UNEP-WCMC Library Non-timber forest products in Uganda Spatial tools supporting sustainable development Olivier Cottray, Lera Miles and Adrian Newton UNEP WCMC DNAS Digitized by the Internet Archive in 2010 with funding from UNEP-WCMC, Cambridge http://www.archive.org/details/nontimberforestp06cott Non-timber forest products In ganda K\ UNE! EP WCMC UNEP World Conservation Monitoring Centre 219 Huntingdon Road Cambridge CB3 ODL, UK Tel: +44 (0) 1223 277314 Fax: +44 (0) 1223 277136 Email: info(@unep-wemc.org Website: www.unep-wcmc.org Director: Jon Hutton THE UNITED NaTIONS ENVIRONMENT PROGRAMME WoRLD CONSERVATION MoniTorING CENTRE [UNEP-WCMC} is the biodiversity assessment and policy implementation arm of the United Nations Environment Programme [UNEP], the world’s foremost intergovernmental environmental organization. The Centre has been in operation for over 25 years, combining scientific research with practical policy advice. UNEP-WCMC provides objective, scientifically rigorous products and services to help decision makers recognize the value of biodiversity and apply this knowledge to all that they do. Its core business is managing data about ecosystems and biodiversity, interpreting and analysing that data to provide assessments and policy analysis, and making the results available to national and international decision makers and businesses. SPONSOR This publication is an output of a research project funded by the United Kingdom Department for International Development (DFID) for the benefit of developing countries. The views expressed are not necessarily those of DFID. ZF0177 Forestry Research Programme. CONTRIBUTORS enneull Olivier Cottray OryxMapping GIS Consultancy 11 Grande Rue, 95470 Fosses, France UNEP-WCMC 219 Huntingdon Road Cambridge CB3 ODL, UK Adrian Newton WITH THE KIND COLLABORATION OF Simon Bolwig and Jordan Chamberlin International Food Policy Research Institute, 2033 K IFPRI Street, NW, Washington, DC 20006-1002, USA < Derek Pomeroy, Herbert Tushabe and Marjorie Nakibuka Makerere University Institute of Environment and Natural Resources Makerere University, PO Box 7062 Kampala, Uganda PHOTOGRAPHS UNEP WCMC Lera Miles UNEP-WCMC 219 Huntingdon Road Cambridge CB3 ODL, UK School of Conservation Sciences, Bournemouth University < Talbot Campus, Poole, Dorset BH12 5BB, UK LTS Mike Harrison NO maxon LTS International Pentlands Science Park, Bush Loan, Penicuik, Nr Edinburgh EH26 OPH, UK Joseph Obua Department of Forest Biology and Ecosystem Management Makerere University, PO Box 7062 Kampala, Uganda 0. Cottray and U. Allis, except, front cover and title page: far left, Heather Angel/Natural Visions; centre, Moctar Sacande; far right, Gordon Miller/IRF © DFID 2006 Citation: Cottray, O., Miles, L., Newton, A. 2006. Non-timber forest products in Uganda. Spatial tools supporting sustainable development. UNEP-WCMC, Cambridge, UK URL: http://www.unep-wemc.org/resources/publications/UNEP_WCMC_bio_series/18.htm A Banson production Printed in the UK by Swaingrove Imaging using digital technology on chlorine-free Forest Stewardship Council certified paper produced from sustainably managed sources with water-based coatings The contents of this report do not necessarily reflect the views or policies of UNEP or contributory organizations. The designations employed and the presentations do not imply the expressions of any opinion whatsoever on the part of UNEP or contributory organizations concerning the legal status of any country, territory, city or area or its authority, or concerning the delimitation of its frontiers or boundaries ere (aice uman wellbeing relies on our ability to exploit our diverse and often fragile natural environment sustainably and into the far distant future. If there is no such thing as environmentally neutral economic growth, there is certainly an increasing number of options for sustainable human and social development. Such new approaches are essential to the achievement of the United Nations Millennium Development Goals. Non-timber forest products (NTFPs} have been a particular focus of development interest in recent years. The hope is that forest-dependent people can gain new income-generating opportunities with minimal environ- mental costs. Fruit, basketry, honey and medicinal plants are just a few examples of economically and socially valuable products that can be produced from a sus- tainably managed natural resource base. To offer a long-term source of income, NTFP production will still require careful planning, manage- ment and monitoring. Ideally, NTFP commercialization should raise the standard of living for the poorest communities whilst protecting vulnerable ecosystems and their biodiversity. Spatial analysis can support these Non-timber forest products in Uganda dual objectives by informing locational decisions, directing external support to areas with the greatest prospect of success. This project demonstrates that powerful spatial analysis tools now allow the combination of relevant social, economic and environmental data into a common analytical framework. The results offer a strong indication of the most appropriate sites for the sustainable development of NTFP harvesting and commercialization. Such ‘expert systems’ can be made accessible to any number of stakeholders, providing a truly participatory and inclusive tool for the sound management of our common natural heritage. Jon Hutton Director UNEP World Conservation Monitoring Centre Non-timber forest products in Uganda Coles Brefacey tres. srrenkeh cot this errteie s.r tetra, 18 EEE: res ene Rc ila te Bel as ev bedened, Aces. is Ban ihe ee 3 (nthoductionvamd Sco pe isee cies ceec Sa cack veece ace castes coeac eh oso Soe 5 Non=timbemforestproductsand sustainableidevelopment:.....ceeestescecmcsees: fess -escreesseeee-ceseresscesec so 7 Spatial tools for sustainable development: the ‘development domain’ CONCEPL .......c.cccecccsssesesescesescseecsecscecseteveceetecseseaceree 9 DevelopmentidomainsifoniNiRPs inWganda — Methodology -..c-cctrscscsssarteeessesscs.ciesvcsesveresc suet eee ee 11 Distributionyofthe: NEP nesourceibasety.«.7-.2.sae-t Mees -c.4 teeny weiner A eee, ee 11 Rrotectedsakeas s-nwyet nent te sn een Se a i 7 Bin RS AAG. RU. oe ne 11 Marketa cce ss) teats c.c.csccrstcore: Severe wert meee ee peat ust Oe ERR Ron shee a oe ea ee ee 15 PAOV EN; Cy Aes cec acs soe ates varas. sicssy sta seucucses Sosesves teens seaee sided puede fo ge eee RO He ied ee 15 Devetopmentidomains:—iResultssandidiscussionMice..:.2c2c200ace-ceeesesreeecs es -sssscisteesee eae eee ee 19 BaSKetinyseesrreccacccvnctra.s0sssresstec elect scstssvaced ont onset gees et cot aeteecs tetaetioee settee eee oe Gum arabic WomelUSiOMs tex ssrscnteene carcgeicss sy tun cee eset ert es os cose sass She mare eek te ene i cc eee 25 Resource: baseramoraGeeSSi Gig it Siac. tee sce-ceecercnse secs: a ee entree ers code anes sy aae ren scone ose eee 25 Income-generatingipotemtialgamalmaGhetsr-.csmet cer cre: etee coerce eer mrt ce ee ese ean erage 25 ANDDETEVICCI OMS eseess ew See ote oes ca cia ea tea tet aue dacs esos Sao VR et at aa nook ois Sete ee Zi, PREG M CE Speers ecw as ty nseey asc az von cessaethaussckes sucess sss assseu snes avasciessseseespeanbe dasta cuore elstee sus Seep 28 Non-timber forest products in Uganda Introduction and scope rural livelihoods is increasingly being recognized in national and international policy. For example, achievement of all the Millennium Development Goals will depend on maintaining the environmental goods and services that are key to human productivity. Approaches to development are therefore required that enable incomes to be derived from natural resources, while supporting the effective conservation of these resources. Non-timber forest products (NTFPs) offer an important example of how such goals may be achieved in practice. Many rural livelihoods are based on the collection and sale of products derived from forest resources, including fruits, nuts, fibre and resins. Trade in NTFPs can act as an incentive for forest conservation by providing a source of income from resources that might otherwise appear to have little financial value. In addition, the environmental impact of harvesting NTFPs is generally much lower than typically results from timber harvesting. As a consequence, many rural development initiatives are now supporting the commercialization of NTFP resources. Rural communities often require external financial and technical support for successful com- mercialization of NTFP resources. Some national governments, aid agencies and non-governmental organizations are providing such support, to assist with the process of rural development and environmental conservation. However, exploiting NTFP resources may not be an appropriate option for sustainable development in all areas, as some rural communities are located far |= importance of natural resources in supporting from potential markets, or do not have access to appropriate forest resources. Tools are therefore needed that could be used to direct external support to those areas with the highest potential for success. This report describes an approach to define where NTFP resources offer an appropriate option for sustain- able development. This is achieved using Uganda as a case study. Uganda's rich natural heritage and its position as one of the most rapidly growing economies in Africa serve to highlight the conflict between national develop- ment efforts and the need for a globally responsible approach to natural resource conservation. The report addresses the following questions: 4 How can NTFP commercialization contribute both to rural poverty eradication and forest conservation in Uganda? 4 How do spatial factors affect these two goals? Based on currently available information, where do specific NTFPs have the highest chances of being successfully developed and commercialized in an economically and environmentally sustainable way? 4 How can this spatial analysis be refined in the future to give a more complete picture? oO Research into NTFP commercialization is still relatively recent and many of the datasets necessary for a thorough analysis of these questions are still lacking. However, sufficient information was available to build a demons- tration model using geographic information systems (GIS), providing a useful preview of the benefits of such an approach. Non-timber forest products in Uganda Figure 1: Forest cover and environmentally vulnerable areas in Uganda re Forest reserve Game reserve = Important bird area al Important plant area eal National park [ee] 0% forest cover a 0-10% forest cover ee 10-40% forest cover aa | 40-60% forest cover ae 60-100% forest cover Water bodies e Major towns Source: Protected area data (2002), provided by MUIENR, 2003 Forest cover data are based on remote-sensing data distributed by Earth System Science Interdisciplinary Center, University of Maryland, USA (Hansen et al. 2003'). Non-timber forest products in Uganda Non-timber forest products and Sustainable development natural resource base and a number of important protected areas hosting extremely high levels of biodiversity (Figure 1). Agricultural productivity is low, and Uganda has one of the highest population growth rates in Africa. In recent decades, political instability, unregulated agricultural expansion and limited ins- titutional capacity have contributed to a depletion of natural resources across the country’. Sustainable solutions to Uganda’s economic development are urgen- tly required if its natural resource base and biodiversity are to be conserved in the future. NTFPs, broadly defined as any forest-derived tradable products other than commercial timber, have been widely regarded as a potential meeting point between conservation and rural development priorities’. Common examples of NTFPs in Uganda include medicinal plants, handicrafts, musical instruments, honey and _ light construction material. Their production is usually less destructive than timber harvesting, and offers good opportunities for improving livelihoods as NTFPs are generally easily accessible to the rural poor and little capital investment is needed for collection, processing and marketing”. Several studies have demonstrated the success of NTFPs in providing this so-called ‘win-win’ solution to development and conservation’. Despite diffi- culties in assessing the total economic value of this sector, the Forestry Department of Uganda estimates that NTFP commercialization contributes UShséé6 billion {approximately US$33 million) per year to national income, worth 17 per cent of the forest sector's contribution to gross domestic product (GDP)’. It is therefore suggested that policies geared towards increasing the economic return of NTFPs will lead to an internalization of forest resource values and an increased incentive for conservation through local resource management. However, this assumption is controversial. Not all NTFPs remain ‘environmentally benign’ when extracted on a large scale, and not all resources remain accessible to poor, landless producers once their value becomes apparent to more powerful stakeholders’. The conser- vation value of NTFP promotion may be just as scale- dependent as any other form of forest exploitation such as timber harvesting or palm oil production. It may simply be that the relatively low demand for, and low investment in, | rt has an exceptionally rich and varied NTFPs explains their reputation as environmentally sound, people-friendly products’. If NTFPs sometimes fail to make a positive contribution to sustainable development then there is a need to analyse the ecological, socio-economic and cultural factors that determine the success of NTFP commercialization’. NTFPs can undoubtedly provide many potential benefits to people and the environment, if managed carefully. Programmes such as SAFIRE (Southern Alliance for Indigenous Resources} and CAMPFIRE (Communal Areas Management Programme for Indigenous Resources Project) operating in southern Africa have shown that conservation through use can be a realistic and successful option’. The purpose of this report is to demonstrate how development zones for non-timber forest products may be delineated through spatial analysis, with an explicit focus on environmentally sustainable income-generating oppor- tunities for the poorest sections of Ugandan society. These people, by definition, lack access to financial capital, land and labour, and have limited geographical mobility. We therefore suggest that the specific NTFPs under consideration should fulfil the following requirements: 1. Be readily accessible to poor rural communities. The resources should be available in forests with communal access rights [therefore usually natural or semi-natural as opposed to plantation forests), and collection points as well as market points should be within manageable distance. While beyond the scope of this project, issues surrounding land tenure and its effect on resource exploitation should be taken into consideration. 2. Harvesting should cause minimal environmental disruption. 3. Production should remain non-exclusive once com- mercial value has been demonstrated, and populations local to the resource base should retain the benefits of commercialization. A discussion of the possible policy options to ensure implementation of requirement 3 is beyond the scope of this project (for further information, see studies such as Marshall et al. 2006"). However, requirements 1 and 2 lend themselves well to spatial analysis. The methods by which this can be achieved constitute the main focus of this report. Non-timber forest products in Uganda Figure 2. Rural development domains in Uganda based on spatial variations in market access, agroclimate and population density Population Market Rainfed agricultural density access potential Low Bimodal-low, Unimodal-low Low Bimodal-med Low Unimodal-low Low Bimodal-med Low Unimodal-med, Unimodal-high Low Unimodal-high Low Bimodal-high Low Bimodal-high Low Unimodal-very low High Bimodal-med, Bimodal-high, Highland High Bimodal-low, med, high, Unimodal-very low, low, med, high, Highland High Unimodal-high High Unimodal-med Low Unimodal-low, Bimodal-low High Unimodal-low, Bimodal-low Regional boundaries No data District boundaries Lakes and waterways Source: Bolwig et al (2002) Non-timber forest products in Uganda Spatial tools for sustainable development: the development domain concept overty is one of the main concerns of Ugandan Prot initiatives today. Development and mainten- ance of human welfare and capital is a fundamental component of the country’s long-term development strategy. The Ugandan Poverty Eradication Action Plan 2000 (PEAP) has set out four main pillars of action’: 4 Pillar 1: Economic growth and transformation 4 Pillar 2: Governance and security 4 Pillar 3: Ability of the poor to raise incomes 4 Pillar 4: Quality of life. Whilst all four goals are linked, it is towards pillar 3 that research into NTFP development might most directly con- tribute. More than 80 per cent of Ugandans are employed in agricultural activities and earn less than 50 per cent of the gross national income’. There is growing recognition both nationally and within the international donor community of the linkages between economic development and the spatial pattern and quality of the natural resource base. Sustainable land use is essential to economic growth and poverty eradication in rural Uganda, where the vast majority of people depend on natural resources for their livelihoods. In light of this, USAID/Uganda asked the International Food Policy Research Institute (IFPRI) to prepare a planning framework for rural land-use develop- ment, entitled Strategic Criteria for Rural Investment in Productivity (SCRIP). It adopts land use as the unifying factor to integrate agricultural growth and rural livelihood needs with responsible environmental management, including biodiversity conservation’. Land-use analysis produces location-specific outcomes from environmental variables (soils, landscape, climate, natural plants and animals), their distribution over space and their inter- action with socio-economic factors (population density, income distribution and infrastructure]. IFPRI's approach developed previous work by the Center for Development Research (ZEF, Germany) which has been instrumental in implementing the concept of development pathways. Development pathways are defined as common patterns of change in farmers’ livelihood strategies, associated with their causal and conditioning factors”. Many natural and socio-economic factors may determine development pathways and these depend heavily on the specific location of a particular study. This view is replacing the blanket’ approach of implementing development policies at a national level irrespective of regional and local heterogeneity. Previous research on agricultural development has indicated that certain natural resource and socio-economic factors are of particular importance. In Africa, the most significant include population density, access to markets and agri- cultural potential''. IFPRI integrated these factors by spatial analysis using GIS to map ‘development domains for the whole area of Uganda (Figure 2). A development domain is a stratified model of spatial parameters used to identify which development pathway has the highest chance of success in a specific area. Development domains can be used to delineate areas within which a given pathway or, in the case of this project, the development of a given NTFP, may most successfully result in poverty reduction. The range of NTFP definitions in the literature’, the high number of potential NTFPs found in Uganda and their wide-ranging resource base mean that a generic NIFP development domain would cover almost all of Uganda, making it of little use as a policy-targeting tool. Our approach therefore adopts a product-specific modelling process, rather than computing a set of generic parameter combinations to fit multiple products. Non-timber forest products in Uganda Figure 3: Vegetation types and areas under cultivation High altitude moorland and heath HBB Moist Acacia savannah Ea High altitude forest BB Moist Combretum savannah HB Medium altitude moist evergreen forest Bray Butyrospermum savannah HB Medium altitude moist semi-deciduous forest "9 ) Palm savannah HE Woodland ME Dry Combretum savannah QB Forest/savannah mosaic GME Dry Acacia savannah MN Tree and shrub steppe MN Grass savannah a Swamp Water bodies WB Grass steppe M9 Bushland ME Moist thickets I) Dry thickets Ps | Post-cultivation communities HN Communities on sites with impeded drainage jo) Swamp forest Areas under cultivation (NBS) Source: Vegetation types derived from Langdale-Brown et al (1964)'°. Areas under cultivation derived from the National Biomass Study 11996)" Data provided by MUIENR 10 Non-timber forest products in Uganda Development domains tor NIFPs in Uganda - Methodology section of spatially distributed parameters within a GIS. In its SCRIP analysis of agricultural develop- ment domains for Uganda, IFPRI used parameters identi- fied as critically important to agricultural development in East Africa” '’. These were agricultural potential, access to markets and population density. Together these represent a mix of absolute and comparative advantages in produc- tion found in different geographic areas. In the context of NTFPs, agricultural potential is replaced by the potential occurrence of species used in NTFP production. Market access is used as a comparative advantage indicator in a similar approach to that of the SCRIP model. However, data regarding the effect of population density on the compara- tive advantage of producing specific NTFPs, rather than alternative products, are not yet available. Two additional factors were included here in estimating appropriate areas for NTFP production. These were poverty distribution and areas of importance for biodiversity. The following sections review these various datasets and their integration to derive development domains for NTFPs. [ Jessen domains are derived from the inter- DISTRIBUTION OF THE NTFP RESOURCE BASE The base layer of an NTFP development domain is the spatial distribution of habitat classes associated with the occurrence of species required for the NTFP’s production. This can be understood as the maximum potential distribution of its resource base: locations where the nec- essary species may be found in the wild, or where ecological factors would favour their cultivation. Products were related to a spatial distribution of their resource base by integrating the results of two contributory studies to the development of IFPRI’s SCRIP framework. As part of SCRIP, an analysis was undertaken of a representative selection of commercially valuable NTFPs along with their associated species, and the environmental impact of harvesting was estimated’. Fifty-five species were selected that displayed high potential for commercial NTFP development as well as minimal environmental impact in exploitation. The term NTFP was defined in a broad sense and includes products derived from trees which may occur in fragmented woodlands as opposed to forests per se. A second report produced as part of SCRIP by the Makerere University Institute of Environment and Natural Resources (MUIENR] aimed to characterize the major re- maining natural and semi-natural ecosystems of Uganda’. It identified those which are most threatened, and where they are located, and considered how people might receive greater benefits from the rernaining natural resources. MUIENR’s analysis is based on the 22-class Langdale-Brown (L-B) vegetation map of Uganda produced in 1964". There is no evidence of any major change in the classification of these ecosystems since the 1960s, besides those identified as land currently under cultivation by the National Biomass Study of 1996 (NBS]'. So the L-B classification can be seen as a map of potential vegetation over much of the country”. In order to highlight only those natural and semi-natural areas that may be accessible to the landless poor, NBS agricultural areas were subtracted from the L-B cover (Figure 3]. MUIENR researchers related the results of their analysis to the 55 sample species by tabulating the likely occurrence of these within each L-B ecosystem class. A broad overview of which natural resources are to be expec- ted where, and whether they are likely to be sufficiently common or extensive to justify any recommendations for further use, is provided. On the basis of these data, it was then possible to generate a database relating NTFPs to L-B classes via their shared species. This database was in turn linked to the GIS model described here, effectively providing an interactive mapping facility for the maximum potential resource base of selected NTFPs (Figure 4). PROTECTED AREAS If NTFPs are to promote sustainable management and con- servation of natural resources, their development should be appropriate to the protection level and conservation value of affected ecosystems. Debate continues about the relative merits of exclusionary conservation practices and the more inclusive and participatory ‘conservation through manage- ment’ approach. We assume here that some areas do require absolute protection, but that in many areas properly managed NTFP harvesting might be appropriate. In its study, MUIENR ranked protected area (PA) sites according to their conservation importance and vulnerability. This value was derived by combining several parameters such as areal extent, occurrence of IUCN-listed Red Data species, and current level of protection and management by either the Forest Authority or the Uganda Wildlife Authority. Non-timber forest products in Uganda Figure 4: Relational database structure for determining the resource base of selected NTFPs in Uganda Basketry Shea butter Langdale-Brown vegetation class Species distribution Basketry Gum arabic Shea butter Bea Potential species distribution le =] Species absent Water bodies 12 Non-timber forest products in Uganda Figure 5: Protected area vulnerability and forest integrity index Protected area vulnerability Forest integrity index (2 High GB High [] Forest cover <40% /) Medium (9 Medium Water bodies [-) Low Low Source: Protected area vulnerability derived from MUIENR [2002] ” Forest integrity derived from Kapos et al (2000)'° 13 roducts in Uganda Non-timber forest Figure 6: Market towns and road network d speed (kms/hr) Estimated ro. by Bolwig et al. (2002)*. Road network and market data (2002), provided by IFPRI rce: Road network classification oe Market towns 0) AS oa ° a i o 2 = he et 14 The highest-ranked PAs are least suitable for extractive use. They are small, already being exploited, or unprotected, despite having high conservation importance”. We used this ranking to exclude unsuitable areas from NTFP development domains. PAs with a high {H) ranking were excluded from consideration. It is assumed that in the remaining PAs appropriate multiple-use zones can be implemented. In some instances these may already be in place. In addition to MUIENR’s ranking of protected areas, we used a backdrop layer of forest integrity. These data were derived from MODIS satellite imagery to give an indication of forest fragmentation in areas selected for NTFP production”? (Figure 5). MARKET ACCESS The Uganda Participatory Poverty Assessment” states that distance to markets is seen as one of the most important causes of rural poverty after poor health and disease: “..people living in rural areas, especially those communities distant from a major town, complained of lack of proximal, frequent markets ... Local people cited long distances, impassable roads, and lack of affordable transport, especially in the rainy season, as barriers to accessing markets. Women complained of time wasted walking long distances, while men mentioned the transport problems associated with reaching general markets in the town. While in other sites, local people said distance and lack of road access restricted the frequency of market attendance, led to goods being damaged in motorized transit, and theft of profits on return from markets...” This concern is echoed in Uganda's Plan for Mod- ernization of Agriculture {PMA)"”, which defines rural infrastructure and market access as two of its seven priority action areas. As an input to NTFP development domain identif- ication, a map was generated to estimate travel times from any point in Uganda to its nearest marketplace. The surface was generated using road and urban settlement data {Figure 6). Roads were differentiated into five classes and an approximate average travel speed allocated to each class as shown in the legend to Figure 6. For the computation, points not associated with roads were assumed to be accessible only on foot and the associated speed of travel was set at an average of 1 kilo- metre per hour to account for difficult terrain and product load. These figures are based on best available estimates and travel experience in Uganda, but should be subjected to more rigorous survey in future work. Additionally, the Poverty Eradication Action Plan (PEAP] recommends the promotion of affordable, alternative means of transport such as ox carts and bicycles for the rural poor. This suggests substantial variation in speeds on a particular Non-timber forest products in Uganda class of road as the means of transport may vary widely from one NTFP trader to the next. The text on page 19 reviews a way of coping with this sort of uncertainty. Additionally, no distinction is made here between products that are processed on site and products sent to processing facilities. We are making the broad assumption that from the point of view of the NTFP resource collector there is no difference between constraints faced when sell- ing to the end-user (market) and those faced when selling to secondary processors. What is more, a number of NTFPs, such as honey and charcoal, are sold at the nearest roadside as opposed to town markets’. For these prod- ucts, a more accurate description of travel times would therefore be based on distance to busy roads. These finer points require data on the processing and marketing methods of specific NTFPs, information that could be acquired in future surveys. The resulting travel-time surface is illustrated in Figure 7. Although the results should be seen as estimates at this stage, preliminary checks against actual travel time between various towns show that they are within the correct order of magnitude. They suggest that the average travel time from any point in the country to the nearest market town is around six and a half hours. This surface can be integrated into the development domain identification procedure to exclude areas too far from possible markets for NTFP production to be econ- omically worthwhile or feasible. The exclusion mask can be varied according to travel-time threshold values appropriate for the product in question. A thorough survey of appropriate threshold values was beyond the scope of this project, and so they were arbitrarily allocated according to the estimated market potential of the products under study. Products sold in local markets were given a five-hour travel time threshold whereas products sold on international markets [via Kampala] were given a higher threshold of ten hours. POVERTY Most Ugandans are self-employed, mainly in agriculture, where over 80 per cent of the population earns less than half the national income. If NTFP development is to benefit the poorest sections of society then attention should be focused in those areas with the highest rates of poverty. At the time of the present analysis, several poverty mapping projects were underway at global and national scales. Further details can be found on http://povertymap.net. Here we use IFPRI's district-level data, where poverty incidence is defined as the share of households that fall under a given expenditure-based poverty line’”. The data are illustrated in Figure 8 and show wide variations in poverty levels throughout the country, with the highest incidences occurring mainly in the north and east. 15 Non-timber forest products in Uganda Figure 7: Travel time to markets within Uganda Travel time to nearest market town HB Within 10 hours @ = Market town MO Within 5 hours ' Water bodies ~ Within 1 hour 5-hour contours 16 timber forest products in Uganda Non Figure 8: Poverty incidence EP SBS _ “ 53 Bes Poverty incidence Water bodies 3) 15-30% MS 30-50% GB 50-70% MM 70-90% Source: Bolwig et al. (2002)°. 23] No data available 17 Non-timber forest products in Uganda Figure 9: Intersection of ten-hour access layer with resource base layer to compute development domains for gum arabic Ten-hour market access layer Resource base for gum arabic ee] Areas within ten hours of nearest market eS] Water bodies 18 Non-timber forest products in Uganda Development domains — Results aid GiSsCussion basketry products, gum arabic and shea butter. Figure 9 illustrates the intersection procedure used to define their development domains. The following sections review the results for each product along with summary information regarding related species, distribution, market potential and environmental impact. This information is derived from Baldascini’s report to IFPRI (2002)°. | = NTFPs were selected for this demonstration: BASKETRY Basketry products are one of the main handicrafts produced in Uganda. The main forest or woodland species used include raffia (Raphia farinifera (Gaertn.) Hylander], sisal (Agave sisalana Perr.) and bamboo (Arundinaria alpina K. Schum.]. Products include baskets (UShs2 500-6 000), mats (UShs10 000-20 000), table mat sets (UShs4 000-20 000), hats (UShs4 000-7 000), chairs (UShs35 000), tables (UShs40 000) and lampshades (UShs5 000-10 000) (there are approximately UShs2 000 to one US dollar). The products are sold locally as well as more widely to tourists. A five-hour travel time threshold value was allocated for development domain computation (Figure 10). GUM ARABIC The main source of gum arabic is three-thorned acacia (Acacia senegal (L.) Willd.) although gum arabic can also be extracted from white-galled acacia (Acacia seyal Del.), woman's tongue tree [Albizia lebbeck [L.) Benth.) and saman tree (Albizia saman F. Muell.). It is obtained by tapping or exudation, a process with minimal environ- mental impact if properly managed. Gum arabic is used in confectionery, soft and alcoholic drinks, pharmaceuticals, and in the printing, ceramics and textile industries. It is also used locally as an adhesive or as an ingredient of traditional medicines. It has an established international market, fetching up to US$5 000 per tonne. A travel-time threshold of ten hours was used for development domain computation. In addition, Acacia is an excellent plant for afforestation of arid tracts and soil reclamation, and so planting gum arabic trees could serve the dual purposes of environmental restoration and income generation. By superimposing development domains for gum arabic on a surface of forest integrity, it is possible to identify areas that might benefit from afforestation while offering income opportunities for the local population (Figure 11). SHEA BUTTER Shea butter is derived from the nuts of the shea tree (Vitellaria paradoxa C.F. Gaertn.), found in the savannah of eastern and northern Uganda. It is used in Europe, Japan and Russia primarily in cosmetics as a basis for soaps, creams, moisturizers, hair conditioners and shampoos, and also as an ingredient in chocolate products. Due to its extensive international market a ten-hour travel-time threshold was used to identify its development domains (Figure 12). DEALING WITH UNCERTAINTY Research into NTFP development and commercialization is still in its early stages. However, it is already apparent that strict rules as to NTFP land suitability are difficult to define because of the large number of potential NTFPs, the many methods of processing and commercialization, and the number of stakeholders involved in, and affected by, forestry decisions. What attributes should be taken into consideration? Whose interests should be represented, and by what parameter values? As different segments of society {smail-scale producers, owners of large farms, national authorities, non-governmental organizations) often differ on what is acceptable, it seems desirable to identify and account for several sets of acceptable conditions, representing the views of these different groups. We suggest that explicitly accounting for this uncertainty offers a realistic, if not a deterministic, approach to spatial analysis. Geographic error and uncertainty should be seen as an integral part of human knowledge and understanding concerning reality. Ideally, information should include well-informed assessments of uncertainty” rather than being presented at face value. A potential method of achieving this goal, based on concepts such as fuzzy logic and Bayesian inference, is briefly introduced here. It is suggested that further work be carried out to evaluate empirically the usefulness of such an approach in identifying NTFP development domains. Fuzzy logic can be applied to the development of environmental indices to resolve many of the problems addressed above, such as incompatible observations and implicit value judgements”. It can bridge the gap between 19 Non-timber forest products in Uganda Figure 10: Development domains for basketry products within five hours of closest market Development domains (EB High forest integrity ea Medium forest integrity Low forest integrity 20 Poverty incidence = 70-90% GM Environmentally vulnerable areas [_] 50-70% {__] 30-50% "— Water bodies 15-30% 0-15% No data Non-timber forest products in Uganda Figure 11: Development domains for gum arabic within ten hours of closest market Development domains a High forest integrity aa Medium forest integrity | Low forest integrity Poverty incidence [J 70-90% HE Environmentally vulnerable areas {__] 50-70% =a 30-50% fea Water bodies 15-30% — -0-15% "No data 21 Non-timber forest products in Uganda Figure 12: Development domains for shea butter within ten hours of closest market Development domains Poverty incidence [EB High forest integrity [_] 70-90% a] Medium forest integrity [aa] 50-70% Low forest integrity fea 30-50% ag Water bodies Ea Environmentally vulnerable areas 15-30% 0-15% No data 22 scientific measurement and the fulfilment of social objec- tives and provide a way to translate a wide variety of infor- mation - objective data, qualitative information, subjective opinions and social needs - into a common language for characterizing environmental effects. In addition, the process of acquiring the necessary information to set up such a system is participatory by nature and contributes to a much more transparent decision-making process. Fuzzy logic is based on the premise that a statement, instead of necessarily being either true or false, may have a degree of truth. Computationally, this translates into setting the level of truth of a value at a number between 0 and 1. Traditional ‘crisp’ logic only allows for values to be either 0 (false) or 1 (true). In the case of NIFP development domain parameters, for example, stating that a certain species occurs within a given L-B vegetation type requires some sort of subjective judgement as to the statements truth. Typically, if there are insufficient empirical data to compute probabilities, this will be resolved by expert opinion with a 0 or 1 type answer and any uncertainty involved in attaining this solution is lost in subsequent stages of analysis. Similarly, the threshold distance for feasible market access may be assessed completely differently by an economist using a financial cost-benefit analysis and by a producer balancing other priorities such as alternative income-generating activities and family obligations. Both opinions have merit and deserve to be reflected appropriately in the final analysis. One way of achieving this is to use the distribution of stakeholder opinion. Databases can readily be developed that provide non- experts with an interface allowing them to input their opinions. When this approach is applied to the various layers of a development domain GIS model, each representing one parameter, the result can be visualized as a density gradient representing the level of truth or belief in the suitability of a location to the development of a given NTFP. There exist a number of fuzzy and Bayesian algorithms for combining these parameters-, the simplest of which is a straightforward multiplicative method as illustrated in Figure 13. Such a tool would help land management authorities to appreciate and take account of the variety of opinions and standpoints that resulted in the planning map. An alternative method of exploring uncertain data is provided by Bayesian belief networks (BBNs]. A BBN is a network of linked nodes, each of which is associated with a probability function. The nodes represent either variables with a defined number of states, or variables with a continuous distribution. The relationships between the nodes are represented by the links. Non-tirnber forest products in Uganda BBNs offer a means of analysing probabilistic data, through the use of Bayes’ theorem. In particular, they provide a tool for inferring the probability of the state of a given variable, given evidence about other variables. In this way, they provide a valuable tool for exploring uncertain data. With respect to development domains, many of the issues discussed above in relation to fuzzy logic could similarly be addressed using BBNs. Different forms of information, such as quantitative data, qualitative information and subjective opinions, can readily be integrated and analysed using BBNs, by ex- pressing them as probabilities associated with different states of categorical variables. One of the main applications of BBNs to date has been in the develop- ment of ‘expert systems’, or decision-support tools incorporating expert knowledge, much of which may be subjective in nature. With respect to NTFP development domain para- meters, the occurrence of a certain species within a given L-B vegetation type could be accorded a probability, with an associated degree of error or uncertainty, based on available quantitative or qualitative data, or even expert opinion. Similarly, contrasting assessments of the threshold distance for feasible market access from different stakeholders could be explicitly analysed, for example by representing the probability distribution of stakeholder opinion. As with fuzzy logic approaches, interfaces can readily be developed that provide non- experts with user-friendly access to the tool. BBNs have been applied to assess the factors influencing success of NTFP commercialization, and the impact of such commercialization on rural livelihoods, in a research project funded by the DFID Forestry Research Programme. The project has developed a decision- support tool to enable those NTFPs with high potential for successful commercialization to be identified. Further details of the project, entitled CEPFOR, are available from the following website: http://www.unep-wemc.org/ forest/NTFP/. One of the key advantages of an approach that focuses on addressing trade-offs is that it seeks long- term solutions that explicitly reflect (rather than minimize or ignore) the diversity of views among the various community, agency and technical participants’. The relatively recent concept of NTFP development domains and the wide variety of stakeholders potentially affected by NTFP issues provide a welcome opportunity to develop decision-support tools that are truly grounded in stakeholder values and maintain transparency in the analysis process. It is hoped that this pilot project might serve to demonstrate this need and generate new research initiatives in this direction. 23 Non-timber forest products in Uganda Figure 13: Comparison of ‘fuzzy’ and ‘crisp’ development domains Accessibility ‘belief level’ Hypothetical probability of species occurrence (Hypothetical proportion of respondents agreeing that location ‘X’ is within reasonable access of markets} Hypothetical proportion ee : Hypothetical probability High: 100 High: 100 Water Water Low: 0 bodies Low: 0 bodies Fuzzy development domains for gum arabic Crisp development domains for gum arabic Belief index High: 100 ‘noc Development domains Water Ris bodies Low: 0 bodies 24 Sonclusrons by which seemingly disparate data can be integrated to show their spatial relationship and reveal new information to support land management decision-making. This report has outlined a theoretical GIS-based method to derive development domains for non-timber forest products. Development domains are intended to guide development efforts by mapping areas of comparative advantage in the commercialization of a given product based on four major parameters: occur- rence of necessary species (the resource base), access to markets, poverty and environmental vulnerability. At present, the model reveals the potential resource base within an acceptable distance of the market for selected NTFPs. It eliminates highly vulnerable protected areas and signposts domains that coincide with high poverty incidence. It is argued that the successful commercial- ization of such NTFPs may help provide income- generation opportunities to the poorest sections of Ugandan society while providing incentives for the long- term management and preservation of environmental resources. Although only three representative products were selected for this demonstration, the method should be applicable to any NTFP under study. However, only a broad and preliminary methodology is presented here. The accuracy of the development do- mains is dependent mainly on the accuracy or reliability of the available data. Meanwhile, their precision is a function of how many aspects of the problem we can represent spatially —- the more aspects we can integrate into the GIS model, the more focused are the resulting development domains. To refine the picture requires several additional datasets that are not available in this phase of the project but are suggested in the following sub-sections as the focus of future lines of research. Additionally, it is argued that an integrated GIS tool could be developed for use by land managers and decision-makers that readily accounts for differing value judgements by various stakeholder groups. Such a tool would allow for participatory and transparent decision-making by explicitly recognizing the uncertainty and subjectivity inherent in land management choices. G eographic information systems (GIS) provide a tool RESOURCE BASE AND ACCESS RIGHTS The model in this project uses data on the potential resource base of NTFP species rather than recorded distributions. The data were derived from a MUIENR analysis of the potential occurrence of species within L-B vegetation classes. This would be a valid basis on which to Non-timber forest products in Uganda construct the development domain model, if one allowed for possible domestication of NTFP species in plantation forests or on agricultural land. However, this highlights the issue of resource access rights — the poorest sections of society are dependent on communally accessed land with very little guarantee of tenure. Long-term investment in tree plantations or forest restoration as a method of NTFP domestication may be seen as too risky in comparison with alternative means of income such as agricultural labour. Hence, as a prerequisite for understanding the potential contribution to poverty alleviation, we need to answer the question: ‘What guarantee does current land tenure legislation provide to landless populations who wish to invest in planting on common-access land?’ The assumption that people will be in a position to take advantage of NTFP-related opportunities touches on issues beyond the scope of GIS models, such as gender relations and the role of local elites and hierarchical social structures. However, combining the answer to the above question with spatial data on land tenure would allow the GIS model to highlight those areas within the currently selected development domains that also provide sufficient security of access for investment in NTFP domestication. A related point concerns how best to safeguard natural resources from unregulated access leading to overexploitation. Excessive harvesting of NTFP primary products may be just as damaging to the environment as other forms oi harvesting such as timber. Sustainable yields have yet to be estimated for many NTFPs, but resource management usually requires regulations or other agree- ments on access. A development domain GIS model might contribute to the planning of these regulations by com- paring the environmental impact of NTFP production with ecosystem vulnerability. The question that follows from this, and that requires further research, is: What environmental impacts result from the production of specific NTFPs?” INCOME-GENERATING POTENTIAL AND MARKETS The available quantitative data on the income-generating potential of specific NTFPs are limited. The theoretical framework for research on this issue is largely in development and is complicated by the wide variety of products and services which could be defined as NTFPs, as well as by the various levels of processing that they require. This makes cost-benefit analyses difficult to apply at this stage. In this project, we based our market access analysis on a set of hypothetical travel-time thresholds as 25 Non-timber forest products in Uganda well as estimated travel time by road category. Actual data would allow us to account for stakeholder opinions and could be readily available through surveys and observation. However, the analysis would also benefit from a more quantitative computation of travel costs versus income potential. This should include factors such as quantity-to-weight and quantity-to-bulk ratios as well as market selling price. Such travel-cost data can be integrated into the GIS model in much the same way as travel time, and could additionally account for the nature of the target market. A product attracting an international market and high prices can be derived from a resource base further afield than products attracting national or local demand. GIS is ideally suited to represent such logical rules and can be used to adapt the development domain of a product to its market potential. This analysis requires information such as NTFP market prices, and the quantity-to-weight and quantity-to-bulk ratios of an NTFP during transport. To conclude, this report briefly reviewed the merit in accounting for subjectivity and uncertainty in spatial analysis. NTFP market analysis is in its early stages and offers an opportunity to incorporate this novel approach. Interactive tools which allow for the input of various stakeholder opinions on all the questions addressed above and retain an indication of the resulting uncertainty throughout the analytical process can now readily be built. Such tools could use only that information which is relevant to a given stakeholder’s decision process or combine various stakeholder opinions; their output can display a development domain for a single NTFP or can show those 26 areas with greatest development potential taking all selected NTFPs into consideration. More importantly, they can bring the functionality of expert systems within the reach and influence of non-experts thereby providing participatory and transparent decision-support tools. Non-timber forest products in Uganda Abbreviations BBN CAMPFIRE DFID FRP FAO GDP GIS IFPRI IUCN L-B MODIS MUIENR NBS NTFP PA PEAP SAFIRE SCRIP UNEP UNEP-WCMC USAID UShs US$ Bayesian belief network Communal Areas Management Programme for Indigenous Resources Project Department for International Development Forestry Research Programme (UK) Food and Agriculture Organization of the United Nations Gross domestic product Geographic information systems International Food Policy Research Institute IUCN-The World Conservation Union Langdale-Brown vegetation classification Moderate resolution imaging spectroradiometer Makerere University Institute of Environment and Natural Resources National Biomass Study Non-timber forest product Protected area Poverty Eradication Action Plan Southern Alliance for Indigenous Resources Strategic Criteria for Rural Investment in Productivity United Nations Environment Programme UNEP World Conservation Monitoring Centre United States Agency for International Development Ugandan shillings United States dollar 27 Non-timber forest products in Uganda References N wo es ou oO a foe} ~o . Hansen, M., DeFries, R., Townshend, J. R. G., Carroll, M., Dimiceli, C. and Sohlberg, R. (2003). MOD44B: Vegetation Continuous Fields. Collection 3, Version 3.0.0. User Guide. Available from: http://modis.umiacs.umd.edu/documents/ MOD44B_User_Guide_v3.0.0.pdf . Bolwig, S., Hazell, P. and Wood, S. (2002). Strategic Assessment of Land Use Options for Uganda - Strategic Criteria for Rural Investment in Productivity (SCRIP) Phase I. International Food Policy Research Institute, Washington DC, USA. Belcher, B. (2003). What isn’t an NTFP? International Forestry Review 5[2): 161-167. Neumann, R. P. and Hirsch, E. (2000). Commercialisation of Non-Timber Forest Products: Review and Analysis of Research. Center for International Forestry Research, Bogor, Indonesia. . Marshall, E., Schreckenberg, K., Newton, A.C. eds) 2006 Commercialization of Non-timber Forest Products: Factors Influencing Success. Lessons Learned from Mexico and Bolivia and Policy Implications for Decision-makers. UNEP World Conservation Monitoring Centre, Cambridge, UK. . Southgate, D., Coles-Ritchie, M. and Salazar-Canelos, P. [1996]. Can Tropical Forests be Saved by Harvesting Non- timber Products? Centre for Social and Economic Research on the Global Environment (CSERGE] Working Paper GEC 96-2, UK. Uganda Ministry of Water, Lands and Environment (2002). The National Forest Plan. Kampala, Uganda. . Baldascini, A. (2002). Income Generating Opportunities Arising from Natural Ecosystems in Uganda. \nternational Food Policy Research Institute, Washington DC, USA. . Uganda Ministry of Finance, Planning and Economic Development (2001). Poverty Eradication Action Plan. Kampala, Uganda. 10. Ruecker, G. R., Park, S. J., Ssali, H. and Pender, J. (2003). Community Resource Mapping for Regional Land Quality Assessment in Uganda. Center for Development Research (ZEF], Bonn, Germany. 11. Pender, J., Place, F. and Ehui, S. (1999). Strategies for Sustainable Agricultural Development in the East 28 1 1 1 2 2 2 2 “y 8. 9 0. N w African Highlands. |International Food Policy Research Institute, Washington DC, USA. . Makerere University Institute of Environmental and Natural Resources (2002). Uganda Ecosystem and Protected Area Characterisation. International Food Policy Research Institute, Washington DC, USA. Langdale-Brown, |., Osmaston, H. A. and Wilson, J. G. (1964). The Vegetation of Uganda and Its Bearing on Land-use. Government of Uganda, Entebbe, Uganda. . National Biomass Study (1996). Uganda: Land Cover Stratification (Vegetation). Uganda Forest Department, Kampala, Uganda. Kapos, V., Lysenko, I. and Lesslie, R. (2000). Assessing Forest Integrity and Naturalness in Relation to Biodiversity. FAO/UNEP World Conservation Monitoring Centre, Cambridge, UK. . Uganda Ministry of Finance, Planning and Economic Development (2000). Uganda Participatory Poverty Assessment Report. Kampala, Uganda. . Uganda Ministry of Agriculture, Animal Industry and Fisheries and Uganda Ministry of Finance, Planning and Economic Development (2000). Plan for Modernisation of Agriculture: Eradicating Poverty in Uganda. Kampala, Uganda. Pomeroy, D. (2003). Pers. comm., 2 October. . Bolwig, S., Wood, S. and Chamberlin, J. (2003). A Spatially- based Planning Framework for Sustainable Rural Livelihoods and Land Uses in Uganda - SCRIP Phase Il. International Food Policy Research Institute, Washington DC, USA. Zhang, J. and Goodchild, M. (2002). Uncertainty in Geographical Information. Research Monographs in Geographic Information Systems (eds: Fisher, P. and Raper, J.). Taylor & Francis, London, UK. pp 1-14. . Silvert, W. (2000). Fuzzy indices of environmental conditions. Ecological Modelling 130: 111-119 . MathWorks (1998). Fuzzy Logic Toolbox. The MathWorks Inc., Natick MA, USA. . Gregory, R. S. (2002). Incorporating value trade-offs into community-based environmental risk decisions. Environmental Values 11(4): 461-488. Non-timber forest products in Uganda Spatial tools supporting sustainable development One path to sustainable development for forest-dependent populations is the commercialization of non-timber forest products (NTFPs). Fruit, baskets, honey and medicinal plants are just a few examples of everyday products that can be harvested from a sustainably managed natural resource base. This report describes a new map-based approach to defining areas best suited for NTFP commercialization. Uganda is used as a case study. As one of the most rapidly growing economies in Africa, its rich natural heritage highlights the conflict between national development efforts and the need for a globally responsible approach to biodiversity conservation. The report addresses the following questions: : 4 How can NTFP commercialization contribute both to rural poverty eappe scree sieteuees eradication and forest conservation in Uganda? vere 1 How do spatial factors affect these two goals? Where do specific NTFPs have the highest chances of being successfully developed and commercialized? 3 How can this analysis be refined in the future to give a more complete picture? This project demonstrates that powerful spatial analysis tools now facilitate the integration of social, economic and environmental data, in support of better decision-making. Such ‘expert system’ tools could be made accessible to any number of stakeholders, providing a truly participatory and inclusive model for the sound management of our common natural heritage. www.unep.org United Nations Environment Programme P.O. Box 30552, Nairobi, Kenya Tel: +254 (0) 20 7621234 Fax: +254 [0] 20 7623927 Email: uneppub@unep.org Website: www.unep.org UNEP World Conservation Monitoring Centre 219 Huntingdon Road, Cambridge UNEP-WCMC Biodiversity Series No 18 CB3 ODL, United Kingdom Tel: +44 (0) 1223 277314 ISBN: 92-807-2364-2 Fax: +44 (0) 1223 277136 Email: info@unep-wemc.org Website: www.unep-wcmc.org August 2006 DEW/0832/CA