1 Elsevier Editorial System(tm) for Agricultural Systems Manuscript Draft Manuscript Number: Title: Local knowledge about how ecosystem services and biodiversity conservation are related to productivity in coffee agroforestry Article Type: Research Paper Keywords: farmers' knowledge; shade-grown coffee; Costa Rica; Central America; AKT software Corresponding Author: Mr. Carlos Roberto Cerdán, Ph. D. candidate Corresponding Author's Institution: CATIE First Author: Carlos Roberto Cerdán, Ph. D. candidate Order of Authors: Carlos Roberto Cerdán, Ph. D. candidate; Maria C Rebolledo; Gabriela Soto; Bruno Rapidel; Fergus L Sinclair Abstract: The potential for tree components of coffee agroforestry systems to provide ecosystem services is increasingly recognized in certification schemes and direct payments for environmental services. Management practices are a key factor in the amount and quality of ecosystem services provided. There is relatively abundant information about ecosystem services provision within coffee agroforestry systems, but comparatively scant information about how farmers manage their plantations, the factors influencing their farming practices and the extent to which farmers' local knowledge - as opposed to global scientific understanding - underpins management decisions. This research explores farmers' knowledge about how trees affect coffee productivity and ecosystem services in Costa Rica. Farmers' knowledge about the effects of trees on coffee productivity was compared with that of other knowledge sources: coffee processors, technicians and scientists. Farmers were shown to have detailed knowledge about ecosystem services that their coffee agroforestry systems provide as well as about the interactions between trees and coffee productivity. Farmers classified trees according to the services that they provided, such as water protection, soil formation, or contribution to biodiversity conservation. Comparison of knowledge about coffee productivity from different knowledge sources revealed considerable complementarity and little contradiction. The effects of shade trees on biophysical conditions and their interactions with coffee productivity were well understood by farmers. They recorded and classified shade trees as 'fresh' (suitable for integration with coffee) or 'hot' (unsuitable) based on their leaf texture and size, foliage density, crown shape, and root system attributes. This classification was widely used by farmers but not by coffee technicians. Farmers were also aware of the influence of these ecosystem services on crop productivity. Generally, management decisions were made to maintain coffee productivity rather than ecosystem services. Technical interventions addressing the improvement of coffee plantations should take into account not only the scientific information about agroforestry interactions but also the knowledge possessed by farmers.
2 Suggested Reviewers: Marney Isaac Ph.D. Faculty of Forestry, University of Toronto Dr Isaac has published on agroforestry systems and local knowledge. Her methods and results are complementary with ours. Daniel H Walker Ph.D. CSIRO Dr Walker has published on the methodology that we used. He has not been involved in this research and should therefore have a thorough insight on our work. Nicole Sibelet Ph.D. CIRAD Dr Sibelet is a sociologist and an agronomist. She uses approaches in collection of local knowledge that are very different from the one we have used in this paper. Lorena Soto-Pinto Ph.D. ECOSUR Dr Soto-Pinto works in coffee agroforestry systems in Mexico and is well acquainted with local knowledge about the provision of ecosystem services and trade-offs between production and conservation.
3 Cover Letter The Editorial Office of Agricultural Systems Elsevier Turrialba, Costa Rica, May 11, 2011 Dear Editor, Please find attached a manuscript intended for publication in Agricultural Systems. It presents the results of a study on farmers knowledge about coffee productivity, ecosystem services and biodiversity in Costa Rica, Central America. A systematic approach was chosen to store and analyse the knowledge statements: the interviews were entered into knowledge bases using AKT5 software (Agro-ecological Knowledge Toolkit). Farmers knowledge was compared with scientific literature, as well as knowledge of coffee technicians and processors. The ecosystem services reported by farmers at farm and landscape scale were soil formation, erosion reduction and water conservation. Furthermore, farmers had several local functional classifications for trees regarding both services and coffee production that we could related to trees attributes. The effects of shade trees on the biophysical environment of the coffee fields and their interactions with coffee productivity were explained using farmers knowledge. Shade trees were recorded and classified as fresh (suitable) or hot (unsuitable), this classification was based on tree attributes, and functions. All these attributes are reported. The advantages of using AKT are discussed. The AKT knowledge base will be posted on the Web, so that interested stakeholders will be able to access and use it. Best regards, Carlos R. Cerdán PhD student, CATIE- Bangor University Joint Programme
4 *Highlights Highlights Farmers had local functional classifications for trees according to services provisioning, such as water protection, soil formation, pollination, medicinal use or quality of firewood, and also for coffee shading. Classifications were based on leaf texture and size; foliage density, crown shape, root system attributes, and their overall interaction with coffee plants and soil conditions. In Costa Rica, a middle income country where the knowledge about coffee productivity was expected to be widely shared among stakeholders, we found new knowledge among farmers on the effect of specific tree species on coffee, not detained by other stakeholders, nor reported in the literature. Few of the farmers knowledge new findings are contradictory with the published knowledge, rather they are complementary
5 *Manuscript Click here to view linked References 1 2 Local knowledge about how ecosystem services and biodiversity conservation are related to productivity in coffee agroforestry 3 4 C. R. Cerdán a-b, *, M. C. Rebolledo c, G. Soto a, B. Rapidel a-c and F.L. Sinclair b-d a CATIE, Tropical Agricultural Research and Higher Education Centre, Costa Rica b Bangor University, Bangor, Gwynedd, LL57 2UW, Wales, UK c CIRAD, UMR SYSTEM, Turrialba, Costa Rica d ICRAF, World Agroforestry Centre, Nairobi, Kenya * Author for correspondence ( Abstract The potential for tree components of coffee agroforestry systems to provide ecosystem services is increasingly recognized in certification schemes and direct payments for environmental services. Management practices are a key factor in the amount and quality of ecosystem services provided. There is relatively abundant information about ecosystem services provision within coffee agroforestry systems, but comparatively scant information about how farmers manage their plantations, the factors influencing their farming practices and the extent to which farmers local knowledge as opposed to global scientific understanding underpins management decisions. This research explores farmers knowledge about how trees affect coffee productivity and ecosystem services in Costa Rica. Farmers knowledge about the effects of trees on coffee
6 25 26 productivity was compared with that of other knowledge sources: coffee processors, technicians and scientists Farmers were shown to have detailed knowledge about ecosystem services that their coffee agroforestry systems provide as well as about the interactions between trees and coffee productivity. Farmers classified trees according to the services that they provided, such as water protection, soil formation, or contribution to biodiversity conservation. Comparison of knowledge about coffee productivity from different knowledge sources revealed considerable complementarity and little contradiction The effects of shade trees on biophysical conditions and their interactions with coffee productivity were well understood by farmers. They recorded and classified shade trees as fresh (suitable for integration with coffee) or hot (unsuitable) based on their leaf texture and size, foliage density, crown shape, and root system attributes. This classification was widely used by farmers but not by coffee technicians Farmers were also aware of the influence of these ecosystem services on crop productivity. Generally, management decisions were made to maintain coffee productivity rather than ecosystem services. Technical interventions addressing the improvement of coffee plantations should take into account not only the scientific information about agroforestry interactions but also the knowledge possessed by farmers Keywords: Farmers knowledge; shade-grown coffee; Costa Rica; Central America, AKT software
7 Introduction Agro-ecosystems, under certain management and in specific locations, are providing important goods and services that contribute to human wellbeing, economic development and poverty alleviation across the globe. Efficient and effective management of these agro-ecosystems can sustain the provision of vital ecosystem services such as climate stabilization, drinking water supply, flood regulation, crop pollination, recreation opportunities and amenity and cultural assets (Millennium Ecosystem Assessment, 2005). According to both the Millennium Ecosystem Assessment (2005) and the International Assessment of Agricultural Science and Technology for Development (2008), externalities provided by agro-ecosystems should be taken into account in their assessment. Nowadays, there is a great deal of interest in providing financial benefits to landowners and farmers for land-use practices that supply valuable environmental services to the human population (FAO, 2007) Agroforestry systems are increasingly being viewed as significant providers of ecosystem services, including environmental benefits (Harvey et al., 2006) and economic commodities, as part of a multifunctional working landscape (Perfecto and Vandermeer, 2006). The integration of trees and agricultural crops and/or animals into an agroforestry system has the potential to enhance soil fertility, reduce erosion, improve water quality, enhance biodiversity, increase aesthetics and sequester carbon (Garrett and McGraw, 2000; Garrity, 2004; Williams-Guillén et al., 2008; Nair et al., 2009). It has been well-recognized that the services and benefits provided by agroforestry systems occur over a range of spatial and temporal scales (Izac, 2003).
8 Coffee is a very important crop in Central America, both economically and culturally. It is mainly grown as an agroforestry system. The role of coffee production areas in providing ecosystem services is important not only because of the area covered by coffee plantations but also because coffee farms are frequently overlapping with priority areas for biodiversity conservation (Moguel and Toledo, 1999). Biodiversity conservation (Moguel and Toledo, 1999), carbon sequestration (Albrecht and Kandji, 2003), and erosion control (Beer et al., 1998) are some of the benefits derived from trees within coffee plantations. A number of schemes have incentivised coffee farmers to conserve biodiversity and provide ecosystem services whilst producing coffee, such as local and national programmes for payment of ecosystem services or coffee certification schemes (LeCoq et al., 2011) Ecosystem services and biodiversity conservation in coffee agroforestry systems have frequently been studied in isolation from coffee productivity. Although it is becoming increasingly clear that a diverse and abundant selection of trees contribute to biodiversity conservation, it is also clear that farmers perceive that too many trees within their coffee plots decreases coffee yields. Available scientific literature on the relationships between canopy cover, coffee yields and profits show contradictory results. Some studies found significant increases in yields when shade is removed (Matoso et al., 2004; damatta, 2004), whereas others found no effect of shade over yields (Romero et al., 2002) or maximum yields at intermediate levels of canopy cover (Perfecto et al., 2005). It seems that, under certain conditions, shade trees favour the coffee crop, increasing its productivity (Soto-Pinto et al., 2000) with yields greatest under 35 65% shade cover (Staver et al., 2001; Perfecto et al., 2005). Additional
9 products of coffee agroforestry systems, like timber, can be sold, increasing the farmers profit and favouring the use of shade trees (Beer et al., 1998) The trade-offs between coffee profitability, ecosystem services and biodiversity clearly depend on the specific local conditions, such as the altitude and orientation of slope, climate and soil conditions, coffee prices and local wages. Usually, farmers understand quite well the consequences of their management practices in their particular environment, and how this may affect their livelihoods. Through years of experience and observations, they also probably know how the trees in their fields affect the income or goods that they obtain from their farms Farmers have a leading role as ecosystem managers: the provision of ecosystem services from coffee agroforestry systems will depend on their decisions for farm management. These decisions will, in turn, depend on their knowledge of both the ecosystem services provided by these plantations - in particular, by the trees they contain - and the tradeoffs between shade trees and coffee productivity in their specific context. Surprisingly, little is reported in the literature about the knowledge possessed by coffee farmers; the literature available mainly discusses knowledge of attributes and classes of trees (Albertin and Nair, 2004; Soto-Pinto et al., 2007).The authors found no available information on farmer knowledge of the interactions between trees and ecosystem services and how they are related with coffee productivity The objective of this study was to capture coffee farmers knowledge about the trees present in their farms, the ecosystem services they provide, their value for biodiversity
10 conservation and the relationship they have with coffee productivity in Turrialba, Costa Rica Methodology Local knowledge was studied using a systematic methodology: Agroecological Knowledge Toolkit AKT (Walker and Sinclair 1998, Sinclair and Walker 1998). AKT is both a methodology to research local ecological knowledge and also a softwareprogramme to store and analyse the collected knowledge. The methodology consists in presenting the knowledge so that may be stored in a computer readable form, and the software allows the analysis of this data by performing searches and synthetizing causeand-effect relationships. The formal AKT methodology comprises four steps: scoping, definition, compilation and generalisation Scoping This first step is set up in order to refine the objectives of knowledge acquisition. It is a period of familiarization with the context and orientation of the study purposes. Meetings were held with two members of the main organic coffee cooperative, three technicians of the Costa Rican Coffee Institute, the manager of a large coffee estate and several scientists from CATIE. The purpose was to identify possible informants for the compilation stage of the research as well as to list the factors they believed may modify the knowledge detained by these informants. Such factors could be the farm management system (organic or conventional), the location within the study area, and the size of the farm. The size of the farm was highlighted as an important factor to
11 classify differences in knowledge of Costa Rican coffee farmers due to the link with the socio-economic status of the farmer, which is in turn related to the intensity of coffee management; it is believed that larger farms are managed more intensively than smaller ones Definition This step concerns the boundaries and the terminology of the interview protocol. One organic and one conventional farmer were interviewed in order to adjust the research questions according to farmers understanding. In addition, this provided an insight into how farmers expressed their knowledge. It was particularly important during this step to become familiar with the local names of trees and also with specific terms used by farmers. All interviews were recorded (with the farmers approval) using a digital recording device so that the interviews could be more fluent and allow the farmers to express themselves more freely Knowledge compilation This third step in the AKT methodology is the collection and analysis of local knowledge. This step includes interviewing people, creating a comprehensive knowledge base, and analysing the knowledge Interviews 172
12 Semi-structured interviews were held with informants selected during the scoping stage. Each informant, either farmer, technician or processor, is referred to as a source of knowledge. The sources were included from all situations and were described according to the variables identified as likely to influence knowledge held by people in the scoping stage (Walker and Sinclair 1998); for farmers, these variables were farming management (organic and conventional), coffee area, and indigenous origin. All organic farmers interviewed were identified by the technicians of the two organic cooperatives. Some conventional farmers were selected by the technicians of the Costa Rican Coffee Institute, and others were randomly selected from coffee localities where technicians did not select anybody. Farmers identified by the technicians were sources of more abundant information than randomly selected farmers A total of 64 sources were interviewed: 50 farmers, 8 technicians and 6 processors (Table 1). Most interviews were carried out in the farmers fields, thus farmers were able to support their assertions on examples from the surrounding environment, putting them into context. Most sources, especially farmers, were visited a second time after the formal interview in order to fill gaps or clarify misunderstandings Two sets of interviews were carried out: the first set compiled farmers knowledge about ecosystem services and biodiversity within coffee farms with 50 farmers as sources; the second set concerned the relationships between ecosystem services and biodiversity on the one hand, and coffee productivity and quality on the other hand. Of the 50 original farmers, 16 were interviewed for the second set, as well as 8 coffee technicians and 6 processors, in order to broaden the sources of knowledge. Differences between these three kinds of sources (farmers, technicians and processors) in their
13 understanding of factors affecting coffee productivity were also investigated. For both sets of interviews, a list of important interview topics was prepared in advance and checked during the interview, but interviews were conducted in a conversational manner. This approach was chosen to allow farmers to feel more comfortable and share their knowledge, and focused primarily on the topics that farmers were most knowledgeable about. Interviews were always initiated with a full description of the purpose of the research In order to gather a complete representation of knowledge on coffee yield and quality in the second set of interviews, a conceptual model of coffee phenological phases related to yields components was used. This focused on the factors (climate, practices, pest and diseases) affecting each component, based on scientific literature. This conceptual model was then used as a guide for interviewing farmers, technicians and processors in searching for local, additional or contradictory knowledge on yield components and their development, as influenced by the presence of shade trees Knowledge base creation The Agroecological Knowledge Toolkit (AKT software) was used to record, manage and represent the knowledge acquired through interviews (Walker and Sinclair 1998). AKT can be freely downloaded from the website Formal representation of knowledge in AKT involves its disaggregation into unitary statements (which cannot be further broken down) and translation into a formal grammar (Dixon et al. 2001). Four different kinds of unitary statements can be entered: causal, comparative, linking and descriptor of attributes and values. These unitary
14 statements are based on formal terms, which are defined by the user and represent single words as objects, natural processes, or actions. This semantic has to be maintained throughout the whole knowledge base. This approach captures definitions, contextual information, and the relationships between formal terms and statements and facilitates the organization of formal terms into hierarchies. Hierarchies are sets of formal terms with the same properties and characteristics; for instance, the hierarchy called rough rooted trees grouped all the tree species that farmers considered as trees with rough textured roots. The use of hierarchies is a means of synthesizing knowledge. Instead of repeating statements referring processes or attributes to each member of a hierarchy, the definition of a meaningful hierarchy, i.e. a hierarchy where the attributes and processes actually shared among the hierarchy members are precisely identified and verified, allows the reduction of the statements to only those referring to the hierarchy itself. In the example of rough-rooted trees, the knowledge base contains the statement rough rooted trees cause a decrease in soil fertility instead of a repetition of the same statement for all individual species of rough rooted trees Knowledge can then be diagrammatically represented as nodes and links. Such visual representations of knowledge can improve clarity and understanding and facilitates simultaneous consideration of many related statements from different sources. Continuous evaluation of acquired knowledge with AKT throughout the collection process helped to identify gaps in understanding and to organize repetitive interviewing of sources, if required Two separate knowledge bases were composed; one for each set of interviews mentioned previously. The first, related to ecosystem services and biodiversity, contains
15 formal terms used in 560 statements from 50 sources (table 2). The second, related to coffee productivity and quality, contains 515 statements from 30 sources Knowledge analysis AKT allows the production of diagrams, in order to control the clarity of the knowledge and look for contradictory statements that require further explanations from farmers. An example of topics for diagramming is the set of factors that farmers consider to affect soil erosion. Contradictions are visualized by a double direction arrow in the diagrams, alerting the need to assess the sources of the contradictory statements to elucidate these contradictions or to ask more direct questions on the subject to others farmers. In the same way, the creation of diagrams showing statements on the same topic from different sources enables the visual comparison of the origin of the knowledge These diagrams can be further used to build conceptual models about a topic, showing the factors affecting the topic and the links between them. Each link can be characterized by statements originated from referred sources. The statements on the selected topic, however, are not automatically diagrammatised by AKT in an easily comprehensible way. As all statements related to the selected topic appear on the diagram, some manual arrangements are needed to improve its presentation and remove obvious or less-useful links Knowledge was categorised as shared, unique or contradictory and a literature review searching for the explanations that farmers provided was carried out. Issues that were mentioned by farmers and that have been published or are widely accepted by
16 technicians and scientists were considered as shared knowledge. Issues not found in any publication were considered as unique knowledge, detained only by farmers. Contradictory knowledge included the issues that are both unique and in apparent opposition to scientific knowledge Knowledge generalisation This fourth step is set up in order to test how widespread this knowledge is across the whole community. A representative sample size of coffee farmers was chosen (n=93) and then randomly selected from the 2003 Costa Rican Coffee Census (ICAFE, 2003). These 93 farmers were surveyed on seven topics chosen because of knowledge not reported elsewhere or contradictions between sources, as well as being considered important to carry out future technical interventions (Table 5). Topic selection was done in a workshop with scientists and coffee technicians Results Two knowledge bases were created, one for the farmers knowledge about ecosystem services and biodiversity within coffee farms, the other one comprises knowledge from farmers, processors, and technicians about coffee productivity and quality. The farmers knowledge base consisted of 579 statements supplied by 50 sources about ecosystem services and biodiversity conservation within coffee farms (Table 1). Almost 70% of the statements were explicitly about causal relationships, indicating considerable explanatory content (Table 2). There were 176 objects defined in the farmers knowledge base, arranged in thirty-five taxonomic hierarchies, for which information
17 was held locally about classes of objects (e.g. all soft-leaved trees, all big-leaved trees, all deep-rooted trees) Tree attributes and tree functional classifications Coffee farmers create functional classifications of trees through the combination of tree attributes (Figure 1), such as leaf size, root depth, growth rate, and canopy. Farmers, for example, determine whether a tree is in a good, neutral or bad for soil fertility, taking into account how much biomass is produced by the tree (leaf production), how big its leaves are, if they are soft (fast-degrading) or hard (slow-degrading), how frequently and in which time of year the leaves fall, and how much the root system competes with the coffee for resources Farmers use a fresh/hot classification for trees that involves many different attributes, and overlaps with classifications relating to soil and water. Trees that were classified as fresh were thought to be good for water conservation, whereas hot trees were strongly related to low conservation of water. Table 3 shows all 36 species mentioned by farmers, with the classes that were formed and their values for the different attributes. As an example, poró (Erythrina poeppigiana) is classed as a fresh, easily managed, non-dripping tree, good for soil and water. These classifications took into account attributes as short height with fast growth and high biomass production; it is easy to prune and the crown is open to let in light; it has large and very soft textured leaves; and its roots are soft and numerous. 321
18 Table 3 orders tree species according to their impact on soil and water, putting those with positive impact at the top, such as Erythrina poeppigiana, Musa paradisiaca, Cordia alliodora and Inga spp., and Ocotea floribunda, Trichilia martiana and Syzygium malaccense at the lower. Farmers showed understanding of which trees were useful in terms of improving soil fertility and protecting water resources. However, the reasons for keeping particular trees in coffee plantations were not always for these purposes; a tree would often have to prove itself useful in additional ways for it to be kept there, for example by providing a food source Farmers knowledge about biodiversity within the coffee farms Coffee farmers identified the usefulness of each tree species present in their farm for the conservation of small mammals and bird diversity and the type of resource each tree provides (Figure 2). Coffee farmers had knowledge about the behaviour of the majority of the birds and mammals related to the trees in their farms, such as feeding patterns and habitat preferences for nesting or protection. Some tree species were considered bad for biodiversity conservation; for example Pinus oocarpa and Eucalyptus deglupta were mentioned as trees with the potential to reduce the presence of animals within coffee plantations. On the other hand, Erythrina poeppigiana was the species most mentioned by farmers as being useful for many faunal species. However, the great dominance of E. poeppigiana in the coffee agroforestry systems within the study area probably increased the positive perception that farmers have of this species. Moreover, even when E. poeppigiana was considered beneficial for the resources given to birds and mammals, farmers recognize that if trees are frequently pruned the benefits for biodiversity will be considerable diminished.
19 Farmers knowledge about water balance within a coffee farm The diagramming capabilities of AKT were utilised to build a conceptual model of the effects of tree presence on water in coffee plantations, according to farmers knowledge (Figure 3). The maintenance of an appropriate level of humidity for optimum growth of coffee was an important aspect of shade tree management, and farmers explained that at different times of year, more or less soil content water is needed according to the coffee phenology Tree canopies played an important role in the conservation of water as they are the medium through which sun and rainfall are filtered. Farmers considered that the capture of rainfall by the tree canopy could decrease the amount of rainfall reaching the soil, therefore preventing the input of water into the system. Regarding water resources, farmers showed an understanding of which tree species are useful for protecting water sources and should therefore be kept close to a water source. However, farmers were more careful and tended not to disturb the natural species composition around these areas to prevent a possible decrease of water supply There were some knowledge differences between organic and conventional farmers. For instance, organic farmers frequently mentioned in their discourse the importance of water provision for human consumption, as well as how water could be polluted through the utilisation of chemical inputs. Similarly, the management of soil moisture balance due to the litter and soil organic matter was mentioned by a higher number of organic farmers than conventional ones. There was a general concern among all the
20 farmers about the soil and water conservation and not using of chemical inputs. Both organic and conventional farmers were worried about the residual effect of herbicides affecting soils, but organic farmers were also more worried than conventional farmers about the effects of chemical fertilisers Farmers knowledge about coffee productivity Farmers knowledge about coffee productivity was addressed by a twofold approach. First, the knowledge base created in the first set of interviews was explored to extract the causal relationship between factors affecting coffee productivity and tree presence; in the second step, new insights from the coffee productivity conceptual model were investigated, using the knowledge base created with the second set of interviews In the first step, knowledge statements regarding trees and coffee productivity were arranged according to five factors: Pests and diseases, weeds, soil erosion, soil fertility and pollination. For each factor, statements directly relating to the factor were searched for, and then followed until reaching a statement involving trees (Table 4). Within the initial sequence of statements in Table 4, the first statement was related to the relationship between coffee growing and the incidence of coffee diseases, followed by statements related to coffee growing The sequences were then sorted into three categories; knowledge that is unique to farmers, knowledge shared among farmers and scientific knowledge, and contradictions between farmers and scientific knowledge (Table 4). With regards to soil fertility, much of the knowledge was found to be shared by farmers and scientists, but much of the