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