary Fieldwork Report April-August 2001. ICRAF SEA. Unpublished.

CHARRAS, M and PAIN, M. (1993). Spontaneous settlements in Indonesia: Agricultural pioneers in Southern Sumatra. ORSTOM, CNRS, Departemen Transmigrasi.

CLEVELAND, D.A. (1998). Comments on Sillitoe (1998). Current Anthropology 39 (2): 237-238.

COLDING, J. and FOLKE, C. (2001). Social taboos: "Invisible" systems of local resource management and biological conservation. Ecological Applications, 11 (2): 584-600.

COOLS, N., DE PAUW, E. and DECKERS, J. (2003). Towards an integration of conventional land evaluation methods and farmers' soil suitability assessment: A case study in northwestern Syria. Agriculture Ecosystems & Environment, 95 (1): 327-342.

COOPER, P.J.M., LEAKEY, R.R.B., RAO, M.R. and REYNOLDS, L. (1996). Agroforestry and the mitigation of land degradation in the humid and sub-humid tropics of Africa. Experimental Agriculture 32, 235-290.

CRITCHLEY, W.R.S., REIJ, C., WILLCOCKS, T. (1994). Indigenous Soil and Water Conservation: A Review of the State of Knowledge and Prospects for Building on Traditions. Land Degradation and Rehabilitation, 5(4):293-314.

CRITCHLEY, W.R.S. (1999). Harnessing traditional knowledge for better land husbandry in Kabale District, Uganda. Mountain Research and Development, 19 (3): 261-272.

CRITCHLEY, W.R.S. (2000). Inquiry, initiative and inventiveness: Farmer innovators in East Africa. Physics and Chemistry of the Earth, Part B - Hydrology Oceans and Atmosphere, 25 (3): 285-288.

CRITCHLEY, W.R.S. and MUTUNGA, K. (2003). Local innovation in a global context: Documenting farmer initiatives in land husbandry through Wocat. Land Degradation & Development, 14 (1): 143-162.

DAVIS, A. and WAGNER, J.R. (2003). Who knows? On the importance of identifying "Experts" when researching local ecological knowledge. Human Ecology, 31 (3): 463-489.

DECKERS, J.A., NACHTERGAELE, F.O., SPAARGAREN, O.C. (1998). World Reference Base For Soil Resources: Introduction. Acco, Leuven. 165 pp.

DICUM G. and LUTTINGER, N. (1999). The Coffee Book: Anatomy of an industry from crop to the last drop. The New Press, New York. 196 pp.

DIXON, H.J., DOORES, J.W., JOSHI, L. and SINCLAIR, F.L. (2001). Agroecological Knowledge Toolkit For Windows: Methodological Guidelines, Computer Software and Manual For AKT5. School of Agricultural and Forest Sciences, University of Wales, Bangor, UK. 127 p.

DRIESSEN, P., DECKERS, J.A., SPAARGAREN, O.C., NACHTERGAELE, F.O. (2001). Lecture notes on the major soils of the world. FAO.

DOVE, M.R. (2002). Hybrid histories and indigenous knowledge among Asian rubber smallholders. International Social Science Journal, 54 (3): 349.

ELLEN, R. (1998). Comments on Sillitoe (1998). Current Anthropology, 39 (2): 238-239.

ELLEN, R. and HARRIS, H (1996). Concepts of indigenous environmental knowledge in scientific and development studies literature - A critical assessment. Draft paper East-West Environmental Linkages Network Workshop 3, Canterbury. Available on http://lucy.ukc.ac.uk/Rainforest/SML_files/Occpap/indigknow.occpap_TOC.html. Accessed: October 19, 2003

ELLEN, R. and HARRIS, H. (2000). Introduction. In: Indigenous Environmental Knowledge and its Transformations: Critical Anthropological Perspectives. Ellen, R., Parkes, P., Bicker, A. (eds.). OPA, Amsterdam. 1-33.

ELLIS-JONES, J. and TENGBERG, A. (2000). The impact of indigenous soil and water conservation practices on soil productivity: Examples from Kenya, Tanzania and Uganda. Land Degradation & Development, 11 (1): 19-36.

ELMHIRST, R. (1996). Transmigration and Local Communities in North Lampung: Exploring Identity Politics and Resource Control in Indonesia. The Association of South East Asian Studies Conference, SOAS, London.

FERRADÁS, C. (1998). Comments on Sillitoe (1998). Current Anthropology, 39 (2): 239-240.

FLAVIER, J.M., DE JESUS, A. and NAVARRO, C. (1995). The regional program for the promotion of indigenous knowledge in Asia. In: The cultural dimension of development: Indigenous knowledge systems. Warren, D.M., Slikkerveer, L.J. and Brokensha, D. (eds). Intermediate Technology Publications, London. 479-487.

FORSYTH, T. (1998). Comments on Sillitoe (1998). Current Anthropology, 39 (2): 240-241.

FUJISAKA, S. (1992). A Case of Farmer Adaptation and Adoption of Contour Hedgerows for Soil Conservation. International Rice Research Institute, Los Banos, Laguna, Philippines.

FURBEE, L. (1989). A folk expert system: soil classification in the Colca Valley, Peru. Anthropological Quarterly 62 (2): 83-102.

GARRITY, D.P., MERCADO JR, A. and SOLERA, C. (1995). The nature of species interference and soil changes in contour hedgerow systems on sloping acidic lands. In: Alley Farming. Kang, B.T. (ed.). International Institute of Tropical Agriculture, Ibadan, Nigeria.

GARRITY, D.P. (1996). Tree-soil-crop interactions on slopes. In: Tree-Crop Interactions: A Physical Approach. ONG, C.K. and HUXLEY, P. (eds.). CAB International, Wallingford, UK, 299-318.

GINTINGS, A.N. (1982). Aliran Permukaan dan Erosi Tanah yang Tertutup Tanaman Kopi dan Hutan Alam di Sumberjaya – Lampung Utara. Laporan No. 399. Balai Penelitian Budidaya, Jember.

GOBIN, A., CAMPLING, P., DECKERS, J. and FEYEN, J. (2000a). Quantifying soil morphology in tropical environments: Methods and application in soil classification. Soil Science Society of America Journal, 64 (4): 1423-1433.

GOBIN, A., CAMPLING, P., DECKERS, J. and FEYEN, J. (2000b). Integrated Toposequence Analyses to combine local and scientific knowledge systems. Geoderma, 97 (1-2): 103-123.

GONZALEZ, R.M. (1995). KBS, GIS and documenting indigenous knowledge. Indigenous Knowledge and Development Monitor 3(1).

Electronic version available on http://www.nuffic.nl/ciran/ikdm. Nuffic-CIRAN, The Hague, Netherlands.

GROSSMAN, J.M. (2003). Exploring farmer knowledge of soil processes in organic coffee systems of Chiapas, Mexico. Geoderma, 111 (3-4): 267-287.

HALL, T.E. and BIGLER-COLE, H. (2001). Sociocultural factors and forest health management. Northwest Science, 75: 208-233.

HAMILTON, L.S. and PEARCE, A.J. (1987). What are the soil and water benefits of planting trees in developing country watersheds? In: Sustainable Resource Development in the Third World. Southgate, D.D. and Disinger, J.F. (eds). Westview, Boulder, Colorado, USA. 39-58.

HARTOBUDOYO, D. (1979). Pemangkasan kopi. Balai Penelitian Perkebunan Bogor, Sub Balai Penelitian Budidaya, Jember.

HOCKING, D., SARWAR, G. and YOUSUF, S.A. (1997). Trees on farms in Bangladesh. 4. Crop yields underneath traditionally managed mature trees. Agroforestry Systems, 29: 313-320.

HOLLE, K.F. (1863). Proefhandleiding voor de kultuur en gewone inlandsche bereiding van Koffij. Batavia. Gedrukt bij W. Ogilvie. 24 p.

HUITEMA, W.K. (1935). De bevolkingskoffiecultuur op Sumatra. Tropische landbouw. Wageningen, landbouwhogeschool. 238 pp.

HUNTINGTON, H.P., BROWN-SCHWALENBERG, P.K., FROST, K.J., FERNANDEZ-GIMENEZ, M.E., NORTON, D.W. and ROSENBERG, D.H. (2002). Observations on the workshop as a means of improving communication between holders of traditional and scientific knowledge. Environmental Management, 30 (6): 778-792.

ICRAF (1998). Agroforestry under pressure: Lampung research planning trip. Unpublished.

ICRAF (2000). ICRAF’s Lampung research, update on current activities and plans, prepared for field trip 7-11 February 2000. Unpublished.

ICRAF (2001). Backgrounds for Sumberjaya 2001 research planning meeting. Unpublished.

JUNIUS, F.J. (1933). Memorie van overgave van de afgetreden Resident der Lampongse Districten. Buitenzorg.

KHASANAH, N.; LUSIANA, B., FARIDA and VAN NOORDWIJK, M. (2002). Simulating Runoff and Soil Erosion using WaNuLCAS: Water, Nutrient and Lights Capture in Agroforestry System. In: Backgrounds for ACIAR Project Planning Meeting. Unpublished.

More information: http://www.worldagroforestrycentre.org/sea/AgroModels/WaNulCAS/index.htm

KIEPE, P. (1995). No Runoff, No Soil Loss: Soil and Water Conservation in Hedgerow Barrier Systems. Tropical Resource Management Papers, Agricultural University, Wageningen, Netherlands.

KIMMERER, R.W. (2002). Weaving traditional ecological knowledge into biological education: A call to action. Bioscience, 52 (5): 432-438.

KRASIL'NIKOV, V. (1999). Early studies on folk soil terminology. Eurasian Soil Science, 32 (10): 1147-1150.

KRASIL’NIKOV, P.V. and TABOR, J.A. (2003). Perspectives on utilitarian ethnopedology. Geoderma, 111 (3-4): 197-215.

KUSWORO, A. (2000). Community-forest Interactions and Forestry Policies in Sumber Jaya, Lampung. ICRAF’s Lampung research, update on current activities and plans, prepared for field trip 7-11 February 2000.

KUSWORO, A. (2000). Dispute over the Forest Zone in Sumber Jaya Lampung: Case Studies in Dwikora and Sumber Jaya. ICRAF’s Lampung research, update on current activities and plans, prepared for field trip 7-11 February 2000.

KUSWORO, A. (2000). Perambah Hutan atau Kambing Hitam? Portret Sengketa Kawasan Hutan di Lampung. Bogor, Pustaka Latin.

LAL, R., SANCHEZ, P.A. and CUMMINGS, R. (1986). Land Clearing and Development in the Tropics. Balkema, Rotterdam, Netherlands.

LEAKEY, R. (1996). Definition of agroforestry revisited. Agroforestry Today, 8 (1): 5-7.

LEAKEY, R. (1997). Redefining agroforestry – and opening Pandora’s box? Agroforestry Today, 9 (1): 5.

LI, T.M. (2000). Locating Indigenous Environmental Knowledge in Indonesia. In: Indigenous Environmental Knowledge and its Transformations: Critical Anthropological Perspectives. Ellen, R., Parkes, P., Bicker, A. (eds.). OPA, Amsterdam.121-149.

LISWANTI, N. (1999). Sumberjaya Biodiversity Survey. 2-10 September 1999. ICRAF. Draft Report. Unpublished.

LUNDGREN, L. and LUNDGREN, B. (1979). Rainfall, interception and evaporation in the Mazumbai Forest reserve, West Usambara Mountains, Tanzania and their importance in the assessment of land potential. Geografiska Annaler, 61A: 157-178.

MACKINSON, S. (2001). Integrating local and scientific knowledge: An example in fisheries science. Environmental Management, 27 (4): 533-545.

MARTIN, A. (2003). On knowing what trees to plant: local and expert perspectives in the Western Ghats of Karnataka. Geoforum, 34 (1): 57-69.

MOUGEOT, E. and LEVANG, P. (1990). Marketing of rice, cassave and coffee in Lampung, Indonesia. Bogor, CGPRT Centre. 123 p.

MUEHLEBACH, A. (2001). "Making place" at the United Nations: Indigenous cultural politics at the UN Working Group on Indigenous Populations. Cultural Anthropology, 16 (3): 415-448.

NIEMEIJER, K. (1995). Indigenous soil classifications: complications and considerations.

Indigenous Knowledge and Development Monitor 3(1). Nuffic-CIRAN, The Hague, Netherlands.

NIEMEIJER, D. and MAZZUCATO, V. (2003). Moving beyond indigenous soil taxonomies: local theories of soils for sustainable development. Geoderma, 111 (3-4): 403-424.

NORTON, J.B., PAWLUK, R.R. and SANDOR, J.A. (1998). Observation and experience linking science and indigenous knowledge at Zuni, New Mexico. Journal of Arid Environments, 39 (2): 331-340.

NYGREN, A. (1999). Local knowledge in the environment-development - Discourse from dichotomies to situated knowledges. Critique of Anthropology, 19 (3): 267-288.

OLSSON, P. and FOLKE, C. (2001). Local ecological knowledge and institutional dynamics for ecosystem management: A study of Lake Racken Watershed, Sweden. Ecosystems, 4 (2): 85-104.

PANDEY, D.N. (2003). Cultural resources for conservation science. Conservation Biology, 17 (2): 633-635.

POSEY, D.A. (1998). Comments on Sillitoe (1998). Current Anthropology, 39 (2): 241-242.

PRETTY, J.N. and SHAH, P. (1997). Making soil and water conservation sustainable: From coercion and control to partnerships and participation. Land Degradation & Development, 8 (1): 39-58.

PURCELL, T.W. (1998). Indigenous knowledge and applied anthropology: Questions of definition and direction. Human Organization, 57 (3): 258-272.

PUSAT PENELITIAN TANAH (1989). Buku Keterangan Satuan Lahan dan Tanah Lembar Kotaagung, Sumatra. Pusat Penelitian Tanah, Bogor.

ROBBINS, P. (2000). The practical politics of knowing: State environmental knowledge and local political economy. Economic Geography, 76 (2): 126-144.

ROBBINS, P. (2003). Beyond ground truth: GIS and the environmental knowledge of herders, professional foresters, and other traditional communities. Human Ecology, 31 (2): 233-253.

ROSS, A. and PICKERING, K. (2002). The politics of reintegrating Australian Aboriginal and American Indian indigenous knowledge into resource management: The dynamics of resource appropriation and cultural revival. Human Ecology, 30 (2): 187-214.

RUINARD, J. (1951). Het gebruik van schaduwbomen by de cultuur van koffie, thee en cacao in Indonesie. Irs-scriptie voor Tropische Houtteelt. 43 p.

RYDER, R. (2003). Local soil knowledge and site suitability evaluation in the Dominican Republic. Geoderma, 111 (3-4): 289-305.

SAJJAPONGSE, A. (1992). Management of sloping lands for sustainable agriculture in Asia: An overview. In: Technical Report on the Management of sloping lands for Sustainable Agriculture in Asia: Ohase I (1988-1991). Network Document No 2, International Board for Soil Research and Management, Bangkok, Thailand. 1-15.

SANDOR, J.A. and FURBEE, L. (1996). Indigenous knowledge and classification of soils in the Andes of Southern Peru. Soil Science Society of America Journal, 60 (5): 1502-1512.

SILLITOE, P. (1998). The development of indigenous knowledge - A new applied anthropology. Current Anthropology, 39 (2): 223-252.

SMITHSONIAN MIGRATORY BIRD CENTER (1997). Why Migratory Birds are Crazy for Coffee. Washington DC.

SNOEP, W. (1932). Over cultuurmaatregelen betreffende de bodembehandeling en schaduw bij koffie. De Bergcultures, jrg.6.

SRIYANI, N., SUPRAPTO, H., SUSANTO, H., LUBIS, A.T. and OKI, Y. (1997). Wood population dynamics in Coffee Plantation Managed by Different Soil Conservation Techniques. Faculty of Agriculture, University of Lampung, and Faculty of Environmental Science and Technology, Okayama University, Japan.

STIRRAT, R.L. (1998). Comments on Sillitoe (1998). Current Anthropology 39 (2): 242-243

STONE, P. (1998). Comments on Sillitoe (1998). Current Anthropology, 39 (2): 243.

SUYANTO, S. (2000). The Underlying Causes and Impacts of fires in Southeast Asia: Fire, deforestation and land tenure in the northeastern fringes of Bukit Barisan Selatan National Park, Lampung. Socio-economic Report Site 1. Sekincau, Lampung Province, Indonesia. Bogor, ICRAF-SEA. 25.

SYAM T., NISHIDE, H., SALAM, A. K., UTOMO, M., MAHI, A.K., LUMBANRAJA J., NUGOHO, S. G. and KIMURA, M. (1997). Land Use and Cover Changes in a Hilly Area of South Sumatra, Indonesia (from 1970 to 1990). Soil Science and Plant Nutrition, 43 (3), 587-599.

TANDEAN, M. (2000). Preliminary results of analysis of population data of sub-district Sumberjaya. Report of a 2-month internship at ICRAF. Unpublished.

TENGBERG, A., MURIITHI, L. and OKOBA, B. (1999). Land management on semi-arid hillsides in eastern Kenya: Learning from farmers' diverse practices. Mountain Research and Development, 19 (4): 354-363.

THOMSON, A.J. (2000). Elicitation and representation of Traditional Ecological Knowledge, for use in forest management. Computers and Electronics in Agriculture, 27 (1-3): 155-165.

TICK, A.W. (1993). Editorial. Indigenous Knowledge and Development Monitor 1(3). Nuffic-CIRAN, The Hague, Netherlands.

TURKELBOOM, F., ONGPRASERT, S. and TAEJAJAI, U. (1993). Alley cropping on steep slopes: Soil fertility gradients and sustainability. Paper presented at the International Workshop on Sustainable Agricultural Development: Concepts and Measures. Asian Institute of Technology, Bangkok, December 14-17. 16 p.

ULTEE, Dr. J.A. (1949). Koffiecultuur der ondernemingen. In: ‘De Landbouw in de Indische Archipel’ uitgegeven onder redactie van Dr. C.J.J. van Hall en C. van de Koppel. Deel II_B: genotmiddelen en specerijen. Van Hoeve ,‘s Gravenhage. 7-88.

USHER, P.J. (2000). Traditional ecological knowledge in environmental assessment and management. Arctic, 53 (2): 183-193.

VAN DER VEEN, R. (1935). Wortelconcurrentie in de koffie- en rubbertuinen. Archief voor de koffiecultuur, jrg. 9.

VAN NOORDWIJK, M., TOMICH, T.P., DE FORESTA, H. and MICHON, G. (1997). To segregate or to integrate? The question of balance between production and biodiversity conservation in complex agroforestry systems. Agroforestry Today, 9: 6-9.

VERBIST, B. (2000). Landscape context of Sumberjaya. ICRAF’s Lampung research, update on current activities and plans, prepared for field trip 7-11 February 2000. Unpublished.

VERBIST, B. (2001). Land use and its changes in Sumberjaya. Backgrounds for Sumberjaya 2001 research planning meeting. ICRAF. Unpublished.

VERBIST, B., PUTRA, A.E.D. and BUDIDARSONO, S. (2002). Background paper: Sumberjaya land use change, history and its driving factors. ICRAF – backgrounds for ACIAR project planning meeting. Unpublished.

VISSER, S.M., LEENDERS, J.K. and LEEUWIS, M. (2003). Farmers’ perceptions of erosion by wind and water in northern Burkina Faso. Land Degradation & Development 14 (1): 123-132.

WAKINDIKI, I.I.C. and BEN-HUR, M. (2002). Indigenous soil and water conservation techniques: effects on runoff, erosion, and crop yields under semi-arid conditions. Australian Journal of Soil Research, 40 (3): 367-379.

WARREN, D. M. (1991). Using Indigenous Knowledge in Agricultural Development. World Bank Discussion Paper No.127. The World Bank, Washington, D.C.

WARREN, D.M. (1998). Comments on Sillitoe (1998). Current Anthropology, 39 (2): 244-245.

WERNER, O. and SCHOEPFLE, G.M. (1987). Systematic Fieldwork Volume 2: Ethnographic Analysis and Data Management. Sage Publications.

WIDIANTO and SUPRAYOGO, D. (2002). Effects of land use change on the hydrological functions of Way Besai Catchment in Sumberjaya, West Lampung. Jurusan tanah, Fakultas Pertanian Universitas Brawijaya. Progress report for the ACIAR Meeting, October 2002. Unpublished

WIDIANTO, NOVERAS, H., SUPRAYOGO, D., WIDODO, R.H., PURNOMOSIDI, P. and VAN NOORDWIJK, M. (2002). Konversi hutan menjadi lahan pertanian: Apakah fungsi hidrologis hutan dapat digantikan agroforestri berbasis kopi? Seminar HITI NTB, Mataram tanggal 27-28 Mei 2002.

WILSON, D.C. (2003). Examining the two cultures theory of fisheries knowledge: The case of bluefish management. Society & Natural Resources, 16 (6): 491-508.

WINKLERPRINS, A.M.G.A. (1999). Local soil knowledge: A tool for sustainable land management. Society & Natural Resources, 12 (2): 151-161.

WINKLERPRINS, A.M.G.A. and SANDOR, J.A. (2003). Local soil knowledge: insights, applications, and challenges – Preface. Geoderma, 111 (3-4): 165-170.

WULANDARI, R. (2002). Land Use Change in the Riparian Filterstrip of Sumberjaya, Lampung. Progress report for ICRAF – backgrounds for ACIAR project planning meeting. Unpublished.

YOUNG, A. (1997). Agroforestry for Soil Management, 2^nd edition. CAB International, Wallingford, UK. 320 p.

ZUBERI, M.I. (1998). Comments on Sillitoe (1998). Current Anthropology, 39 (2): 245-246.

ZURAYK, R., EL-AWAR, F., HAMADEH, S., TALHOUK, S., SAYEGH, C., CHEHAB, A.G. and AL SHAB, K. (2001). Using indigenous knowledge in land use investigations: a participatory study in a semi-arid mountainous region of Lebanon. Agriculture Ecosystems & Environment, 86 (3): 247-262.

Nederlandstalige samenvatting

Lokale ecologische kennis van bodem- en waterfuncties bij boeren in Sumberjaya, Sumatra, Indonesië

Sumberjaya is een valleiengebied in de bergen van de meest zuidelijke Sumatraanse provincie Lampung. Ondanks bewijzen van prehistorische aanwezigheid is het hele gebied tot 50 jaar geleden relatief onbewoond gebleven.

Rond het begin van de 20^ste eeuw immigreerden Semendo boeren naar de hoofdzakelijk nog met primair regenwoud bedekte streek om er ‘shifting cultivation’ toe te passen. Sinds de jaren 1950 echter steeg de bevolking van Sumberjaya aanzienlijk, zowel door geprogrammeerde als door spontane migratie, vooral vanuit het dichtbevolkte en nabijgelegen Java. De voornamelijk Sundaanse en Javaanse immigranten kapten het regenwoud en installeerde kleine koffietuinen op de hellingen en geïrrigeerde rijstvelden in de valleien. In de jaren 1970 was de bevolkingsaangroei zo groot en de ontbossing zo massaal dat er van het oorspronkelijke regenwoud weinig overbleef. Het Indonesische bosbouwministerie dat 80% van het Indonesische grondgebied beschouwde als staatsbos waarop niemand aanspraak mocht maken, voelde zich genoodzaakt in te grijpen. In de jaren ’80 en ’90, verdreef het Indonesische leger herhaaldelijke keren duizenden illegale koffieboeren uit Sumberjaya, terwijl ze de koffietuinen met de grond gelijkmaakten en herbebossingsprogramma’s initieerden. Begin jaren 1990 was de bouw van een dam en waterkrachtcentrale en bijgevolg de nood aan een regelmatige aanvoer van zuiver water een extra reden om het krachtdadige bewind door te zetten. Maar de aanpak had weinig effect. Gedreven door armoede en werkloosheid, kwamen de boeren terug naar de vruchtbare gronden van Sumberjaya. De verzwakte positie van de Indonesische regering na de economische en monetaire crisissen eind jaren 1990 en de toenemende internationale kritiek op het repressieve Indonesische beleid gaven aanleiding tot onderhandelingen tussen regering en boeren. Ondertussen is de regering aarzelend begonnen met enkele programma’s rond bosbeheer door de lokale gemeenschappen (community forestry), maar de toestand blijft onzeker voor de boeren.

Aangezien ecologische kennis van bepaalde gemeenschappen niet steeds autochtoon is, maar eerder een mengeling van autochtone en vreemde elementen, en vaak ook niet traditioneel maar eerder dynamische en innoverend, zijn de vaakgebruikte begrippen ‘Autochtone’ en ‘Traditionele’ kennissystemen niet helemaal consistent. Beter is het gewoon te spreken van ‘Lokale’ Ecologische Kennis (LEK). Deze kennis komt voort uit dagelijkse ervaringen en intiem contact met de lokale omgeving, en is dus eerder impliciet aanwezig in handelingen en gebruiken dan expliciet te verwoorden. Lokale kennis is een holistisch systeem met aanzienlijke coherentie en complexe verbondenheid met sociale en culturele wortels. Door globaliserende en homogeniserende trends komen lokale kennissystemen echter steeds meer in verdrukking.

Het wetenschappelijke vooruitgangsdenken en het concept van technologieoverdracht kon duidelijk niet tegemoetkomen aan de noden van kleine boeren in de derde wereld, en stilaan groeide het besef dat lokale kennis en waarden een grote rol moesten spelen in ontwikkelingsprojecten. Dit nieuwe ontwikkelingsparadigma vertrekt van de mensen zelf, met nadruk op eigen rechten en zelfbeschikking en als doel mensen in staat te stellen hun leven in eigen handen te nemen.

Vaak worden lokale kennis en universele, wetenschappelijke kennis beschouwd als twee fundamenteel verschillende systemen met andere epistemologische beginselen. Zo zou wetenschappelijke kennis gezien worden als empirisch, objectief en waardevrij, terwijl lokale kennis sterk verweven is met cultuur en waarden. Nochtans, steeds wordt duidelijker dat deze dichotomie een illusie is. Lokale kennis wordt steeds meer beschouwd als wetenschappelijk, en tegelijkertijd beseft men dat elk wereldbeeld, ook dat van de westerse wetenschap, sterk sociaal, cultureel en politiek bepaald is. Verder is het belangrijk te beseffen dat geen enkele kennis volledig endogeen of exogeen is, maar eerder een dynamische en heterogene hybride.

Mede omdat evaluatie van lokale ecologische kennissystemen plaatsvindt in de schemerzone tussen positieve en sociale wetenschappen, is er weinig consensus over welke methodologieën het meest efficiënt zijn. Het feit dat LEK vaak sterk lokaal en cultureel getint is, aanwezig in anekdotische vorm, niet expliciteerbaar in woorden, maar eerder pragmatisch en symbolisch, bemoeilijkt systematische analyse en classificatie. Verder blijkt ook dat inzichten uit LEK onderzoek zelden gebruikt worden in de praktijk van ontwikkelingssamenwerking. Tot groot ongenoegen van sociale wetenschappers, extraheren vele LEK onderzoekers kleine eenheden van lokale kennis uit hun culturele context, om het op die manier naar eigen goeddunken, als waren het onafhankelijke technische feiten, in hun eigen vertrouwde wetenschappelijke context te plaatsen. Deze uit haar context gerukte kennis draagt enkel bij tot de instandhouding van het ‘topdown’ ontwikkelingsparadigma van technologieoverdracht en doet dus vaak meer kwaad dan goed.

Diepgaand inzicht in de lokale cultuur is absoluut vereist om de impact van ontwikkelingsprojecten in te kunnen schatten, maar dit is allesbehalve eenvoudig. Ethnografisch onderzoek kost veel tijd, wat problematisch is voor de meeste ontwikkelingsprojecten. De wetenschap moet openstaan voor culturele, sociale en politieke reflecties. Aanvaarding van alternatieve denkwijzen en interculturele perspectieven zal bijdragen tot minder kritiekloze aanname van het eigen westers wetenschappelijke paradigma. Aan de andere kant zijn vele ethnografische studies te abstract, steriel en doorgaans weinig toegankelijk voor wetenschappers en ontwikkelingswerkers.

Lokale ecologische kennis wordt vaak bekeken als waardevol economisch goed en de controle erover is commercieel interessant. Verkeerde aanwending van deze kennis heeft vaak averechtse effecten. Autochtone rechten zijn in dit perspectief essentieel voor het cultureel overleven van de volkeren die deze kennis genereerden en onderhouden.

Evaluatie van de lokale ecologische kennis van boeren betreffende bodem- en waterfuncties in Sumberjaya is het onderwerp van onderzoek. Deze kennis ligt mede aan de basis van lokale besluitvorming over landbeheer en de documentatie en analyse ervan kunnen dus bijdragen tot een efficiëntere ondersteuning van boeren als beheerder van natuurlijke hulpbronnen. Meer specifiek spitst dit onderzoek zich toe op lokale kennis van erosieprocessen, functies van het waterbekken en de gevolgen voor waterkwaliteit. Deze processen interageren op landschapsniveau waarbij de verschillende landschapselementen zoals koffietuinen, bosrestanten en geïrrigeerde rijstvelden elk een eigen rol spelen.

- Inzichten over waterkwaliteit in historisch perspectief: oorzaken van achteruitgang en oplossingen

- Lokale landgeschiktheidsevaluatie: welke criteria zijn belangrijk voor een bepaald landgebruik

- Vergelijking van lokale ecologische kennis met conventionele wetenschappelijke kennis en met de kennis van andere lokale betrokkenen zoals dorpsbewoners die niet leven van de landbouw, de landbouwvoorlichting, de agenten van de bosbouwdienst en de bestuurders van de waterkrachtdam.

Een geschikte onderzoeksstrategie werd opgezet tijdens een periode van introductie in het onderzoeksgebied. Besloten werd om eerst een hele reeks diepte-interviews af te nemen aan de hand van een opgestelde thematische lijst. Tegelijkertijd dienden deze interviews geanalyseerd worden en verwerkt met aangepaste software. Op het moment dat de geconstrueerde kennisdatabank vrij volledig geacht werd, moest een vragenlijst opgesteld worden om meer informatie te verzamelen over enkele controversiële punten en om het kennissysteem in het algemeen te verifiëren en te testen naar representativiteit onder de boerenbevolking.

Om de diepte-interviews uit te voeren was het belangrijk boeren uit alle delen van de gemeenschap te betrekken. Boeren werden geselecteerd uit 2 kleinere waterbekkens, waarvan het landschap pragmatisch werd opgedeeld in 3 eenheden: de stroomafwaarts gelegen zone, het stroomopwaarts gesitueerde deel, en het gebied er tussenin. Eerder onderzoek bevatte aanwijzingen dat boeren van verschillende etnische groepen een andere ecologische kennis demonstreerden, dus werd ervoor gezorgd evenveel boeren van elk van de drie dominante etnische groepen te betrekken (javaans, sundaans en semendonees). Andere discriminerende factoren waren veldtype, ouderdom en type van de koffietuin, landeigendomssituatie en ervaring van de boer.

De literatuur suggereert verschillende methodes om informanten te selecteren. Samen met de leiders van lokale boerengroepen en met de lokale ICRAF-staf werden geschikte interviewkandidaten geselecteerd. Besloten werd om met diepte-interviews door te gaan tot verdere interviews weinig nieuwe informatie bijbrachten. Dit verzadigingspunt werd bereikt na ongeveer 26 interviews.

De semi-gestructureerde diepte-interviews laten de geïnterviewde vrij zijn kennis in zijn eigen woorden uit te drukken en uit te wijden over aspecten die hijzelf belangrijk acht. Als leidraad werd een lijst gebruikt met alle onderwerpen relevant voor het onderzoek. Tijdens de interviews was het vooral de opdracht om de barrière tussen boer en onderzoeker zo veel mogelijk te reduceren. De onderzoeker tracht zich op te stellen als leergierige student, hierbij proberend eigen invloed en elk oordeel tijdens het interview uit te schakelen. De interviews vonden plaats in het Indonesisch, duurden gemiddeld een halfuur tot een uur, en werden integraal opgenomen met een bandopnemer.

Direct na de interviews werd dit opgenomen materiaal geanalyseerd. Alle nuttige informatie werd geëxtraheerd in de vorm van ondubbelzinnige beweringen in de oorspronkelijke bewoordingen van de boer. Deze beweringen werden nadien ingebouwd in een systematische reconstructie van het kennissysteem. Na 17 interviews werd deze geconstrueerde kennisdatabank geëvalueerd, werden hiaten in de verzamelde kennis geïdentificeerd en werd de interviewlijst bijgewerkt. 9 interviews later werd het verzadigingspunt benaderd en konden we overgaan naar de volgende fase van het onderzoek.

Het is niet eenvoudig een doeltreffende methodologie te vinden om lokale ecologische kennis in al zijn aspecten te vatten en weer te geven. Computers kunnen een belangrijk hulpmiddel zijn bij het systematisch formaliseren van descriptieve kennis in kennisdatabanken. Voor dit onderzoek werd beroep gedaan op de computersoftware AKT5 (Agroforestry Knowledge Toolkit for Windows, speciaal ontworpen door de universiteit van Wales in Bangor (UK) voor het opslaan en weergeven van complexe lokale kennissystemen). Hierbij dienen beweringen in een natuurlijke taal omgezet te worden in formele ‘statements’ die gebruik maken van een gecodeerde taal. De software laat diepgaande exploratie van het lokale kennissysteem en representatie van de verschillende verbanden tussen kennissubeenheden toe en vergemakkelijkt dus de evaluatie van kennissystemen en de identificatie van verschillen en hiaten.

Hoewel er onenigheid bestaat over hoe representatief bepaalde beweringen moeten zijn om beschouwd te kunnen worden als deel van het lokale kennissysteem, toch is het nuttig om de geconstrueerde kennisdatabank uitgebreid te testen. Op die manier kan nagegaan worden hoe wijdverbreid de verzamelde kennis is onder de leden van de lokale gemeenschap, en details die nog niet opgemerkt waren kunnen aan de kennisdatabank toegevoegd worden. Zo werd een vragenlijst opgesteld met 68 beweringen waarvan de helft omgekeerd werd om de boeren niet in een bepaalde richting te sturen. Er werd besloten om een bewering als deel van de algemene lokale kennis te beschouwen van zodra minstens 75% van de ondervraagde boeren ermee akkoord ging. Zo werden nog eens 28 boeren ondervraagd, opnieuw geselecteerd volgens waterbekken (ditmaal drie), positie, etnische groep en veldtype.

Tenslotte werden nog enkele andere betrokkenen geïnterviewd, in het bijzonder dorpsbewoners die niet leven van de landbouw, de landbouwvoorlichting, de agenten van de bosbouwdienst en de bestuurders van de waterkrachtdam.

De kennisdatabank en de AKT5 software laten toe de kennis in zijn geheel te beschouwen en verbanden te leggen die nooit ten volle uit een gesprek naar voren kunnen komen. De sterk gesystematiseerde benadering slaagt erin de complexe verbondenheid van het gehele kennissysteem voor te stellen en verschaft op die manier de ideale basis ter beschrijving en interpretatie van lokale ecologische kennis.

De kennis wordt besproken aan de hand van de formele statements die zich in de databank bevinden, en dit een logische volgorde. Er wordt begonnen met het woud en zijn functies, en de gevolgen van ontbossing. Vervolgens wordt dieper ingegaan op twee belangrijke consequenties: overstromingen en waterkwaliteit. Bodemerosie, een essentieel gelokaliseerd fenomeen, is cruciaal in de ecologisch processen die aan de basis liggen van bekkendegradatie. Er wordt ook aandacht besteed aan de impact van landbouw- en conserveringstechnieken en invloed van verschillende landschapselementen en hun positie op de uiteindelijke degradatie van de functies van het waterbekken. En tenslotte worden de lokale bodemkennis (etnopedologie) en het verband met criteria voor een lokale landgeschiktheidsevaluatie besproken.

Ontbossing wordt gezien als de fundamentele oorzaak van een hele reeks landschapsecologische problemen. De voornaamste hiervan zijn bodemerosie, met zijn negatieve gevolgen voor bodemvruchtbaarheid en helderheid van rivierwater, en verlies van constante waterafvoer door de rivier, met overstromingen in het regenseizoen en watertekort in droge periodes. Boeren weten zeer goed dat donkere bosbodems uiterst vruchtbaar zijn en veel organisch materiaal bevatten, en zijn zich ervan bewust dat deze vruchtbaarheid achteruitgaat door langdurige bewerking en vooral door bodemerosie. Een woud wordt gedefinieerd als dichte vegetatie van grote bomen, via dewelke het bos zijn regulerende functie kan uitoefenen. De dichte begroeiing verhindert rechtstreekse impact van regendruppels op de bodem, en de overvloedige bodembedekking zorgt voor toegenomen waterretentie in de bodem. Boomwortels en bladafval verhinderen erosie door bevordering van infiltratie van regenwater, waardoor afspoeling (runoff) vermindert. De bodem wordt beter vastgehouden en kan niet zomaar met het runoff-water meegespoeld worden. Bijgevolg blijft het rivierwater helder, zelfs in regenachtige periodes. Aangezien runoff sterk verminderd wordt, verhindert het woud dat rivieren overstromen en dat aardverschuivingen plaatsvinden. Dus, de boeren beschouwen het bos als een spons die het water bijhoudt en bijdraagt tot een constant waterdebiet in de rivieren. Bossen beschermen ook de natuurlijke waterbronnen. Deze boerenkennis is sterk vergelijkbaar met wetenschappelijke kennis. Ook de andere betrokkenen zijn zich sterk bewust van deze bosfuncties.

Overstromingen worden dus veroorzaakt door de massale ontbossing. Minder bos betekent volgens de boeren ook dat de overstromingen sneller na het begin van de regenbui plaatsvinden en dat ze minder lang duren. Tijdens overstromingen is het rivierwater altijd zeer troebel, zowel door bodemerosie door regenwater en runoff als door de krachtige, destructieve rivierstroming zelf. De rivier sleurt alles mee en erodeert de oevers en de rivierbodem. Aardverschuivingen vinden plaats en kunnen zelfs de rivierbedding van plaats doen veranderen. Het effect van overstromingen op geïrrigeerde rijstvelden hangt sterk af van de kracht van de waterstroming, de overstromingsduur en het ontwikkelingsstadium van de rijst. Krachtige overstromingen hebben zeer negatieve effecten, sleuren de rijstplanten mee, eroderen de bodem en kunnen gehele velden vernietigen. Tragere stroming veroorzaakt sedimentatie van bodempartikels en afval en verhoogt op die manier de vruchtbaarheid van de velden, hoewel excessieve sedimentatie de rijstplanten kan bedekken en doden. Te lange overstromingen doen de rijst rotten. Overstromingen brengen dus naast een aantal voordelen een hele reeks risico’s met zich mee. Deze lokale kennis is consistent met wetenschappelijke bevindingen. Ook de andere betrokkenen zijn zich bewust van de destructieve effecten van overstromingen.

Troebelheid is het voornaamste probleem voor waterkwaliteit volgens de boeren. Deze troebelheid resulteert uit erosieprocessen en wordt verminderd door sedimentatie van bodempartikels en filtering van runoff-water. Ontbossing, landgebruik en vegetatietype op de rivieroevers hebben een uitgesproken effect op de helderheid van rivierwater. Meer stroomopwaarts blijft het water minder lang troebel na een regenbui en volgens de boeren is bronwater altijd helder. Troebelheid vermindert sterk de geschiktheid van het water voor consumptie. Sommige mensen filteren het of laten de bodemdeeltjes bezinken, anderen gebruiken put- of bronwater. Naast troebelheid zijn ook vervuiling door menselijk afval – vooral stroomafwaarts van de dorpen – en pollutie door pesticidenresidu’s toenemende problemen voor het gebruik van rivierwater voor huishoudelijke doeleinden.

Bodemerosie wordt veroorzaakt door regenval op hellend land. Volgens de boeren hebben landbedekking en vegetatie een sterke invloed op erosie door hun effect op verscheidene processen zoals het breken van de val van regendruppels, waterretentie, infiltratie, runoff, verankering van de ‘topsoil’, enz. Boeren beweren ook dat een verminderde bodembedekking de bodem sneller doet uitdrogen en barsten waardoor de erosiegevoeligheid toeneemt. De boeren hebben ook een besef van verschillen in erosiegevoeligheid volgens bodemtype. Bodemerosie leidt tot een verlies aan bodemvruchtbaarheid omdat vooral de bovenste bodemlagen rijk aan organisch materiaal weggespoeld worden. Erosie kan steile velden snel doen degraderen. De boeren beweren ook dat aan de voet van de helling het weggespoelde bodemmateriaal weer afgezet wordt, waardoor hier de bodemvruchtbaarheid verhoogt. Hetzelfde vindt plaats op grotere schaal: in stroomopwaartse gebieden vindt erosie plaats, terwijl stroomafwaarts vooral sedimentatie optreedt. De boeren getuigen ook over het optreden van grote aardverschuivingen. Weerom blijken er weinig indiscrepanties te bestaan tussen boerenkennis enerzijds en wetenschappelijke inzichten en percepties van andere lokale betrokkenen anderzijds.

De verschillende landbouwkundige technieken kunnen een sterke invloed hebben op bodem- en waterconservering. Boeren weten dat het vooral zaak is de bodem niet onbedekt te laten, zeker als de koffie nog jong is. De aanleg van terrassen, geulen en infiltratieputten, het gebruik van schaduwbomen, en de manier van onkruid wieden bepalen sterk in welke mate runoff en bodemerosie optreden.

Terrassen bestaan in vele vormen en worden beschouwd als een van de meest doeltreffende maatregelen tegen erosie, hoewel ze veel arbeid vereisen voor aanleg en onderhoud. Het regenwater kan niet meer direct naar beneden stromen, wordt gestopt op de terrassen en kan infiltreren, waardoor de vruchtbare bodem en het regenwater beter ter plaatse gehouden worden. Dit heeft een sterke invloed op de uiteindelijke opbrengsten. Geulen hebben een gelijkaardige functie; ze vangen het runoff-water op, vertragen het en voeren het af. In de geulen vindt bijgevolg sedimentatie plaats en dit vruchtbare sediment kan gebruikt worden om de velden opnieuw te bemesten. Infiltratieputten (of bemestingsgaten) spelen ook een rol in het verhinderen van directe afspoeling door runoff-water, en hebben en bijkomende functie in de verbetering van de bodemvruchtbaarheid, waardoor ze ook toegepast worden op vlakke velden. Plantafval en organisch materiaal worden verzameld in de putten om afbraak en humusvorming te bevorderen.

Schaduwbomen zijn multifunctioneel in de koffietuinen, wat ook al blijkt uit de vele verschillende begrippen die boeren gebruiken om schaduwbomen te duiden. Schaduwbomen doen de koffietuinen meer op het bos lijken, en door de aanwezigheid van vele wortels en bladafval verhinderen ze op dezelfde wijze runoff en erosie. Boeren weten dat schaduwbomen ook bijdragen tot bodemvruchtbaarheid door hun bladafval en door het feit dat ze nutriënten uit diepere bodemlagen weten te recycleren. Bijproducten zoals hout en vruchten worden sterk gewaardeerd. Daarnaast hebben schaduwbomen ook een sterke invloed op koffieproductie op zich. Schaduwbomen beschermen de koffie en de bodem tegen te veel zon zodat de temperaturen niet te hoog worden en de bodem niet uitdroogt. Koffie heeft wel wat zon nodig om te bloeien, dus de schaduwbomen moeten regelmatig gesnoeid worden. Boeren beschouwen schaduwbomen dus als essentiële componenten van gezonde koffietuinen en hebben een uitgebreide kennis over de eigenschappen van verschillende soorten.

Onkruid heeft een dubbel effect: enerzijds houdt het de bodem bedekt en vermindert het zo runoff en erosie, maar anderzijds heeft het een negatieve invloed op de koffieproductie door sterke competitie voor nutriënten en water. Wieden maakt de bodem losser en verlucht de bodem, wat positief is voor de koffie. Dus boeren oordelen dat het zeker nodig is om te wieden, maar best niet te vaak omdat dat te veel erosie zou veroorzaken. Onkruid wordt ook gebruikt in een aantal bodemconserveringstechnieken zoals ringwieden en terrassen met stroken onkruid. Wieden kan gebeuren volgens verschillende technieken: afsnijden met een sikkel, verwijderen met een hak, sproeien van herbiciden, al dan niet gecombineerd met het vorken van de bodem. Momenteel zijn er ook experimenten aan de gang om een bodembedekker aan te planten als erosiemaatregel.

De lokale kennis van de effecten van deze landbouwkundige technieken komt sterk overeen met wat beschreven staat in de wetenschappelijke literatuur. Andere dorpelingen en de landbouwkundige extensie vertellen grosso modo hetzelfde verhaal, terwijl de bosbouwdienst en de bestuurders van de dam een veel minder gedetailleerde kennis tentoonspreiden.

De boeren weten heel goed dat herbebossing de functies van het waterbekken zou herstellen. Meer bomen, maar ook andere vegetatietypes zoals struiken en hoge grassen kunnen verdere degradatie verhinderen. Voor de koffietuinen is het vooral belangrijk de juiste landbouwtechnieken toe te passen. Schaduwbomen en terrassen zorgen ervoor dat de koffietuinen de bosfuncties ten dele overnemen door hun positief effect op bodem- en waterconservering.

De vegetatie op de oevers van de rivieren is volgens de boeren ook een heel belangrijk landschapselement dat de degradatie van het waterbekken sterk beïnvloedt. Bos, bomen, struikgewas, bamboe of hoge grassen op de oevers zorgen ervoor dat het runoff-water tegengehouden wordt, gefilterd wordt of infiltreert voor het de rivier bereikt, zodat het rivierwater veel minder troebel is. De vegetatie houdt ook de oevers beter vast zodat minder aardverschuivingen en oevererosie plaatsvinden, vooral tijdens overstromingen.

De geïrrigeerde rijstvelden langs de rivieren hebben ook een sterke invloed op de helderheid van rivierwater. Bij irrigatie met troebel rivierwater vindt sedimentatie plaats, waardoor de troebelheid van het water vermindert. Het is wel belangrijk dat de irrigatie goed gereguleerd is, anders kan de soms krachtige waterstroom erosie en vernieling veroorzaken. Hetzelfde geldt voor runoff-water komende van hoger gelegen koffietuinen. Sedimentatie kan plaatsvinden, maar het water kan ook destructieve effecten hebben, dus ook hier is regulatie nodig. Dit gebeurt meestal door een geul tussen de koffietuinen en het rijstveld, die het teveel aan runoff-water naar de rivier afvoert. De boeren beweren dus dat de rijstvelden inderdaad een filterfunctie kunnen vervullen. Anderzijds, bewerking van de rijstvelden met de hak en het planten van de rijst tijdens regenperiodes of irrigatie brengen veel bodemmateriaal in de rivier. De wetenschappelijke kennis over de rol van landschapselementen is gelijkaardig, waarbij de boerenkennis over rijstvelden als filterelement zeker een nuttige aanvulling is voor wetenschappelijke hypotheses. De andere betrokkenen zijn allemaal overtuigd van het nut van oevervegetatie, maar behalve voor de dorpelingen is de filterfunctie van rijstvelden een stuk minder gekend.

Kennis van bodemtypes en hun eigenschappen is essentieel in de bepaling van de geschiktheid van land voor landbouwkundige doeleinden. Daarom werd gepeild naar de perceptie van de boeren betreffende bodems en hun eigenschappen. Etnopedologie is echter geen eenvoudige aangelegenheid. Het opstellen van lokale taxonomieën behelst het gebruik van een hele reeks antropologische procedures en etnografische technieken. Daarbij is het vaak onmogelijk het resultaat eenduidig te vergelijken met wetenschappelijke taxonomieën omdat beiden zich baseren op conceptueel verschillende fundamenten. Hoewel het in het gegeven tijdsbestek onmogelijk was om een uitgebreide studie naar lokale bodemkennis te doen, konden toch enkele algemene tendensen opgetekend worden. Net als op andere plaatsen in de wereld, blijkt de lokale bodemindeling vooral gebaseerd op kleur en textuur. Boeren in Sumberjaya gebruiken ook vaak de begrippen ‘koud’ en ‘warm’ om bodemgeschiktheid te beschrijven.

Het is duidelijk dat de lokale kennis van de boeren in Sumberjaya het resultaat is van een heleboel invloeden. Over de jaren heen is de traditionele kennis van de verschillende etnische groeperingen aangevuld geweest met andere inzichten door onderling contact, door landbouwvoorlichting, door het vormen van boerengroepen, door het sterk betwiste optreden van de regering en in het algemeen door toegenomen contacten met de westerse wetenschappelijke buitenwereld.

Uit de testresultaten blijkt dat meer dan 90% van de beweringen gedragen wordt door minstens 75% van de boerenbevolking. De 6 niet-gevalideerde beweringen waren deels al controversieel in de databank en waren deels slecht geformuleerd in de vragenlijst. Dus alle boeren blijken een gedetailleerde en diepgaande kennis te bezitten over de ecologische processen die zich in hun omgeving afspelen. Statistische analyse van de data onthulde geen verschillen tussen de etnische groeperingen zoals gesuggereerd werd door vroeger onderzoek, noch tussen de verschillende kleinere rivierbekkens en de types van velden (boeren met en boeren zonder rijstvelden). Alleen bleek dat meer stroomopwaarts wonende boeren zich net iets meer bewust zijn van al deze ecologische processen dan hun collega’s meer stroomafwaarts.

Ondanks het feit dat boeren zeer goed weten hoe belangrijk het is verdere degradatie van hun natuurlijke leefomgeving tegen te gaan, en zelfs tot in detail kunnen uitleggen hoe dit moet gebeuren, toch blijkt dat weinig boeren hun kennis in de praktijk toepassen. Vele koffietuinen bevatten geen schaduwbomen noch terrassen en de rivieroevers zijn vaak onbegroeid. De boeren zelf zeggen dat de meeste van hen te lui zijn en dat het veel werk en geld vereist om conserverende maatregelen in de praktijk te brengen. De onzekere landeigendomssituatie verhindert de boeren om op lange termijn te denken. Daarnaast speelt de socio-economische situatie een grote rol. Sinds de economische crisis in Indonesië eind jaren 1990 en door de huidige lage koffieprijzen, worden de boeren gedwongen in de eerste plaats te denken aan overleven op de korte termijn. Ongunstige socio-economische en politieke omstandigheden worden overal ter wereld gezien als de grootste hinderpaal voor duurzaam landgebruik. Naast lokale ecologische kennis beïnvloeden vele andere factoren het besluitvormingsproces van boeren betreffende landgebruik en conserveringstechnieken. De leiders van de boerengroepen benadrukten ook het belang van collectieve actie. Voor hen was het duidelijk dat ze alleen samen weerstand konden bieden aan de toenemende degradatieproblemen.

De verzamelde lokale ecologische kennis komt zeer goed overeen met wetenschappelijke kennis, hoewel de gebruikte terminologie een verschilpunt blijft. Van de andere betrokkenen staan de dorpelingen duidelijk heel dicht bij de boeren. Ook het hoofd van de landbouwvoorlichting heeft een vrij gedetailleerde kennis die sterk overeenkomt met die van de boeren. Het regionaal hoofd van de bosbouwdienst en de directeur van de waterkrachtdam zijn duidelijk veel minder op de hoogte van landbouwkwesties, maar weten toch veel over de functies van het bos.

- De boeren van Sumberjaya bezitten een uitgebreide kennis van bodem- en waterfuncties en hoe deze beschermd kunnen worden, sterk vergelijkbaar met wetenschappelijke inzichten;

- ondanks deze verregaande kennis missen de boeren de motivatie om hun kennis in de praktijk te implementeren;

- de kennis van andere betrokkenen contrasteert niet sterk met deze van de boeren, ze is alleen op sommige punten minder gedetailleerd.

Annexes

All annexes can be consulted on the website http://wim.schalenbourg.be/ and on the cd-rom that is included with this written dissertation.

Annex XVIII: Statistical analysis and interpretation of heterogeneity of test results

[1] See Annex I: More detailed review of the history of the Sumberjaya watershed

[2] See Annex XIX for a topographic map of Sumberjaya.

[3] A coffee cultivation typology is presented in Annex III

[4] More about current farming in Sumberjaya can be found in Annex II.

[5] For example, social taboos are one of the systems of local resource management and biological conservation. These ‘resource and habitat taboos’ are considered as important informal institutions that guide human conduct toward the natural environment (Colding and Folke, 2001).

[6] See Annex IV for all the topic lists used as basis for interviews.

[7] In Annex V, a list of all interviewed farmers can be found.

[8] See Annex IV for all topic lists used as basis for interviews.

[9] The questionnaire with test statements and results can be consulted in Annex VI.

[10] In Annex V, a list of all interviewed farmers can be found.

[11] See Annex IV for all the topic lists used as basis for interviews.

[12] See diagram ‘forest clearing and fertility’ in Annex VIII ‘Diagrams of farmer knowledge’

[13] See Annex I: More detailed review of the history of the Sumberjaya watershed

[14] See diagram ‘tree presence effects’ in Annex VIII ‘Diagrams of farmer knowledge’

[15] See diagram ‘forest cover effects’ in Annex VIII ‘Diagrams of farmer knowledge’

[16] A note about the lack of local awareness of biodiversity issues is included in Annex XX.

[17] See diagram ‘flooding causes’ in Annex VIII ‘Diagrams of farmer knowledge’

[18] See diagram ‘flooding of river effects’ in Annex VIII ‘Diagrams of farmer knowledge’

[19] See diagram ‘paddy flooding effects’ in Annex VIII ‘Diagrams of farmer knowledge’

[20] Soil erodibility is discussed more thoroughly in Annex IX: Local soil types and properties in Sumberjaya.

[21] See diagram ‘river water turbidity causes’ in Annex VIII ‘Diagrams of farmer knowledge’

[22] See diagram ‘water quality’ in Annex VIII ‘Diagrams of farmer knowledge’

[23] See diagram ‘erosion causes’ in Annex VIII ‘Diagrams of farmer knowledge’

[24] See diagram ‘erosion effects’ in Annex VIII ‘Diagrams of farmer knowledge’

[25] See diagram ‘landslide’ in Annex VIII ‘Diagrams of farmer knowledge’

[26] See diagram ‘actions that prevent erosion’ in Annex VIII ‘Diagrams of farmer knowledge’

[27] More about farmer knowledge and cover crops can be found in Annex XI.

[28] See diagram ‘terracing effects’ in Annex VIII ‘Diagrams of farmer knowledge’

[29] See diagram ‘furrow in coffee garden effects’ in Annex VIII ‘Diagrams of farmer knowledge’

[30] See diagram ‘lubang effects’ in Annex VIII ‘Diagrams of farmer knowledge’

[31] See diagram ‘tree presence effects’ in Annex VIII ‘Diagrams of farmer knowledge’

[32] See diagram ‘shade tree planting effects’ in Annex VIII ‘Diagrams of farmer knowledge’

[33] See diagram ‘shade tree sunshine effects’ in Annex VIII ‘Diagrams of farmer knowledge’

[34] See diagram ‘shade tree presence effects’ in Annex VIII ‘Diagrams of farmer knowledge’

[35] More about farmer knowledge and cover crops can be found in Annex XI.

[36] See diagram ‘weed effects’ in Annex VIII ‘Diagrams of farmer knowledge’

[37] See diagram ‘weeding effects’ in Annex VIII ‘Diagrams of farmer knowledge’

[38] See diagram ‘riverside forest and tree effects’ in Annex VIII ‘Diagrams of farmer knowledge’

[39] See diagram ‘riverside belukar effects’ in Annex VIII ‘Diagrams of farmer knowledge’

[40] Annex X contains more information on the effect of furrows within the paddy fields.

[41] See diagram ‘paddy water regulation effects’ in Annex VIII ‘Diagrams of farmer knowledge’

[42] See diagram ‘paddy water turbidity causes’ in Annex VIII ‘Diagrams of farmer knowledge’

[43] See chapter 3 for more information on the testing methodology.

[44] In Annex VI, all test statements and results can be found.

Farmers’ term for shade tree	English translation
Kayu	Wood
(Pohon) pelindung	Protection (tree)
(Pohon) bayang	Shade (tree)
(Pohon) naungan	Shelter (tree)
(Pohon) penghijauan	Green (tree)
Pohon tanam tumbuh	Tree planted to grow