MIOMBO COMMUNITY LAND USE & CARBON MANAGEMENT: N’HAMBITA PILOT PROJECT Annual Report 1 August 2006 – 30 November 2007 CONTRACT No. B7-6200/2002/063-241/MZ Co-ordinator: The University of Edinburgh, Prof. John Grace Report prepared by John Grace, with assistance from colleagues at the Edinburgh Centre for Carbon Management and Envirotrade. Researchers at Edinburgh University during this reporting period have been: John Grace, Matthew Williams, Casey Ryan, Silvia Flaherty, Sarah Carter, Alistair Herd. Project administration: Joanne Pennie. Contact: Tel: +44 131 650 5400 E-mail: jgrace@ed.ac.uk N’hambita Project Report 2006/07 1. Introduction We begin this annual report with a résumé of highlights of the year, and then we list progress against stated goals. As the project is in its final year, we have listed the stated aims of the original work-plan against which achievement can be assessed and through which gaps can be identified. Finally, we present some key results obtained this year. 2. Highlights of the year • • • • • • • At the beginning of this reporting period, the project site was visited by John Grace and Joanne Pennie, Richard Tipper from ECCM, a delegation from the EU mission in Maputo and two external advisors. This was a good opportunity for the project to be discussed in detail with the sponsors. Grace and Pennie returned in September 2007 to complete discussions on several aspects of the project. This year a project database has been created by ECCM, which is a positive step for the project as it increases the traceability of the carbon from the producer the purchaser. Will Garrett from ECCM and Joanne Pennie have trained local staff to use it, and a follow up in early 2008 by ECCM will complete the process. New satellite imagery of the area was acquired (SPOT and MODIS), and this has enabled us to prepare a forest map and estimate deforestation rates. A paper summarising the carbon stocks and biodiversity at abandoned machambas in N’hambita is now in the journal Forest Ecology and Management. Over 70% of the N’hambita community are now involved in project activities. In 2006-7, over 500 new farmers joined the project, and approximately 1000 ha of new land will be managed for project activities. As of April 2007, 8000 ha of land was under forest management as part of project activities. By June 2007, there should be 500,000 trees planted as part of project activities. The learning outcomes of this pilot study so far are being used to guide the roll-out of this work to other areas. In 2007 Envirotrade launched a new project at Cheringoma in the buffer zone of the Marrameu Reserve, as a joint initiative with Zambezi Delta land-use Project. Envirotrade also launched a project in the Quirimbas National Park in Northern Mozambique. In this report, Richard Tipper of ECCM presents a financial analysis of the roll-out process. John Grace presented an overview of the project by invitation of the British Ecological Society at its Annual Meeting in Glasgow. (http://www.miombo.org.uk/BESJohnGrace07.pdf) 2 N’hambita Project Report 2006/07 • In his association with the Global Canopy Programme, John Grace commented on, and signed, the draft Forests Now declaration (www.ForestsNow.org), which calls upon national governments to “Ensure that carbon emissions from deforestation and the protection of standing trees are included in all national and international carbon markets”. The declaration was released on Sept 12th 2007, and received publicity in advance of the UN Climate Change meeting in Bali, December 2007. 3. Progress against stated aims In this section we reproduce verbatim the stated objectives, then in italics we outline progress, noting where appropriate any deviation. Activity 1. Forest management 1a Establishment of community forest association The Provincial Forestry and Wildlife Department and a national Mozambican NGO, ORAM, will assist in the formation of a community forest association. This forest association will represent the community on the Regional Trust Fund with respect to carbon crediting and Plan Vivo registration. This was achieved early in the life of the project (i.e. registering the community association, obtaining the "Duat" and legalising the community’s claim to the land and its products). We have a set of documents that describes the registration of the N’hambita Community with ORAM (Organizacao Rural de Ajuda Mutua, Rural Organisation for Mutual Help). The Forestry Association operates within the Community Association, and incorporates a forest users’ forum, dealing with diverse matters ranging from fire control, the sawmill and the carpentry shop. 1b Training of community forestry workers Community forest workers will be given training in aspects of miombo forest management, including the selection of tree species, collection of seed, planting and maintenance of seedlings, planting and maintenance of trees and forest mensuration and inventory techniques. Community forest workers will input to the forest management planning process and be encouraged to cultivate a sense of community ownership throughout the project area. This has been ongoing since the start of the project; much of the effort in the current year has been put into fire control, as it has become evident that this is one of the most critical aspects of the work (without fire control, the carbon stocks are in jeopardy). In the coming year attention will be place in management of area A- see the satellite image on page 18 of this report). 1c Forest inventory Forest inventory will be carried out by community forest workers across the community forest area and will involve stratified random sample plotting to record species, basal areas, top heights and tree condition. The inventory will inform; timber utilisation through volume estimation and predicted yields, forest management through estimation of fire damage, seedling mortality and biodiversity loss and carbon conservation potential through tree volume data. Forest inventory work has been on-going since 2004 in connection with the evaluation of carbon stocks, and the inventory report is appended to this document. We found 164 woody species. The 21 most significant (in terms of biomass) species in the sample plots collectively account for over three quarters of the total biomass and include the defining species of miombo woodland. Some rare and valuable species were found. Satellite remote sensing has been used to measure the area of forest, and thus to estimate the total biomass and total carbon stocks available. 3 N’hambita Project Report 2006/07 1d Community forest nursery establishment The community forest nursery will be established with the agro-forestry nursery. Rough wooden shelters with wooden work surfaces will be constructed adjacent to raised beds for seedling transplants. It will be important to consult women within the community regarding the siting and construction of the nursery as they will constitute the bulk of the nursery workforce. This has been successful; the nursery sites are illustrated on our website (http://www.miombo.org.uk/LandUse/Nursery.html). Early problems of irrigation were solved by the installation of a bore hole. The first nursery worked very well, and now two others have been established. The nursery workers include females although in general, for cultural reasons, they do not constitute the bulk of the work-force. JG visited one of the new nurseries on 25/9/07: a staff of six workers was producing 40,000 seedlings per year; seedlings are ready for use at the ‘four true leaves’ stage after about three months. 1e Production of seedlings Seed will be collected from selected tree species and will be planted in containers. Seed collection will be done by children and seed will be stratified and then sown and labelled within the nursery. Seedlings for most miombo woodland tree species take one year to reach the stage where they are ready for planting out. Training in tree and seedling identification, nursery techniques and plant handling will be given throughout the programme. The community will be encouraged to regard the nursery as a community resource. The current model, as outlined in 1d, can be replicated wherever seedlings are needed. Workers have been trained to collect the seeds of preferred species on a regular annual cycle, and so far there have been no signs of poor germination. Schools have been involved in tree planting around school houses and two schools have their own micro-nurseries. 1f Management planning A management plan for the forest area will be informed by community input, use existing data, and information gathered by the forest inventory and will be updated over the course of the project. Simple, verifiable outputs will be identified and community agreement required for the different management units. Management planning of the forest area is in the hands of a very experienced Mozambican forester employed by Envirotrade, Eng. Antonio Serra, formerly of Centro de Experimentação Florestal, Maputo. He does much of the strategic planning and organises the workforce, guided by the document Draft Forest Management Policy for N’hambita Regulado (Appendix 1 on Page 24). In the coming year he will prepare a management plan for fire control in area A (see satellite map on page 18) 1g Establishment of Permanent Sample Plots (PSPs) Permanent Sample Plots will be used for long term monitoring of the miombo forest condition and quantification of project impacts through forest management. Sample plots will be located by forest workers and marked with durable plot identifiers and recorded with GIS location co-ordinates. Fifteen PSPs of one-hectare each were established in randomly stratified locations within the main land cover types that had been identified in the initial survey of land cover based on LANDSAT imagery (Spadevecchia et al. 2004). These PSPs are geo-referenced. All woody plants with a stem diameter ≥ 5 cm were recorded and mapped, and their diameters at 1.3 m above the ground (‘diameter at breast height’, dbh) were measured in 2004, 2005, 2006, 2007. By 2007, all trees were tagged with a metal label bearing a serial number which can withstand fire (when a tree is completely destroyed by fire, the tag can be recovered from the ash and recorded). By early 2007, some individual trees were additionally instrumented with metal girth bands and vernier gauges for precision-measurement of growth, although by late 2007 it became evident that the gauges do not always withstand fire. 1h Timber extraction Timber extraction will be carried out once inventory data provides sustainable yield figures, in respect of volume and species. A felling licence (or concession) will be obtained from the Forest Service, on approval of the sustainable yield and management plan. Community forest workers will be trained in basic tree felling techniques and felling will be carried out with hand tools. It is not deemed desirable to 4 N’hambita Project Report 2006/07 train and equip the community with motor manual power saw systems during the first five years of the project. Extraction will be on prescribed and predetermined extraction routes. Approximately 87 m3 of dead logs were extracted from the woodlands and taken to the carpentry shop. New authorisation from the Agriculture Department was obtained to collect dead logs. About 21 m3 of dead logs from different species were collected and turned into about 9 m3 of planks. The community received authorisation from the Forest Department to cut 20 m3 of Pau Preto, 30 m3 of Panga Panga, 25 m3 of Umbila and 25 m3 of Umbaua. After the community has paid the tax on the above species, they are allowed to cut the authorised amount. 1i Replanting and enrichment planting Planting of indigenous miombo tree species will be carried out in forest areas that have been high graded of particular species, clear felled in patches for fuel wood, and will be selectively felled as part of the project. Replanting and enrichment planting are in the Policy document although this has not been required so far; however, the Management Plan states the principles as follows. The rehabilitation or reforestation of denuded forest is to be undertaken in blocks or strips in a systematic manner according to a pre-set plan. It must be linked to the sale of carbon thereby giving the incentive to the community to undertake such a venture. By implication, the species selected must be chosen, in addition to other reasons, for their potential for carbon fixing. The sites for reforestation may require a different approach according to the scale of denudation; heavily deforested sites will require pioneer species to be planted prior to more valuable species being re-introduced. Those sites that still maintain some tree growth could be enriched with selected species from the area that will add value to the future forest in terms of lumber, fruits or medicine. In order to jump start the process of getting the young saplings out of the fire danger height (suffrutex height), the advantage of truncheons or cuttings will not be overlooked. These selected species could be pre-struck in the nursery or planted directly into the field, with the probability of a reduced strike rate. Planting spacement will be 5 – 10 meters apart thus delivering a population of 100 – 200 trees per hectare. In the case of having to re-establish a tree stand totally with the use of pioneer species, one will need to increase the initial population. 1j Measurement of PSPs The permanent sample plots will be measured on an annual basis for the first five years. Information regarding stocking density, tree condition, fire loads, regeneration and tree species will be recorded by community forest workers. Where feasible, biodiversity recording of observed flora and fauna will be carried out. This is covered in 1g above, and results are in the inventory report (Appendix 2 on Page 47). Systematic observations of fauna have not been done, although the floristic data have been used to calculate the Shannon Index of biodiversity. Activity 2. Timber Utilisation The timber utilisation operation will be set up as a community business. Community members will be paid for work carried out, wages will be set at regional rates and the selection of workers will be decided by community workers with co-operation from project staff. The income generated by the timber utilisation will be re-invested in the business and into community funds. Community members will be provided with training in the use and operation of mobile sawmills. A wood workshop will be set up to produce furniture both for local use and sale. The wood workshop will be modelled on a similar successful initiative operating in the area. An existing community facility is operational near Dondo 5 N’hambita Project Report 2006/07 (Sofala) and this has been successfully producing and marketing indigenous timber goods for the last three years. 2a Establishment of community timber utilisation association The community timber utilisation association will be formed in order to facilitate community ownership of the equipment and basic facility that is established as part of the project. Representatives will be drawn from the community to serve on the association. This association will be responsible for the disbursement of financial surpluses made from timber utilisation. It is therefore important to constitute this association in such a way as to channel future funding into community wide projects such as healthcare and education. The community business is well established, and the workshop has the capability to convert fallen timber into a variety of goods, ranging from utility items (beds, stools, chairs, beehives, coffins, doors and shutters) to high quality products such as tables made from attractive woods finished to a high standard. The refurbishment of the Gorongosa National Park continues to generate orders for custom made cabinets and other furniture. 2b Provision of equipment (saw mill and carpentry) Equipment will be selected on the basis of ease of operation, robustness and reliable working. It is important in such a remote community to provide appropriate technology that requires little maintenance. Examples of this type of equipment can be seen operating in Sofala Province. The saw mill which was purchased quite early on in the life of the project continues to give good service. Basic tools were initially a problem as those purchased locally were of low quality; it was necessary to purchase second hand tools of high quality in South Africa instead. These have given good service. Accounts of the saw mill show a useful surplus. 2c Training of community workers Workers will be trained in basic carpentry tool use, tool maintenance and carpentry. From experience elsewhere in Sofala Province it takes 4-5 years to fully train woodworkers to a standard where they are producing high quality timber products. Two of the technicians will be trained in mobile saw-milling. Basic business training and bookkeeping will also be taught in the expectation that the community will own and run the timber utilisation business after five years. Training has been successful, as can be judged from the standard of the products. The project was able to recruit a trained and apprenticed carpenter from the Maforga Project (which trains craftsmen and artisans) who manages the micro-business. Improvements will continue to be made, and new recruits to the carpentry trade are being taken on, so training is on-going. 2d Production of sawn timber After felling and extraction, timber will be sawn on the mobile saw and left to air dry (6 to 18 months depending on tree species). Timber may then be sold as planked timber or have further value added by conversion to furniture. The basic principles of air-drying are now well understood by the carpentry shop; sales of planked timber have been made, but most of the output has been in finished goods. 1n 2006/7 About 38 m3 of sawn timber was produced. 2e Manufacture of furniture and other products Community workers, with training and guidance from the timber utilisation co-ordinator, will carry out the manufacture of furniture and other wooden products. The project has access to existing furniture and wooden product designs through links with other Mozambican timber projects and community workers will be encouraged to develop design and prototyping skills. The project has developed a range of products and their design continues to be improved as new skills, materials and tools become available. The original emphasis was on durability, but with the assistance of consultants and visitors to the project, improvements in design and ergonomics are being effected. The Maforga Project near Chimoio has played an important role in assisting in this process. 6 N’hambita Project Report 2006/07 2f Marketing of products The market for furniture and wooden products will, in the first instance, be located in the city of Beira and Zimbabwe. The economy of this part of Sofala Province is expanding and there is a demand for simple and robust, locally manufactured wooden products. The development of markets in South Africa and beyond will be carried out in partnership with ICRAF (World Agroforestry Centre based in Zimbabwe). Furniture and beehives were sold to Gorongosa National Park, the Agriculture Department in Manica and Sofala, and various customers in Gorongosa and Beira. Vegetables were sold at the markets in Gorongosa, Chimoio and recently Gorongosa National Park. In February 2007 the first five of twelve (high value) special tables from the community carpentry were exported to South Africa. The list of items manufactured and sold in the reporting period is as follows: doors, doorframes, special chairs, normal chairs, windows, beds, bee hives, cages, tables, sink unit, towel holder, cabata, safe, stools, room divider, ironing table, dressing tables, carpentry table, coffins, cupboard, wardrobe, chest of drawers. Activity 3. Agroforestry In the current pattern of land use, each farmer has from 3 to 6 ha of land in use; approximately one third is used for crops and the remaining lies as fallow. The crop is rotated each 2 to 3 years and once the land is exhausted the farmer clears a new area of land from the forest. Agroforestry activities will aim to maintain soil fertility in agricultural lands so that the need for new land from within the community is reduced. 3a Training of farmers Farmers will be given training in all aspects of agroforestry techniques, including the selection of species, collection of seed, production of seedlings, planting and maintenance of trees and the best way to manage agroforestry systems to maximise yields. Training will be ongoing and will continue through the programme. Appropriate use of exchange visits to other areas where farmers have already started using agroforestry activities will be used to allow farmers to discuss the pros and cons of systems with other farmers. Over 70% of the N’hambita community are now involved in project activities. In 2006 / 7, over 500 new farmers joined the project, and approximately 1000 ha of new land will be managed for project activities. 3b Propagation of seedlings The first steps will be the construction of field nurseries and collection of seed. Seed will be collected from areas where other agroforestry projects have taken place and where mature trees are available. Other species such as banana will be propagated vegetatively, again collecting material from farmers who have previously been involved in agroforestry projects. The aim is to foster an attitude of community exchange where farmers who have been supported in the past provide new farmers with access to seed sources. If necessary, seed can be bought from regional institutes such as ICRAF in Zimbabwe or Malawi. Women will be involved in the planning and implementation of nursery work. Two models of local (village) seedling production units have been trialled at Mucombeze Ponte and Mucombeze 1. In one, the contract for the annual production of 40,000 seedlings has been awarded to a team of six people who produce the seedlings according to a strict protocol and schedule. In the other, the contract is with an individual who then subcontracts a team of seven. In both cases, they collect the seeds, obtain the soil, sow the seeds in 4 inch pots under shade and provide irrigation. Seedlings are sold at the four-leaf stage (3 months from sowing) at 16 US cents each. The annual income from this is about 1000 US dollars, exceeding the 400 dollars that can be made by producing charcoal. 3c Intercropping Intercropping involves the planting of nitrogen fixing trees into fields being used for annual crops. The trees may be planted in lines (alley cropping) or randomly dispersed in the field according to the farmer’s preference. The trees are cut back regularly (twice per year) and the green matter incorporated 7 N’hambita Project Report 2006/07 into the soil. Previous projects in the Gorongosa region, building on research carried out by ICRAF, have demonstrated the effectiveness of various tree species and these will be used as the base for promotion. However, farmers will be encouraged to experiment with various tree species on their land to ensure that systems are well adapted to local conditions. Early impressions are that many farmers opt for the Plan Vivo option with the least intervention, which is boundary planting. However, the gains from intercropping are substantial and we need to establish a demonstration plot to show farmers what can be achieved. 3d Improved fallows Improved fallow involves using a mixture of native and exotic soil improving tree species in the fallow to increase the accumulation of soil nutrients before the land is reused for annual crops. Many native species have very low growth rates and the selection of appropriate species can greatly increase the rate of soil recovery from 5-6 years to 2-3 years. This activity is central to the work of the project and extension support to farmers in introducing additional crops and techniques is aimed at the fallow cycle and soil nutrition. Evidence of the changes that have resulted can be found in the improved yields from existing crops (maize and sorghum) and the considerable expansion of “new” crops like “pigeon pea”. This is being achieved through agroforestry techniques using recognised species. 3e Reforestation with fruit trees Abandoned fallow of less than 10 years in age are recognised as being the property of an individual farmer. Farmers can diversify their production systems and possibly increase their incomes by using these areas for growing fruit trees and, if combined with bee keeping, by planting flowering trees and plants as bee fodder. A number of species of fruit tree are native to miombo woodland and these, together with other fruiting trees such as Papaya, will be planted on abandoned fallow land. The project has imported and sourced high quality grafting material for the nurseries. The community has secured additional charitable funding to import six species of high value fruit trees from South Africa and orchards to supply grafting materials are being planted at each nursery. This will ensure that the nurseries are supplied with the necessary grafting material to grow large numbers of higher value fruit trees in the future. This is part of the sustainability strategy for these micro-businesses. 3f Planting of riverine areas Planting of river-banks with a mixture of bananas, sugar cane and trees helps reduce soil loss through runoff and river bank erosion. Planting takes place at high density and provides a new source of production from bananas and sugar cane. This forms part of the extension support given to the farmers participating in the project. Extensive planting of the river banks has taken place and banana production has been significantly increased by the introduction of other cultivars and new techniques. During last year’s rainy season when the lowlands were inundated, those areas which had been rehabilitated were able to escape damage from the flooding. 3g Participatory analysis of results The agroforestry component will involve a participatory analysis of results as part of demonstrating the advantages of new techniques to farmers and encouraging increased uptake. Soil samples are taken from fields and fallows each year and a record of crop yields recorded (see Methodology). Results will be discussed and reviewed with farmers along with observations on factors such as soil erosion and performance of different systems or species. This has happened through the “Escola da mashamba”whicht is an important hub in the project. The area is planted with demonstration plots which include crops, trees, orchards, vegetables and other plantings. This area is used for extension support training, for teaching and demonstration and produces products that provide another income stream for the community fund. The micro-credit scheme to supply drip-irrigation units is also linked to the school and all farmers who purchase these systems come and work and are trained in the school. These plots are monitored and studied by our staff and yields and techniques are monitored with lessons learned being incorporated into extension 8 N’hambita Project Report 2006/07 support. Specific demonstrations of techniques to deal with erosion and flooding are present at the school. 3h Extension of techniques Of the initial group of farmers, a number of community extension workers (2 to 4 people) will be identified and employed by the project to promote agroforestry techniques. Community extension workers will be trained and equipped with bicycles. Community extension workers will work together with project staff to train farmers. Women will be involved in the training of other members of the community in respect of nursery techniques. The community extension programme is conducted in close collaboration with the Department of Agriculture and involves ongoing interaction with the community at different levels to encourage the uptake of new techniques and technology. The community extension officers are trained in the project at courses conducted by the project staff under the guidance of Eng. Serra. The project has developed a training syllabus and suitable training material and successful trainees are equipped with the skills to provide low level extension support in the community. These individuals are provided with bicycles and form part of the project staff imbedded in the community. Activity 4. Non-Timber Forest Products The aim of promotion of NTFPs will be to increase local income, diversify production systems and reduce pressure on forest resources. The management of NTFPs also has the potential to involve other groups of society not involved in forest management and timber utilisation. Two main activities have been identified through discussions with the community and local extension workers as having the potential to achieve these aims: beekeeping and captive rearing of cane rats. 4a Training and provision of equipment Individuals to take part in training activities will be selected based on discussions in the community identifying those people with particular interest and capacity, for example people who already practice bee keeping will be targets for hive management training. Training will be ongoing involving small groups of around 10 people. This has taken place. 4b Bee keeping Bee keeping is already practised in the community, for the most part, using traditional log hives. While these techniques produce honey suitable for local consumption, the quality is not high enough to facilitate sale in regional markets. Other common problems with this sort of management are the loss of bee colonies during harvest and the spread of fire from the smoking process used to calm bees before harvesting honey. The project will work with bee keepers to improve hive management and increase the quality of honey. Although it is not necessary to try to replace traditional hives as these still have a role in domestic production, the use of Kenyan top bar hives can be used to increase honey quality through better hive management. Extraction techniques are also important in production quality and will be a focus of project work. Bee keepers will be provided with bee suits, hives and extractors on interest-free credit and given training in hive management and honey extraction. We have introduced the Kenyan bar top bee hive, an improved model which means honey can be harvested without the destruction of the hive. It is used in the in the community bee-keeping association (of over 70 members, and 295 hives, in N’hambita, Munhanganha, Bue Maria and Pungwe. Beekeeping is becoming an increasingly more secure investment as fires are reduced throughout the project area. The implementation of new technology has meant that 'clean' honey can be sold on the wider market. In September 2007, the project was visited by Michael Schmolke, one of Africa’s most well-known apiarists. He and his son gave advice to the project and ran a training course. However, some people persist with ‘old style’ hives, made from totally bark-ringing mature Marula trees (Sclerocarya birrea) which then die. Hives are usually placed high in the canopy, so that they withstand burning. Each hive contains several hundreds of thousands of bees. During one three-month period for which records are available, about 260 kg of honey was produced from which 220 kg was purchased by a merchant for sale in town. Comb-honey is likely to be the most lucrative. African bees are harder-to-manage than European bees, although they are the same species. So apiculture is often 9 N’hambita Project Report 2006/07 challenging, especially as people look after their bees as a ‘spare time’ activity and probably underproduce through lack of attention to the habits of the bees. Only a few of the bee-keepers are women, although women are supposed to be good with bees. To assist management of the hives, smoke generators are being manufactured in the metalwork shop next to the carpentry shop. 4c Cane rat production Hunting often involves the use of fire that can present a threat to forest resources but provides an important source of protein to rural communities. Cane rats (Thryonomys swinderianus) are large spineless porcupines commonly hunted by local people. Work in Ghana has demonstrated that cane rats can be successfully reared in captivity for meat production; animals do not require large areas and can be raised in back yards and be fed on grass and household vegetable waste. Cane rats do, however, require strong wire cages increasing the start up costs. Families will be provided with cages on interestfree credit, and given support in rearing animals. Cane rats proved to be too delicate and resistant to domestication and suffered high mortality rates when bred in captivity, and so domesticated Guinea Fowl have been used instead. A householder is provided with 50 fertile eggs which are hatched by hens (but hens cannot be the mainstay of meat production because they are prone to disease). Once they are raised, eggs produced are ‘paid back’ to the community and given to another household, to disseminate the process. 4d Marketing Marketing will initially focus on honey, wax and other hive products. As the quality of these products is improved it will be possible to help bee keepers target different markets and achieve better prices. Key problems in accessing markets are containers for the honey and transportation. By organising larger groups of beekeepers it will be possible to achieve economies of scale in transport costs and the project will invest in large durable containers for honey. The project will seek to negotiate agreements with local and regional market places to allow the sale of honey through these outlets. The project will also explore the possibility of marketing other hive products such as propolis, pollen and wax. Wax will be supplied to the community carpentry enterprise. The community has secured the de-mining of the abandoned church at the entrance to the village. Currently a budget is being sought for the rehabilitation of the building for use as a shop to sell the NTFPs and all other produce from the project. The shop is well placed to attract custom from people travelling to the Gorongosa National Park. It will provide a much needed outlet for producers, craftsmen and women in the community. 4e Extension People given training will be encouraged to disseminate techniques learnt and help in the training of subsequent groups in the community. Individuals supplied with training and equipment will be expected to help supply new trainees with breeding stock (queen bees, swarms and cane rats) as part of the programme of community exchange. Dissemination has occurred through links with government and non-government agencies. At district level, agricultural interests have been involved since the beginning of the project. We have been collaborating on tree dissemination in other areas considered to be a priority by the district government. We provided them with seedlings of Jatropha and Macadamia to be planted in Gorongosa village and at Gorongosa Mountain. Another example of collaboration is training, especially in Plan Vivo and nursery procedures. In return, they assisted the community with advice and seeds of sorghum as well as pineapple seedlings. Collaboration with the Health Department has occurred in training and provision of basic equipment. The Education Department has been involved in advising the Community Association, and have helped to involve schools. The idea is that each school will have at least one Plan Vivo system, supporting the government plan of "one student one tree". We also have two schools in our beekeeping programme. Apart from this, and in collaboration also with WWF, we are doing some Environment Education. As a result of the experience, the government is negotiating with other institutions to disseminate environmental education to other schools in the Gorongosa buffer zone. Provincial Government Departments. The Head of the Forestry Department and the Head of the Land Department have visited the project. It was decided to exchange seeds and jointly participate in 10 N’hambita Project Report 2006/07 training on nursery issues. Authorization to collect dead logs has been obtained and this year the forest department agreed to give a logging licence. At national level, the Envirotrade team travelled to Maputo to visit the Minister of Environment, also the focal point on Climate Change, the national directorate of Land, Forest and Wildlife, the National directorate of Extension Services (these two from Ministry of Agriculture), and senior staff from Eduardo Mondlane University. Other institutions with whom discussions and collaborations have been undertaken are: the agriculture research centre in Manica province with whom we collaborate on seed collection, training on fruit tree grafting and Jatropha diseases and pests. We have had important discussions with the Gorongosa National Park authority, supporting and advising the community, including negotiations to define conservation areas. Activity 5. Regional carbon management research Research will be carried out to quantify carbon benefits that can be generated from land use activities. This component of the project will produce outputs that that will be applicable to land use projects throughout the miombo ecosystem. Various land use systems that have been identified as having the potential to provide socio-economic benefits to rural communities. Research will involve: 5a Literature review Literature review will be used to identify what information is currently available on biomass accumulation by miombo woodlands. This will be used to produce preliminary estimates of carbon offset potential and to plan biomass field surveys. The literature review on the carbon stocks and fluxes for savannas in general was published by Grace et al. (2006) and for miombo in particular we reviewed the literature for our recent publication (Williams et al. 2007). 5b Training of community technicians (ICRAF) Community technicians will be trained to carry out biomass surveys under the supervision of researchers and field technicians. This is no longer the role of ICRAF, but was taken over by Envirotrade and by the University of Edinburgh. An elite group of community workers have been trained to measure the trees in the PSPs. 5c Biomass surveys Biomass surveys will be carried out to quantify the standing carbon stock of different land use systems in the project area and the rate of accumulation of carbon by these systems. Biomass surveys will measure biomass in various carbon pools such as timber, foliage, roots, soil carbon etc. The design of surveys will ensure carbon pools expected to change will be measured. We have found that the carbon stocks in the biomass vary between the vegetation types from less than 10 tC ha-1 in machambas to over 40 tC ha-1 in riverine vegetation. The average for 87 plots is 21 tC ha1 above ground with a further 5 tC ha-1 estimated to be below-ground biomass and another 18–140 tC ha-1 in the first 30 cm of the soil, as organic matter. These values are consistent with data reported by Williams et al. (2007b) in a synthesis study for the whole of Africa. Interpolating their graph we obtain biomass carbon stocks of 35 tC ha-1 and soil carbon stocks of 80 tC ha-1. Another estimate may be made from Frost (1996) who related biomass of miombo woodland to annual rainfall. From his regression equation, taking the annual rainfall as the 1990-2005 average (749 mm/annum) we get a biomass of 30 tC ha-1. The methodology and further results are described in the appended inventory report. 5d Regional baseline analysis The potential carbon offset produced by any land use is always measured in comparison to a baseline case – the amount of carbon which would have been accumulated or stored if no project activity was implemented. For activities which aim to conserve carbon stocks through the avoidance of 11 N’hambita Project Report 2006/07 deforestation the baseline involves an analysis of how the rate of deforestation would have proceeded in the future. The project will use methodologies developed by the DFID (UK Department for International Development) funded CLIMAFOR project in Mexico to construct regional baselines for deforestation (see Methodologies). This method involves an analysis of the relation between historical land use trends in the region and socio-economic factors. The end result is a baseline matrix that gives the predicted deforestation rate for a combination of predisposing and driving factors (for example distance to towns and type of land use). For carbon uptake on machambas where trees are planted, the biomass carbon at the outset is effectively zero, so it is sufficient to measure the trees, and to calculate the carbon from the equation developed by our researcher, henceforth we call this the Casey equation: y = 0.0267d2.5996 ; where d, diameter at 1.3 m above the ground, is measured in cm, and the biomass y is in kg C. When various forms of agroforestry such as intercropping are practised it will be necessary to measure soil carbon at 2-5-year intervals, to estimate the C accumulation rate. We have calculated the current deforestation rate of the whole area from SPOT images from 1999 to 2007. The value varies over the region as a whole, but the overall deforestation rate is 1.63% per year. 5e Carbon modelling The uptake and storage of carbon by the various land use systems will be calculated by carbon modelling. A number of models are available for the purpose of calculating the carbon offset potential of land use activities. Modelling will take account of the accumulation of carbon in various biomass pools in the forest ecosystem (timber, foliage, roots, leaf litter and soil organic matter) and the effect on these pools of management practices (thinning, harvesting, timber utilisation, etc.). Existing models are not appropriate as they do not address the processes at an appropriate scale, and are usually over-developed in some respects and insufficiently developed in other respects. The model needs to be able to simulate the response of miombo to changes in fire pressure, and to the likely increases in harvesting (fuel wood, charcoal) which will happen in the ‘without project’ case. Here we outline the general principle and give a preliminary result. It also needs to be able to take into account ‘story-lines’ about possible economic development in Africa (Alcamo et al. 2006). The carbon balance on a per year basis, Δ, expressed as tonnes of carbon per hectare per year, can be represented as: Δ = Pn - Fr - Ff Where Pn is the net primary productivity which, to a first approximation, is a linear function of rainfall (Frost 1996), and different for ‘managed’ systems on machambas where N-fixation has a stimulatory effect; Fr is the decomposition rate of the organic matter produced, attributable to the respiration of microbial organisms, and normally a function of moisture and temperature; Ff is the flux of carbon lost through fire. This latter term is the most important to be able to model, and to understand its dependency on socioeconomic factors. It includes a number of terms which depend on human activity as follows: Ff = f (he, hf, hc …) Where he, hf and hc are human influences represent encroachment, fuel usage and charcoal making respectively. These are dependent on the human population, but in different ways: he and hf depend on local population density but hc depends on the demand for charcoal, which previously was low but is currently increasing (Mangue 2000, Brouwer and Falcao 2004, Millenium Ecosystem Assessment 2005). This increase in demand comes from townspeople. It is also dependent on distance from town and distance from the road. Here we include a preliminary result to show the modelling idea (Fig 1). It should be possible to estimate the rate at which the carbon stocks would disappear in the absence of a project. This preliminary model, not surprisingly, shows great sensitivity to human population size. 12 N’hambita Project Report 2006/07 Forest area remaining (%) 100 80 60 A 40 C 20 B 0 0 10 20 30 40 50 60 70 Years from now Fig 1; shows the likely loss of forest resulting from three assumptions. A, deforestation rate proceeds at its current rate; B, deforestation rate grows according to the size of the Mozambican population (we assume the population grows at 1.8 %, which is the current rate), and C, in addition, that the per capita demand for wood increases at 3% per year as more people in towns use charcoal. 5f Production of technical specifications The end product of the research will be a series of “technical specifications” describing land use systems identified by the project as having a potential to provide socio-economic benefits to communities in the region. These technical specifications will describe the management requirements necessary to generate a given carbon offset and how this value varies across the range of environmental conditions found in the region. The specifications will include monitoring indicators so that carbon accumulation and storage may be easily verified. The specifications will be evidence-based documents with all information sources and assumptions stated, so that they may be independently verified. Technical specifications will be designed to be user-friendly documents that may be used by forest technicians as well as academics. The research team will work closely with the project technical team to identify which land use systems should be included in the research and to ensure that the outputs are suitable for use by technical support organisations in the region. The technical specifications are typically 3-5 page documents describing the way to achieve carbon sequestration by tree planting or forest protection. Those we have developed so far are as follows: • • • • • • • Boundary planting Dispersed interplanting - Gliricidia Dispersed interplanting - Faidherbia Fruit orchard - Cashew Fruit orchard - Mango Homestead planting Woodlot An important further technical specification is ‘avoided deforestation’, which is in the final stages of preparation as this report is being written (Appendix 4 of the Woodland Inventory report). The technical specifications may be consulted at http://www.geos.ed.ac.uk/miombo/Documents.html. A list of tree species can be found at: http://www.geos.ed.ac.uk/miombo/MiomboTreeSpecies.pdf. By far the 13 N’hambita Project Report 2006/07 greatest take-up by farmers has been boundary planting. This is the least interventionist and presumably reflects caution on the part of the farmer; it is expected that as confidence increases so the take-up of other technical specifications will occur. Payments for the estimated carbon sequestration are spread over seven years: 30 % immediately after planting, then 70% spread over six years. Activity 6. Carbon verification A key activity of the project will be developing the institutional structure necessary to verify carbon offsets on the ground, administer the sale of carbon offsets and provide support to communities in planning land use activities and registering carbon assets. The research and technical capacity building components of the project will be co-ordinated to ensure that research is being correctly orientated and that the trust fund technical team have full opportunity to learn from research activities. Capacity building activities will build on experiences learnt in the DFID funded international pilot carbon-trading project in Mexico. Mexican trust fund staff with experience with Plan Vivo systems will be contracted to carry out training in Mozambique to capitalise the direct field experience of these technicians. 6a Establishment of institutional structure An independent trust fund will be set up that will administer the registration and sale of carbon offsets generated by project activities. The trust fund will keep records of all land use activities implemented in the target community and details of monitoring activities carried out in these lands. The trust fund will serve as a point of exchange between parties interested in purchasing carbon offsets and communities involved in land use activities, providing verification services to purchasers and ensuring transparent accounting of carbon. The aim is that the trust fund will continue to act as a registry of carbon offsets for other communities after the project finishes and will eventually be funded by carbon sales. Note that in the case of Mexico, the trust fund is now a self financing organisation. The Mozambique Carbon Livelihoods Trust (MCLT), hereafter called ‘the Trust Fund’ will become fully operational in late 2007, with the aim of ensuring that the proceeds of carbon offset sales are safeguarded. An organogram is given below to show how the trust fund will work (Fig 2). The Trust Fund will be controlled by a committee which will include the following: a nominee of the N’hambita Community Association, a nominee of Envirotrade Lda and a nominee of WWF. The planned structure and modus operandi are similar to the one which successfully controls funds for the Scolel Te project in Mexico. A bank account has been opened with monies being administered by the Mozambican finance company Contabil. The fund has a balance of USD 92,613.00. The MCLT Trust Fund is distinct from the community fund, which is a bank account which farmers pay into, providing community activities which have included contributions to the schoolhouse and clinic. 14 N’hambita Project Report 2006/07 Fig 2. Trust Fund and Plan Vivo, and operation of carbon payments. Each farmer has an individual Plan Vivo which is an agreement to plant the land according to technical specifications, guidelines, procedures and standards. Adoption of the Plan by a farmer triggers payment from the Trust Fund. Those who buy carbon receive certification from BioClimate Research and Development (http://www.brdt.org). Verification services are available through independent companies such as Smartwood (http://www.smartwood.org). The schedule of payments to farmers for tree planting goes as follows: immediately after planting, 30% of payment, then 12% per year for five years, then a final payment of 10% in the seventh year. Thereafter, the trees are established and yielding sufficient tangible benefits to dissuade the farmer from reverting to slash and burn. Payments for avoided deforestation through fire control go direct to the community fund and not to the Trust Fund. 6b Training of administrative and technical personnel The trust fund will require technical and administrative personnel to carry out verification and registration activities. The trust fund will initially contract personnel from the Provincial Forestry and Wildlife Department to provide these services. The project will work with the Department to build capacity so that they will be able to provide technical support to the trust fund. Personnel will be trained in the use of the ‘Plan Vivo’ systems and this will build on extension work funded by DFID in March 2001. Working with the Provincial Forestry and Wildlife Department in this way has the advantage of providing access to trained technicians and links to other communities in the region. Project staff will work with trust fund staff on each of the following activities. Staff are already trained in basic office techniques and in the use of Plan Vivo. There is, however, a need to further develop skills and to link appropriately with Provincial agencies which may assist in aspects of forest protection. 6c Land use planning Trust fund personnel will work with project managers in charge of promoting land use activities to help communities and individual farmers make simple management plans of their intended land use activities. These plans will be in the form of annotated maps and will provide the information necessary to assess the carbon offset potential of the land use activities. Plans give details of the current situation in terms of vegetation and land use, planned activities in terms of species and management regimes and planned inputs in terms of costs and labour requirements. The plans exist, and details of all of them have been entered into the Plan Vivo data base. Security of the data base is fundamental to the long term success of the project, and the data will be backed up to a secure server. 15 N’hambita Project Report 2006/07 6d Assessment and registration of carbon assets The trust fund will be responsible for assessing land use plans produced by farmers and the community. Plans will be assessed using the technical specifications developed by the research component. Each technical specification will define management requirements and environmental conditions for a given carbon offset. Assessment of management plans will include an assessment of the baseline, and the long-term viability of planned activities. Technicians will work with community members to explain the assessment process and how plans might need to be modified to meet the assessment criteria. Once plans have been assessed the carbon will be registered on the trust fund database. Currently the land use plans are assessed by the project managers, filed as a hard copy and entered into the Plan Vivo database. In future, the trust funds will participate in the process of approval so that the process is transparent and can be seen to be fair. 6e Monitoring and administration of carbon assets Trust fund technicians will monitor land use activities registered with the trust fund. Monitoring indicators set out in the technical specifications will be based on easy to measure aspects of plan implementation such as tree growth and forest protection. Trust fund technicians will train community technicians to help carry out monitoring activities which will make monitoring more cost effective and increase community participation in the process. The results of monitoring will be recorded by the trust fund administration and used to credit saleable carbon offset units to the community and individuals. Sales of carbon will also be recorded on the database and the community will be issued with carbon account books clearly showing how much carbon they have and how much they have sold. These systems will ensure that all carbon is accurately recorded and that no double accounting takes place. Monitoring of carbon assets over future years and decades will be as follows: Level 1, ground survey by farmers and the Trust. Farmers will record the survivorship and growth rates of the trees they have planted, some data on survivorship is needed for storage on the central data base, at the level of the management unit. Permanent sample plots will also be recorded on the ground, as an indication of any changes in the baseline. Level 2, periodic ground monitoring by an independent outside agency. It is not cost effective to do this often, as it incurs a high price, and so it should be done only to settle disputes that occur in the carbon trading activities, and when satellite sensors are changed. Level 3, satellite survey using MODIS and SPOT data. This enables the woodland carbon stocks and the extent of fires and burned area to be monitored. It becomes highly cost effective when several carbon projects are in progress in the same region, and it enables the observation of project leakage (survey areas nearby and regional deforestation rate). It will enable fire monitoring and it will detect changes in the vegetation cover from year to year (examples of these products are shown in this report). 5. Key results from 2006/7 5.1 Biomass survey The activity has produced an inventory report (Appendix 2 on Page 47) which contains the required information to be able to make estimates of the carbon stocks of large areas (thousands of hectares) and to make more exact determinations of carbon stocks in specific parcels of land selected for management. The basic information is shown in Table 1 and discussed more fully in Appendix 2. 16 N’hambita Project Report 2006/07 Table 1 Statistical summary of the five land cover types. (±) refers to the standard deviation, but variables are not normally distributed so this is only an approximate indication of the variability. Trees/ha Basal area (m2/ha) Above-ground biomass mean (tC/ha) Above-ground biomass median (tC/ha) Max. dbh (cm) Shannon index Sum of plot area (ha) Number of plots Riverine Tropical woodland forest Savanna Secondary Machamba woodland 421±167 13.8±3.3 406±253 10.0±3.2 386±275 5.8±3.9 561±255 8.0±2.0 38 2.4 47±18 27±13 14±10 13±9 8 43 24 12 14 8 92 2.2±0.5 2.1 64 2.0±0.4 10.1 61 1.2±0.5 3.8 38 2.1±0.6 8.6 70 N/A 1 6 26 10 17 1 These estimates are made from the equation developed for the local woodlands by our researcher Casey Ryan, who carried out destructive harvests of individual trees. The equation supplants those we have been using on a provisional basis, which came from other researchers in different parts of Africa (Williams et al. 2007a). These data may be compared with the literature survey of Williams et al. (2007b): for the latitude of N’hambita (19 oS) the biomass is 30 tC/ha. Another estimate may be made from Frost (1996) who related biomass of miombo woodland to annual rainfall. From his regression equation, taking the annual rainfall as the 1990-2005 average (749 mm/annum), we get a biomass of 30. 5.2 Remote sensing of forest/non-forest The SPOT imagery shows the state of the woodland cover in 1999 and 2007, over a region of some 60,000 ha. The imagery does not enable us to unequivocally differentiate the different woodland types, so here we have combined them. The deforestation rate in this period (1999-2007) in the region as a whole is 1.63 % per year, and in the managed area A it is 1.06 % per year (Table 2). This may be contrasted with the figures over the period 1991-2000 obtained by our researcher using LANDSAT imagery in nearly the same area (Wallentin 2006). In this earlier period the regional deforestation rate was estimated as 0.15 % per year at the regional level, and 0.03 % per year in the N’hambita area. Comparison with national deforestation rates declared by Mozambique and other miombo-containing countries may also be made from official figures made available by FAO (2007b): Mozambique 0.3 %, Angola 0.2 %, Tanzania 1.0 %, Zambia 0.9 % and Zimbabwe 1.5 %. The increase in rate is believed to be caused by a population increase leading to a higher requirement for machambas, and by an enhanced demand for charcoal. 17 N’hambita Project Report 2006/07 Fig. 2 SPOT 4 images of the project area, showing different land cover types in 1999 and 2007. The total scene has a land area of 67,754 ha. Zones are: A, corresponding to N’hambita, Bue Maria and Posta Da Pungwe; B, containing Pavua and M’Bulawa; C, a buffer zone only sparsely inhabited, and south of the river Pungue; D Mucombeze. The scene contains an additional area E, part Mucombeze and part Pinganganga (the latter falls into Manica province). The right hand panel plots the areas which have been deforested, with a spatial resolution of 20 metres. 18 N’hambita Project Report 2006/07 Table 2 Woodland cover in 1999 and 2007, and the average annual loss in that period, obtained from the analysis of SPOT imagery. A B C D E Total area (ha) 6378 9538 12149 15669 24020 Woodland cover 1999 (ha) 5385 8512 9706 11035 24020 Woodland cover 2007 (ha) 4927 7674 11538 7333 19539 Annual loss of woodland (%) 1.06 1.23 -2.36 4.19 2.33 5.3 Investigation of charcoal making We focused this year on an investigation of the activity of charcoal burners, as they have been identified as the primary agents for deforestation (Herd 2007). It deals with the basic processes, and the socioeconomic aspects. Fig 3 Charcoal manufacture, showing the structure of the kiln, building of the kiln, the aftermath, and the selling of charcoal at the roadside (Herd 2007). Production of charcoal is a time consuming process and is poorly regulated or not regulated at all; it is also inefficient and destroys some of the best woodland (Fig 3). 1 kg of charcoal is obtained from 5.7 kg of wood (kiln conversion efficiency is 17.6 %). Based on annual production rates, an average of 1559 tonnes are produced from an estimated 35 hectares of cleared miombo woodland. The study revealed that charcoal production was principally occurring within a strip 2 km wide to the west of the EN– 1, and is one of the main land use activities contributing towards land use change in the area (Fig 4). For those involved in the activity it is an important livelihood strategy, accounting for 74 % and 59 % of annual incomes for males and females respectively. In spite of this, 95 % of the producers fell below the US$ 1 poverty line, 19 N’hambita Project Report 2006/07 indicating that current production strategies are not capable of pulling people out of poverty. Sustainable charcoal production built upon community participation in the community has therefore been seen as a way to address these issues and the results from this study suggest that it is feasible given the known growth rates and a concession area of approximately 2585 ha. As a follow up to this study the suitability of the concession area proposed by the producers should be assessed through an inventory and written into a management plan for sustainable charcoal production. Fig 4 Location of charcoal kilns and deforested areas 1991-2000 (Herd 2007). The complete dissertation is available at http://www.miombo.org.uk/ARCHerdMSc.pdf. 20 N’hambita Project Report 2006/07 5.5 Technical Specifications A new technical specification has been produced for planting Faidherbia albida. This is a tree which can be used very successfully in agroforestry systems, because it sheds its leaves during the wet season. It also produces fruits which can be eaten, flowers which provide bee fodder, and it can also be used as a live hedge. A new technical specification for avoided deforestation is also currently being written, and will include methods for calculating carbon benefits of forest protection and a discussion on the practical considerations for forests. 5.6 Calculations for financial sustainability Following the mid-project meeting in Mozambique, Richard Tipper (ECCM) worked with the Envirotrade Team to develop financial projections for the development of the project. The main consideration for planning beyond the end of the current EU project is to understand how the project can reach a point where it is self-sustaining in terms of carbon finance. Our understanding is that the intention is to scale-up the project over the next 4 years to become a financially viable programme, based mainly on carbon finance. The key factors in achieving financial sustainability are: - the volume of sales of carbon certificate that the project is able to achieve and the value of VERs, the fixed costs of the core project team and “base camp” infrastructure, the feasibility of bringing new areas of miombo woodland under management, the amount of carbon that the project is required to retain as a “risk buffer” by the Plan Vivo system and the extent to which micro-businesses established by the project become selffinancing (and operationally independent). These factors are still under discussion within the project team. However, based on an assumption that annual project operational costs for establishing sustainable miombo management (through a combination of agroforestry, sustainable woodland management, restoration and fire protection), at scale of operations of 8,000 to 15,000 hectares will be approximately $900,000 per year, we estimate that the project will need to raise annual sales of VERs to around 600,000 tCO2 at $7 per tonne (after sales commission) to become financially self-sustaining. ECCM believes that this is a reasonable basis for planning, given the growth in demand in the voluntary carbon market. Indeed there appears to be a reasonable prospect for prices of Plan Vivo VERs rising above $8 per tCO2 over the next year, or more if the sale of carbon in this case can be coupled to the notion of conservation of biodiversity through avoided deforestation. The table below shows how the sales of carbon credits could be developed over the next 4 years to reach this point (Table 3). It should be stressed that these financial 21 N’hambita Project Report 2006/07 projections are still subject to discussion and require more detailed examination over the next 6 months. Table 3 Financial plan for the upscaling of the pilot project- first estimates. We assume in these calculations that the ‘without project’ scenario would lead to complete removal of woody vegetation within 50 years, and we further assume that the areas brought under management would retain their current carbon content which we assumed to be 30 t/ha. Note that prices are on a per CO2 basis rather than a per carbon basis; CO2 contains 12/44 of carbon. Note also that the price trend of carbon is uncertain. Sales of VERs Price of VERs Year 2006 t CO2 25,000 $ / t CO2 5 2007 50,000 6 2008 100,000 6 2009 300,000 7 2010 550,000 7 Revenue from sales $ 125,000 300,000 600,000 2,100,000 3,850,000 Project operational costs Plan Vivo Certification Cost Payments to community funds & farmers $ / t CO2 2.5 $ 62,500 $ / t CO2 0.25 $ 6,250 $ / t CO2 2.25 Total $ 56,250 455 2.4 120,000 0.25 12,500 3.35 167,500 909 2.2 220,000 0.24 24,000 3.56 356,000 2727 1.8 540,000 0.23 69,000 4.97 1,491,000 5000 1.7 935,000 0.15 82,500 5.15 2,832,500 Area brought under mgt approx ha 227 The required area for this growth rate is available (Fig 2, Table 2), as the entire N’hambita regulado has approximately 35,000 ha of which much is well-stocked (miombo and combretum woodland, 10-30 t biomass C ha-1), collectively containing about 1 million tonnes of biomass carbon, which is equivalent to about 3.5 million tonnes of CO2. 6. Conclusions This period has been one of consolidation and progress. Much has been achieved in the N’hambita community; there have been research advances through field work and satellite analyses, management practices have been developed and enhanced and carbon trading is growing steadily. For the remaining 8 months of the project, the partners will complete those issues outstanding which include carbon modelling, the avoided deforestation technical specification and the final implementation of community sustainability for those enterprises which require it. Looking ahead to June of next year, the planned conference in Edinburgh will bring together scientists from all over the world and from many disciplines with the key objective being to produce a major document which will identify the key questions related to the future of miombo woodland, and outline a plan for sustainable development. It is also hoped that the conference will provide the launch pad for the currently dormant Miombo Network. Around the same time as the Edinburgh conference, Prof. John Grace has been invited to preside at a meeting at the prestigious Royal Society of London. This meeting will concentrate on woodland savannas from a more scientific perspective, and Prof Grace 22 N’hambita Project Report 2006/07 will present the Miombo Project as an example of how they can be managed sustainably. 7. References Alcamo, J, Kok, K, Busch, G, Priess, JA, Eickhout, B, Rounsevell, M and Rothman, DS (2006) Searching for the future of land: scenarios from local to global scale. P137155 In Land-use and land-cover change, eds. EF Lambin & HJ Geist. Springer. Brouwer, R. and Falcao, MP (2004) Wood Fuel Consumption in Maputo, Mozambique. Biomass and Energy, 27: 233-245 Frost, P (1996) The Ecology of Miombo Woodlands. pp. 11-57. In B.M. Campbell (Ed):The Miombo in Transition: Woodlands and Welfare in Africa. Center for International Forestry Research, Bogor, Indonesia. Grace, J et al. (2006) Productivity and carbon fluxes of tropical savannas. Journal of Biogeography (J. Biogeogr.) (2006) 33, 387–400 Herd, A (2007) Exploring the socio-economic role of charcoal production and the potential for sustainable production in the Chicale Regulado Mozambique. Unpublished MSc dissertation, School of GeoSciences, University of Edinburgh. Mangue, MP (2000) Review of the existing studies related to fuel wood and/or charcoal in Mozambique. European Commission, Directorate-General VIII, Brussels. Millenium Ecosystem Assessment (2005) Global Assessments Reports: 1: Current State & Trends, section 9 Timber, Fuel, and Fibre. http://www.millenniumassessment.org/documents/document.278.aspx.pdf Ryan, C, Williams, M and Grace, J (2007). Guidelines for the rapid assessment of vegetation carbon stocks in the N’hambita area. Unpublished project document, included in the inventory report which is appended to this document. Spadavecchia, L et al. (2004) Synthesis of Remote Sensing Products and a GIS Database to Estimate Land Use Change: an Analysis of the N’hambita Community Forest, Mozambique. This paper can be viewed at http://www.miombo.org.uk/LSpSept04.pdf. Williams, M, Ryan, CM., Rees, RM, Sambane, E, Fernando, J and Grace, J (2007a) Carbon sequestration and biodiversity of re-growing miombo woodlands in Mozambique. Forest Ecology and Management. Williams, CA, Hanan, NP, Neff, JC, Scholes, RJ, Berry, J, Scott Denning, A and Baker, DF (2007b) Africa and the Global Carbon Cycle Carbon Balance and Management 2, doc 10.1186/1750-0680-2-3, available as an open-access journal at www.cbmjournal.com/content/2/1/3. 23 APPENDIX 1 – Draft Forest Management Policy, 2007 DRAFT FOREST MANAGEMENT POLICY FOR N’HAMBITA REGULADO GORONGOSA ENVIROTRADE MOÇAMBIQUE APRIL 2007 24 APPENDIX 1 – Draft Forest Management Policy, 2007 CONTENTS INTRODUCTION…………………………………………………… Synopsis……………………………………………………………….. Policy Statement………………………………………………………. 2 2 2 1.0 MANAGEMENT OBJECTIVES…………………………………… 3 2.0 2.1 2.2 2.3 MANAGEMENT PLANNING……………………………………… Maps…………………………………………………………………… Forest Type Classification……………………………………………... Management Classes…………………………………………………... 3 3 4 5 3.0 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 MANAGEMENT PRACTICES (for increasing carbon sink)…….. Forest Rehabilitation by Re-planting………………………………….. Forest Rehabilitation by Regeneration……………………………….. Clearing New Land for Cultivation…………………………………… Plan Vivo Scheme……………………………………………………... Tree Propagation………………………………………………………. Forest Maintenance……………………………………………………. Forest Protection………………………………………………………. Law Enforcement……………………………………………………... Research and Monitoring……………………………………………… 6 6 7 7 8 8 9 12 13 13 4.0 4.1 4.2 4.3 4.4 4.5 FOREST MANAGEMENT BY CLASS (to maintain carbon sink) Timber Utilization Class………………………………………………. Protection and Conservation Class……………………………………. Recreation and Nature Reserve Class…………………………………. Research Class………………………………………………………… Community Class……………………………………………………… 13 13 14 15 16 17 5.0 FOREST RESOURCE UTILISATION……………………………. 18 5.1 Consumptive Use……………………………………………………… 18 5.2 Non-consumptive Use…………………………………………………. 19 6.0 COMMUNITY TRAINING & PARTICIPATION………………... 20 7.0 COMMUNITY EMPOWERMENT………………………………… 21 25 APPENDIX 1 – Draft Forest Management Policy, 2007 INTRODUCTION SYNOPSIS Woodlands provide a wide range of timber and non-timber products, they are inhabited by numerous species of fauna and flora and can be utilised for a variety of recreational and other purposes. The woodlands contain unique biota which will disappear unless specially managed and the woodland itself will be destroyed unless it too is managed. Sustainable resource management and biodiversity conservation must be practised to integrate the requirements of the community with the well being of the woodland. A Policy is required for the N’hambita Regulado which will provide the balance in environmental health, economic opportunities and social needs. This document sets out to give a basis for the above and it is hoped that the necessary impetus will be available to implement the Policy. POLICY STATEMENT In line with the objectives that Envirotrade has envisaged for their holistic approach to help rebuild the N’hambita economy, the rehabilitation of the forests in which the community lives is seen as vital to the successful implementation of the objectives per se. The welfare of the community is paramount, but that wealth is intrinsically tied to the forest. For this reason, the forest has to be managed and protected, to become the sustainable resource available to the very community that could destroy it. The starting point has to be a Management Plan which is dynamic and will evolve as information becomes available, or circumstances change which can dictate a different approach to the issues in question. The Community has to be involved in the implementation of the Policy and only with their ownership of the total concept is there any likelihood of it succeeding. 26 APPENDIX 1 – Draft Forest Management Policy, 2007 1.0 MANAGEMENT OBJECTIVES The objectives are very simple, with no high technical requirements anticipated to come from the beneficiaries of the woodland assets. • • • • • To protect the woodland for the benefit of present and future generations and as a secure carbon sink. To empower and upgrade the communities to become guardians of their woodland. To develop the community to manage the woodland in a sustainable manner. To improve the viability of the natural assets by introducing sustainable systems into the human environment. To secure the resource from outside intervention. The underlying objectives of the Management plan are the long term security and well being of the N’hambita Regulado, which will be enhanced with the sustainable use of the forest resource. 2.0 MANAGEMENT PLANNING Based on the management policy and objectives, planning will be implemented in order to fully integrate all facets of the natural resource with the objectives. The following planning instruments will provide the framework for implementation. 2.1 MAPS Management of a forested area is difficult to implement without the aid of maps which are maintained and updated on a regular basis. The most practical scale for a management map is 1:10000 which will provide the necessary detail to make decisions based on factual knowledge of the area. All features that are important to the management must be indicated, as far as possible, on the maps. It is preferable to have separate maps for different purposes, such as; A. Topographical and physical maps. • Contours and ground features. • Permanent water courses, dry drainage lines and wet areas. • Roads, tracks and paths. • Boundaries and neighbours. B. Forest management. • Roads, tracks and slip paths (when logging commences). 27 APPENDIX 1 – Draft Forest Management Policy, 2007 • • • The sub-division of the area into blocks and compartments. The demarcation of forest type and other natural vegetation types. Research and inventory plots. C. Land use map. • Habitation and social infrastructure details. • Areas for cropping and domestic requirement. • Forest areas and other management options. • Tourism and aesthetic considerations. D. Fire protection. • Fire breaks, traces and fire barriers. • Roads and other access options. • Vegetation and fuel load. • Fire protection features. The availability of maps is limited and the use of GIS and orthophoto mapping may not be available immediately, satellite imagery may be the only option in the first instance. 2.2 FOREST TYPE CLASSIFICATION The classification of forest and vegetation type is important in that it provides the ecological basis for area specific forest management. The miombo woodland of N’hambita is described as containing six groups of different woodland types. (Mushove Report - December 2004) A. Miombo woodland dominated by Brachystegia and Julbernardia species. B. Miombo woodland as above with abundance of Diplorhynchus. C. Miombo woodland with dominance of Pterocarpus rotundifolia, Burkea africana and Erythrophleum africanum. D. Combretum woodland. E. Riverine woodland. F. Combretum and Palm woodland. These groups of vegetation types, in general, are naturally separated into blocks and can be demarcated reasonably accurately, albeit with overlaps and patches within one another. This is of course most applicable to the Riverine class. Provision for the management of these isolated areas will have to be catered for when implementing management practices for the dominant type. The three miombo woodland classes will provide the source of commercial timber while the Riverine woodland group is to be protected as a sensitive area along the 28 APPENDIX 1 – Draft Forest Management Policy, 2007 drainage system. Combretum woodland is of no real value other than as a carbon sink which requires fire management and protection from cutting. 2.3 MANAGEMENT CLASSES Management of the woodland will be allocated to the following classes for purpose of differentiating any forest type specific and management requirements. It is probable that there may be overlaps in geographic boundaries particularly with the Protection and Conservation class and the Research class. An attempt must be made to protect the first four classes of management areas by limiting activities to that which it is designated. This does not include the collection of non timber forest products (NTFP) such as mushrooms, herbs and for bee keeping. A. Timber utilisation class. B. Protection and Conservation class. C. Recreation and Nature Reserve class. D. Research class. E. Community class. The utilisation of the forest for non timber forest products (NTFP) and dry fuel wood, (excluding building material), is not separated into a class as the total area can be utilized for the purpose of domestic collection. 2.3.1 Timber Utilisation Class The three Miombo Woodland areas will provide the concession site for timber harvesting following the best known practices in accordance with the Forestry departments principles of sustainable logging practices. In the first instance, confirmation of the existence of merchantable logs will be requested from SPFFB – Sofala before a harvesting plan is considered. 2.3.2 Protection and Conservation Class This class, in reality, applies only to the Riverine Woodland areas. These occur in sensitive zones protecting the stream and drainage line banks from unnecessary erosion. Furthermore, the area, by nature, is limited so it therefore implies that the species numbers are reduced and require conservation. The areas along stream and river banks are generally richer in nutrients and being close to water are more desirable for the establishment of gardens which adds to the threat. 2.3.3 Recreation and Nature Reserve Class The area surrounding the hot springs and extending to the Pungue River in the south, to the GNP in the north and east and to the Envirotrade camp in the west would be a natural site for a variety of recreational activities and accommodation which could be 29 APPENDIX 1 – Draft Forest Management Policy, 2007 linked to the communities. The area would encompass sections of all vegetation types that would provide added interest for walking trails. A strict control on the hunting of birds and small mammals and the collection of plant material would contribute to the bio-diversity of the area. 2.3.4 Research Class As a sample of all vegetation types are needed for the research input, small (1 ha) plots will be designated in all vegetation classes. These are to be managed in a sensitive manner in accordance with research requirements. 2.3.5 Community Class The allocation of substantial forested areas around the settlements will be made to provide the community with the material to sustain their domestic and structural requirements. 3.0 MANAGEMENT PRACTICES (for increasing carbon sink) Due to the nature of the Envirotrade management philosophy and the lack of resources pertaining to the owners, the N’hambita community, management of the woodland is limited to only the most essential practices to meet the objectives. 3.1 FOREST REHABILITATION BY RE-PLANTING The rehabilitation or reforestation of denuded forest is to be undertaken in blocks or strips in a systematic manner according to a pre-set plan. It must be linked to the sale of carbon thereby giving the incentive to the community to undertake such a venture. By implication, the species selected must be chosen, in addition to other reasons, for their potential for carbon fixing. The sites for reforestation may require a different approach according to the scale of denudation; heavily deforested sites will require pioneer species to be planted prior to more valuable species being re-introduced. Those sites which still maintain some tree growth could be enriched with selected species from the area, adding value to the future forest in terms of lumber, fruits or medicine. In order to jump start the process of getting the young saplings out of the fire danger height (suffrutex height), the advantage of truncheons or cuttings will not be overlooked. These selected species could be pre-struck in the nursery or planted directly into the field, with the probability of a reduced strike rate. Planting spacement will be 5 – 10 meters apart thus delivering a population of 100 – 200 trees per hectare. In the case of having to re-establish a tree stand totally with the use of Pioneer species, one will need to increase the initial population. Fire is part of the ecology of the miombo woodland but in order to maximize the chance for seedlings to establish themselves it is necessary to protect the area from 30 APPENDIX 1 – Draft Forest Management Policy, 2007 fire until such time as the seedlings escape the suffrutex stage which will take up to three seasons from planting. Hence, the requirement for reforestation to be done in orderly blocks so as to manage the fire risk. 3.2 FOREST REHABILITATION BY REGENERATION Farmers will be encouraged to allow selected trees to regenerate in their cropping areas. A. Existing Cultivated Land Although under cultivation, trees have a natural propensity to regenerate on farmland. This can be promoted as follows: • • • • Select and protect 60-100 coppicing trees per ha, evenly spaced at 10-15 m intervals. Thin the shoots from each plant to 1 or 2 dominant stems to promote vertical growth. This avoids development of low bushy growth which makes cultivation difficult, and produces low quantities of wood. Trim the selected saplings as they develop in size. Should there be an insufficient population of trees then re-planting of additional trees will be necessary. B. Regeneration on Fallow or Abandoned Land Many areas are abandoned or fallowed when people move, or when the soils have been depleted. These areas have high potential for tree regeneration. Follow the practices under Existing Cultivated Land above, to encourage wood production, to provide valuable wood for domestic or structural use, and to promote fast tree growth by reducing competition for space, light, nutrients and moisture. • • Use sharp tools to make clean, angled cuts. Thin out coppice shoots to 1-2 main stems. 3.3 CLEARING NEW LAND FOR CULTIVATION Cultivation by most smallholder farmers involves land clearing with excessive felling of trees. With appropriate selection and spacing, a number of native trees could be left on these lands with positive effects on the soils, crops, wood supply, and the general environment. The best trees to retain on agricultural land are fast-growing trees that coppice well and are compatible with crops. These include, but are not limited to species of Acacia, Albizia, Bauhinia, Brachystegia, Combretum, Markhamia, Pericopsis, Pterocarpus, Terminalia and Ziziphus. The density of trees to remain standing in an agricultural system depends mainly on tree size: 31 APPENDIX 1 – Draft Forest Management Policy, 2007 Description Spacing S.P. Ha Large trees Medium trees Small trees 15 – 18 12 – 15 10 40 60 100 3.4 PLAN VIVO SCHEME The Plan Vivo scheme, of giving the farmer a choice of a variety of planting plans for the area of his/her choice, is followed. A farmer is encouraged to buy into the scheme and make his/her own decisions rather than be forced to adopt a scheme which has been forced on them. It is up to the farmer whether they choose one scheme or a combination of several. The schemes include the following planting options; 1 Border Planting 2 Woodlot Planting 3 Homestead Planting 4 Inter-cropping Planting 5 Fruit Orchard with Mangoes 6 Fruit Orchard with Cashew Nuts A farmer is able to add further systems or to expand his system in ensuing years; he is not limited in his options other than to abide by a contract for each new initiative. 3.5 TREE PROPOGATION Envirotrade has committed itself to supplying tree seedlings to the farmers who are contracted to the Plan Vivo scheme. The seedlings should be of the best possible quality from selected trees or, of known genetic origin. The medium term plan is to privatise the nursery into a local business or Club run by community members, however while doing so the quality of seedlings cannot be compromised. A. Indigenous Trees A selection of trees has been made which occur in the area and these trees were selected for a variety of characteristics including; density for carbon storage, domestic usage by the community, value as lumber and growth rate. The seed is collected locally and every effort must be made to identify trees which have superior characteristics in terms of form, fruiting quality and vibrancy. GPS coordinates record the trees thus identified and the fruit is collected annually from them. Obviously as time goes by further selections will be made from selected trees. 32 APPENDIX 1 – Draft Forest Management Policy, 2007 SEED COLLECTION CALENDAR INDIGENOUS Jan Feb Mar Apr May Jun Jul Aug Sep Oct Afzelia quanzensis Albizia versicolor Anacardium occidentale Annona senegalensis Berchemia discolor Brachystegia boehmii Brachystegia spiciformis Cordyla africana Dalbergia melanoxylon Diospyros mespiliformis Faidherbia albida Julbernardia globiflora Lonchocarpus capassa Khaya anthoteca Kigelia africana Millettia stuhlmannii Piliostigma thonningii Pterocarpus angolensis Scelocarya birrea Spirostachys africana Strychnos spinosa Swartzia madagascariensis Tamarindus indica Trichilia emetica Vangueria infausta Ziziphus mucronata In terms of species such as and specifically the Marula (Sclerocarya birrea), it is advantageous to graft a scion from a selected female tree onto a developed root stock. The Marula is dioecious with the result that trees produced from seed may turn out to be male and non fruit producing. The fruit of a Marula can be selected in terms of size and sweetness; therefore, a scion will produce the exact quality of fruit that the parent produces. The added advantage of grafting these trees is the reduced time taken until the production of fruit. Root stock will be produced using the seed collected and grown out in the nursery Trees which can be propagated by vegetative cutting should preferably be done this way and although they will spend longer in the nursery, the tree will have a better chance once planted out. B. Fruit Trees In order to improve the quality and productivity of fruit trees, superior cultivars imported from recognised nurseries and planted into orchards as clone banks, will 33 Nov Dec APPENDIX 1 – Draft Forest Management Policy, 2007 provide an ongoing source of high quality clone material. Root stock using local seed will be grown out to accept the clones thus produced. In the case of avocado pears, phytophera free cultivars will be sought that will improve the survival potential of this highly nutritious fruit. C. Tree Establishment The most critical aspect in the establishment of trees is timing; saplings must be planted out in the field not later than the end of January in order to benefit from the remaining 2-3 months of rainfall. The most common cause of tree mortality is delayed planting and the seedling being unable to cope with the long dry period. All other operations need to be timed to work back from the planting time and this is depicted in this calendar. TREE ESTABLISHMENT CALENDER ACTIVITY RESP Jun Jul Aug Sept Oct Nov Dec Jan Feb Mar Apr NURSERY Train for nursery mng. Envtr Prepare nursery site Club Collect soil Club Source seed & sleeves Envtr Issue & fill pots Club Issue seed, treat & sow Club Maintenance of plants Club FIELD Train in soil preparation Envtr Soil preparation Farmer Train in out-planting Envtr Supply seedlings Club Plant and protect Farmer In filling Farmer Weeding Farmer Fire protection Farmer D. Soil Preparation Proper hand tools, such as a good hoe are all that is needed to prepare the planting position, a pit 50cm in diameter and 30cm in depth should be dug. The soil must be broken up and all roots and rocks removed prior to making a small basin to hold rain water. 34 May APPENDIX 1 – Draft Forest Management Policy, 2007 A tree when planted into a well prepared planting position will respond quickly and the effect is lasting; good soil preparation is vital. Soil preparation must be done well, in good time and cannot delay planting. E. Seedling Planting Plant early; as soon as the first rains have moistened the soil down to 30cm. It will pay off handsomely, with good survival and well grown trees before the dry season. Planting should take place while it is raining or on overcast days following rain, and should be stopped if, for a week, no rain has fallen. The seedlings must be well watered in the nursery prior to dispatch to the field. The most important thing when planting is to minimise disturbance to the ball of soil around the root. Polythene tubes can easily be removed by rolling the tube gently in the hand and then carefully pulling the sleeve over the top of the plant. The sleeve can be re-used. In the case of a pot being used it will most probably be necessary to slit the side to remove the plant. The soil around the roots should be held together by the mass of fine roots which will have developed in a well grown seedling. On no account must the seedling be planted with the polythene sleeve or bag still on. The plant must be positioned in the soil as deep as possible, but not much beyond the level of soil in the pot, this is to ensure that the root plug is in the zone of moisture should there be a dry period. The soil must be pressed firmly around the plant to eliminate all air pockets. F. Seedling Maintenance There is no doubt whatsoever that trees, and even indigenous trees, cannot compete with grass and weeds; inefficient weeding will result in poor growth or even failure. Initially, the minimum cleaning necessary will be to hoe away any growth within a 2 meter diameter of the sapling and to slash to ankle height an area away from the sapling to prevent tall growth from falling over the plant. This residue can be left as mulch around the plant. 3.6. FOREST MAINTENANCE This option is unlikely to be implemented as the cost of moving into these areas is prohibitive. However, if a farmer has a section of woodland that he considers his own, he should do some maintenance for his own benefit. Managing Natural Woodlands • • • In dense stands, selectively thin out scrubby or malformed trees to give space for the development of the dominant and stronger trees or preferred species. Reduce shoots of young regenerating trees to promote vertical growth, leave one or two of the strongest shoots. If a tree is cut, leave at least one the same size within a ten meter radius, under the rule “take one, leave one”. 35 APPENDIX 1 – Draft Forest Management Policy, 2007 3.7. FOREST PROTECTION A. Fire Management Fire management comes in two forms, controlled burns and wild fire. The management of wild fire is dealt with separately in the Fire Protection Plan while controlled burning is planned at Management’s discretion. The miombo woodlands naturally require fire and the balance of the woody species could be seriously affected by the total exclusion of fire. It will be necessary to understand this relationship and to manage the area accordingly. The ecological effect which fire has on the miombo woodland depends on four factors. (The South African Forestry handbook 2000 – Published by the Southern African Institute of Forestry.) • • • • Frequency – how many years between fires (a mean). This will affect the grass species composition, the woody plant population and the fodder availability and quality. Intensity – how hot is the fire, measured in kJ/m/s. A fire with intensity above 3000 kJ/m/s will kill the above ground parts of trees whose canopies are within the flame zone. Season of burn – time of year, dry season, late or early summer. This will influence the impact of fire on various species. Fire type – towards or away from wind, ground or canopy fires. The type of fire will have an impact on the intensity and effect on the vegetation. B. Illegal Wood Cutting The threats of charcoal burners and the cutting of building material by noncommunity members are always there. Intensive extension work by Envirotrade staff together with the community must be exercised to combat this threat. The formation of a Forest Protection Unit under the auspices of the Forest Committee must be implemented and charged to carry out patrols of the vulnerable areas. C. Invader Plant Control Although there is little risk of invasive plants entering the forest, monitoring of this is advisable, particularly along river frontage. Lantana camara is a specific threat to the area. 36 APPENDIX 1 – Draft Forest Management Policy, 2007 3.8. LAW ENFORCEMENT The Community Association is directly responsible for the protection of their forest area; they are to be encouraged to form bye-laws to guard against indiscriminate tree cutting, charcoal burning, uncontrolled hunting and non-sustainable honey gathering. The threat to the forest asset from charcoal burners warrants all the attention possible and this problem has to be tackled in collaboration with the relevant authorities. 3.9. RESEARCH AND MONITORING In collaboration with the Edinburgh Centre for Carbon Management (ECCM), ongoing research is practised as to how the carbon cycle affects the miombo woodland and its ability, by species, to provide a carbon sink. In addition to permanent sample plots (PSPs), measuring growth and foliar change, there is also research activity into the effects of fire on miombo and the trees rooting characteristics amongst other topics. As the N’hambita project is wholly based on the Plan Vivo carbon registry system, there is a requirement that targets for sequestration and poverty alleviation are met. To this end, the University of Edinburgh and the European Union closely monitor and audit the transformation. 4.0. FOREST MANAGEMENT BY CLASS (to maintain carbon sink) The N’hambita Regulado is divided into specific Forest Classes which will be managed according to the requirements of the Class. 4.1. TIMBER UTILISATION CLASS A. Description of Area Physical area (e.g. description of boundaries, size), principal tree species, forest type. B. Dominant Land Use The prescribed area will be managed principally for the sustainable harvest of saw logs, for the supply to and processing by the N’hambita Community saw mill. The collection of other renewable forest products, such as thatching grass, bamboo, mushrooms, fruit, etc. may be practised in a controlled manner. In addition, the sustainable management of bee hives is to be encouraged. 37 APPENDIX 1 – Draft Forest Management Policy, 2007 C. Forest Management Practices Management principals which concur with the requirements of the Department of Forestry will be entrenched. These will entail; • • • • • • The division into felling blocks on a rotation of 30 years. Standard fire protection implementation. Harvesting and extraction practices which subscribe to, and exceed, required environmental standards. Selection of specified species and their minimum size according to prescription. Enrichment planting to replace a minimum of 5 times the number felled. Formalised coppice care of stool. A separate management Plan will be produced for the commercial management of the area entailed. D. Maintenance Requirements An annual maintenance plan and budget will be formalised and carried out to maintain the following; • • • • • All silvicultural requirements such as coppicing, weeding and cleaning. A fire protection system. The development and maintenance of a road system. Noxious weed control. Stream bank management. The maintenance input should be subsidised by sweat equity from the community to develop a feeling of ownership and protection. E. Specific Protection and Monitoring Requirements A rigid control of the Timber Utilisation Class has to be maintained to ensure that the area is not subjected to the pressure of population requirements, which would be to the detriment of the sustainable utilisation of the area. The following needs to be monitored; • • • • • Illegal wood cutting and charcoal making. The destruction of young trees for the collection of bark rope and building material. Illegal hunting and snaring of game. Unlawful inhabitation and cultivation. Bad honey harvesting practices. Again, the community must monitor and police the area to ensure that they will get the maximum benefit from the long term sustainability of the forest. 38 APPENDIX 1 – Draft Forest Management Policy, 2007 4.2. PROTECTION AND CONSERVATION CLASS A. Description of Area List all sites, e.g. graveyards, sacred areas, springs, unusual trees or areas, historic sites, raptor nests. B. Dominant Land Use The areas designated to this class are, by name, protected and conserved while every effort is made to maintain or improve their status in accordance with custom or natural standards. The sites are purely for tribal customary practice, research or conservation due to their unusual characteristics. C. Forest Management Practices No specific management practices are required other than to maintain the surrounding forest in as natural a state as possible. Should the area become degraded, the rehabilitation of the site by enrichment planting will be attempted, care being taken to ensure that the species used conform to the natural surroundings. Should visitors frequent the site, control may be required to ensure that there is no damage through human actions. In the event that harvesting is taking place in the area of the site and more specifically that the site is a raptor nest, care must be taken that no trees are felled in the immediate area and should it be a nest, the tree does not get selected for felling. D. Maintenance Requirements The area will be maintained in accordance with the standard practice for the forest, with the exception of sites that may require a high preservation status from destruction by fire. The standard practice of keeping the area free of noxious weeds and preventing indiscriminate tree cutting will be enforced. E. Specific Protection and Monitoring Requirements Because of the sensitivity of some or all of these sites frequent monitoring may be required to ensure that their status remains intact. The use of Community policing systems should be incorporated into this program. It may be necessary to provide site specific fire protection in certain cases. 39 APPENDIX 1 – Draft Forest Management Policy, 2007 4.3. RECREATION AND NATURE RESERVE CLASS A. Description of Area List all sites, e.g. hot springs, tourist stall, Beau Marie, campsites, lodges, cultural village and river ventures. B. Dominant Land Use These sites are very localised and contained, other than a Nature Reserve should one be designated, they are specifically for the continuing presence of human occupation and these pressures need to be catered for. C. Forest Management Practices Management of the surrounding areas of forest will require specific attention due to the presence of many humans. It may be necessary to assume a buffer zone between the different Management Classes, in some instances. Normal management as specified for the Class of the surrounding area will be instituted. D. Maintenance Requirements Because of the presence of tourists and visitors, adequate facilities, for the site in question, must be available. Sufficient toilets, water points, cooking facilities (braai sites) and rubbish disposal points must not be overlooked. E. Specific Protection and Monitoring Requirements Fire protection will of necessity be critical during the dry season and adequate measures must be in place for this eventuality. Monitoring of the site will be done dependent on the pressure of usage. It must be stressed that because of the public presence, regular monitoring is necessary to maintain a suitable standard. 4.4. RESEARCH CLASS A. Description of Area A number of sites were demarcated as Permanent Sample Plots (PSPs) within the N’hambita forest area. These were selected on a pre determined basis along the roads that traversed the forest area. It was hoped that the sites would encompass all different vegetation types contained within the area. The objective of having PSPs is to characterise and track growth of the woody vegetation in the N’hambita community area. The results are to provide information for land use planning with a view to carbon production, timber utilisation and NTFP development. 40 APPENDIX 1 – Draft Forest Management Policy, 2007 B. Dominant Land Use The requirement for these plots is purely of a research nature and for no other purpose. Additional sites may be included as and when needed for various research requirements, such as fire management programs. C. Forest Management Practices Permanent sample plots have no management requirements, other than the research activities, and as such are left specifically to any event that would occur naturally to the surrounding area. In the case of community woodland, it does include human deprivation. D. Maintenance Requirements No maintenance is necessary and all activities are purely for the use of research. E. Specific Protection and Monitoring Requirements Special arrangements may be formalised to ensure that certain sample plots associated with fire management experiments do not get burnt should they be in unprotected zones. Monitoring arrangements are determined by the research activities and associated personnel. 4.5. COMMUNITY CLASS A. Description of Area Describe area and boundaries. B. Dominant Land Use The objective of the Community Class area is to provide an area of woodland for unlimited access by the community to enable them to utilise wood and other products from the forest. This will include the destructive practice of pole procurement and bark rope gathering, both of which require the community to be trained in sustainable harvesting practices. The other Classes of forest type will of course be available for additional non destructive harvesting practices by the communities. C. Forest Management Practices It would be of enormous benefit if the community were to practice sustainable methods in the gathering of forest products, to this end training in these fundamentals will be necessary. Correct methods of removing bark and of tree felling which will ensure the best opportunity for coppice recovery of the stump are a priority in this area. 41 APPENDIX 1 – Draft Forest Management Policy, 2007 Enrichment planting of popular trees and shrubs used for construction material, domestic products, craft work, food and traditional medicines should be encouraged and possibly insisted upon forcibly. The Committee should be held responsible in this regard. The methods and practice of hunting mammals and birds in order to ensure sustainability should be taught and practised. D. Maintenance Requirements The correct use of fire for fauna and flora protection and forest hygiene will need to be investigated and applied in a planned and controlled system. Coppice maintenance on an annual basis will be of benefit in maximizing growth and future tree quality of valuable timber species. E. Specific Protection and Monitoring Requirements The community need to jealously guard their heritage and ensure that it is safeguarded against illegal and indiscriminate forest users, such as charcoal burners and sellers of poles and other sought after materials in urban areas. This applies to poaching from external sources and from illegal community practices. Wild fires can be the most destructive force and a scientifically based fire management plan which provides a proven system to maintain the forest integrity must be implemented. 5.0. FOREST RESOURCE UTILIZATION The natural resources which are attributable from an area are classified as either consumptive or non-consumptive. Obviously, greater care has to be taken with the consumptive resources to ensure that they remain sustainable. 5.1. CONSUMPTIVE USE A. Wood Trees are the life of the community and used for almost every purpose; housing materials, handcrafts, household items, fuel wood, lumber wood, carving and many more. Appreciation of the finite resource must be hammered home to the community in order that the harvesting of wood is done responsibly and sustainably. • Harvest fuel wood and for many other purposes whenever possible by collecting dead wood or by pruning branches rather than cutting down the tree. 42 APPENDIX 1 – Draft Forest Management Policy, 2007 • • Timber harvesting for saw milling must be done according to the sustainable volume predicted for the area through the implementation of the Forest Department’s mensuration plan. Use wood from large dead trees (standing ring-barked trees or felled trees on the ground) rather than burning or leaving them to rot in situ. Such trees are often not used because they are too large to transport or difficult to cut into smaller pieces. Care should be taken when felling trees in order that the maximum opportunity for coppice re-establishment is available to the stump. B. Honey The harvesting of wild bee colonies for the honey is most often done in a very destructive manner which results in the colony having to relocate and very often having to grow a new queen. This results in a considerable delay in their ability to commence honey production again. Training and participation in working groups to introduce the following must be encouraged; • • Apiary Societies must be formed and all honey production and harvesting, in the community, should be channelled through this body. Traditional destructive means of honey harvesting must be outlawed by the community and sustainable methods utilising the Kenya Top Bar Hive must be practised. Traditional and destructive bark hives must be abandoned. C. Medicines and Edible Plants Collection of these items can very often lead to their disappearance from the area. Some form of control may be necessary to regulate the collection of these resources and the community needs to be made aware of the potential loss to their livelihoods. The following could be implemented in order that the community has some control of their resource. • • • • Form a Society to monitor and control exploitation. Prevent exploitation for commercial gain and from outsiders. Educate populace in sustainable collection methods. Develop systems and gardens to grow the trees and plants for commercial harvest. 5.2. NON-CONSUMPTIVE USE A. Tourism and Recreation The N’hambita community is in a unique situation with their proximity to the potential of sharing in the requirements of the tourists visiting the Gorongosa National Park. The natural assets available and the requirements for accommodation could be exploited in a value adding manner. 43 APPENDIX 1 – Draft Forest Management Policy, 2007 The hot springs, the Pungue River, traditional homes and the forest could be utilised to attract tourists to stay in a variety of situations; camping, traditional accommodation and forest lodge and to partake in canoe rides, hiking, birding and a guided experience to traditional life in a forest environment. • • • • Obtain the use of a Tourism and Marketing Consultant. Form a Tourist Association and develop a plan to meet requirements. Obtain donor funding to subscribe to Community Fund for development. Train management and staff in handling a paying customer. B. Renewable Forest Products The collection of certain regenerating forest products is an ongoing practice and should be encouraged rather than using non-renewable resources, i.e. trees, or the buying in of alternative materials from outside. Items such as bamboo, thatching grass and wild fruit are valuable products and as such need to be utilised and controlled responsibly. Because these products have a high value they are sought after by those who cannot obtain them from elsewhere so this needs to be controlled to prevent illegal collection. These products must be harvested and sold by the community who own them. • • • Forest Management Committee to be enlightened in opportunities. Organize a system of watch dogs to monitor illegal gathering. Arrange for harvesting groups to sell the products with the proceeds benefiting the community. 6.0. COMMUNITY TRAINING & PARTICIPATION As the N’hambita Community Project involves all of the people living in the Regulado, they are ultimately responsible for their forest; therefore, it is imperative that they be given training in numerous fields to strengthen their ability to appreciate and to manage the resources they have. Training in the following areas must be dealt with which should invoke the discipline and responsibility needed; • • • • • • Environmental awareness. Threats to their livelihood, i.e. charcoal burners, poachers, etc. Land planning and use. Forest protection. Tree planting and maintenance. Agro-forestry techniques. 44 APPENDIX 1 – Draft Forest Management Policy, 2007 • • • • Sustainable collection methods of forest products. Health and hygiene. Business development skills. Committee and Financial management The training benefits which are provided by the WWF, in partnership with Envirotrade, deal with several of the ecological and environmental issues and Envirotrade staff handle agricultural and protection matters. Medical issues are dealt with in collaboration with the Carr Foundation and the Department of Health, but this area still requires further attention. External trainers and NGOs which provide training in areas that are lacking are sought after and encouraged to assist. Through the Community Association, the people of N’hambita have to learn to manage their assets in a sustainable way. It is this vehicle which will have to take the message to every inhabitant and ensure that all participate in the total management of their finite resource. The committee are exceedingly important and as such have to be trained and coached in every aspect which involves the community and their assets. 7.0. COMMUNITY EMPOWERMENT In line with the Envirotrade key principle of community upliftment and empowerment, several areas of small business development within the forest environment are encouraged to take hold. The facility of micro business loans is available through Envirotrade or from the Community Association fund to finance approved schemes. A. Tree Nurseries The privatisation of the nurseries providing plants for the Plan Vivo program is a simple and low key business which does not require a major cash outlay. All plants produced, to an order, have a guarantee of purchase and the price fixed in advance. Input from Envirotrade will be necessary to ensure that the species mix and seedling quality is of the required standard and that the plants are ready when required. B. Apiaries The production, processing and marketing of Miombo Honey is an area of untapped opportunity. The commercial production of honey using Kenya Top Bar hives and the purchase of raw honey from other communities has the potential to be rewarding. The major chain store is importing honey for sale. 45 APPENDIX 1 – Draft Forest Management Policy, 2007 C. Craft shop With the development of the Gorongosa National Park, the increased tourist traffic passing through the Regulado means that the potential for a Tourist Banca is good. Local forest products, craft work, wood work and refreshments should have a ready market with passing vehicular traffic. D. Timber supply and Saw mill Having the opportunity to take the processing chain from stump to saw mill and through to a finished product, ensures that the total monetary proceeds remain within the community. Control of the harvesting operation to ensure that the environmental and Forest Plan standards are met is vital for the integrity of the project. The saw mill must endeavour to remain customer driven within the limits of the log supply. E. Carpentry shop The marketing and quality control of the finished products from the carpentry shop are the crux of this business and will require input and guidance for some time. A suitable brand logo to market the furniture under, implying that it is Carbon Neutral and environment friendly, would be desirable and an added marketing advantage. F. Tourism The possibilities of simple tourist lodges with a traditional flavour could become a valuable alternative to the more standard accommodation that the GNP can offer. Added attractions such as a traditional village at work and a forest walk could become sought after by the foreign tourist. The Envirotrade management will continue to be involved in these business ventures to ensure that financial management, marketing and quality control remain intact. REFERENCES: • • • • South African Forestry Handbook 2000 Landcare Practices in Malawi. W.T. Bunderson, Z.D. Jere, I.M. Hayes and H.S.K. Phombeya W.T. Bunderson PhD. Personal Notes Patrick Mushove Preliminary Inventory of N’hambita Community Forest, Gorongosa District, Mozambique May 2004 46 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 An inventory of tree species and carbon stocks for the N’hambita Pilot Project, Sofala Province, Mozambique. John Grace, Casey Ryan, Mat Williams with assistance from Silvia Flaherty, Sarah Carter, Joanne Pennie, and contributions from Evelina Sambane, Roberto Zolho, Joao Fernando, William Garrett, Luke Spadavecchia, and help in the field from staff of Envirotrade (Piet van Zyl and Antonio Serra). 47 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 Contents 1. Introduction ........................................................................................................................................ 49 2. Aims ................................................................................................................................................... 50 3. Methods.............................................................................................................................................. 50 3.1 Study area..................................................................................................................................... 50 3.2 Land Use ...................................................................................................................................... 50 3.3 Soils.............................................................................................................................................. 51 3.4 Methods for evaluating carbon stocks.......................................................................................... 51 3.4.1 Permanent Sample Plots (PSPs).................................................................................................... 52 3.4.2 Allometric relationships ................................................................................................................ 52 3.4.3 Density and carbon content ........................................................................................................... 53 3.4.4 Earth Observation.......................................................................................................................... 53 4 Results ................................................................................................................................................. 53 4.1 Sample Plots................................................................................................................................. 53 4.1.1 Species........................................................................................................................................... 53 4.1.2 Variation between the PSPs .......................................................................................................... 61 4.2 Allometric relationships ............................................................................................................... 62 4.3 Density and carbon contents ........................................................................................................ 63 4.4 Carbon stocks ............................................................................................................................... 63 4.5 Land cover, forest area and rates of change 1999-2007............................................................... 64 5. Summary and recommendations ........................................................................................................ 66 References .............................................................................................................................................. 66 Appendix 1: List of species in the project area ...................................................................................... 68 Appendix 2: Ryan et al. (2007) .............................................................................................................. 72 Appendix 3. Williams et al. (2007a) .......................................................Error! Bookmark not defined. Appendix 4. Tipper and Garrett (2007) ………………………………………………………………..90 48 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 investigation was to obtain an initial impression of the vegetation and its floristic composition. On the basis of a survey of 30 quarter-hectare plots aligned along 15 linear transects at right angles to various roads, he suggested that the N’hambita community forest consists of four main types of vegetation. The miombo type is dominated by species such as Brachystegia boehmii, B. spiciformis, Julbernardia globiflora, Diplorhynchus condylocarpon, Erythrophleum africanum and Burkea africana. Collectively they accounted for over 70% of the basal area in the miombo woodlands. In the second type, Combretum woodland, he found that Combretum apiculatum (29% of basal area), Commiphoa mossambicensis (15%) and P. rotundifolius (15%) dominate the tree layer while C. apiculatum (51%) and P. rotundifolius (36%) dominate the shrub layer. He identified also subtype Combretum/Palm woodland, with a tree layer dominated by C. apiculatum (27%), Philenoptera violacea (25%) and Hyphaene coriacea (15%). Finally, he recognised a third type; a distinct riverine woodland dominated by Adansonia digitata (26% of basal area) Cleistochlamys kirkii (10%), A. nigrescens (8%) and Xeroderris stuhlmannii (6%). C. apiculatum (50%) and Combretum molle (24%) dominate the shrub layer of the riverine woodland. 1. Introduction This inventory is a key deliverable for CONTRACT No. B7-6200/2002/063-241/MZ, Miombo Community Land Use and Carbon Management: N’hambita Pilot Project. The pilot project aims to develop systems for sustainable land use and rural development in the buffer zone around the Gorongosa National Park in Sofala Province, Mozambique, working with the N’hambita community to assess the potential of these activities to generate verifiable carbon emission reductions. It is clear that most of the woodland is miombo. Miombo is a deciduous woodland mosaic of about 3 million km2 covering more or less continuously much of the Central African plateau (Tanzania and the Democratic Republic of the Congo, Zambia, Malawi and eastern Angola, Zimbabwe and Mozambique). It is home to many charismatic animal species as well as 35 million people who live by subsistence farming, utilise slash-andburn agriculture, and use the woodlands extensively for fuel wood and other forest products. A carbon-oriented woodland inventory of the N’hambita is fundamental to the aims of the project, being required as a baseline against which future trends in carbon storage can be measured. Given the expected lifetime of the project as a whole (decades) and the need to monitor regional deforestation rates, this inventory needs to be related to satellite remote sensing so that the areas under management may be assessed rapidly and economically by satellite surveillance. Such an inventory is also needed as a basis for management of sustainable use of the woodlands, especially in relation to the ‘avoided deforestation’ to be achieved through firemanagement, and in relation to other human activities such as harvesting for fire-wood, charcoal and constructional timber. 2. An initial survey of land cover based on LANDSAT imagery (Spadavecchia et al. 2004) and the related establishment of Permanent Sample Plots PSPs. Eight types of land cover were discerned and later identified by ground survey: miombo forest, moist forest, closed canopy forest, palm woodland/herbaceous, sands/riverine, bare earth/defoliated, water and seasonally waterlogged Combretum. This LANDSAT survey confirmed that the location of the Mushove (2003) plots was unbiased. Then, it was used as the basis for setting up 15 onehectare permanent sample plots (PSPs) which we have established in randomly stratified locations within the main land cover types. Individual trees have been tagged with metal labels which can withstand fire. These plots were enumerated in 2004, 2005, 2006 and again in 2007. Here, we report data from 2006. The project has achieved three main data sets as follows: 1. The commissioned ‘Preliminary Inventory’ report of Mushove (2003). The purpose of this 49 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 apparent, we decided to henceforth use SPOT and MODIS imagery. Both cover the period 1999-2007. 3. The Masters thesis of Sambane (2005), now published as Williams et al. (2007a). This work investigated 28 sites of 0.125 ha which had previously been under shifting cultivation as machambas, and a further 14 relatively undisturbed woodland sites of 0.25 ha each from Mushove (2003). The purpose of this was to study the accumulation of carbon and species on abandoned machambas. It was shown that the abandoned sites accumulate carbon at a rate of 0.70 tonnes/ha/annum with an estimated standard deviation of 0.19. In the first 20 years of abandonment, the number of woody individuals per ha increased from less than 100 to 1200 and then declined to a value approaching that of the relatively undisturbed machambas. 3. Methods 3.1 Study area The area is adjacent to the south-west boundary of Gorongosa National Park between coordinates 18º 49’ 30”- 19º 04’ 00” South and 34º 02’ 00” 34º 17’ 30” East, approximately 60 km west of Vila Gorongosa in the Sofala Province, Mozambique, and can be accessed by the national road EN-1, and East-West by the rural road ER-418 that serves as the access to the west gate of the Gorongosa National Park. The Pungue and Vunduzi Rivers are the southern and western boundaries respectively (Fig. 1). In addition, further half-hectare plots were surveyed by project staff in late 2007, aiming to provide detailed information over a specific area identified for conservation. 2. Aims The aims of this carbon-inventory report are to provide (i) an assessment of the woody vegetation in the N’hambita area, including the species and their importance in terms of basal area, (ii) a characterisation of the extent of woodland from the available satellite remote sensing and (iii) estimates of carbon stocks and rates of change of these stocks. Two major difficulties were encountered, and tackling these difficulties has delayed the report. The first relates to the allometric relationships, needed in the estimation of carbon content from the diameter or basal area of trees. Although several such allometric relationships have been reported in the literature for similar vegetation elsewhere (Abbot et al., 1997; Chidumayo, 1997; Frost, 1996) the published relationships vary substantially and so we decided to build our own allometric relationships for the N’hambita area by destructively harvesting a sample of trees. The second related to the unexpected malfunction of the LANDSAT 7 sensor in May 2003 which has never been repaired. The LANDSAT imagery was chosen at the start of the project because it was readily available, easily interpreted, goes back to the 1970s, has a suitably high spatial resolution, and it is cheap to obtain. When the magnitude of the image degradation became Fig. 1. Map of the study area showing the area for immediate management (A, corresponding to N’hambita, Bue Maria and Posta Da Pungwe), and areas for future use, B (containing Pavua and M’Bulawa), C (a buffer zone, only sparsely inhabited), and south of the river Pungue, D (Mucombeze) 3.2 Land Use The N’hambita community lands, which cover an area of 348 km2, can be divided into three major types: protected area, buffer zone and the community land. The protected area is under State ownership; it constitutes the land contained within the boundaries of the Gorongosa National Park. It is currently managed by the Carr 50 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 new machambas are established, the inhabited areas are encroaching fast on the woodland. Foundation for the protection of flora and fauna and for the enjoyment of the public. The buffer zone is the land immediately adjacent to the boundaries and surrounds the entire National Park area. It is co-managed by various stakeholders including government institutions, non-governmental organisations and community associations, and to some extent with the involvement of the private sector. The community land in or outside the buffer zone is the land legalised under the Land Act Number 19/97 and must be managed for the benefit of all members of the community organised in association. Any resource extraction must be subject to management plans approved by the Government. There are settlements, and slashand-burn agriculture, but land clearance is restricted legally. In the west, the community lands are outside the buffer zone, population density is greater, and land clearance is rapidly ongoing. Within the N’hambita community the main land use activities are: 3.3 Soils The combination of the crystalline nature of many of the rocks, low relief, moist climate and high temperatures has produced a highly weathered soil which is often more than 3 m deep on the plateau. Shallow stony soils also occur along escarpments. Loamy sand, sand loam and sand clay loam textures predominate, with a marked increase in clay with depth. The miombo ecosystems generally occur in soils which are predominantly alfisols, oxisols and ultisols; these are highly acidic, low in cation exchange capacity, low total exchangeable bases and low available phosphorus. These soils are formed by a catenary sequence of well drained, deeply weathered soils on higher areas, a narrow zone of sandy soils along the footslopes and poorly drained vertisols in the wet areas i.e. the ‘dambos’ (Desanker et al. 1995). Generally they have low levels of organic matter as a consequence of the abundant termite activities and frequent incidence of fire (Chidumayo, 1997). ● subsistence agriculture (practised by almost all families), ● charcoal production and firewood collection, ● livestock rearing, ● fishing and hunting, ● small scale commerce. 3.4 Methods for evaluating carbon stocks Agriculture is the main land use (but a minor land cover) and constitutes the most important source of subsistence for all families in the community. Charcoal and firewood fuel production are practiced by some members of the community and also by outsiders who arrive in the area, establish kilns and sell bagged charcoal at the roadside (Herd 2007). This activity has increased since land mines were dealt with, and all roadside woodland (0-3 km) is vulnerable. There is evidence from Maputo that charcoal is now being consumed by town-dwellers at rates of 0.9-1.0 m3 yr-1 (Brouwer & Falcão 2004). Large amounts of charcoal and firewood are sold along the main road, EN-1, and in some cases this constitutes an important source of family revenue. Human activities are central to the current dynamics of miombo ecosystems, because many areas are burned on an annual basis as part of the traditional ‘slash and burn’ system of land use. As populations increase, and We evaluate carbon stocks in a four-stage process • • 51 Focus on obtaining inventories of biomass and soil carbon at sample plots. The sample plots cover a representative sample of forest types (as that is where most carbon is found), with some attention to agricultural systems and the re-growth which occurs on abandonment. Develop allometric relationships to enable biomass to be calculated from these inventories (and any detailed inventories that might be made in the future). The relationships are obtained by destructive sampling and weighing of a representative sample of individual trees. APPENDIX 2 - Woodland Inventory, N’hambita, 2007 • • Mountain. Most of the rain falls between November and March, and July to September are the driest months. Rainfall data from ARACentro (The Mozambican water board), collected at the Gorongosa National Park headquarters at Chitengo show a mean annual rainfall of 749 mm for the last seven years, with high inter-annual variability. The average precipitation for 1957-69 was 879 mm. N’hambita is ~100 m higher than Chitengo on the orographic rainfall gradient created by Mount Gorongosa, so rainfall is likely to be slightly higher. Biomass must be converted to carbon content, so the carbon percentage of biomass must be known. Earth observation, to enable the carbon content of thousands of hectares to be estimated from the sample plot data. Table 1. Location of 15 1-hectare permanent sample plots and elevation above sea level (metres).The locus given is the corner nearest to the track. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Latitude S 18˚ 51.846΄ 18˚ 54.050΄ 18˚ 56.536΄ 18˚ 58.556΄ 18˚ 56.133΄ 18˚ 57.815΄ 18˚ 59.738΄ 19˚ 00.376΄ 19˚ 00.347΄ 18˚ 52.790΄ 18˚ 54.814΄ 18˚ 54.840΄ 19˚ 00.937΄ 19˚ 03.216΄ 19˚ 00.937΄ Longitude E 34˚ 05.986΄ 34˚ 07.236΄ 34˚ 08.081΄ 34˚ 06.145΄ 34˚ 08.512΄ 34˚ 10.067΄ 34˚ 12.037΄ 34˚ 14.163΄ 34˚ 16.877΄ 34˚ 05.578΄ 34˚ 05.010΄ 34˚ 06.471΄ 34˚ 13.181΄ 34˚ 13.510΄ 34˚ 13.181΄ m asl 312 206 157 132 181 111 92 76 36 299 268 258 93 53 93 3.4.1 Permanent Sample Plots (PSPs) Fifteen PSPs of one-hectare each were established in randomly stratified locations within the main land cover types that had been identified in the initial survey of land cover based on LANDSAT imagery (Spadavecchia et al. 2004). These PSPs are geo-referenced, their co-ordinates are given in Table 1, and their positions are plotted on Fig. 1. All woody plants with a stem diameter ≥0.05 m were recorded and mapped, and their diameters at 1.3 m above the ground (‘diameter at breast height’, dbh) were measured in 2004, 2005, 2006, 2007. By 2007, all trees were tagged with a metal label bearing a serial number which can withstand fire (when a tree is completely destroyed by fire, the tag can be recovered from the ash and recorded). By early 2007, some individual trees were additionally instrumented with metal girth bands and vernier gauges for precision measurement of growth, although by late 2007 it became evident that the gauges do not always withstand fire. Geologically, the land consists of eroded surfaces of granite and basaltic gneiss complex of Precambrian times, which is heavily weathered, yielding sandy soils that are generally unsuitable for any form of intensive agriculture (Tinley, 1971, 1977). The landform is undulating to incised, with elevations changing from about 40 m on the flank of the rift valley rising towards the west, where elevations are 400 m and more. The drainage is dense and closely spaced and assumes a typically dendritic pattern, oriented to the West, South and East. The smaller streams are seasonal and fast running and the Pungue and Vunduzi are the only perennial rivers. The groundwater levels are generally very shallow and located either in the weathered regolith in valley bottoms or in fractures in the bedrock (Lynam et al., 2003). Species identification follows Coates Palgrave et al. (2003), and vernacular-to-scientific translation of names was checked using De Koning (1993). Notes on the utilisation of species were gleaned from Coates Palgrave et al. (2003), from Sambane (2005) and from trusted internet sources. The climate is typical of the central region of Mozambique, sub-tropical with alternating cooldry winters (April-October) and hot-wet summers (November-March). May-July is the coolest period and October is the hottest month. The mean rainfall of the region is between 600 and 800 mm/yr and is generally influenced by the orographic effect of the Gorongosa 3.4.2 Allometric relationships Samples of 26 trees of seven common species were measured and destructively harvested. The root systems were excavated using an available JCB vehicle which was in use in the Gorongosa Park. Biomass was dried to constant weight by exposure of samples for several weeks in a home-made solar oven. Possible allometric 52 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 Difference Vegetation Index (NDVI) was generated for each image and a five-layer image consisting of bands 1, 2, 3, 4 and NDVI was generated. All these bands were used for the classification. relationships (dependent variable, diameter; independent variable, dry biomass) were explored using Matlab (http://www.mathworks.com/products/matlab/). 3.4.3 Density and carbon content Densities of major species was determined by measurement of dry mass and fresh volume of a small sample. Further information on density was obtained from Bunster (2006). Carbon content was determined by multiplying mass by 0.47. The variation in density is not sufficiently important to require incorporating into the predictive equation for calculating carbon content from diameter. NDVI = (NIR − RED ) (NIR − RED ) For SPOT imagery, we define NDVI as: NDVI = (B3 − B 2 ) (B3 − B 2 ) where B2 is reflectance in the red band and B3 is in the near infra-red. All the SPOT images were processed; however, due to inter-annual variation in the timing of seasons only the 1999 and 2007 images were used for classification and land cover change detection. Cloud masking was not necessary for these two images. 3.4.4 Earth Observation Mechanical failure of the scan line corrector on the LANDSAT satellite in May 2003 has never been repaired and so we investigated other available imagery, including MODIS and SPOT. For the present application we decided to use SPOT, as it can be obtained with the required spatial resolution (20 m). SPOT imagery was obtained as SPOT 4, Level 2A. Several images were acquired but due to seasonal differences and the presence of cloud cover in some cases, only the March-1999 and April-2007 images were selected for study of the land cover and its deforestation rate. The pixel size is 20 m and the spectral bands are: green (B1, 0.50 - 0.59 µm), red (B2, 0.61 - 0.68 µm), near infra-red (B3, 0.78 - 0.89 µm), and mid infra-red (B4, 1.58 - 1.75 µm). All image processing was performed using ERDAS Imagine 8.7. Supervised classification requires a priori knowledge (also known as training data) about the area to be classified. This information is normally derived from, amongst other sources, field work, aerial photographs, maps, or reports. In this case, we relied on geo-referenced field observations made by our own researchers, including the Permanent Sample Plots. The result of ‘supervised training’ is a set of signatures where each signature belongs to one class. Pixels were finally classified as ‘forest’ or ‘nonforest’ and maps with latitude/longitude coordinates were made for use in the field, and as the basis for locating geo-referenced project activities. Absolute areas of forest and non-forest were obtained by counting the 20 m x 20 m pixels. SPOT 4 LEVEL 2A images were radiometrically corrected to remove distortions due to differences in sensitivity of the elementary detectors of the viewing instrument. They are also geometrically corrected and referenced to cartographic projection (UTM WGS84) not tied to ground control points. Further geometric corrections were performed by referencing the images to the 2000 Landsat 7 ETM+ image used previously in this project by Wallentin (2006). 4 Results 4.1 Sample Plots In this section we report the main findings from our sample plots. Raw data were converted to radiances first and then into top-of-atmosphere reflectances (TOA). The atmospheric correction was performed by subtracting the minimum reflectance value (histogram minimum method). A Normalised 4.1.1 Species The 21 most significant species in the sample plots collectively account for over three quarters 53 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 the 15 PSPs in the present study is given as Table 2; the only species from FAO (2007a) which was not found was Friesodielsia obovata which may be out of its natural range here. of the total basal area (Fig. 2). Half of the total basal area is accounted for by only seven species as follows. Brachystegia boehmii is a defining species of miombo woodland. It has dense wood which blunts tools, and is used for firewood, charcoal and, in medicine, for constipation and lumbago. Combretum adenogonium is a small to medium-sized tree, frequent in dry areas over much of southern Africa, and sometimes used as an antiseptic (Maregesi et al. 2006). Diplorhynchus condylocarpon is a fire-resistant shrub or small, deciduous, multi-stemmed tree up to 8 m high, containing latex which can be used as glue; its wood is good for carving. Pterocarpus rotundfolius is another important miombo species, much-visited by bees, and highly fire resistant (Zolho 2005). Julbernadia globiflora is another signature species of miombo woodland; it is a well branched tall tree up to 18 m, and also has dense wood. It has the reputation of being the best bee tree in Africa. Burkea africana is a widespread tree in southern Africa. The heartwood is extremely durable and the bark contains high concentrations of antioxidants. The tree has distinct annual growth rings which have enabled researchers to study the impact of El Niño-related changes in rainfall on its growth (Fichtler et al. 2004). Sclerocarya birrea is also a large deciduous tree and a miombo species, used to make Marula oil for skin care and the liquer Amarula. In Israel, this tree has been domesticated for its fruits. The fruits have 8 times the concentration of vitamin C as that found in oranges, and when fermented they have a high alcohol content. Altogether 164 woody species were found, of which 124 have been identified to species level, eight to genus only and 32 minor species remain unidentified (Table 2). The species may be compared with a recently-compiled list of tree species in Mozambican miombo woodlands (FAO 2007a). This FAO study found considerable regional variation but the most common in Sofala province were Annona senegalensis, Bauhinia thoningii (synonym Piliostigma thoningii), Brachystegia spp., Crossopterix febrifuga, Diplorhynchus condylocarpon, Friesodielsia obovata, Julbernardia globiflora, Millettia stuhlmannii, Pseudolachnostylis maprouneifolia, Sterculia Africana and Terminalia sericea. The list from 54 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 Table 2. Woody species found as in Palgrave (2002). †Species not found in Palgrave (2002). Scientific Name Combretum adenogonium Brachystegia boehmii Diplorhynchus condylocarpon Local Name(s) Fiti, Fit Mefuti Nham'tomole, Mutoa, Mutowa Family Combretaceae Caesalpinioideae Wood density 0.55 0.52 Contribution (%Biomass) 12.20% 8.90% Apocynaceae 0.60 8.70% Erythrophleum africanum Messimbe Caesalpinioideae 0.63 6.40% Sclerocarya birrea Marula, Mfula Anacardiaceae 0.48 5.30% Julbernardia globiflora Muhimbe Caesalpinioideae 0.63 5.00% Brachystegia spiciformis Messassa Caesalpinioideae 0.63 4.80% Pterocarpus rotundfolius Miómba, Miombué Papilionoideae 0.65 4.00% Burkea africana Pseudolachnostylis maprouneifolia Mucarate, Mucarati Caesalpinioideae 0.64 3.20% Mundoto, Mussonzoa Euphorbiaceae Philenoptera violacea Millettia stuhlmannii Mpakhasi, Pacassa Panga Panga Mbila,Mulombe,Mukwa, Mulombwa,Muconambira Papilionoideae Papilionoideae Mulonde Murumangama, Murumanyama Muungussi, Munguza Munhangolo Papilionoideae Dzvototo Burseraceae Fabaceae, Mimosoideae Rhamnaceae Papilionoideae Pterocarpus angolensis Xeroderris stuhlmannii Cassia abbreviata Bombax rhodognaphalon Cleistochlamys kirkii Commiphora mossambicensis Acacia nigrescens Berchemia discolor Dalbergia boehmii Guhu, Ngungo Chiglamademo Chimanda,Ximanda 2.80% 0.56 0.65 Papilionoideae Caesalpinioideae Bombacaceae Annonaceae 2.40% 2.20% 2.20% Uses and Products Good general purpose timber if treated against borer attack, Browse (game and livestock), fence posts, live fences, latex from roots Hard heavy and tough wood, construction, bark extract tanning. Shade, fruit, timber: Light, soft timber suitable for crafts and furniture Bees, shade, low quality wood but durable so suitable for general purpose Shade, fodder, ropes, apiculture, fuel wood, charcoal, inferior timber Timber easily worked but not very durable - general use, bee fodder Heavy, durable and hard timber - furniture and general purpose. Charcoal Browse/fodder (fruits), bee fodder, smooth timber handicrafts, fuel wood Browse (game and stock), hard and heavy timber, traditional medicine Shade, bee fodder, fruit, hard timber of excellent quality High value, medicinal, apiculture, fodder, live fence, tannins, nitrogen fixing Dye and tannins (sap), browse (seed pods - game and stock), 0.52 1.70% Traditional medicines, small timber volume - carving 0.71 1.40% 1.20% 1.20% 0.47 1.20% 0.46 1.10% 1.10% 1.10% Wood for household utensils Strong, dark, close grained heartwood is good for carpentry, bees, medicinal Timber has attractive grain so suitable for furniture 0.64 55 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 Scientific Name Local Name(s) Faidherbia albida Terminalia stenostachya Mussadze Mucoudò Albizia lebbeck Bauhinia petersiana Lecaniodiscus fraxinifolius Markhamia zanzibarica Combretum sp. Lannea schimperi Tanga Tanga Munhando Family Fabaceae; Mimosoideae Combretaceae Fabaceae; Mimosoideae Caesalpinoideae Wood Density Contribution (%Biomass) 1.00% 1.00% 0.63 0.90% 0.90% Sapindaceae Bignoniaceae Combretaceae Anacardiaceae 0.80% 0.80% 0.70% 0.70% Dalbergia melanoxylon Mutarara Nhamukabari, Nhacabare Nhantsangu, M'tshanga Mumbumbu Paupreto, Mpingue, Mphingue Papilionoideae 0.60% Grewia monticola Tadza, Thadza Tiliaceae 0.60% Combretum hereroense Munangare, Muchenalore Combretaceae 0.50% Strychnos innocua Muteme Strychnaceae 0.50% Cordyla africana Mutondo Papilionoideae 0.40% Crossopteryx febrifuga Entandrophragma caudatum Muchombêgo Rubiaceae Mulolo Mussequessa, Mussequesse Muchane, Muchachane meliaceae Piliostigma thonningii Tabernaemontana elegans Caesalpinioideae Apocynaceae 0.51 Browse (commonly by game) Valuable dark timber suitable for carving and musical instruments Attractive timber, walking sticks and tool handles, food (fruits) Fodder and browse for cattle, timber - tool handles, drink (fruits) Fuel wood, fodder, food (fruit). Timber – low quality and medicine Food (fruits have high vitamin C), heartwood for building and drums Hard, durable timber suitable for woodwork, carving and building. Dark wood - furniture, boat building, tannins and dyes (bark) Fruit, bees, fodder, fibre, general timber, intercropping, erosion control 0.40% 0.40% 0.53 Uses and Products Fodder (fruits for cattle), hard and durable timber - poles and tool handles Strong timber - furniture, cupboards and general use Shade, nitrogen fixing, fuel wood, timber - poles, fodder (livestock), fuel wood Drink, timber - small items of furniture/carvings, medicinal 0.40% 0.40% Azanza garckeana Strychnos madagascariensis Mutohue, Mutogue Malvaceae 0.30% Wood is of little value, implement handles, handicrafts. Bark fibre Mutunduru 0.30% Food (fruit - not very pleasant taste though) Acacia karroo Nsangarassa, Sangarassa Strychnaceae Fabaceae, Mimosoideae 0.20% Combretum molle Fithidondo Combretaceae 0.20% Diospyros mespiliformis Mufuma Ebenaceae 0.20% Fodder, fuel wood, bee fodder, gum and tannins from bark Termite resistant timber is suitable for implement handles and fence posts Good timber, canoes, pestles, furniture and flooring. Food (fruits) 56 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 Scientific Name Hyphaene coriacea Rhus chirindensis Rhus sp. Local Name(s) Mucheu Muchanfu Chinamazize Chicumbite, Mussiquiribanda, Mussiquiri Nhacassadzi, Nomba, Nhomba Family Arecaceae Anacardiaceae Anacardiaceae Wood density Contribution (%Biomass) 0.20% 0.20% 0.20% Munjonjota Munhenza Sapomdaceae Fabaceae, Mimosoideae Fabaceae, Mimosoideae Fabaceae; Mimosoideae Capparaceae Chibaramadona Ochnaceae Mussengue Cuacuacho Suruza, Tsurudzo Cumbanzóo Umbaua Mussunganhemba Araliaceae Boraginaceae Heteropyxidaeceae Apocynaceae Mekuaceae Combretaceae Polygalaceae Combretaceae 0.10% 0.10% Ziziphus abyssinica Mupupo, Mpumpu Mussussu Mussau-sanga, Mussautanga, Mussaotsanga 0.10% Acacia robusta Mussodze Acacia sieberiana Gunga Acacia sp. Adansonia digitata Mulumanhama Mulamba, Mbuyu, Mulambe Rhamnaceae Fabaceae, Mimosoideae Fabaceae, Mimosoideae Fabaceae, Mimosoideae Albizia amara Mussorora, Mussorola Zanha africana Acacia nilotica Acacia polyacantha Amblygonocarpus andongensis Boscia salicifolia Brackenridgea zanguebarica Cussonia arborea Ehretia amoena Heteropyxis dehniae Holarrhena pubescens Khaya anthotheca Pteleopsis myrtifolia Securidaca longipedunculata Terminalia sericea Guoe, M'guoe Bombacaceae Fabaceae; Mimosoideae 0.20% 0.45 Strong, attractive timber. 0.10% Timber - rough work e.g. Planks, construction Fuel wood, timber suitable for fence posts, live fences, gum, medicinal Traditional medicine, featureless timber - construction, tannins (bark) 0.10% 0.10% Light coloured timber - furniture and shelving, food (roots) 0.10% 0.63 Uses and Products 0.10% Low value timber, but can be used for traditional musical instruments Strong timber – fence posts, pestles Smooth timber - small carpentry items. Soft timber. Hard timber but easily worked for furniture, dugout canoes. 0.10% 0.10% 0.10% 0.10% 0.10% 0.10% Soap (bark), fibre (bark), low value timber, bee fodder Hard timber - furniture and general purpose Traditional medicines Soft timber, fuel wood, charcoal, apiculture, fodder, shelter Bark fibre (rope and floor mats), Fruit (food and drink) Soap, decorative timber, furniture, fuel wood, charcoal, shade 0.58 57 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 Wood density Scientific Name Local Name(s) Albizia brevifolia Chiteta, Chithetha Albizia harveyi Chissio Family Fabaceae; Mimosoideae Fabaceae; Mimosoideae Anacardium occidentale Cashew Anacardiaceae 0.52 Annona senegalensis Antidesma venosum Calliandra sp. Carica papaya† Miloro, Mulembe Mundangaramunhu Calliandra Papaeira Annonaceae Euphorbiaceae Caesalpinioideae Caricaceae 0.48 0.13 Combretum apiculatum Corchorus junodii† Cussonia spicata M'fithi Chimbiati, Chimbia Muchendje Dichrostachys cinerea Diospyros kirkii Diospyros usambarensis Diospyros sp. Dombeya shupangae Mupangara Mucula Nhamudima Mulala Thoa Entada abyssinica Erythrina sp. Erythroxylum emarginatum Excoecaria bussei Ficus sycomorus Chidzedze Thithi Combretaceae Tiliaceae Araliaceae Fabaceae; Mimosoideae Ebenaceae Ebenaceae Ebenaceae Sterculiaceae Fabaceae; Mimosoideae Papilionoideae Flacourtia indica Gardenia ternifolia Gliricidia sp.† Grewia sulcata Grewia transzambesica Mucosse, Mukossa Mucombacore Mushamvu Mutumbotumbo, M'tema, Mitema Munomoro Gliricidia Sene Thetha, Theja Flacourtiaceae Rubiaceae Fabaceae Tiliaceae Tiliaceae Jatropha curcas Karomia tettensis Kigelia africana Jatropha Nhacabale Mevungute Euphorbiaceae Lamiaceae Bignoniaceae Contribution (%Biomass) Uses and Products Cashews, fruit, hard timber, furniture, building, firewood, charcoal Food (fruit), timber - construction, traditional medicines, dye (bark) Food (fruit), latex/rubber, medicines Timber, large pieces are difficult to obtain so suitable for fence posts. Food (roots), fodder for livestock, coarse timber. Fodder, hard timber - hard, durable, fence posts, tool handles Highly esteemed fruits Hard timber, dyes from timber. Weak and hard to work timber, bee fodder, decoration (seeds) 0.51 Erythroxylaceae Euphorbiaceae Moraceae Close-grained, durable heartwood - attractive furniture Food (fruits are valued) Medium weight timber - small carpentry items, bees, food, medicinal Fine grained timber 0.61 Drink (fruits) Biodiesel, live-fence, intercropping, soil improver (green manure) 0.28 Famine food, fodder, apiculture, general use timber 58 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 Wood density Scientific Name Local Name(s) Family Mangifera indica Markhamia obtusifolia Millettia mossambicensis Monodora stenopetala Monotes engleri Ozoroa obovata Ozoroa sp. Peltophorum africanum Pericopsis angolensis Premna senensis Pterocarpus lucens Rhynchosia pseudoviscosa† Scolopia stolzii Senna petersiana Sterculia africana Sterculia appendiculata Mangeira Phewa, Feba Mussungarossa Muchinge Xithethe, M'tongolo Chifuca, Chifupa Cantassaro Muzeze Xouanga Mhacanucanunga Muanga Anacardiaceae Bignoniaceae Papilionoideae Annonaceae Dipterocarpaceae Anacardiaceae Anacardiaceae Caesalpinioideae Papilionoideae Lamiaceae Papilionoideae Nhamperepera Comacamba Buembacor Munjale, Mucuna Bicancula Cherecheti, Goza, Tcherechete Papilionoideae Flacourtiaceae Caesalpinioideae Sterculiaceae Sterculiaceae 0.54 Sterculiaceae 0.35 Potanzoe Panda Muthupa Bignoniaceae Strychnaceae Strychnaceae Papilionoideae Combretaceae Combretaceae Trichilia emetica Vangueria infausta Vitex doniana Vitex sp. Voacanga thouarsii Ximenia americana Ximenia caffra Pao Ferro Mucodomue Mundangamunho M'sequeira, Massaniqueira Munziro, Mufula Mucuvu Mucuno, Mucuna Nhaponda Mudogodogo M'tenguene, Mutenguene Meliacea Rubiaceae Lamiaceae Lamiaceae Apocynaceae Olacaceae Olacaceae Ziziphus mucronata Muchecheni Rhamnaceae Sterculia quinqueloba Stereospermum kunthianum Strychnos henningsii Strychnos potatorum Swartzia madagascariensis Terminalia brachystemma Terminalia sambesiaca Contribution (%Biomass) 0.40 Uses and Products Fruit, fodder, fuel wood, charcoal, timber, shade, intercropping when young Fuel wood, fibre (bark rope) 0.47 Shade, homestead planting, carving, furniture, bees Fine grained timber - flooring blocks, panelling, fence posts Fuel wood, browse (commonly game) Hard timber does not split - tool handles. Gum, soft timber, bark fibre - mats and rope Timber light but strong - general construction and furniture, gum Shade Durable timber - fence posts and tool handles Close grained wood, carving, fodder, bees, medicinal, poison Timber - tool handles Easy to work timber - household utensils, shelving, dugout canoes. Food (fruits) Food (fruits) Latex Oils, drink, timber - small craftwork only Oils (sap - leather softener, cosmetic use), drink (fruits) Fodder, timber non-durable - rough/temporary fencing/planks 0.65 59 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 Scientific Name Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Local Name(s) Balamadona Cataussarro Chiriacamba Guacha, Guaquacho, Guakuacho Mbubunhanga Muchambu Muchanvo Muchinguena Mucongontua Mucuiramhondoro Mugaramanjiva Mukhodone Mukunku, Mukungo, Mucuncu Muleme Mulonde Mumudendo Munganzo Munhomba Munhongoro, Munhungoro Munomolo Murara Muthumdolo Mutindi Mutunguricua Nhacafupa Nhampulurue Nhancamba Nhanganzo Ruronde Senge Tubo Tsanga Family Wood density Contribution (%Biomass) 0.28 0.73 0.61 0.65 60 Uses and Products APPENDIX 2 - Woodland Inventory, N’hambita, 2007 Fig. 2. Basal area of the 21 most significant woody species in the N’hambita plots Strychnos. cocculoides and S. madagascariensis producing edible fruit and the seed germinates easily, trees are fast growing, and there are success stories in Zambia; Vangueria infausta, a good fruit tree/shrub. Of the tree species present at N’hambita, our agroforestry expert (Sambane, 2005) identified many that are potentially useful in the context of sustainable land use: Acacia nigrescens, grows well under the natural conditions in Gorongosa; Acacia nilotica, used for fuelwood, charcoal, fodder and construction; Adansonia digitata, fruit and leaves are edible, seed germinates fairly easily; Albizia lebbeck, a fast growing multipurpose species and N-fixing; Amblygonocarpus andongensis, Khaya anthoteca, Kigelia africana, all likely to be important carbon fixers as they form huge trees if not destroyed by fire; Annona senegalensi, has medicinal properties and edible fruit; Commiphora mossambicensis, likely to propagate well from stem cuttings; Faidherbia albida, well-known agroforestry tree; Peltophorum africanum a fast growing tree with potential to fix carbon, has medicinal properties, seed germinates well; Pseudolachnostylis maprouneifolia, a good honey tree, fire resistant, responds well to cultivation; Pterocarpus rotundfolius, potentially a good carbon fixer given its relative abundance in the Gorongosa area; Sclerocarya birrea, one of the most highly valued multipurpose fruit trees but monoecious; Herd (2007) investigated charcoal making in the area, interviewing a sample of charcoal-makers on-site. The respondents stated that their top five preferred species in order of importance were; Brachystegia boehmii, B. spiciformis, Julbernardia globiflora, Pterocarpus rotundfolius and Burkea africana. In reality, charcoal burners were using a wider variety of tree species, but the top four used were the same as the preferred species, with the exception of Swartzia madagascariensis. 4.1.2 Variation between the PSPs The number of woody species per hectare varies from nine to 53. We have also calculated the Shannon index H’, one of several indices used by conservationists to estimate biodiversity: 61 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 Where S is the number of species, pi is the relative abundance of each species, calculated as pi = ni/N where ni is the number of individuals in each species and N is the total number of individuals. The index varies from 1.19 in the savanna to 2.19 in the riverine forest (Table 3). We have not attempted to relate the species list to those from other studies in the region, such as Kanschik & Becker (2001) working in Zimbabwe and Bacheus et al. (2006) in Tanzania. Based on all 87 plots which include data on species, tree diameter, height and wood density, we divide the land cover into: Fig. 3. Relationship between diameter at breast height (dbh) and carbon content of the above ground biomass, obtained from destructive sampling of trees in the N’hambita area. The data are shown as black symbols and black line. Other lines are the corresponding relationship obtained by other authors working in similar vegetation types. Maximum dbh values of individual trees recorded on tropical woodland, savanna and secondary woodland were 64, 61 and 38 cm respectively. Only riverine woodland contained large stems exceeding the limits on this graph. - Riverine or Riparian Forest. The top five species by biomass ranking are: Sclerocarya birrea, Khaya anthoteca, Cleistochlamys kirkii, Acacia nigrescens and Pterocarpus rotundfolius. - Tropical Woodland, including, but not limited to that dominated by the miombo species. The top five species by biomass ranking are: Brachystegia boehmii, Diplorhynchus condylocarpon, Pterocarpus rotundfolius, Burkea africana, Brachystegia spiciformis. - Savanna, dominated by grass, but with sparse woodland of the genera Combretum or Acacia. The top five species are: Combretum adenogonium, Combretum apiculatum, Combretum hereroense, Commiphora mossambicensis, Pterocarpus rotundfolius. - Secondary Woodland, including abandoned machambas and degraded woodland. The top five species are: Brachystegia boehmii, Julbernardia globiflora, Brachystegia spiciformis, Diplorhynchus condylocarpon, Burkea Africana. - Machambas (agricultural plots). The top five species are: Sclerocarya birrea, Diplorhynchus condylocarpon, Pterocarpus angolensis, Burkea Africana, Pseudolachnostylis maprouneifolia. 4.2 Allometric relationships The relationship between diameter and dry biomass is best described by a power law (Fig. 3). The species used to build this relationship were amongst the most common, including the ‘top seven’ in terms of basal areas, and care was taken to obtain a wide range of stem diameters. For above ground biomass, the best fit was obtained by: y = 0.0267d2.5996 ; R2 = 0.93; n = 29 where d, diameter at 1.3 m above the ground is measured in cm, and the biomass is in kg C. The below-ground fraction of the total biomass declines somewhat as the trees grow, but here the fraction was generally lower than most published values from woodland savannas globally (Grace et al. 2005). For the purposes of a carbon inventory, a conservative assumption is that the below ground biomass is 0.25 of the above ground biomass. It is possible that other floristically distinct types exist but have been missed by the sampling, such as the protected ‘sacred’ area south-east of the N’hambita village, which has dense and nearlyimpenetrable woodland with little grass. 62 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 Savanna Secondary woodland Machamba Basal area (m2/ha) Aboveground biomass mean (tC/ha) Aboveground biomass median (tC/ha) Max. dbh (cm) Shannon index Sum of plot area (ha) Number of plots Tropical woodland Trees/ha Riverine forest Table 3. Statistical summary of the five land cover types. (±) refers to the standard deviation, but variables are not normally distributed so this is only an approximate indication of the variability. 421 ±167 13.8 ±3.3 406 ±253 10.0 ±3.2 386 ±275 5.8 ±3.9 561 ±255 8.0 ±2.0 38 47 ±18 27 ±13 14 ±10 13 ±9 8 43 24 12 14 8 92 64 61 38 70 2.2 ±0.5 2.1 2.0 ±0.4 10.1 1.2 ±0.5 3.8 2.1 ±0.6 8.6 N/A 6 26 10 17 1 4.3 Density and carbon content Wood densities vary from 0.13 to 0.73 (Table 2, and Bunster 2006). In the work on C aggradation following abandonment, Williams et al. (2007a) found a trend of increasing wood density with time of abandonment. 4.4 Carbon stocks 2.4 Above and below-ground carbon stocks in the biomass vary with the vegetation types between less than 10 tC ha-1 in machambas to over 40 tC ha-1 in riverine vegetation (Fig 4). These values are consistent with data reported by Williams et al. (2007b) in a synthesis study for the whole of Africa. Interpolating their graph we obtain biomass carbon stocks of 35 tC ha-1 and soil carbon stocks of 80 tC ha-1. Another estimate may be made from Frost (1996) who related biomass of miombo woodland to annual rainfall. From his regression equation, taking the annual rainfall as the 1990-2005 average (749 mm/annum) we get a biomass of 30 tC ha-1. 1 Fig. 4. Stem biomass of land cover types. The boxes contain 50% of the values, horizontal lines in the boxes are the median values, whiskers include 95% of the values and diamonds show outliers. 63 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 Fig. 5 Frequency distribution of stem biomass carbon, as tonnes per hectare, for all sample plots. the soil surface before it becomes substantially incorporated into the soil organic matter. Combining the data from many plots shows a skewed distribution for above-ground carbon, with a mode of 20-30 tC ha-1, a mean of 21 tC ha1 and a long ‘tail’ which reaches 70-80 tC ha-1 (Fig 5). There is a further component of biomass not reported here: below ground biomass is at least 25 % of that above-ground, so we may safely assume an average (above- and belowground) of 26 tC ha-1. 4.5 Land cover, forest area and rates of change 1999-2007. Satellite imagery provides the means of mapping the woodland cover, and repeated survey enables the rates of deforestation to be measured. However, the vegetation undergoes seasonal changes which substantially influence the spectral reflectance. In the context of savanna woodland, one of the major difficulties is separating the contribution made by the grassy and herbaceous components of the vegetation. To achieve this, we selected imagery from the time of year when the grass has died down yet the trees remain green. This is not necessarily the same week in every year, since the seasons vary from year to year, and so phenological events vary in their timing. A further difficulty is that images are often impaired by cloud cover. Biomass carbon in the machambas is negligible (less than 2 tC ha-1 in biomass) unless tree planting and protection have been successfully practised. As for carbon stocks in the soil, these are reported as medians in Williams et al. (2007a): woodlands, 58 tC ha-1; abandoned machambas 44.9tC ha-1. Soil C stocks in the top 0.3 m on abandoned land had a narrower range (21–74tC ha-1) than stocks in woodland soils (18– 140tC ha-1), and rather surprisingly soil carbon did not increase at all with time from abandonment (Williams et al. 2007a), in contrast to what has often been reported for humid rain forests and indeed for temperate systems. However, arid regions behave differently from the humid tropics, as was also found by Post and Kwon (2000) who reported annual rates of C accumulation in the soil of as little as 0.03tC ha-1 year-1 in arid locations. We believe this is because the annual or near-annual fire causes combustion of the organic matter deposited on In this study, we compared the satellite imagery in 1999 with that of 2007, using SPOT imagery with a spatial resolution of 20 m (Fig 6). This high level of resolution is required so that the PSPs and machambas can be unequivocally observed. We also used recent MODIS data to investigate seasonal variations in NDVI. It has 36 spectral bands but a spatial resolution of only 1 km in the relevant wavebands, yet it is useful for 64 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 the detection of seasonal and interannual variability, and also for the timing and spatial variation of fires (http://modis-fire.umd.edu/). Fig. 6. SPOT 4 images of the project area, showing different land cover types in 1999 and 2007. Areas are: A, corresponding to N’hambita, Bue Maria and Posta Da Pungwe; B, containing Pavua and M’Bulawa; C, a buffer zone only sparsely inhabited, and south of the river Pungue; D Mucombeze. The scene contains an additional area E, part Mucombeze and part Pinganganga (the latter falls into Manica province). The map derived from SPOT imagery shows the contrast between the different parts of the region; the protected area (Gorongosa National Park), the buffer zone and the community land (Fig. 6). The community land is being rapidly deforested (Table 4), and the buffer zone also shows substantial deforestation. To the north and south of the image, we see the impact of the spreading and expanding population of the towns, Gorongosa to the north and Inchope to the south. The deforestation rate in this period (1999-2007) may be contrasted with that over the period 19912000 obtained using LANDSAT imagery in nearly the same area (Wallentin 2006). In this earlier period the regional deforestation rate was estimated as 0.15 % per year at the regional level, and 0.03 % per year in the N’hambita area. Comparison with national deforestation rates declared by Mozambique and other miombocontaining countries may also be made from official figures made available by FAO (2007b): Mozambique 0.3 %, Angola 0.2 %, Tanzania 1.0 %, Zambia 0.9 % and Zimbabwe 1.5 %. The increase in rate is believed to be caused by a population increase leading to a higher requirement for machambas, and by an enhanced demand for charcoal. Table 4. Woodland cover in 1999 and 2007, and the average annual loss in that period, obtained from the analysis of SPOT imagery A B C D E Total area (ha) Woodland cover 1999 (ha) Woodland cover 2007 (ha) 6378 9538 12149 15669 24020 5385 8512 9706 11035 24020 4927 7674 11538 7333 19539 Annual loss of woodland (%) 0.87 1.09 -1.09 2.95 0.76 The drivers of deforestation are principally slashand-burn agriculture and clearance of land during charcoal making. It appears that both charcoal making and fuelwood harvesting in the area has increased rapidly as a result of an increased demand from towndwellers (Brouwer & Falcao 2004). Production of charcoal is a time consuming process and is poorly regulated or not regulated at all; it is also inefficient. Herd (2007) investigated charcoal The carbon stocks at a regional level, and their changes over the period 1999-2007, are summarised in Table 4. This calculation ignores any recent gains in the machambas, brought about by project-related tree-planting, as it is estimated that so far this is less than 1% of the carbon in the region. 65 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 exact evaluation for carbon payments, a survey of sample plots is required in the specific designated areas using the protocol outlined by Ryan et al. (2007), provided here as Appendix 1. making in the region. He found 1 kg of charcoal is obtained from 5.7 kg of wood (kiln conversion efficiency is 17.6 %). His study revealed that charcoal production was principally occurring within a strip 2 km wide to the west of the EN–1. For those involved in the activity it is an important livelihood strategy accounting for 74 % and 59 % of annual incomes for males and females respectively. References Abbot P, Lowore J, Werren M (1997) Models for the estimation of single tree volume in four Miombo woodland types. Forest Ecology and Management 97, 25–37. 5. Summary and recommendations Backeus I, Pettersson B, Stromquist L and Ruffo C (2006) Tree communities and structural dynamics in miombo (Brachystegia–Julbernardia) woodland, Tanzania Forest Ecology and Management 230, 171-178. 1) The woodlands are diverse, with the distinguishing species of the miombo formation. The woodland mosaic in the study area includes Combretum savanna and riverine woodlands. Brouwer R and Falcão MP (2004) Wood fuel consumption in Maputo, Mozambique Biomass and Bioenergy 27, 233 – 245. 2) The woodlands are under threat, being lost at rates which have increased in recent years and now stand as high as 2.95% per year in one area (D), and 0.87% per year in the immediate area of N’hambita village, where encroachment is occurring. The pattern of deforestation is related to the designation of the land and the use of the land by a variable and generally increasing population of humans. Bunster J (2006) Commercial timbers of MozambiqueTechnological Catalogue, 2nd Edition. Traforest Lda, Maputo. Chidumayo EN (1997) Miombo Ecology and Management: An Introduction. IT Publications in association with the Stockholm Environment Institute, London. Coates Palgrave K, Coates Palgrave K & Drummond, RB et al. (2003) Trees of Southern Africa. Struik Publishers, Cape Town. 3) An important means of protecting the woodland is through international payments for ‘avoided deforestation’. Elsewhere, we have produced a technical specification of how this might be achieved (Tipper & Garrett 2007). If the current rate of deforestation could be halted or reduced, carbon payments and/or biodiversity payments would be highly significant and they could be used to assist sustainable development. de Koning, J (1993) Checklist of Vernacular Plant Names in Mozambique. Wageningen Agricultral University, Wageningen. Desanker, PV, Frost PGH, Justice CO, and Scholes RJ (eds.). (1995) The Miombo Network: Framework for a Terrestrial Transect Study of Land-Use and Land-Cover Change in the Miombo Ecosystems of Central Africa. – IGBP Report 41. FAO (2007a) Background to Miombo woodlands in Mozambique http://www.fao.org/docrep/008/j5251e/J5251E%20-02.htm, undated report, accessed 18.11.07. 4) To avoid deforestation, charcoal-making needs to be restricted; this is best done by establishing designated areas for production of timber for charcoal. The encroachment of machambas onto forested areas should also be restricted; this is best done by increasing the productivity of the already-cultivated land by using agroforestry techniques which can enhance the fertility of soils and improve crop production. FAO (2007b) State of the World’s Forests. FAO, Rome. Available on-line at http://www.fao.org/icatalog/intere.htm. Fichtler E, Trouet V, Beeckman H, Coppin P, and Worbes M (2004) Climatic signals in tree rings of Burkea africana and Pterocarpus angolensis from semi-arid forests in Namibia, Trees, 18, 422-451. 5) Carbon stocks of above ground biomass vary widely. For an initial conservative estimation of carbon in above-ground biomass we may assume 21 tonnes carbon per hectare, with a further 5 tonnes in below-ground biomass. To obtain an Frost P (1996) The Ecology of Miombo Woodlands. pp. 11-57. In B.M. Campbell (Ed):The Miombo in Transition: Woodlands and Welfare in Africa. Center for International Forestry Research, Bogor, Indonesia. 66 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 Grace J, San Jose J, Meir P, et al.. (2006) Productivity and carbon fluxes of tropical savannas. Journal of Biogeography 33, 387-400. Williams M, Ryan CM, Rees RM, Sambane E, Fernando J and Grace J (2007a) Carbon sequestration and biodiversity of re-growing miombo woodlands in Mozambique Forest Ecology and Management Herd A (2007) Exploring the socio-economic role of charcoal production and the potential for sustainable production in the Chicale Regulado Mozambique. Unpublished MSc dissertation, School of GeoSciences, University of Edinburgh. Williams CA, Hanan NP, Neff JC, Scholes RJ, Berry J, Scott Denning A and Baker DF (2007b) Africa and the Global Carbon Cycle Carbon Balance and Management 2, doc 10.1186/1750-0680-2-3, available as an open-access journal at www.cbmjournal.com/content/2/1/3. Kanschik K and Becker B (2001) Dry miombo – ecology of its major plant species and their potential use as bioindicators. Plant Ecology 155, 139-146. Zolho R (2005) Effect of fire frequency on the regeneration of Miombo woodland in N’hambita, Mozambique. Unpublished MSc dissertation, School of GeoSciences, University of Edinburgh. Lynam T, Cunliffe R, Mapaure I and Bwerinofa I (2003) Assessment of the value of woodland landscape function to local communities in Gorongosa and Muanza Districts, Sofala Province, Mozambique CIFOR, Indonesia. Maregesi SM, Ngassapa OD, Pieters L and Vlietinck J (2006) Ethnopharmacological survey of the Bunda district, Tanzania: Plants used to treat infectious diseases Journal of Ethnopharmacology 113, 457-470. Mushove P (2003) Preliminary inventory of N’hambita community forest, Gorongosa District, Mozambique. Commissioned report, with additions from M Williams, School of GeoSciences, University of Edinburgh. Post WM and Kwon KC (2000) Soil carbon sequestration and land-use change: processes and potential. Global Change Biology 6, 317–327. Ryan, Williams and Grace (2007) Guidelines for the rapid assessment of vegetation carbon stocks in the N’hambita area. Unpublished document. School of GeoSciences, University of Edinburgh. (Appendix 2). Sambane E (2005) Above-ground biomass accumulation in fallow fields at the N’hambita community, Mozambique. Unpublished MSc dissertation, School of GeoSciences, University of Edinburgh. Spadavecchia L (2004) Synthesis of Remote Sensing Products and a GIS Database to Estimate Land Usage Change an Analysis of the N’hambita Community Forest, Commissioned report, School of GeoSciences, University of Edinburgh. Tinley K (1977) Framework of the Gorongosa Ecosystem. Ph.D. Thesis, University of the Witwatersrand, South Africa. Tipper and Garrett (2007) Conservation of miombo woodland in central Mozambique. Unpublished document School of GeoSciences, University of Edinburgh. (Appendix 4). Wallentin G (2006) Carbon change rate and assessment of its drivers in N’hambita, Mozambique. Commissioned report, School of GeoSciences, University of Edinburgh. 67 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 Appendix 1: List of species in the project area Scientific Name Acacia nigrescens Acacia nilotica Acacia polyacantha Acacia robusta Acacia sieberiana Acacia sp. Adansonia digitata Albizia amara Albizia brevifolia Albizia harveyi Albizia lebbeck Amblygonocarpus andongensis Anacardium occidentale Annona senegalensis Antidesma venosum Azanza garckeana Bauhinia petersiana Berchemia discolor Bombax rhodognaphalon Boscia salicifolia Brachystegia boehmii Brachystegia spiciformis rackenridgea zanguebarica Burkea africana Calliandra sp. Carica papaya† Cassia abbreviata Cleistochlamys kirkii Combretum adenogonium Combretum apiculatum Combretum hereroense Combretum molle Combretum sp. Commiphora mossambicensis Corchorus junodii† Cordyla africana Crossopteryx febrifuga Cussonia arborea Cussonia spicata Dalbergia boehmii Dalbergia melanoxylon Dichrostachys cinerea Diospyros kirkii Diospyros mespiliformis Diospyros sp. Local Name(s) Guhu, Ngungo Nhacassadzi, Nomba, Nhomba Guoe, M'guoe Mussodze Gunga Mulumanhama Mulamba, Mbuyu, Mulambe Mussorora, Mussorola Chiteta, Chithetha Chissio Tanga Tanga Munjonjota Cashew Miloro, Mulembe Mundangaramunhu Mutohue, Mutogue Munhando Chiglamademo Muungussi, Munguza Munhenza Mefuti Messassa Chibaramadona Mucarate, Mucarati Calliandra Papaeira Murumangama, Murumanyama Munhangolo Fiti, Fit M'fithi Munangare, Muchenalore Fithidondo Nhantsangu, M'tshanga Dzvototo Chimbiati, Chimbia Mutondo Muchombêgo Mussengue Muchendje Chimanda,Ximanda Paupreto, Mpingue, Mphingue Mupangara Mucula Mufuma Mulala 68 Family Fabaceae, Mimosoideae Fabaceae, Mimosoideae Fabaceae, Mimosoideae Fabaceae, Mimosoideae Fabaceae, Mimosoideae Fabaceae, Mimosoideae Bombacaceae Fabaceae; Mimosoideae Fabaceae; Mimosoideae Fabaceae; Mimosoideae Fabaceae; Mimosoideae Fabaceae; Mimosoideae Anacardiaceae Annonaceae Euphorbiaceae Malvaceae Caesalpinoideae Rhamnaceae Bombacaceae Capparaceae Caesalpinioideae Caesalpinioideae Ochnaceae Caesalpinioideae Caesalpinioideae Caricaceae Caesalpinioideae Annonaceae Combretaceae Combretaceae Combretaceae Combretaceae Combretaceae Burseraceae Tiliaceae Papilionoideae Rubiaceae Araliaceae Araliaceae Papilionoideae Papilionoideae Fabaceae; Mimosoideae Ebenaceae Ebenaceae Ebenaceae APPENDIX 2 - Woodland Inventory, N’hambita, 2007 Scientific Name Diospyros usambarensis Diplorhynchus condylocarpon Dombeya shupangae Ehretia amoena Entada abyssinica Entandrophragma caudatum Erythrina sp. Erythrophleum africanum Erythroxylum emarginatum Excoecaria bussei Faidherbia albida Ficus sycomorus Flacourtia indica Gardenia ternifolia Gliricidia sp.† Grewia monticola Grewia sulcata Grewia transzambesica Heteropyxis dehniae Holarrhena pubescens Hyphaene coriacea Jatropha curcas Julbernardia globiflora Karomia tettensis Khaya anthotheca Kigelia africana Lannea schimperi Lecaniodiscus fraxinifolius Mangifera indica Markhamia obtusifolia Markhamia zanzibarica Millettia mossambicensis Millettia stuhlmannii Monodora stenopetala Monotes engleri Ozoroa obovata Ozoroa sp. Peltophorum africanum Pericopsis angolensis Philenoptera violacea Piliostigma thonningii Premna senensis Pseudolachnostylis maprouneifolia Pteleopsis myrtifolia Pterocarpus angolensis Pterocarpus lucens Pterocarpus rotundfolius Rhus chirindensis Rhus sp. Rhynchosia pseudoviscosa† Local Name(s) Nhamudima Nham'tomole, Mutoa, Mutowa Thoa Cuacuacho Chidzedze Mulolo Thithi Messimbe Mucosse, Mukossa Mucombacore Mussadze Mushamvu Mutumbotumbo, M'tema, Mitema Munomoro Gliricidia Tadza, Thadza Sene Thetha, Theja Suruza, Tsurudzo Cumbanzóo Mucheu Jatropha Muhimbe Nhacabale Umbaua Mevungute Mumbumbu Mutarara Mangeira Phewa, Feba Nhamukabari, Nhacabare Mussungarossa Panga Panga Muchinge Xithethe, M'tongolo Chifuca, Chifupa Cantassaro Muzeze Xouanga Mpakhasi, Pacassa Mussequessa, Mussequesse Mhacanucanunga Family Ebenaceae Apocynaceae Sterculiaceae Boraginaceae Fabaceae; Mimosoideae meliaceae Papilionoideae Caesalpinioideae Erythroxylaceae Euphorbiaceae Fabaceae; Mimosoideae Moraceae Flacourtiaceae Rubiaceae Fabaceae Tiliaceae Tiliaceae Tiliaceae Heteropyxidaeceae Apocynaceae Arecaceae Euphorbiaceae Caesalpinioideae Lamiaceae Mekuaceae Bignoniaceae Anacardiaceae Sapindaceae Anacardiaceae Bignoniaceae Bignoniaceae Papilionoideae Papilionoideae Annonaceae Dipterocarpaceae Anacardiaceae Anacardiaceae Caesalpinioideae Papilionoideae Papilionoideae Caesalpinioideae Lamiaceae Mundoto, Mussonzoa Mussunganhemba Mbila,Mulombe,Mukwa,Mulombwa,Muconambira Muanga Miómba, Miombué Muchanfu Chinamazize Nhamperepera Euphorbiaceae Combretaceae Papilionoideae Papilionoideae Papilionoideae Anacardiaceae Anacardiaceae Papilionoideae 69 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 Scientific Name Sclerocarya birrea Scolopia stolzii Securidaca longipedunculata Senna petersiana Sterculia africana Sterculia appendiculata Sterculia quinqueloba Stereospermum kunthianum Strychnos henningsii Strychnos innocua Strychnos madagascariensis Strychnos potatorum Swartzia madagascariensis Tabernaemontana elegans Terminalia brachystemma Terminalia sambesiaca Terminalia sericea Terminalia stenostachya Trichilia emetica Vangueria infausta Vitex doniana Vitex sp. Voacanga thouarsii Xeroderris stuhlmannii Ximenia americana Ximenia caffra Zanha africana Ziziphus abyssinica Ziziphus mucronata Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Local Name(s) Marula, Mfula Comacamba Mupupo, Mpumpu Buembacor Munjale, Mucuna Bicancula Cherecheti, Goza, Tcherechete Potanzoe Panda Muteme Mutunduru Muthupa Pao Ferro Muchane, Muchachane Mucodomue Mundangamunho Mussussu Mucoudò M'sequeira, Massaniqueira Munziro, Mufula Mucuvu Mucuno, Mucuna Nhaponda Mulonde Mudogodogo M'tenguene, Mutenguene Chicumbite, Mussiquiribanda, Mussiquiri Mussau-sanga, Mussautanga, Mussaotsanga Muchecheni Balamadona Cataussarro Chiriacamba Guacha, Guaquacho, Guakuacho Mbubunhanga Muchambu Muchanvo Muchinguena Mucongontua Mucuiramhondoro Mugaramanjiva Mukhodone Mukunku, Mukungo, Mucuncu Muleme Mulonde Mumudendo Munganzo Munhomba Munhongoro, Munhungoro Munomolo Murara Muthumdolo 70 Family Anacardiaceae Flacourtiaceae Polygalaceae Caesalpinioideae Sterculiaceae Sterculiaceae Sterculiaceae Bignoniaceae Strychnaceae Strychnaceae Strychnaceae Strychnaceae Papilionoideae Apocynaceae Combretaceae Combretaceae Combretaceae Combretaceae Meliacea Rubiaceae Lamiaceae Lamiaceae Apocynaceae Papilionoideae Olacaceae Olacaceae Sapomdaceae Rhamnaceae Rhamnaceae APPENDIX 2 - Woodland Inventory, N’hambita, 2007 Scientific Name Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Unidentified Local Name(s) Mutindi Mutunguricua Nhacafupa Nhampulurue Nhancamba Nhanganzo Ruronde Senge Tubo Tsanga Family 71 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 Appendix 2: Ryan et al. (2007) Guidelines for the rapid assessment of vegetation carbon stocks in the N’hambita area By Casey Ryan, Mat Williams and John Grace, University of Edinburgh, School of GeoSciences This document describes the stratification, according to carbon stock, of the key land cover types around the N’hambita regulado, Sofala Province, Central Mozambique. The methodology and stratification is based on a set of inventories carried out as part of the N’hambita Community Carbon Project (Grace et al., 2007). Our objective is to provide a set of guidelines to allow rapid estimation of vegetation carbon (C) stocks in areas of similar ecosystem types. It should be broadly applicable to the dry miombo region (<1000 mm mean annual rainfall, sensu Frost (1996)) and covers ecosystems found in association with miombo woodlands, such as savannas and riverine woodlands, as well as miombo impacted by human activity. The descriptions of each ecosystem type, and the estimate of carbon stock is based on 87 sample plots (surveying >28 ha, and ~15,000 stems) and a newly developed allometric relationship between tree diameter and C stock. The allometric relationship was developed using the 4 dominant species (by basal area) in N’hambita, Sofala, but the plot data comes from Sofala, Zambezia and Cabo Delgado provinces of Mozambique. Based on the inventories, which include data on species, tree diameter, height and wood density, we divide the land cover into: - Riverine or Riparian Forest - Tropical Woodland, including, but not limited to that dominated by the miombo species - Savanna, dominated by grass, but with sparse woodland of the genera Combretum or Acacia - Secondary Woodland, including abandoned machambas and degraded woodland - Machambas (agricultural plots) - Method for estimating area of each land cover type Walk and mark the boundaries with a GPS track for the smaller land cover types, such as, ‘riverine forest’ and ‘machambas’. For the other classes, transects or point sampling may be suitable depending on the size of the area under consideration. The aim is to estimate the proportion of the total area under each class. ‘Savanna’ can grade gently into ‘tropical woodland’ and this transition can be difficult to demarcate. Choose the mid point so that the dividing line leaves equal areas of uncertain cover in each class. There are simple methods to determine the area bounded by GPS points, that work by dividing the polygon into triangles. An example is included as an Excel spreadsheet with this report. Classification scheme for each land cover type 1.Riverine forest Thicket at ground level, climbers and aerial roots may be present. No grass, many trees at least 30m in height, some greater than 1m DBH (diameter at breast height or 1.3 m), closed canopy. 72 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 2. Tropical woodland, including miombo Deciduous trees, grass up to 2m high in wet season, canopy cover 40 – 90%, a few 70cm DBH trees per ha. 3. Savanna, with sparse woodland Mostly grass, scattered trees (>10m apart) most trees <10m high. Soil is often not sand, but a black or dark grey ‘cotton’ soil with clay and little sand. 4. Secondary woodland, including abandoned machambas and degraded woodland Tree clearance, stumps, re-sprouts, few trees >50cm DBH. A lack of the key miombo species (Brachystegia, Julbernardia, etc.) is a good indicator of abandoned agriculture, but not necessary. 5. Machambas Crops or residue. Occasional trees, often large ones left during clearance for agriculture. Carbon stocks The representative characteristics of each land cover type have been generated from the sample plot surveys (Tables 1-3). Stocking density, basal area and stem size statistics were generated directly from the stem diameter surveys for each cover type (Table 1). Total woody C stocks (Table 1) were generated from the site-specific allometric relationships and the stem diameter surveys. The dominant species by basal area (Table 2) and stem numbers (Table 3) were determined for each land cover type by species identification. There is considerable variation in biomass within and between these land cover types (Figure 1), with the greatest values in riverine forest and tropical woodlands. The large variation in the biomass of tropical woodlands is caused by site specific variables such as rainfall, soil and disturbance history. Hence these values should be used as a rough estimate only, and inventories carried out where possible. Based on the data from N’hambita, the largest pool of carbon is normally the soil, with median values ~58 tC ha-1 (Williams et al 2007.) Roots contribute a significant amount of live biomass and can be assumed to be at least 25% of above ground carbon. Grassland can be considered to have negligible carbon stocks. Estimates for Palm woodland biomass are unreliable because of the lack of both wood density data and an appropriate allometric. Calculating regional carbon stocks Once the area of each land cover type has been determined for the management area, the total C stock can be determined by the following equation: i =5 C total = ∑i =1 Ai C i where Ctotal is the total C for the management area (tC), Ai is the area of the land cover type (ha, where i=1 to 5, see Table 1) and Ci is the mean C stock for land cover type i (tC ha-1, see Table 1). Acknowledgements Some of data described here were collected by Patrick Mushove and Will Garret, as well as the authors. Envirotrade Ltd supported the collection of the data in Zambezia and Cabo Delgado. This work is funded by NERC, UK and the EU. 73 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 Figure 1. Stem biomass of land cover types. Box contains 50% of the values, horizontal line indicates the median, whiskers include 95% of the values and diamonds show outliers. References Frost, P. (1996). The ecology of Miombo woodlands. The Miombo in transition : woodlands and welfare in Africa. B. M. Campbell. Bogor, Indonesia, Center for International Forestry Research: 11-55. Grace, J., C. M. Ryan and M. Williams (2007) An inventory of tree species and carbon stocks for the N’hambita Pilot Project, Sofala Province, Mozambique, Edinburgh, UK Williams, M., C. M. Ryan, R. M. Rees, E. Sambane, J. Fernando and J. Grace "Carbon sequestration and biodiversity of re-growing miombo woodlands in Mozambique." Forest Ecology and Management In Press, Corrected Proof. 74 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 Table 1. Mean attributes of different land cover types. SD is one standard deviation; nd = no data; *NB data is not normally distributed so SD should be seen as an indicator of variation only. Sum of median Max. plot no. Stocking Density Basal Area Aboveground stem mass DBH DBH Shannon Index, H areas plots Units (id) (trees/ha) (m2/ha) (tC/ha) (cm) (cm) (ha) Mean SD* Mean SD* Mean SD* Median Mean SD* (1)Riverine Forest 421 167 13.8 3.3 47 18 43 12 92 2.19 0.53 2.1 6 (2) Tropical Woodland 406 253 10 3.2 27 13 24 11 64 2.04 0.39 10.1 26 (3) Savanna 386 275 5.8 3.9 14 10 12 10 61 1.19 0.54 3.8 10 (4) Secondary Woodland 561 255 8 2 13 9 14 8 38 2.11 0.6 8.6 17 (5) Machamba 38 2.4 8 8 14 70 N/A 1 1 Table 2. Top 5 woody species in each land cover class, by contribution to biomass 1 Riverine Sclerocarya Khaya anthoteca Cleistochlamys birrea kirkii 2 Woodland Brachystegia Diplorhynchus Pterocarpus boehmii condylocarpon rotundfolius 3 Savannah Combretum Combretum Combretum adenogonium apiculatum hereroense 4 Secondary Brachystegia Julbernardia Brachystegia boehmii globiflora spiciformis 5 Machamba Sclerocarya Diplorhynchus Pterocarpus birrea condylocarpon angolensis Acacia nigrescens Burkea africana Commiphora mossambicensis Diplorhynchus condylocarpon Burkea africana Table 3. Top 5 woody species in each land cover class, by contribution to total stem numbers. 1 Riverine Cleistochlamys Lecaniodiscus Diplorhynchus Pterocarpus kirkii fraxinifolius condylocarpon rotundfolius 2 Woodland Diplorhynchus Pterocarpus Brachystegia Combretum condylocarpon rotundfolius boehmii adenogonium 3 Savannah Philenoptera Pterocarpus Combretum Combretum violacea rotundfolius apiculatum adenogonium 4 Secondary Brachystegia Philenoptera Julbernardia Diplorhynchus violacea globiflora condylocarpon boehmii 5 Machamba Diplorhynchus Pseudolachnostylis Pterocarpus unknown condylocarpon maprouneifolia angolensis 75 Pterocarpus rotundfolius Brachystegia spiciformis Pterocarpus rotundfolius Burkea africana Pseudolachnostylis maprouneifolia Combretum apiculatum Bauhinia petersiana Xeroderris stuhlmannii Bauhinia petersiana Albizia lebbeck APPENDIX 2 - Woodland Inventory, N’hambita, 2007 A PHOTOGRAPHIC GUIDE TO LAND COVER TYPES Riverine Forest Tropical Woodland, including that dominated by the miombo species 76 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 Savanna, dominated by grass, but with sparse woodland of the genera Combretum or Acacia Secondary Woodland is characterised by a lack of large stems, or the presence of large stumps. 77 APPENDIX 2 - Woodland Inventory, N’hambita, 2007 Machambas are agricultural areas, but some trees may remain Palm woodland – No C stock data available for these species. 78 + Models FORECO-10564; No of Pages 11 Forest Ecology and Management xxx (2007) xxx–xxx www.elsevier.com/locate/foreco Carbon sequestration and biodiversity of re-growing miombo woodlands in Mozambique M. Williams a,*, C.M. Ryan a, R.M. Rees b, E. Sambane a, J. Fernando a, J. Grace a a School of GeoSciences, IAES, University of Edinburgh, Edinburgh EH9 3JN, UK b SAC, West Mains Road, Edinburgh EH9 3JG, UK Received 20 March 2007; received in revised form 24 July 2007; accepted 31 July 2007 Abstract Land management in tropical woodlands is being used to sequester carbon (C), alleviate poverty and protect biodiversity, among other benefits. Our objective was to determine how slash-and-burn agriculture affected vegetation and soil C stocks and biodiversity on an area of miombo woodland in Mozambique, and how C stocks and biodiversity responded once agriculture was abandoned. We sampled twenty-eight 0.125 ha plots that had previously been cleared for subsistence agriculture and had been left to re-grow for 2 to 25 years, and fourteen 0.25 ha plots of protected woodlands, recording stem diameter distributions and species, collecting wood for density determination, and soil from 0 to 0.3 m for determination of %C and bulk density. Clearance for agriculture reduced stem wood C stocks by 19.0 t C ha1. There were significant relationships between period of re-growth and basal area, stem numbers and stem biomass. During re-growth, wood C stocks accumulated at 0.7 t C ha1 year1. There was no significant difference in stem C stocks on woodlands and on abandoned farmland 20–30 years old. Soil C stocks in the top 0.3 m on abandoned land had a narrower range (21–74 t C ha1) than stocks in woodland soils (18–140 t C ha1). There was no discernible increase in soil C stocks with period of re-growth, suggesting that the rate of accumulation of organic matter in these soils was very slow. The re-growing plots did not contain the defining miombo species, and total stem numbers were significantly greater than in woodland plots, but species richness and diversity were similar in older abandonments and miombo woodlands. Wood C stocks on abandoned farmland were capable of recovery within 2–3 decades, but soil C stocks did not change on this time-scale. Woodland soils were capable of storing >100 t C ha1, whereas no soil on a re-growing area exceeded 74 t C ha1, so there is a potential for C sequestration in soils on abandoned farmland. Management should focus on identifying C-rich soils, conserving remaining woodlands to protect soil C and preserve defining miombo species, and on investigating whether fire control on recovering woodland can stimulate accumulation of soil C and greater tree biomass, and restore defining miombo species. # 2007 Elsevier B.V. All rights reserved. Keywords: Savanna; Tropical deciduous forest; Basal area; Allometric equations; Wood density; Soil bulk density; Soil C stock; Biodiversity; Shannon index; Chronosequence 1. Introduction The C cycle of tropical open woodlands is relatively understudied compared to other biomes. These woodlands are subject to frequent disturbance via fires and land clearance. Such woodland degradation threatens terrestrial carbon stocks (Chidumayo, 2002; Frost, 1996) but is little monitored or modelled. Climate change mitigation initiatives resulting from the United Nations Framework Convention on Climate Change are now managing tropical woodlands to sequester carbon * Corresponding author. E-mail address: mat.williams@ed.ac.uk (M. Williams). (Silver et al., 2004), including 19 current projects in subSaharan Africa (Jindal, 2006). Alongside the potential to generate income through sales of C offsets, woodland management is likely to benefit other ecosystem services and biodiversity. However, the optimal management approaches are not yet clear, due to a lack of data on vegetation and soil C stocks, as well as biodiversity indices, on the dominant land use types. Here we quantify changing C stocks and biodiversity along a chronosequence of abandoned farmland in Mozambique, and in nearby protected miombo woodlands. Miombo is the vernacular term for the seasonally dry deciduous woodlands that are widespread across southern Africa, dominated primarily by genera Brachystegia, Julber- 0378-1127/$ – see front matter # 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.foreco.2007.07.033 Please cite this article in press as: Williams, M., et al., Carbon sequestration and biodiversity of re-growing miombo woodlands in Mozambique, Forest Ecol. Manage. (2007), doi:10.1016/j.foreco.2007.07.033 + Models FORECO-10564; No of Pages 11 2 M. Williams et al. / Forest Ecology and Management xxx (2007) xxx–xxx nadia and/or Isoberlinia (Campbell, 1996). These open woodlands extend across 2.7 million km2 of some of the world’s poorest countries (Campbell, 1996; Frost, 1996). Low per capita income and high population growth rates in southern Africa mean that subsistence slash-and-burn farming within the miombo zone is the predominant way of life. Growing populations are increasing pressure from slash-and-burn. Loss of miombo woodland is also driven by increasing demand for fuel-wood (Abbot and Homewood, 1999). Our objective was to determine how slash-and-burn agriculture affected soil and vegetation C stocks on an area of miombo woodland, and how C stocks recovered once agriculture was abandoned. We hypothesised that C stocks in both soils and vegetation of abandoned slash-and-burn plots would be lower than in woodland plots. Farmland is abandoned after a few years as soil fertility declines, and we expected this fertility to recover slowly following abandonment. Thus, we hypothesised that C stocks would recover more rapidly in vegetation than in soils. We hypothesised that the species dynamics in recovering plots would indicate a return towards the dominant and defining species of local miombo woodland. We also hypothesised that this successional change in species would result in an increase in mean wood density as pioneer species were replaced with slower growing, denser miombo dominants. This study is unique in collecting and analysing data on soil and stem C (the largest stocks of C in miombo), wood density and biodiversity on a range of different land use types within a small region of miombo woodland. 2. Study site This case study is located in the small community of N’hambita in Sofala Province, Mozambique, located around the operational centre of an EU-funded C sequestration pilot project at 188580 4400 S, 348100 3700 E. This is an area with little or no infrastructure, and a community still recovering from decades of war. Recent population growth means that more land is coming under the traditional slash-and-burn agricultural system. Local households clear machambas (slash-and-burn plots covering 1–3 ha) by felling trees and then burning them, and then raise crops, which are planted with a stick, or primitive entrenching tool. After a few years of tillage, yields fall and the machambas are abandoned to woodland re-growth. The N’hambita community is located on the western escarpment of the southern limit of the Rift Valley, near the Gorongosa National Park, and its ecology is comprehensively described by Tinley (1977). The area receives 690 mm mean annual precipitation (ranging from 407 to 1219 mm), 96% of which falls between October and April, based on data from 1999 to 2005 (Mozambique Central Water Board, ARA-Centro, 2005). Soils are highly weathered and generally freely drained sandy loams or sandy silt loams. Fire is a frequent, generally annual, disturbance agent; during June–October 2006 most natural miombo vegetation in the area was burned (personal observation). The N’hambita community lands, which covers an area of 348 km2, can be divided into three zones. In the east, the community lands lie within the Gorongosa National Park, and there is no settlement, agriculture or land clearance. In the centre, the community lands lie within the Park buffer zone. There are settlements, and slash-and-burn agriculture, but land clearance is restricted legally. In the west, the community lands are outside the buffer zone, population density is greater, and land clearance is ongoing. 3. Methods We surveyed tree biodiversity, above-ground woody biomass, wood density and soil C across an area of largely undisturbed primary woodlands and a series of abandoned machambas of differing ages, during 2004–2005. All survey plots were located in central N’hambita, within the buffer zone of Gorongosa National Park, to avoid areas with heavy disturbance. 3.1. Primary woodlands During December 2004, surveys were undertaken of the woody vegetation in the part of the N’hambita community that lies within the buffer zone of the Gorongosa National Park. Plots were not surveyed if they showed signs of previous cultivation or charcoal burning, or if local informants knew them to have been utilised. Fourteen 0.25 ha plots were established in pairs at seven randomly selected locations, equally spaced along the road and track network within the community buffer zone. At the seven locations, two 50 m  50 m plots were set out at 200, 450, 700 or 950 m from the road, along a transect line perpendicular to the road. The distances for the paired plots were selected randomly from these four options. All living woody specimens >0.05 m diameter at breast height (DBH) were measured, recording species local name (provided by a local informant knowledgeable in botany), species botanical name (Coates Palgrave et al., 2002; de Koning, 1993; Van Wyk and Van Wyk, 1997; Van Wyk, 1993), and DBH (m) using diameter callipers or tape. On trees forking below 1.3 m from the ground level, each stem was measured and recorded separately. Trees forking above 1.3 m were measured at breast height. In one plot there was a single very large (2.17 m DBH) baobab tree (Adansonia digitata), a species with an unusual shape that does not conform to generic allometric relationships. No other baobabs were found anywhere else in the study and this single specimen was excluded from plot calculations as it heavily skewed the analyses for its plot compared to others. 3.2. Abandoned machambas In June 2005, 28 abandoned agricultural fields were surveyed, each identified through talking with the farmers of the N’hambita community. Farmers were asked to locate machambas abandoned over past years and decades. The shortest period of re-growth was 2 years, and the longest exceeded 20 years. It was easier to obtain precise age estimates Please cite this article in press as: Williams, M., et al., Carbon sequestration and biodiversity of re-growing miombo woodlands in Mozambique, Forest Ecol. Manage. (2007), doi:10.1016/j.foreco.2007.07.033 + Models FORECO-10564; No of Pages 11 M. Williams et al. / Forest Ecology and Management xxx (2007) xxx–xxx for more recently abandoned sites (up to 19 years). Site ages were confirmed by the secretary of the community association. Robertson (1984) used this method to obtain the fallow age of plots in Malawi and found it agreed well with estimates from aerial photographs. Several sites had been abandoned >20 years ago, but since independence in 1975, and so the ages for these sites were between 20 and 30 years. Abandoned machambas were divided into four classes, depending on fallow age: class 1, 1–5 years; class 2, 6–10 years; class 3, 11– 19 years; and class 4, 20–30 years. Individual plots on age classes 1–3 have been aged precisely, but those on age class 4 have not, and may be aged between 20 and 30 years. To sample the abandoned machambas, which usually extended over an area of 1–3 ha, each was divided into four approximately equal parts, and each of these parts was sampled by one 10 m radius subplot. Each subplot was randomly located, but locations were discounted that caused overlap between quadrants, or that placed subplots over the agreed boundary of the abandoned machamba. The total area of the four subplots was thus 1256 m2, 1/8 ha. Within each subplot, species and DBH for all stems >0.05 m DBH were recorded, a total of 1955 stems. The abandoned machambas were not randomly distributed in the area with respect to their age, but were clustered in groups of similar age. 3.3. Stem wood density Dry bulk stem density was measured for the 21 most populous species found on the woodland and abandoned machamba plots. These species accounted for 980 of the 3166 stems enumerated, and 72% of the basal area. For the remaining, un-sampled species, each was assigned the average density, weighted by basal area, of species sampled on plots of the same age class. To determine dry bulk density, during November 2005 one or more branches between 0.01 and 0.02 m diameter were taken from three or more trees and cut to 0.15 m length, and the bark removed. Fresh volume was determined by displacement of water in a measuring cylinder (precision 1000 mm3) and also by calculation from three measurements of diameter and one of length (precision 0.1 mm). The samples were returned to the lab 3 and dried at 100 8C to constant weight. Density was calculated from the average of the fresh volume derived from the two methods, divided by the dry weight. This method only determines the dry bulk density of the sapwood of the tree, which may differ from the density of the heartwood, which is present in most miombo species. However, sampling heartwood would have required the destruction of a large numbers of trees, so a sapwood–heartwood comparison was limited to five species, which had been felled fortuitously. The heartwood samples were much larger than the sapwood samples, either circular cross sections of the bole at different heights (0.1–0.2 m thick) or cubes sawn from the centre of a cross-section (0.15 m  0.15 m  0.15 m). Volume was calculated both by measurement of diameter and thickness and also by measuring the weight of water displaced when placed into a full bucket. The wood samples were dried by episodic heating in a microwave to constant weight. 3.4. Stem C stocks Stem biomass was calculated using allometric relationships (Table 1) developed in the miombo woodlands of nearby countries (Abbot et al., 1997; Chidumayo, 1997; Frost, 1996) and also a generic equation for the tropics (Brown et al., 1989). Two of the equations relate DBH or basal area to estimate volume, and then our wood density data were used to estimate biomass. The other two equations estimated biomass directly from DBH data. Multiple approaches to biomass estimation allowed realistic uncertainties to be generated. Wood biomass was assumed to be 50% C (Nabuurs et al., 2003). 3.5. Soil properties and C stocks At all 28 abandoned machambas soil samples were collected from four subplots. In each subplot ten soil samples were collected from the 0–0.03 m horizon and mixed to form a single composite sample for analysis. In one of the subplots soil samples from four depths were collected (0–0.03 m, 0.03– 0.1 m, 0.1–0.2 m and 0.2–0.3 m). Soils data were similarly collected from nine subplots in an area of 1 ha within 500 m of the woodland plots. At five of the subplots a depth profile was Table 1 Allometric equations used in this paper Reference Abbot et al. (1997) Equation(s) (4.22+2.76 log D) V = 10 Source country Notes Malawi For canopy trees, assumed to be >4 m in height for this study. Phuyu site Understorey species, assumed to be trees less than 4 m high for our study. Phuyu site V = (0.057 + 0.000918D2)2 Frost (1996) V = 6.18A0.86 Zaire, Malawi, Zambia and Zimbabwe Equation applied on a stand basis Chidumayo (1997) B = 3.01D  7.48 B = 20.02D  203.37 Zambia For trees <0.1 m DBH For trees >0.1 m DBH Brown et al. (1989) B = 34.47  8.067D + 0.659D2 Dry tropics Not miombo specific, developed in ‘‘the dry tropics’’ B, biomass (t); V, Volume (m3); A, basal area (m2 ha1); D, diameter (cm) at 1.3 m (DBH). Note that for Abbot and Frost equations, wood density values determined at the site were used to calculate biomass from volume (Table 3). Please cite this article in press as: Williams, M., et al., Carbon sequestration and biodiversity of re-growing miombo woodlands in Mozambique, Forest Ecol. Manage. (2007), doi:10.1016/j.foreco.2007.07.033 + Models FORECO-10564; No of Pages 11 4 M. Williams et al. / Forest Ecology and Management xxx (2007) xxx–xxx sampled (0, 0.05, 0.15, 0.25 m depth) and at the remaining four, only a surface sample was taken. In the field, all soil textures were determined by hand texturing (Rural Development Service, 2006). Soil samples were dried, sieved (2 mm sieve) and ballmilled to produce a fine flour. Percentage soil C was determined on dried samples using a Carbo-Erba/400 automated CN analyser. At each soil plot, soil bulk density measurements were determined using steel rings of known volume. Before weighing, the soils samples were dried in an oven at 40 8C for 24 h. Bulk density values were calculated by dividing the mass (t) of the soil sample by volume of the cylinder (m3). Total soil C stock was determined by stepwise integration of the profile data of soil C content from 0 to 0.3 m. Textural analysis on soils of abandoned machambas and woodland sites indicated that all were dominated by sand loams and sand loam silts. Single factor ANOVA revealed that there was no significant relationship between site age and bulk density (BD) on machambas, and so the data were pooled to provide a single bulk density estimate (1.26 t m3, identical to the mean woodland BD) which was applied at all sites. The pooled BD data were combined with soil %C data, and multiplied by the mass fraction of soil remaining after sieving, to generate estimates of total soil C stock for each soil layer. Fig. 1. Measured basal area plotted against age for all abandoned machambas (left panel), and for woodland plots (right panel, diamond indicates mean), with a linear regression assuming oldest abandoned machambas, age class 4, are 25 years. The age of the oldest abandoned machambas is given as 25 years, but lies between 20 and 30 years. Regression parameters are: y = 0.27x + 1.13; P < 0.001; r2 = 0.55 (age class 4 = 30 years), y = 0.35x + 0.48; P < 0.001; r2 = 0.61 (age class 4 = 25 years), y = 0.47x  0.41; P < 0.001; r2 = 0.68 (age class 4 = 20 years). 4. Results 3.6. Biodiversity The Shannon index (H0 ) is a measure of biodiversity calculated from the relative abundance of species in a community: H0 ¼  S X pi ln pi i¼1 where pi = ni/N, ni is the number of individuals present of species i, N the total number of individuals, and S is the total number of species. The Shannon index was determined for woody species >0.05 m DHB for each abandoned machamba and woodland plot. The Jaccard similarity coefficient (J) is a statistic used for comparing the similarity and diversity of sample sets. We used J to determine the degree of similarity of species composition of different age classes. The Jaccard coefficient is defined as the size of the intersection divided by the size of the union of the sample sets: JðA; BÞ ¼ jA \ Bj jA [ Bj where A and B are the binary descriptions of species presence/ absence in given age classes. A value of 1 indicates complete similarity, while 0 indicates complete dissimilarity. We used a list of 161 species, comprising all those recorded in surveys within the N’hambita community on woodlands and abandoned machambas, during 2003–2005. 4.1. Vegetation structure and C stocks There were clear changes in vegetation characteristics along the chronosequence of machamba abandonments. Basal area was significantly correlated (P < 0.001) with time since abandonment (Fig. 1). A linear regression was able to explain 55–68% of the observed variability (uncertainty in ages of age class 4 abandonments, which might be 20–30 years old, accounts for the range of R2 values reported here). The slopes of the regressions indicated that basal area increment was 0.25– 0.47 m2 ha1 year1 (again the uncertainty here reflects uncertainty in age of the oldest abandonments). Basal area recorded in the woodland plots varied from 2.4 to 13.1 m2 ha1, reflecting the highly heterogeneous forest cover in this area. The mean value  standard deviation, S.D., for the woodland plots was 8.2  3.0 m2 ha1. However, there was no significant difference between basal area recorded in the oldest abandonments (>20 years, mean = 8.2 m2 ha1) and the woodland plots (t-test, two sample assuming equal variances, P > 0.05). For the woodland plots there was no significant relationship between basal area and distance from the road (ANOVA, P = 0.59). Stocking density (number of tree stems >0.05 m DBH ha1) also varied along the chronosequence, initially increasing with age, peaking after 10–20 years of abandonment, and then declining (Fig. 2). A third-order polynomial fit was able to explain 57% of observed variability, and revealed a highly significant relationship between stocking density and age (P < 0.001). There was a significant difference between the stocking density of machambas abandoned for >20 years and the Please cite this article in press as: Williams, M., et al., Carbon sequestration and biodiversity of re-growing miombo woodlands in Mozambique, Forest Ecol. Manage. (2007), doi:10.1016/j.foreco.2007.07.033 + Models FORECO-10564; No of Pages 11 M. Williams et al. / Forest Ecology and Management xxx (2007) xxx–xxx Fig. 2. Measured stem stocking density (stems per ha) plotted against age (years since abandonment) for all abandoned machambas (left panel), and for woodland plots (right panel, diamond indicates mean). The age of the oldest abandoned machambas is given as 25 years, but lies between 20 and 30 years. A third-order polynomial is fitted to the data, S.D. = 49.3 + 31.2t + 7.0t2  0.30t3, where t is the time in years. woodland plots (means are 574 and 373 trees ha1, respectively, t-test, two sample assuming equal variances, P < 0.01). Combining all four allometric equations, the mean  S.D. estimated stem C stock estimates in the woodland plots was 19.0  8.0 t C ha1. Nine of the 14 woodland plots had mean biomass estimates from 12 to 24 t C ha1, but plot values ranged from 4.3 to 33.4 t C ha1. The mean estimated stem C stock for the oldest abandoned machambas (>20 years) was 15.7  3.9 t C ha1, ranging from 10.1 to 22.2 t C ha1. There was no significant difference (t = 1.21, P = 0.11) in the stem C stock estimates between abandoned machambas >20 years old and woodlands (t-test, two sample assuming equal variances). Stem C stock estimates in the abandoned machambas were significantly and positively correlated with time since abandonment (P < 0.001, Fig. 3). Using linear regressions for the four allometric equations, and assuming that the oldest abandonment were 25 years old, resulted in estimates of biomass accumulation rates varying from 0.43 to 0.87 t C ha1 year1, and a mean  S.D. of 0.70  0.19. The uncertainty in productivity estimates introduced by uncertainty in ages of oldest abandonments is important. Using the Brown equation (Table 1) and assuming that trees in age class 4 were 30 years old resulted in productivity estimates of 0.57 t C ha1 year1, but 0.74 t C ha1 year1 if age class 4 trees were 25 years old, and 1.00 t C ha1 year1 for 20 year old trees. 4.2. Wood density Wood density for individual species ranged from 0.40 to 0.71 t m3 (Table 2). The mean density  S.D. was 0.56  0.08 t m3. The lower bulk densities were associated with fruiting species such as marula (Sclerocarya birrea) and 5 Fig. 3. Estimate wood C stock plotted against age for all abandoned machambas (left panel), and for woodland plots (right panel). The age of the oldest abandoned machambas is given as 25 years, but lies between 20 and 30 years. C stock is calculated from basal area, wood density and four different allometric relationships. The data here are the mean estimates from the four different allometric equations. The linear regressions determined from each individual relationship are plotted to indicate the uncertainty in biomass estimates. The legend indicates the author of the allometric relationship and the slope of the C stock vs. age relationship (i.e. annual wood C productivity, t C ha1 year1). mango (Mangifera indica). The three defining miombo species (Brachystegia boehmii, Brachystegia spiciformis and Julbernadia globiflora) in our sites had density values of 0.52, 0.63 and 0.63 t m3, slightly below or above the average. The average wood density for each plot was determined from the specific measurements (Table 2), species composition and weighted by the cube of the each stem’s DBH. The lowest values were found in the most recently abandoned machambas, while values for all abandoned machambas >7 years old were 0.55 t m3 (Fig. 4), very close to the mean value from the tree samples. A non-linear curve fit using a saturation equation (D = xt/(e + t), where t is the time since abandonment and x and e are the parameters) was able to explain 30% of observed variation with a root-mean-square error (RMSE) of 0.006. For the five species with both sapwood and heartwood samples, it was possible to make comparisons of wood density estimates from both types of wood (Table 3). The mean wood density was 0.57 t m3 for heartwood and 0.55 t m3 for sapwood. A t-test was used to determine if the heartwood and sapwood densities were significantly different. Only one species of the five had significantly different heartwood and sapwood density (P < 0.05, t-test, two sample assuming equal variances). 4.3. Soil C stocks There were no clear trends in soil C stocks in the top 0.3 m along the abandoned machamba chronosequence (Fig. 6). The frequency distribution of soil C stocks differed between abandoned machambas and woodlands. The Shapiro–Wilks test (IMSL Stats Library) indicated that the abandoned machamba soil C stocks were normally distributed (n = 28, P > 0.05) while the woodland plot soil C stocks were not (n = 25, P < 0.001). The mean soil C stock (S.D.) for abandoned Please cite this article in press as: Williams, M., et al., Carbon sequestration and biodiversity of re-growing miombo woodlands in Mozambique, Forest Ecol. Manage. (2007), doi:10.1016/j.foreco.2007.07.033 + Models FORECO-10564; No of Pages 11 6 M. Williams et al. / Forest Ecology and Management xxx (2007) xxx–xxx Table 2 Stem sapwood dry bulk density data (mean and standard deviation) from the 21 most populous species found on the combined woodland and abandoned machamba plots Species n Mean dry bulk density (t m3) Standard deviation (t m3) Mangifera indica Khaya anthoteca Acacia nigrescens Commiphora mossambicensis Sclerocarya birrea Entada abyssinica Brachystegia boehmii Xeroderris stuhlmannii Piliostigma thonningii Combretum apiculatum Philenoptera violacea Albizia amara Diplorhynchus condylocarpon Julbernardia globiflora Brachystegia spiciformis Albizia lebbeck Burkea africana Erythrophleum africanum Millettia stuhlmannii Pterocarpus rotundifolius subsp. rotundifolius Cleistochlamys kirkii 4 3 5 3 12 18 6 3 4 10 24 10 9 8 8 4 3 3 12 8 3 0.40 0.45 0.46 0.47 0.47 0.51 0.52 0.52 0.53 0.55 0.55 0.57 0.60 0.63 0.63 0.63 0.63 0.64 0.68 0.65 0.71 0.07 0.02 0.03 0.06 0.12 0.11 0.09 0.02 0.05 0.03 0.10 0.11 0.07 0.07 0.04 0.15 0.03 0.15 0.07 0.04 0.11 Samples were collected in N’hambita in 2006, and are sorted by increasing mean bulk wood density. n indicates number of sample collected per species. machambas was 45.2 (14.1) t C ha1. The woodland plots had a clear bimodal distribution of soil C (Fig. 7), with 9 of the 28 plots having values from 10 to 30 t C ha1 and another 9 having values from 60 to 90 t C ha1. The median soil C stocks were 57.9 t C ha1 for woodlands and 44.9 t C ha1 for abandoned machambas. The highest (140 t C ha1) and lowest (18 t C ha1) soil C stocks found in all soil profiles were in woodland plots. A Wilcoxon rank sum test (IMSL Stats Fig. 4. Wood density (D, t m3) estimates for all abandoned machambas plotted against age, and for all woodland plots (right panel, diamond indicates mean). Wood density of 21 dominant species was determined and combined with species data on each plot. A non-linear curve fit for the chronosequence of abandoned machambas is shown, using a saturation equation, D = xt/(e + t), where t is the time since abandonment (years) and x and e are the parameters. For the best fit x = 0.58 and e = 1.00. Library), the nonparametric equivalent of the two-sample t-test, indicated there was no significant difference between the woodland and abandoned machamba soil C stock estimates (P > 0.05). 4.4. Biodiversity There were 69 different woody species (DBH > 0.05 m) in the 14 woodland plots (total survey area 3.5 ha), which contained in total 1211 stems >0.05 m DBH. Five species contributed 54% of the total stem count, and five species contributed 46% of the total basal area (Tables 4 and 5). Diplorhynchus condylocarpon and B. boehmii were the dominant species (by stocking density and basal area, respectively). The dominant species of the woodland plots are typical of those for dry miombo (Kanschik and Becker, 2001). We identified 67 woody species (DBH > 0.05 m) in the 28 abandoned machamba plots (total survey area 3.5 ha), which contained in total 1955 stems >0.05 m DBH. The dominant species in abandoned machambas differed according to time since abandonment (Tables 4 and 5), and according to whether stocking density or basal area were used as measures of dominance. The youngest plots (age class 1) were dominated by fruit trees, some exotic, such as papaya (Carica papaya) marula (S. birrea) and mango (M. indica), or trees with other domestic uses, such as monkey bread (Bauhinia thonningii). In older abandonments (age-classes 2–4), the trees were dominated by small-to-medium sized native trees such as Philenoptera violacea and Combretum apiculatum. The dominant species in the abandoned machambas largely differed from those in the woodland plots. Comparing the oldest abandonments to the Please cite this article in press as: Williams, M., et al., Carbon sequestration and biodiversity of re-growing miombo woodlands in Mozambique, Forest Ecol. Manage. (2007), doi:10.1016/j.foreco.2007.07.033 + Models FORECO-10564; No of Pages 11 7 M. Williams et al. / Forest Ecology and Management xxx (2007) xxx–xxx Table 3 Heartwood and sapwood dry bulk densities for five dominant species Species ID Sclerocarya birrea Heart n Mean (t m3) S.D. t-Test P value a Sap 4 0.50 0.03 a 12 0.47 0.12 Millettia stuhlmannii Heart a Sap 5 0.56 0.05 a 12 0.68 0.07 Philenoptera violacea Heart Sap 3 0.61 0.04 0.002* 0.383 a a 24 0.55 0.10 Albizia amara Heart a Sap 2 0.64 n/a 0.116 Entada abyssinica a 10 0.57 0.11 Hearta Sapa 5 0.55 0.03 13 0.49 0.12 n/a 0.131 For each species are listed the number of branch samples measured (n), the mean wood density of the samples, and their standard deviation (S.D.). A t-test was used to determine if the heartwood and sapwood densities were significantly different; the t-test P values are given in the final row. There was only one significant difference (marked *) at the 5% level. a Wood type. Table 4 The five most dominant species, ranked by stocking density in each age class of abandoned machamba and in woodland, are listed in order Order 1 (1–5 years) 2 (6–10 years) 3 (11–20 years) 4 (20–30 years) Woodland 1 2 Sclerocarya birrea Philenoptera violacea Entandrophragma caudatum Combretum apiculatum Piliostigma thonningii Entandrophragma caudatum Philenoptera violacea Philenoptera violacea 3 4 5 Mangifera indica Acacia nigrescens Philenoptera violacea Piliostigma thonningii Albizia lebbeck Sclerocarya birrea Piliostigma thonningii Combretum apiculatum Sterculia appendiculata Diplorhynchus condylocarpon Pterocarpus rotundifolius rotundifolius Dalbergia boehmii Taub. Combretum apiculatum Commiphora mossambicensis Brachystegia boehmii Vitex doniana Sweet Cleistochlamys kirkii Percent of total n Species richness J 60% 67% 57% 51% 54% 7 5.2 7 7.4 4 17.5 10 13.8 14 15.1 0.15 0.19 0.19 0.31 N/A Also shown are the percentage of total stems made up by the five dominant species, the number of plots sampled in each age class (n), and mean species richness. The final row shows the Jaccard similarity coefficient (J) for species composition of abandoned machambas of different ages (classes 1–4) and that of woodland. woodlands, the dominant five species by basal area had no species in common, while the dominant five by stocking density had just one in common (Tables 4 and 5). The Jaccard similarity coefficient for species composition of abandoned machambas compared to woodland ranged from a minimum of 0.15 between recent abandonments and woodland plots, to a maximum of 0.31 between the oldest abandonments and woodland plots (Table 4). The defining miombo species found in the woodland plots (B. boehmii, B. spiciformis, J. globiflora) were completely absent from abandoned machambas of all ages in stems >0.05 m DBH. The lowest Shannon indices were found in the most recently abandoned machambas, and increased with time since abandonment, but then saturated at greater ages (Fig. 5). A non-linear curve fit using a saturation equation (H0 = at/(b + t), where t is the time since abandonment and a and b are the parameters) was able to explain 62% of observed variation with a root-mean-square error (RMSE) of 0.12. Species richness also increased with time since abandonment, saturating with a similar pattern to the Shannon Index (Table 4). There were no significant differences between abandoned machambas >20 years old and the woodland plots for either Shannon indices (P = 0.15) or mean species richness (P = 0.22, t-test, two sample assuming equal variances). 5. Discussion Because of access issues and chronosequence measurement, the sampling strategy employed was not completely randomised, and so some caution is required in interpreting the results. There was no significant effect of distance from road on woodland structure, so locating plots within easy walking Table 5 Dominant species type ranked by basal area in all abandoned machambas, and in woodland Rank order Age class 1 Age class 2 Age class 3 Age class 4 Woodland 1 2 3 4 5 Carica papaya Trichilia emetica Mangifera indica Acacia nigrescens Sclerocarya birrea Philenoptera violacea Piliostigma thonningii Albizia lebbeck Entandrophragma caudatum Sclerocarya birrea Albizia lebbeck Philenoptera violacea Piliostigma thonningii Combretum apiculatum Sclerocarya birrea Combretum apiculatum Philenoptera violacea Commiphora mossambicensis Faidherbia albida Albizia amara Brachystegia boehmii Acacia nigrescens Diplorhynchus condylocarpon Brachystegia spiciformis Erythrophleum africanum Percent of total 69% 66% 67% 55% 46% The five most abundant species are listed in order. The final row shows the percentage of basal area made up by the five dominant species. Please cite this article in press as: Williams, M., et al., Carbon sequestration and biodiversity of re-growing miombo woodlands in Mozambique, Forest Ecol. Manage. (2007), doi:10.1016/j.foreco.2007.07.033 + Models FORECO-10564; No of Pages 11 8 M. Williams et al. / Forest Ecology and Management xxx (2007) xxx–xxx 5.1. C stocks in vegetation Fig. 5. Shannon index (H0 ) of species diversity plotted against age for all abandoned machambas (left panel), and for woodland plots (right panel, diamond indicates mean). The age of the oldest abandoned machambas is given as 25 years, but lies between 20 and 30 years. A non-linear curve fit for the chronosequence of abandoned machambas is shown, using a saturation equation, H0 = at/(b + t), where t is the time since abandonment (years) and a and b are the parameters. For the best fit a = 2.51 and b = 2.79. distance of tracks is unlikely to have affected the results. Woodland plots were randomly located along the road/track network, but the abandoned machambas were selected on the basis of farmer interviews. The abandoned machambas were all located around the N’hambita village, and are clustered in age groups, according to the historical development of the community. Due to their proximity to human settlements, it is likely that the abandoned machambas were more disturbed than woodland plots, for example by fires, which are often started near settlements or roads, or fuel-wood harvesting. It is also possible that farmers selected woodland areas with richer soils for clearance. Space-for-time studies have been criticised for producing artefacts because of non-random site selection (Frost, 1996). If the cleared and abandoned land was never originally miombo, then the chronosequence approach is compromised. The balance of evidence suggests that these lands were most likely dry miombo. Firstly, the nearest woodland plots (three were within 2.5 km to the NW, N and SE of the cleared areas) were all dominated by the defining dry miombo species. Local seed sources were thus available, and the local climate was suitable. Secondly, there were no clear topographical or soil textural differences between these three woodland plots and the abandoned machambas lying between them. However, the lack of information concerning the natural vegetation of the machamba areas is a major, but unavoidable, cause of uncertainty. We hypothesised that C stocks in vegetation and soils of abandoned machambas would be lower than in woodland plots, and that C stocks would accumulate more rapidly after abandonment in vegetation than in soils. The data for vegetation largely supported these hypotheses, but the results for soils are less clear. For the abandoned machambas, there was a clear relationship between time since abandonment and increased wood C stocks (Figs. 1 and 3). While the oldest abandoned machambas (>20 years) and woodland plots had similar basal area and biomass, suggesting a large degree of recovery, there were still significant differences in stocking density (Fig. 2), indicating that structural differences remained. The woodland plots had a larger variation in biomass than the abandoned machambas, which we attribute to either variable natural disturbance generating a mosaic of differently structured woodland stands (Fig. 7) or perhaps local variations in hydrology. However, further work is required to determine whether differences in soils or hydrology can explain these variations, and whether wood C stocks of low biomass woodlands are aggrading at the same rate as on abandoned machambas. After abandonment, wood C stocks increased by 0.7 t C ha1 year1. The mean annual increment (MAI) of above-ground woody C was similar to other figures for dry miombo, 0.9 t C ha1 year1 over 35 years in Zambia (Chidumayo, 1997), 0.5 t C ha1 year1 in 16 years old coppiced miombo woodland in northern Zambia (Stromgaard, 1985) and 0.75 t C ha1 year1 over 50 years, calculated from data in Frost (1996). There is uncertainty in the productivity estimate (0.2 t C ha1 year1), due to uncertainty in aging the oldest abandoned machambas, and in the allometric relationships used to generate biomass from stem diameter and wood density measurements. These two uncertainties are roughly equal—the range in productivity estimates is 0.44 t C ha1 year1 due to uncertainty in allometric equations, and 0.43 t C ha1 year1 due to uncertainty in time since abandonment. Thus, improvements to the productivity estimates at this site will require both locally determined allometric equations and a more thorough investigation of site histories and constraints on date of abandonment. 5.2. C stocks in soils Disturbance of soils associated with cultivation generally leads to a rapid decline in soil organic C content as a consequence of enhanced microbial respiration (King and Campbell, 1994; Schlesinger, 1986). In a global survey Guo and Gifford (2002) found C losses of 42% due to land use conversion to agriculture from native forest. Walker and Desanker (2004) observed C reductions of 40% after conversion to agriculture from Malawian miombo woodland. In the present study, abandoned agricultural land had a median C stock 23% lower than the surrounding woodlands. There was no significant difference between the pooled abandoned machamba soil C stocks and those of woodlands, so there was no clear support for our expectation of a widespread drop in soil C stocks after slash-and-burn. However, it is possible that farmers selected areas with richer soils for clearance, in which case a real loss of C stocks may be obscured by the comparison with randomly sampled woodland. The broad range of soil C stocks in woodland Please cite this article in press as: Williams, M., et al., Carbon sequestration and biodiversity of re-growing miombo woodlands in Mozambique, Forest Ecol. Manage. (2007), doi:10.1016/j.foreco.2007.07.033 + Models FORECO-10564; No of Pages 11 M. Williams et al. / Forest Ecology and Management xxx (2007) xxx–xxx Fig. 6. Total soil C content in surface 0.3 m for all sites plotted against age for all abandoned machambas (left panel) and for woodlands (right panel). The age of the oldest abandoned machambas is given as 25 years, but lies between 20 and 30 years. suggests major variations in organic matter and probably fertility (Figs. 6 and 7). Subsequent reforestation often leads to a steady but slower accumulation of organic C as plant C inputs accumulate in the surface soil horizons (Guo and Gifford, 2002; Jarecki and Lal, 2003). In a survey of re-growing tropical forests, mostly from the moist and wet tropics, Silver et al. (2000) found that soil C increased at 1.3 t ha1 year1 in the first 20 years after abandonment. However, in a review of rates of C sequestration following land use change, Post and Kwon (2000) reported annual rates of C accumulation of as little as 0.03 t C ha1 year1 in arid locations. These low rates of recovery are consistent with the lack of any identifiable change 9 in soil C on older abandoned machambas at our dry tropical sites, and suggest very slow additions of organic matter to the soil. In a study of miombo woodlands in Malawi, Walker and Desanker (2004) also found that soil C stocks did not increase after abandonment of agriculture. This lack of accumulation is likely a result of frequent fire disturbance (Bird et al., 2000; Chidumayo and Kwibisa, 2003), or perhaps termite activity (Konate et al., 2003). The soil C stocks of the woodland plots are notable for their bimodal distribution (Fig. 7). The significant departure from normality suggests a complex spatial pattern of soil C stocks, which is not easily explained given the lack of textural or bulk density differences among soil samples. Site variables such as drainage are known to influence soil organic C concentrations, but were not characterised in this study. It is possible that disturbance history, particularly frequency of fires, is a critical factor in development of soil C stocks through vegetation-firefuel-load feedbacks (Frost, 1996; Stromgaard, 1985, 1986). Further sampling of soils and other potentially linked environmental controls are required to test whether the bimodal distribution is widespread and to understand its cause. The management implications are clear for sequestering C in wood. Natural regeneration will restore stem C stocks in 2–3 decades, but rates of accumulation are low and total stem C stocks in the natural vegetation are relatively small compared to soil C stocks. The potential to sequester C in soils is less clear. There is no increase in soil C stocks along the chronosequence, which indicates that inputs of organic matter to soils in regrowing miombo are very small. Woodland soil C stocks vary across almost an order of magnitude (Fig. 7). If boosting the C storage of miombo soils were possible (Jarecki and Lal, 2003; Lal, 2003), this would provide a valuable means to sequester C, with an equal or greater potential for C sequestration than restoring abandoned machambas to miombo woodland under current disturbance conditions. Whether a reduction in Fig. 7. Frequency distributions of C stocks (t C ha1) in abandoned machamba soils (top panel), woodland soils (middle) and woodland stem C (bottom panel). Please cite this article in press as: Williams, M., et al., Carbon sequestration and biodiversity of re-growing miombo woodlands in Mozambique, Forest Ecol. Manage. (2007), doi:10.1016/j.foreco.2007.07.033 + Models FORECO-10564; No of Pages 11 10 M. Williams et al. / Forest Ecology and Management xxx (2007) xxx–xxx disturbance (i.e. fire intensity and frequency) would increase long-term wood C stocks is also a critical management issue (Bond et al., 2005; Trapnell, 1959). 5.3. Changes in stem wood density We hypothesised that the successional change in species would result in an increase in mean wood density as pioneer species were replaced with slower growing miombo dominants. The evidence supports this hypothesis, but only for the very early part of the succession (<10 years) where fruit trees are dominant (Fig. 4). These trees have the lowest wood density values (Table 2), but the data suggest that fruit trees are rapidly lost from re-growing woodlands. There was little difference in wood density calculated for sapwood and heartwood from a selection of common species (Table 3) suggesting that the biomass estimates derived from sapwood density data alone were reasonable. However, we were not able to sample the heartwood of larger trees, including the defining miombo species, due to logistical difficulties. It is possible that using sapwood estimates of wood density for these species has resulted in an under-estimate of wood biomass and C stocks. 5.4. Species dynamics We hypothesised that the species dynamics in recovering plots would indicate a return towards the dominant species of local miombo woodland. We reject this hypothesis. None of the defining miombo species are present in any of the abandoned machambas (Tables 4 and 5), as was also found by Stromgaard (1986). However, the secondary dominant species of miombo are found on abandoned machambas, and there is greater similarity in species composition between older abandonments and woodlands (Table 4). Even though most species differed between abandoned machambas and woodlands (as indicated by low Jaccard similarity coefficients), the biodiversity of woody species (i.e. Shannon index and species richness) of abandoned machambas >10 years old was similar to that found in woodlands (Fig. 5). Overall tree biodiversity has not been degraded by the slash-and-burn disturbance, but the time-scale of recovery of defining miombo species is unclear. Miombo species are known to be able to survive the destruction of their above-ground parts (Chidumayo, 1997; Frost, 1996; Nyerges, 1989; Robertson, 1984). They are generally good at re-sprouting and can reproduce from root suckers; 15 years of mattocking were required to kill Brachystegia spp. (Robertson, 1984), and re-sprouting is a common response to destruction by fire. The defining miombo species are, however, thought to be relatively fire-tender (Cauldwell and Zieger, 2000; Trapnell et al., 1976). We have observed that fires are commonplace in the N’hambita community, recurring every year or two. It is possible that in the open, early successional areas, the frequent fires, high grass biomass, and thus high fuels loads, mean that only fire-tolerant species can re-establish (e.g. P. violacea, D. condylocarpon, Combretum spp.). The data show that fruit trees are replaced by fire-tolerant miombo species, but the defining miombo species do not establish over 2–3 decades. Further work is required to study the woody biomass belowground in these sites, which is potentially significant given the prodigious sprouting behaviour from root stock of many miombo species. The bimodal distribution of soil C stocks in woodlands needs to be tested by further detailed surveys including analyses of nutrient dynamics and soil hydrology. Finally, the role of fire in miombo systems is significant and the C dynamics of these woodlands can only be fully understood and predicted within the context of fire disturbance. 6. Conclusions Our objective was to determine how slash-and-burn agriculture affected soil and vegetation C stocks, and biodiversity on an area of miombo woodland, and how C stocks and species changed once agriculture was abandoned. We have shown that clearance for agriculture reduces stem wood C stocks by 19.0 t C ha1 and the years following abandonment wood C stocks accumulated at 0.7 t C ha1 year1. However, the regrowing areas do not contain the defining miombo species, and stem numbers are significantly greater than in woodland plots. Whether typical miombo species regain dominance is less clear, so conserving existing miombo woodlands is vital for maintaining the defining species, and their rich associated fauna. If fire disturbance on the abandoned machambas is heightened by proximity to human settlements (as is the case here) then more fire resistant species may dominate instead. Because woodland soil C stocks vary so much, it is not clear whether slash-and-burn reduces soil C stocks. Analysis of the abandonment chronosequence shows that there is no accumulation of soil C after 20–30 years of abandonment. This suggests that the rate of accumulation of organic matter in these soils has been very slow. Our study emphasises the importance of measuring and monitoring ecosystem C stocks when assessing the potential for C sequestration. While the woodland stem C stocks are capable of recovery within a few decades, the soil stocks do not accumulate following abandonment over a few decades. Woodland soils were capable of storing >100 t C ha1, whereas no abandoned machamba soil exceeded 74 t C ha1, and no stem wood stock exceeded 33 t C ha1, so there is a potential for C sequestration in woodland soils that should be investigated as a management option, probably through experiments with fire exclusion during woodland re-growth (Chidumayo, 1997). Management for C sequestration should also focus on identifying C-rich soils and conserving remaining woodlands to protect soil C. To ensure that local communities can meet agricultural needs without permanent loss of woodlands, land management must provide approaches to increase crop output at low cost, e.g. agro-forestry and intercropping. The major challenge for C management is to understand the observed variability in vegetation and soil C stocks in woodlands, and use this understanding to manage existing forests and re-growing areas for greater C storage. Please cite this article in press as: Williams, M., et al., Carbon sequestration and biodiversity of re-growing miombo woodlands in Mozambique, Forest Ecol. Manage. (2007), doi:10.1016/j.foreco.2007.07.033 + Models FORECO-10564; No of Pages 11 M. Williams et al. / Forest Ecology and Management xxx (2007) xxx–xxx Acknowledgements We dedicate this paper to the late Patrick Mushove, of Harare, whose assistance with the sampling and analysis of the woodland plots was vital. We also thank Piet van Zyl and his colleagues at Envirotrade for their assistance with field operations. We thank the N’hambita community for their hospitality and assistance. Funding for this work was provided by the European Development Fund, EuropeAid, and NERC. We thank Paul Stoy for his comments. M.W. wrote the paper and analysed data. M.W., J.G. and B.R. devised the experiments and observations. E.S. and J.F. collected part of the data as a component of their MSc theses. C.M.R. collected further data, and analysed the complete data set, and contributed to the writing. J.G. and B.R. provided input to the writing. References Abbot, J.I.O., Homewood, K., 1999. A history of change: causes of miombo woodland decline in a protected area in Malawi. J. Appl. Ecol. 36, 422–433. Abbot, P., Lowore, J., Werren, M., 1997. Models for the estimation of single tree volume in four Miombo woodland types. For. Ecol. Manage. 97, 25–37. Bird, M.I., Veenendaal, E.M., Moyo, C., Lloyd, J., Frost, P., 2000. Effect of fire and soil texture on soil carbon in a subhumid savanna (Matopos, Zimbabwe). Geoderma 94, 71–90. Bond, W.J., Woodward, F.I., Midgley, G.F., 2005. The global distribution of ecosystems in a world without fire. New Phytol. 165, 525–538. Brown, Gillespie, R., Lugo, E., 1989. Biomass estimation methods for tropical forests with applications to forest inventory data. For. Sci. 35, 881–902. Campbell, B.M., 1996. The Miombo in Transition: Woodlands and Welfare in Africa. Center for International Forestry Research, Bogor, Indonesia. Cauldwell, A.E., Zieger, U., 2000. A reassessment of the fire-tolerance of some miombo woody species in the Central Province, Zambia. Afr. J. Ecol. 38 138–146>. Chidumayo, E., 2002. Changes in miombo woodland structure under different land tenure and use systems in central Zambia. J. Biogeogr. 29, 1619–1626. Chidumayo, E.N., 1997. Miombo Ecology and Management: An Introduction. IT Publications in association with the Stockholm Environment Institute, London. Chidumayo, E.N., Kwibisa, L., 2003. Effects of deforestation on grass biomass and soil nutrient status in miombo woodland, Zambia. Agric. Ecosyst. Environ. 96, 97–105. Coates Palgrave, K., Drummond, R.B., Moll, E.J., Coates Palgrave, M., 2002. Trees of southern Africa, 3rd ed./New ed. revised and updated by Meg Coates Palgrave (Ed.). Struik Publishers, Cape Town, Great Britain. de Koning, J., 1993. Checklist of Vernacular Plant Names in Mozambique. Wageningen Agricultral University, Wageningen. Frost, P., 1996. The ecology of miombo woodlands. In: Campbell, B. (Ed.), The Miombo in Transition: Woodlands and Welfare in Africa. CIFOR, Bogor. Guo, L.B., Gifford, R.M., 2002. Soil carbon stocks and land use change: a meta analysis. Global Change Biol. 8, 345–360. 11 Jarecki, M.K., Lal, R., 2003. Crop management for soil carbon sequestration. Crit. Rev. Plant Sci. 22, 471–502. Jindal, R., 2006. Carbon Sequestration Projects in Africa: Potential Benefits and Challenges to Scaling Up. EarthTrends, World Resources Institute. Kanschik, W., Becker, B., 2001. Dry miombo—ecology of its major plant species and their potential use as bio-indicators. Plant Ecol. 155, 139–146. King, J.A., Campbell, B.M., 1994. Soil organic matter relations in five land cover types in the Miombo Region (Zimbabwe). For. Ecol. Manage. 67, 225–239. Konate, S., Le Roux, X., Verdier, B., Lepage, M., 2003. Effect of underground fungus-growing termites on carbon dioxide emission at the point- and landscape-scales in an African savanna. Funct. Ecol. 17, 305–314. Lal, R., 2003. Offsetting global CO2 emissions by restoration of degraded soils and intensification of world agriculture and forestry. Land Degrad. Dev. 14, 309–322. Nabuurs, G.-J., Ravindranath, N.H., Paustian, K., Freibauer, A., Hohenstein, W., Makundi, W., 2003. Forest land. In: Penman, J., Gytarsky, M., Hiraishi, T., Krug, T., Kruger, D., Pipatti, R., Buendia, L., Miwa, K., Ngara, T., Tanabe, K., Wagner, F. (Eds.), Good Practice Guidance for Land Use, Land-Use Change and Forestry, Intergovernmental Panel on Climate Change. UNEP, Kanagawa. Nyerges, A., 1989. Coppice swidden fallows in tropical deciduous forest: biological, technological, and sociocultural determinants of secondary forest successions. Hum. Ecol. 17, 379–400. Post, W.M., Kwon, K.C., 2000. Soil carbon sequestration and land-use change: processes and potential. Global Change Biol. 6, 317–327. Robertson, E.F., 1984. Regrowth of two African woodland types after shifting cultivation. Ph.D. Thesis. University of Aberdeen, Aberdeen. Rural Development Service, 2006. Soil Texture, Technical Advice Note 52. DEFRA, London. Schlesinger, W.H., 1986. Changes in soil carbon storage and associated properties with disturbance and recovery. In: Trabalka, J.R., Reichle, D.E. (Eds.), The Changing Carbon Cycle. A Global Analysis. Springer Verlag, New York, pp. 194–220. Silver, W.L., Kueppers, L.M., Lugo, A.E., Ostertag, R., Matzek, V., 2004. Carbon sequestration and plant community dynamics following reforestation of tropical pasture. Ecol. Appl. 14, 1115–1127. Silver, W.L., Ostertag, R., Lugo, A.E., 2000. The potential for carbon sequestration through reforestation of abandoned tropical agricultural and pasture lands. Restorat. Ecol. 8, 394–407. Stromgaard, P., 1985. Biomass, growth, and burning of woodland in a shifting cultivation area of South Central Africa. For. Ecol. Manage. 12, 163–178. Stromgaard, P., 1986. Early secondary succession on abandoned shifting cultivator’s plots in the miombo of South Central Africa. Biotropica 18, 97–106. Tinley, K., 1977. Framework of the Gorongosa Ecosystem. Univeristy of Pretoria, Pretoria. Trapnell, C.G., 1959. Ecological results of woodland burning experiments in Northern Rhodesia. J. Ecol. 47, 129. Trapnell, C.G., Friend, M.T., Chamberlain, G.T., Birch, H.F., 1976. Effects of fire and termites on a Zambian woodland soil. J. Ecol. 64, 577–588. Van Wyk, B., Van Wyk, P., 1997. Field Guide to Trees of Southern Africa. Struik Publishers, Cape Town. Van Wyk, P., 1993. Southern African Trees. New Holland, London. Walker, S., Desanker, P., 2004. The impact of land use on soil carbon in Miombo Woodlands of Malawi. For. Ecol. Manage. 203, 354–360. Please cite this article in press as: Williams, M., et al., Carbon sequestration and biodiversity of re-growing miombo woodlands in Mozambique, Forest Ecol. Manage. (2007), doi:10.1016/j.foreco.2007.07.033 APPENDIX 2 - Woodland Inventory, 2007 Appendix 4: Tipper and Garrett (2007) Draft Plan Vivo Technical Specification Conservation of miombo woodland in central Mozambique Miombo woodland in Sofala Province, Mozambique Editor: Richard Tipper, Edinburgh Centre for Carbon Management Contributors: William Garrett, Edinburgh Centre for Carbon Management 90 APPENDIX 2 - Woodland Inventory, 2007 Version Control Version Date Author Status 1.0 draft 6th December 2007 William Garrett, ECCM Working document Acknowledgements: Acknowledgements to all the staff of Envirotrade Lda in Mozambique, the University of Edinburgh and in particular to the N’hambita Community. 91 APPENDIX 2 - Woodland Inventory, 2007 Contents Page Appendix 4: Tipper and Garrett (2007).................................................................................................90 1. Introduction......................................................................................................................................93 1.1 Description of the ecosystem and the area ................................................................................93 1.2 Description of the threats to the ecosystem ...............................................................................94 2. Approach to community based miombo woodland conservation ....................................................95 2.1 General approach.......................................................................................................................95 2.2 Description of additionality of community conservation areas for miombo woodland protection in central Mozambique .........................................................................................................................95 2.3 Description of the environmental and social benefits from miombo woodland conservation in central Mozambique .........................................................................................................................95 3. Baseline Carbon Emissions ............................................................................................................96 3.1 Initial stock of vulnerable carbon in miombo woodland ..............................................................96 3.2 Rate of Forest Loss in absence of conservation ........................................................................98 4. Leakage Assessment ......................................................................................................................99 5. Management Plan Requirements..................................................................................................100 Maps of area ..................................................................................................................................100 a. Governance plan ....................................................................................................................100 b. Activity plan ............................................................................................................................100 c. Monitoring plan .......................................................................................................................101 6. Calculation of credits for woodland conservation areas ................................................................102 6.1 Project Crediting Procedure .....................................................................................................102 6.2 Monitoring and indicators for crediting .....................................................................................102 6.3 Verification of the Monitoring System.......................................................................................103 References ........................................................................................................................................... 104 Appendix 1 ……………………...…………..………………………………………………………..…...... 105 92 APPENDIX 2 - Woodland Inventory, 2007 1. Introduction This “technical specification” has been developed for use by Plan Vivo projects involving communities participating in central Mozambique (Sofala Province). Through the Plan Vivo system communities may be able to access carbon finance to assist with the conservation and restoration of forests. This technical specification sets out the methods that should be used to estimate the carbon benefits from conserving miombo woodland in central Mozambique, the requirements for a Plan Vivo – management plan, and the indicators to be used for monitoring the delivery of the carbon benefit. It aims to summarise best available evidence about the environmental benefits associated with sustainable management and conservation of these valuable ecosystems. Further information and research is welcome and will be incorporated periodically. 1.1 Description of the ecosystem and the area Mozambique is located along the southeastern coast of Africa. The miombo community land use and carbon management – N’hambita pilot project is located in the buffer zone of the Gorongosa National Park (GNP) in Sofala Province, Central Mozambique. The pilot project covers an area of approximately 20,000 hectares (known as the Chicale Regulado). This technical specification is based on work done as part of this pilot project. The climate in this area is characterised as sub-tropical. There are two distinct seasons each year. The dry season occurs between May to October, and the wet season occurs between November to April. Based on weather data from ARA-Centro (The Mozambican water board) at Chitengo (in the Gorongosa National Park) over the past seven years mean annual precipitation is 749mm distributed mainly between November to April, but with high inter-annual variability. Elevation within the study area ranges from 35 to 330 m.a.s.l. The soils in the Chicale Regulado are generally poor, highly weathered and freely draining sandy loams on the higher ridges and sandy silt loams along stream and river margins. According to Mushove (2003) there are several floristic associations in and around the Chicale Regulado i.e. in the buffer zone of the Gorongosa National Park, Sofala Province (Table 1.1). These include miombo woodland, of which there are three classes, Combretum woodland, riverine woodland and Combretum / Palm woodland. However miombo is the most common woodland type dominated by genera such as Brachystegia, Julbernardia, Erythrophleum, Burkea, Diplorhynchus, and Pterocarpus. Woody Vegetation Group Combretum Riverine Miombo with Combretum/ Woodland Woodland Palm significant Woodland dominance from Pterocarpus rotundifolius, Burkea Africana, and Erythrophleum africanum Table 1.1 Woody vegetation classes (groups) found in N’hambita community forest (from Mushove 2003) Miombo dominated by Brachystegia and Julbernardia Miombo as in Group 1 but with abundance of Diplorhynchus 93 APPENDIX 2 - Woodland Inventory, 2007 Miombo woodlands extend across approximately 2.8 million km2 of the southern sub-humid tropical zone from Tanzania and the Democratic Republic of Congo in the north, through Zambia, Malawi and eastern Angola, to Zimbabwe and Mozambique in the south. The distribution of miombo woodland largely coincides with the flat to gently undulating surfaces that form the Central African plateau (Appendix 1). WWF (2002) calculate that there are approximately 440,000 km2 of miombo woodland in Mozambique. The miombo ecosystem is an open canopy deciduous woodland type dominated by a few species of trees. Crowns of trees are typically NOT interlocking. Generally grasses are found below the trees making this a fire prone system. Miombo woodland is characterised by dispersed vegetation with few large trees. Most miombo species are relatively slow growing. Many of these species are deciduous loosing their leaves during the long dry spell. Typically Miombo species are very drought tolerant and frost sensitive. The extent to which this technical specification may be used the miombo eco-region has yet to be defined. 1.2 Description of the threats to the ecosystem The key threats to this ecosystem are: • Encroachment – land clearances for agriculture. This is observed to occur throughout the area in particular on low lying ground in proximity to water sources. • Charcoal production. Illegal charcoal production is often cited as the key driver of deforestation throughout sub-Saharan Africa. Local markets for charcoal include Beira, Chimoio, Gorongosa and Inchope. Herd (2007) discovered that the majority of charcoal production occurs within 2 km of main roads. • Burning. Prior to the introduction of a fire management regime (which commenced within the managed area in 2005) almost the entire project area was burnt annually (by uncontrolled fires). Frequent burning will hinder natural regeneration (and hence stand recovery) and the accumulation of carbon both in biomass as well as in the soils. • Logging. The majority of logging (cherry picking of the most valuable timber tree species) occurred within this Chicale Regulado prior to the 1980’s. 94 APPENDIX 2 - Woodland Inventory, 2007 2. Approach to community based miombo woodland conservation 2.1 General approach The overall approach to community based miombo woodland conservation is one of working in a collaborative manner with interested communities to establish effective, long-term conservation management based on existing community structures. The main pillars of this approach are as follows: - Development of an understanding within the community that the long-term benefits of conservation will outweigh the short-term costs of protection - Building of effective local governance structure to set rules necessary for protection and to assign key responsibilities to individuals - Establish effective teams to monitor the area; undertake fire protection activities and promote complementary economic actions to prevent or reduce any “leakage” effects associated with the protection of the area. - Provide financial support through carbon finance to cover the costs of protection. 2.2 Description of additionality of community conservation areas for miombo woodland protection in central Mozambique The recent history of miombo woodland loss in central Mozambique is strong evidence that additional efforts to protect this habitat are required. While some protection is given to areas in designated protection areas (such as nature Reserves, National Parks etc.) there is strong evidence that further efforts are required to conserve the remaining fragments of miombo woodland in all areas in central Mozambique. The protection of miombo woodland involves a number of short-term costs that communities or individual farmers with land in adjacent areas must bear: - opportunity cost of not cultivating land - opportunity cost of not extracting wood fuel or timber or live plants - additional effort to control fires - some negative effects of forest wildlife on crops - cost of organising community conservation efforts (governance, monitoring, etc) As there are no formal means by which communities can access funding to cover these costs, the effect of Plan Vivo carbon finance is strongly additional. 2.3 Description of the environmental and social benefits from miombo woodland conservation in central Mozambique While communities and individual farmers living around miombo woodlands incur some short-term costs associated with the conservation of the ecosystem there are long-term benefits that accrue to local communities, downstream populations and society in general. - Soil conservation - particularly the prevention of soil erosion associated with heavy rainfall events and siltation of water courses (climate change adaptation benefit) - Hydrological benefit – harvesting of incidental moisture and improved water flows which will help to reduce catastrophic flooding (climate change adaptation benefit) - Biodiversity benefit – through the protection of wildlife habitat. This may have a long economic impact attracting visitors to the Gorongosa National Park. - NTFP – beekeeping, medicines, fruits etc. 95 APPENDIX 2 - Woodland Inventory, 2007 3. Baseline Carbon Emissions This section sets out the procedure to be used to quantify the “baseline” carbon emissions that would be associated with loss of cloud forest expected in the absence of conservation measures. 3.1 Initial stock of vulnerable carbon in miombo woodland This technical specification refers to the conservation of carbon stocks in woodlands in Sofala province, Mozambique. Throughout this region there tends to be a complex matrix of different land use categories (Table 3.1). The base carbon stocks are very variable between the different vegetation categories. Therefore in order to quantify the initial carbon stock (across an area of land supporting a matrix of different vegetation categories), the initial carbon stocks in each vegetation category should be calculated separately. The initial stock of carbon may be estimated using the following methods (biomass survey or default factors): - Biomass survey This method will only be applicable to areas on miombo woodland i.e. it may not be used to quantify carbon stocks in combretum savanna and riverine woodlands where the default factor should be used. Above ground biomass of miombo woodland can be calculated using the following allometric equation y = 0.0267 d 2.5996 ; R 2 = 0.93; n = 29 where d, diameter at 1.3 m above the ground is measured in cm, and the biomass is in kg C (Ryan et al., 2007). Assume below ground biomass to 0.25 of above ground biomass (Grace et al., 2005). - Default factor: Based on Grace et al. (2007) default factors for base carbon (tC ha-1) may be applied to miombo woodland (where no biomass survey is done) and to combretum savanna and riverine woodlands. 96 APPENDIX 2 - Woodland Inventory, 2007 Vegetation category Description Tropical woodland Tropical woodland including, but not limited to that dominated by the miombo species. The top five species by biomass ranking are: Brachystegia boehmii, Diplorhynchus condylocarpon, Pterocarpus rotundfolius, Burkea Africana, Brachystegia spiciformis. This is a relatively less dense vegetation, with more open spaces between trees and few large trees. Savanna is dominated by grass, but with sparse woodland of the genera Combretum or Acacia. The top five species are: Combretum adenogonium, Combretum apiculatum, Combretum hereroense, Commiphora mossambicensis, Pterocarpus rotundfolius. Dense, high woodland adjacent to watercourses. The top five species by biomass ranking are: Sclerocarya birrea, Khaya anthoteca, Cleistochlamys kirkii, Acacia nigrescens and Pterocarpus rotundfolius. This includes abandoned machambas and degraded woodland. The top five species are: Brachystegia boehmii, Julbernardia globiflora, Brachystegia spiciformis, Diplorhynchus condylocarpon, Burkea Africana. Machambas (agricultural plots). The top five species are: Sclerocarya birrea, Di plorhynchus condylocarpon, Pterocarpus angolensis, Burkea africana, Pseudolachnostylis maprouneifolia. Savanna Riverine or riparian forest Secondary Woodland Machambas Default carbon stocks 26 tC ha-1 12 tC ha-1 43 tC ha-1 14 tC ha-1 8 tC ha-1 Table 3.1 Vegetation categories and default base carbon in Sofala province, Mozambique (Grace et al., 2007) 97 APPENDIX 2 - Woodland Inventory, 2007 3.2 Rate of Forest Loss in absence of conservation This section sets out the method to be used to define the baseline rate of woodland loss in Sofala Province, Mozambique. Evidence of the historic rate of loss of miombo woodland in Sofala Province is limited. The method used to estimate baseline emissions is based on area analysis of historical satellite images combined with ground truthing of vegetation categories. This process of measuring historical changes in land use has been complicated in Sofala as a result of the depopulation that occurred within the study area during the 1980’s (resulting from war) and subsequent re-population that has occurred since the mid 1990’s. This effectively restricts the timeline available for analysis of historical land use change to no more than 10 years. The CLIMAFOR method provides an objective risk-based estimate of future carbon emissions from land use change based upon the proximity of a given area of woodland to “risk factors” such as roads, settlements and existing agriculture. In the case of this study the risk factor considered is proximity to main roads (EN-1) and the ‘protection’ factor for woodland located in designated protection areas (Gorongosa National Park and buffer zone). The dispersed spread of settlements and agriculture within the study area make meaningful analysis of these risk factors using the CLIMAFOR method very difficult. 98 APPENDIX 2 - Woodland Inventory, 2007 4. Leakage Assessment Leakage is unintended loss of carbon stocks outside the boundaries of a project resulting directly from the project activity. In the case of woodland conservation in central Mozambique it is possible that this could displace some activities such as fuel wood collection from the conserved area of woodland to other woodland areas. The Plan Vivo system requires that potential displacement of activities within the community should be considered and that activities should be planned to minimise the risk of any negative leakage. These actions should include: - Protection / sustainable management of any woodland areas within the community Implementation of agroforestry measures to provide products such as fuel wood or poles that may no longer be available from within the conserved woodland A plan to monitor leakage on specific woodland areas outside of the woodland conservation area. Where communities have a satisfactory plan for managing leakage risk resulting from the conservation of woodlands there should be no assumption of leakage. 99 APPENDIX 2 - Woodland Inventory, 2007 5. Management Plan Requirements This section sets out the requirements for a Plan Vivo management plan for the conservation of woodlands in Sofala Province, Mozambique. The requirements reflect the general principles of the Plan Vivo system that management plans should be: - Based on local needs and capabilities - Developed through participatory approaches - Agreed by relevant community authorities (where on communal lands) - Simple enough to be understood by the community - Practical to implement with local resources The management plan should contain the following information: Maps of area The map(s) should show the following features: - Location and extent of vegetation categories within the woodland area - Location and extent of other vegetation types within the community boundaries - Elevation - Ownership boundaries (Regulado, communities, private land, public land) - Roads and tracks - Rivers, streams and lakes - Co-ordinate points - Directions and location within the province - Delineation of compartments or divisions within the woodland. If there are compartments or divisions within the woodland that are to be managed for different purposes (e.g. sustainable charcoal production or strict conservation) then these should be marked. a. Governance plan The governance plan should explain who controls the area and how the management of the area will be governed. The governance plan should contain the following information. - Management agreement / community agreement stating that the area of protected woodland is to be established as a community reserve. The management agreement should include a statement relating to the protection of any other woodland areas outside the boundaries of the agreement over which the community has direct control. - Responsible people: a list of the people responsible for the conservation / management of the area and representatives with whom the project operator should communicate. - If possible, a letter of agreement or recognition from the Provincial authorities. b. Activity plan The activity plan will list the activities to be undertaken to manage / conserve the forest area. It should contain the following information. c. List of activities with estimates of time inputs for the protection of the woodland area d. List of activities to protect and restore stocks of carbon within other woodland / vegetation within the community (in order to prevent leakage resulting from the displacement of activities from the protected area of woodland). e. Estimate of cost of implementation f. Estimates of any income from forest products or other outputs (excluding carbon) g. Fire control plan. (separate guidelines to be provided for Fire control plans) 100 APPENDIX 2 - Woodland Inventory, 2007 c. Monitoring plan The monitoring plan will list the indicators to be used to monitor progress with the conservation of the forest area. This should include: a. Annual boundary inspection: a project representative shall patrol the boundary of the community reserve no less than once per year to inspect fire breaks, incursions and integrity of the boundary controls b. Remote sensing plan (annual visual inspection of MODIS NDVI for the area) c. Ecological indicators: a plan for monitoring the presence of key indicator species should be developed to d. Functioning management / governance: the governing committee shall produce a report summarising their activities for the year, problems encountered and e. including fire breaks f. Restoration activities 101 APPENDIX 2 - Woodland Inventory, 2007 6. Calculation of credits for woodland conservation areas 6.1 Project Crediting Procedure The annual crediting of carbon from protected areas approved under the Plan Vivo system should take place following monitoring. Method 1 should be used where a regional risk factor is available. Method 2 – simplified method for forests under high risk, should only be used with permission of BR&D / Earth Carbon Foundation: ∑ Cv(riskcategory) ∗ R(riskcategory) ∗ (44 / 12 ) Method 1. Application of regional baseline risk factors (where a regional risk factor is available): Where: Cv = stock of vulnerable carbon R = rate of deforestation 44/12 = conversion factor from carbon to CO2 Total carbon credited over time should not exceed Cv. The baseline may be re-assessed and case for further crediting reviewed after 25 years. Cv ÷ 25 Method 2. Application of a simplified baseline for forests at high risk. This method should only be used with permission of BR&D / Earth Carbon Foundation: Where: Cv = stock of vulnerable carbon 6.2 Monitoring and indicators for crediting Monitoring should be undertaken by the project operator annually. Crediting for forest conservation may continue as long as project indicators remain Green (see table 6.1). If a project indicator turns amber then crediting should be delayed by 50% for that year until corrective actions have been implemented to the satisfaction of the operator. If a project indicator turns red then crediting should be suspended until issues have been resolved and corrective actions implemented. 102 APPENDIX 2 - Woodland Inventory, 2007 Crediting Continue Governance Governance working effectively Activities Protection activities implemented as per plan Physical damage Woodland / forest conservation consistent with management plan Delay 50% until CARs implemented Suspend crediting until issues resolved Significant breakdown in governance Governance not functioning Protection activities not properly implemented No effective protection activities Loss of woodland / forest at 50% of baseline rate Loss of woodland / forest proceeding at or above baseline scenario Table 6.1. Determinates for crediting carbon to forests. 6.3 Verification of the Monitoring System Independent verification of the monitoring system should be undertaken to ensure that the monitoring of indicators is being carried out to the required Plan Vivo Standard (Plan Vivo, 2007). 103 APPENDIX 2 - Woodland Inventory, 2007 References De Jong, B. H. J. A. Hellier, M. A. Castillo-Santiago and R. Tipper (2005) Application of the ‘Climafor’ Approach to Estimate Baseline Carbon Emissions of a Forest Conservation Project in the Selva Lacandona, Chiapas, Mexico. Mitigation and Adaptation Strategies for Global Change. Vol. 10 (2). De Jong, B. H. J. (2006) Application of the “Climafor” baseline to determine leakage: the case of Scolel Té. Mitigation and Adaptation Strategies for Global Change Grace J, San Jose J, Meir P, et al. (2006) Productivity and carbon fluxes of tropical savannas. Journal of Biogeography 33, 387-400. Grace J, Ryan C, Williams M, with assistance from Silvia Flaherty, Sarah Carter, Joanne Carrie, and contributions from Evelina Sambane, Roberto Zolho, Joao Fernando, William Garrett, Luke Spaddevechia and help from Envirotrade (2007). An inventory of tree species and carbon stocks for the N’hambita Pilot Project, Sofala Province, Mozambique Herd, A. R. C. (2007) Exploring the socio-economic role of charcoal and the potential for sustainable production in the Chicale Regulado, Mozambique. University of Edinburgh Masters Degree dissertation. Unpublished. Laguna, RR., Pérez, J. J., Calderón, O. A., Trevino Garza, E. J., (2006) Estimación del carbono almacenado en un bosque de niebla en Tamaulipas, México. CIENCIA UANL / VOL. IX, No. 2, ABRIL-JUNIO Mushove P (2003) Preliminary inventory of N’hambita community forest, Gorongosa District, Mozambique. Commissioned report, with additions from M Williams, School of GeoSciences, University of Edinburgh. Plan Vivo (2007) Plan Vivo Manual http://www.planvivo.org/fx.planvivo/scheme/manual.aspx [Accessed 19.12.2007] Ryan, Williams and Grace (2007) Guidelines for the rapid assessment of vegetation carbon stocks in the N’hambita area. Unpublished report, School of Geosciences, University of Edinburgh. Sambane E (2005) Above-ground biomass accumulation in fallow fields at the N’hambita community, Mozambique. Unpublished MSc dissertation, School of GeoSciences, University of Edinburgh. 104 APPENDIX 2 - Woodland Inventory, 2007 Appendix 1. Lake Victoria Dar Es Salam Gorongosa National Park, Sofala Province Maputo Figure 1. Distribution of miombo woodland across southern and eastern Africa. 105