Mbalmayo, Cameroon
“Best bet” Land-use Systems
Thematic reports
Impact of different land uses on biodiversity
Biodiversity and Productivity Assessment for Sustainable
Agroforest Ecosystems
Unique id: IDA1AIXB
Source file: D:\Projects\ASB\ASB Country and Thematic
reports\Above ground biodiversity assessmet WG\CamRep4.xml
Authors: A.N. Gillison
--------------------------------------------------------------------------
Period: August 1996 to November 1999
Funding agency: DANIDA; CIFOR
code: R-BIO-14-1-DNK01
Summary
This study is part of an integrated approach to assessing
the impact of land use on biodiversity across tropical forested landscapes in
ecoregional study sites in the Western Amazon basin, Thailand,
Indonesia and Cameroon
in humid tropical West Africa. Activities were focussed
initially to respond to a contractual requirement for the GEF. More recently,
this has changed to an examination of links between biodiversity and
profitability. In each study area land use intensity gradients were chosen to match
as closely as possible similar land uses ranging from closed canopy forest,
agroforestry plantations with and without indigenous species, monoculture
plantations and agricultural cropping systems such as Cassava, through to
degraded grasslands. Comprehensive gradients of this kind provide a more
efficient context for simulating the influence of land use on the performance
of certain sets of key biota than intensive studies of isolated, individual
land use types (LUTs).The aim of the study was to provide managers and policy
planners with the requisite tools to rapidly assess the natural resource as a
means of deciding which land use options might provide the most acceptable
tradeoffs between sustainable biodiversity and economic returns - rrather than to
suggest a single alternative solution to regressive slash and burn. The same generic methods of plant-based
biodiversity assessment were used in each country. In Cameroon
the study area was originally concerned with humid tropical lowland forested landscapes
at Mbalmayo in the south. This was later extended to include sub-humid savanna
in the Makam III area about 200 km north of Yaounde.
The addition of sites reflecting land use in savannas as well as humid forests,
enhanced the interpretation of biodiversity response to human-induced impacts
within Cameroon
and facilitated comparison between the ecoregional gradients. For the Cameroon
study as with other regions, a digital elevation model (DEM) was prepared by
CIFOR to assist with the selection of representative sites. The mapping section
of ONADEF in Yaounde also assisted
with the provision of vegetation maps
and satellite (SPOT) coverage. Following a reconnaissance visit in August 1996
when three sites were documented, a subsequent training workshop on
biodiversity assessments with 20 participants was held in Mbalmayo in June
1997. Following this workshop 18 additional sites were recorded along the
Mbalmayo-Makam III rainfall gradient. As with the other ecoregional sites, the
survey used a proforma designed specifically for rapid assessment of site
physical features and vegetation (structure, species and Plant Functional Types
– PFTs). Results from the survey tend to
confirm that vegetation responds in a similar way to similar land use in
different countries. This is reflected in ranking vegetation according to an index (the "V" index) that is
computed from a multi-dimensional scaling (MDS) analysis of vegetation elements that are known to vary
significantly with the distribution of certain animal groups, soil nutrient
availability and above-ground carbon. The analysis provided a useful context
for assessing the role of early fallows in slash and burn dynamics following
forest removal and under long-term, short-cycle fallows. In Cameroon,
fallows tend to be dominated by the asteraceous, exotic weed Chromolaena odorata. While aggressive
and difficult to control, the study suggests C. odorata may contribute to increased litter production and soil
amelioration. Its effect on biodiversity is uncertain. This phenomenon has a
parallel in indomalesia and Latin America and the
ecological role of this and other similar Asteraceae in early fallows requires
additional research. Appropriate
alternatives to slash and burn in more-or-less forested landscapes suggest Cacao plantations managed with minimum
removal of other forest species may offer an appropriate tradeoff between
sustaining biodiversity and maintaining an acceptable economic return including
sequestration of carbon. This is analagous to the ‘Jungle Rubber’ in Indonesia
and certain mixed agroforests in the WesternAmazonBasin
in Brazil and Peru.
While other farming and cropping systems involving Oil Palm, timber plantations
and extractive forest reserves remain to be studied in more detail in Cameroon,
results indicate the survey method can be applied with limited of personnel.
The information acquired can be applied to characterise the existing resource
and also to forecast the likely impacts of specific land uses on biodiversity
and potential profitability. As in other countries, the “V” index and the
values of the attributes used to generate this index may be usefully incorporated into a Policy Analysis Matrix.
This study has provided a potentially cost-effective operational framework for
examining linkages between biodiversity and profitability both in Cameroon
and other areas of humid and sub-humid, tropical, Central and West
Africa.
Introduction
The loss of plant and animal habitat in many tropical
countries continues unabated because as yet, there is no direct means of
attaching a value to biological diversity and thereby balancing its loss
against economic gain. Despite global
concern, there is no operational definition for biodiversity and no ready means
of measuring biodiversity in tropical landscapes where most of the world's
terrestrial plant and animal species reside. The present study is part of an
integrated series of investigations in other tropical lands (mainly Indonesia,
Thailand and the Western Amazon Basin
of Perú and Brazil).
By using the same methods of data collection to study vegetation response along
similar gradients of land use intensity within regional environments it has
been possible to identify similar trends that reflect impact of land use on
biodiversity. Most land managers and land owners aim to maximise economic
return with minimum effort and minimum impact on the resource that supports
that activity. While most are concerned about the destruction of habitat and
declining harvests, there is rarely any incentive to modify land use intensity
unless it can be shown that there is a commensurate economic tradeoff. Research in mainly developed countries tends
to support the concept that productivity in farming systems is enhanced by
sustaining biodiversity. This is also reflected in improvements in integrated
pest management and in the maintenance of farming systems that aim to maximise
the number of species. The challenge for research is to demonstrate the
consequence of differing land use practices on biodiversity and related
productivity for human needs and to provide tools that can be used by managers
and planners to adapt management to achieve an acceptable balance can be
maintained between profitability and the natural resource. Many of the external
factors that influence farmers behaviour can be unpredictable ( el niño events,
market shifts, changes in government…). Under such circumstances farmers must
be able to respond quickly to such changes in order to maximise profitability
and maintain quality of life. If there are tools available to rapidly evaluate
the natural resource then this will greatly assist planners and managers alike.
The aim of this study is to explore methods of rapid resource appraisal with
respect to biodiversity and to provide a cost efficient method that can be
readily transferred to the stakeholders concerned with integrating natural
resource management.
Training workshop
In order to assist in-country partners and other associated
interested parties a training workshop was held from 27-29 May 1997 that
included 20 participants from various locations and institutions in Cameroon (Annex I). Course work included training in
aspects of survey design, site location recording of site physical features,
vegetation structure and plant species and PFTs using the rapid survey proforma
described belowThe vegetation mosaic around the Mbalmayo station provided ample
conditions for field training. As with a previous workshop in Pucallpa,
Perú, the participants were divided into four groups and collected data
independently. While the sites used (plantation forest with secondary growth; a
forest-cocoa plantation and a recent subsistence garden) were by their very
nature quite different, the groups showed an acceptable level of convergence as
indicated by subsequent pattern analysis that improved when some obvious
inconsistencies in the data were removed. Since the workshop the University
of Yaounde via Dr Zapfack Louis has
implemented the technique in a regional vegetation survey.
Methods
A detailed account of the methods used in this survey is
described by Gillison and Liswanti (1999). In general terms the survey
technique uses gradient-based transects or gradsects (Gillison and Brewer,
1985; Wessels et al., 1998) in which
a hierarchy of physical environments is used to frame site location (e.g.
rainfall seasonality, parent rock type, drainage patterns, land cover, soil
catenary sequences, land use pattern..). It has been shown elsewhere (Gillison
and Brewer, 1985; Wessels et al.,
1998) that gradsect sampling improves the efficiency of recovering information
about the distribution of biota. The primary ecoregional gradient within Cameroon
is dominated by rainfall and secondly by soil type and drainage patterns that
in turn influence land use pattern. This has parallels in other study areas and
provided the framework for the present study. In the present study additional
physical environmental information was provided by IITA and ONADEF about land
use and land cover including vegetation. For the Mbalmayo benchmark site
developed by IITA (Box 1) there is SPOT Imagery available that has been
classified according to perceived LUSs and land cover classes. In order to
provide added environmental context for biodiversity assessment the benchmark site
was extended to include the sub-humid savannas approximately 200km north of Yaounde
towards the sahelian zone. Using topograpic maps supplied by IITA, CIFOR
compiled a digital elevation model (DEM) that was also used to assess site
representativeness and as an initial spatially-referenced platform for the
acquisition of additional physical environmental data. A copy of the DEM
database has been lodged with IITA and a master copy is available at CIFOR.
The following method of recording site physical features and
vegetation was applied: At each site a 40 x 5m transect is laid out in which
all vascular plant species are recorded together with plant functional types
(PFTs) (formal combinations of specific plant functional attributes or PFAs
that are primarily morphological adaptations of plant rsponse to environment).
The recording of PFTs in addition to species complements the taxonomic
information in a way that helps to interpret vegetation response to
environment. Because PFTs are independent of species (more than one PFT can
occur within a species and more than one species can occur as the same PFT),
from a biodiversity standpoint they provide potentially useful, additional
information that indicates a genetically based response to environment (see also
Vanclay et al., 1997). By recording
PFTs using a generic protocol, it becomes possible to compare for example, data
from geographically remote locations where the species my differ but where
environments and plant response may be similar. The underlying ecophysiological
rationale and supporting theory for the use of PFTs of functional modi is described by Gillison and
Carpenter (1997). In addition to species and PFTs vegetation structure (mean
canopy height, crown cover %, litter depth, furcation index, cover-abundance of
woody plants <1.5m tall and cover-abundance of bryophytes) is also recorded.
Site physical variables include geo-reference with a portable, global
positioning system, slope percent, aspect (deg.), elevation (m), parent rock
type, soil type and soil depth.
Additional notes include names of observers, relevant site history, and
a profile sketch of the vegetation along the 40m transect.
At the time of the survey a computer software program FUNDAT
was used to collate and store field data. FUNDAT was subsequently replaced with
PFAPro, a more efficient, Windows-based
program, that includes an error-checking protocol, a facility for
tabulation of data and graphs and the estimation of several ecological diversity
measures based on PFTs. Photographs of each site were taken and filed together
with the cross-referenced data.
Visits
were made to field sites at Mengomo, Akok, Mbalmayo, Awae, Nkol Foulou,
Nkometou, Bafia and Batoum II. A range of about 500km from humid forest to
savanna. I was accompanied by three Cameroonian botanists: Dr Zapfack Louis (YaoundeUniv.),
Mr Kaufani Anacletus (YaoundeUniv.
retired ) and Dr Bonaventure Sonke (DjaForest
reserve project), Mr Nicodeme Tchamou and Ms Martine Ngobo (both research
assistants to Stephan Weise). We recorded 18 plots focusing mainly on a range
from Imperata grasslands and Lophira & Butyrospermum shrub savanna through different slash and burn fallow
sequences to closed-canopy (mostly secondary) forest. Included were ‘Jungle’ Cocoa
plantations and sedentary fallows dominated by Chromolaena odorata. To
these were added three plots recorded from an earlier reconnaissance visit in
1996. Site physical and locational data are listed in Table 1, plant species,
PFTs and PFT diversity indices (Giillison and Carpenter, unpubl. 1999) are
listed in Table 2 and vegetation structure in Table 3. The first ten sites were
co-located with Dr Cheryl Palm (TSBF) who made assessments of above ground
carbon. Listings of all species and PFTS collected for each site are contained
in Annex II. All plot data have been stored in MS ACCESS format and a copy of
field sheets was left at IITA Nkolbisson together with copies in electronic
media. Copies of the more recent data conversions via PFAPro have been emailed
to these repositories.
Data
analyses included standard regression measures and exploratory data analysis
using the PATN program (Belbin, 1992). In addition to these a single index that
represented key elements of vegetation structure, total plant species, total
PFTs per plot and their ratios, was extracted using multi-dimensional scaling
(MDS) as described by Gillison (1999) and Box 2 below. This is termed a
"V" index and is an exploratory attempt to seek a relative ranking of
vegetation that may have the potential to serve as a useful correlate for
biodiversity and site productivity potential or carbon sequestration. To assist
in the construction of a Policy Analysis Matrix (PAM) data used in the
compilation of the "V" index
were supplied to Dr J. Gockowski (see Cameroon
country report). These included vegetation structure, species and PFTs and
V-Indices arranged according to site (Table 4) and according to the ranked
"V" index itself (Table 5). Cumulative species/area, PFT/area and
spp/PFT ratio/area curves were plotted for each contiguous 5x5 metre quadrat
along the 40m transect using PFAPro. These are displayed in Annex III. The
curves provide additional insight into the level of representativeness of a
40x5m plot in each of the LUTs.
|
|
Box 1
The IITA Benchmark program
The IITA benchmark
concept was initiated by Doyle Baker (IITA). After a countrywide study
about 2/3rds of the current Mbalmayo benchmark template was in place by
1993. ASB provided the stimulus to follow through and this resulted in
IITA/IRA collaboration and use by ASB (mainly environmental). EPHTA (Ecoregional Program for Humid and
Sub-Humid Tropics of Sub-Saharan Africa) is coordinated by IITA now
cooperates with ASB on benchmark research. Both programs are interested in
developing alternative farming systems (not necessarily with environment or
biodiversity in mind). They aim at maintaining a resource base for
production systems – e.g. managing short-fallow systems and community
forestry. Another area of rsearch is concerned with diversification of land
use systems and conservation of the national resource base. They are
targetting specific LUSs and are especially focussed on:
Renewal of traditional
plantations
Development of home
garden systems
Annual and perennial crop
multistrata systems
Northern Guinea savanna
(N. Nigeria)
Southern Guinea savanna
(Côte d’Ivoire)
Coastal savanna (S.Benin)
Forest pilot sites will
be located in Gabon, Côte d’Ivoire, Republic of the Congo, Congo
(Brazzaville) and for ASB possibly another BM in Ghana. GIS plays a
critical role in the ecoregional programs and uses area-based sampling on a
10’X10’ (80km x 80km) grid cell for ASB. Approximately 30-40 villages are
typically involved (one village per grid cell). These are characterised
according to ‘resource use domains’. From these are selected representative
sites for more detailed investigation at higher spatial resolution
according to farmer circumstances. The studies are entirely correlative
with no mechanistic or dynamic models. Linear programming is used to
manipulate tradeoffs between biophysical and socioeconomic conditions
|
|
Results
Both
classification and ordination (MDS) (Figs 1a & 1b) reveal three readily
distinguishable clusters of LUTs. The first is represented by closed canopy
forest including both 'jungle' Cacao
and 'Plantation' Cacao. The second is characterised by
the Chromolaena - dominated fallow
systems with sub-structure suggesting some differentiation according to time
since the original forest was 'opened' by slash and burn. The final category
consists of the savanna woodland sites including a Cassava garden in savanna
and two newly established food gardens within the Mbalmayo rain forest zone. If
total species richness is regressed on total richness of PFTs the result is a
highly significant statistical correlation (Fig. 2) where plot distribution
reflects land use intensity within a broad climate gradient. It is informative
to note that relatively high plant biodiversity is contained in both the Cacao plots. The fact that the
'plantation' Cacao also figures
highly is due to the presence of a number of ground-dwelling, semi-herbaceous,
weedy species and close proximity to surrounding, highly disturbed secondary
forest.
The
cumulative species, PFT and ratio /area curves (Annex III) give an indication
of the level of representativeness of each LUT. Steep curves suggest additional
samples are needed. Of particular interest are the relatively depressed curves
of the long-term, short-cycle fallows compared with those of the richer fallows
that immediately follow forest clearance. The ratios of species to PFTs show
that whereas in the more complex forests more species collapse into PFTs, in
the recent gardens and savannas these are flattened. This reflects the greater
varuiability of available ecological niches per species as well as the higher
number of adaptive combinations or PFTs. In the savanna plots within the first quadrat
the ratio becomes stable and this may be an indication of the relative
equilibrium of savanna dynamics compared with those of the more complex,
disturbed forests and older fallows. The ecological meaning of these profiles
requires further investigation.
The
'V' index (Fig. 3) illustrates the relative position of each plot in terms of
increasing structural complexity and richness in both species and PFTs. It is
of interest to note that in this instance, 'Plantation'
Cacao is much lower in the sequence (see also Box
2). When V Index values are regressed against above-ground
average carbon (C. Palm unpubl. 1999) there is a very significant statistical
relationship (Fig.4). This is also mirrored in similarly high correlations
between above-ground carbon, mean canopy height and basal area.
Discussion
The
sequences revealed very similar patterns in vascular plant species richness to
those found in Latin America (Yucatan and Perú /Brazil).
However richness in functional groups appears to be less than in either the Americas
or Indonesia. As with the other benchmarks relative to the
more mature forests, there appears to be a surprising increase in species and
functional richness in the early (1-3 year) fallows dominated by Asteraceae
(most notably by the exotic C. odorata).
This is of potential interest from a farming systems viewpoint as it suggests
that despite claims that C. odorata
is a repressive exotic weed it may have a positive contribution to plant-based
biodiversity and may enhance soil nutrient availabilty through increased litter
deposition and reduced soil exposure.
Exploratory data analysis combined with the V index gives a strong indication
that Cacao grown within secondary
forest is closely associated with high plant biodiversity. The extent to which
this may depress yield relative to more efficiently tended Cacao plantations needs to be considered when assessing trade-offs
and in identifying best-bet alternatives to slash and burn. There appear to be
close parallels between levels of biodiversity in Cacao grown in so-called 'jungle' conditions and that of 'Jungle
Rubber' of Sumatra
All
the forest types examined in this study were heavily used by the local people
for hunting, fuel and medicinal resources. From this point of view they are
highly regarded as a potentially rich source of extractable
non-timber-forest-products. The inclusion of savanna LUTs has improved the
ecological and environmental context needed to assess biodiversity overall and
provided an additional, spatially-referenced framework for spatial modelling of
actual and potential land use impact on plant-based biodiversity. Should there
be a need to consider the likely effects of climate change then this extended
gradient will be of potential use in modelling different climate change
scenarios. A recent report of a biodiversity baseline study (Lawton et al., 1999) found little evidence to
support the use of one taxon to predict the presence of another. That study was
restricted to a localised, rain forest land use mosaic in Mbalmayo and did not
use plant indicators. Because most or all animal taxa depend on plants for survival
and because the distribution of many taxa extend beyond the immediate bounds of
closed forest, it is likely that predictive performance could have been
considerably improved had plants been included and had the samples been
extended to a wider array of LUTs as in the present study.
Where
estimates of above-ground carbon are logistically demanding evidence from this
study and from the Sumatran benchmark study (Gillison and Liswanti, 1999)
suggest that alternative, easy to measure surrogates may be available through
the use of mean canopy height and basal area. Similarly a V index may be
potentially useful but unlike simple estimates of structure this requires much
more sophisticated computation and more variables.
The
training workshop held at Mbalmayo was considered successful by all who
attended. Feedback from participants also helped to refine the framework for
subsequent workshops that have been held in Latin America, Thailand
and Vietnam.
Conclusions
The
rapid survey method of recording site physical features and vegetation has
successfully identified a variety of land use impacts on plant-based
biodiversity. Results from more intensive multi-taxa and soil studies in Indonesia
and Thailand
suggest that a more cost effective approach for such surveys should include soil
analyses wherever possible. In each of the ecoregional sites including Cameroon,
the close correspondence between the measures used to characterise vegetation
and estimates of above-ground carbon (Gillison and Liswanti, 1999;C. Palm,
1999; Hairiah and van Noordwijk, 1999) are consistent with other correlations
between soil nutrient availability, vegetation and anilmal habitat. Indicators
based on combinations of vegetation structure (mean canopy height and basal
area), total plant species richness, richness of PFTs and ratios of these
richness estimates can be used to quickly identify relative impacts of land use
on biodiversity along a land use intensity gradient. The 'V' index provides one
relative measure of this impact that appears to be more or less consistent
between the ecoregional benchmarks. In circumstances where botanical assistance
is unavailable or where survey time is limited, simple estimates of canopy
height and basal area will provide most of the information needed about the
'heath' status of a site and its likely response to variations in land use. But
whereas this may be locally appropriate, sites with similar vegetation
structure between geographically remote areas are not likely to be ecologically
equivalent. For such inter-regional comparisons additional species and PFT data
are necessary.
Acknowledgements
References
Gillison, A.N.
and Brewer, K.R.W. (1985). The use of gradient directed transects or gradsects
in natural resource surveys. Journal of
Environmental Management20: 103-127.
Gillison, A.N.
and Carpenter G. (1997). A generic plant functional attribute set and grammar
for dynamic vegetation description and analysis. Functional Ecology11:
775-783.
Gillison, A.N. (1997). In 1997 ASB Annual Review Meeting
report. Unpubl.
Gillison, A.N. and Liswanti, N. (compilers). An intensive
biodiversity baseline study in Jambi province, Central Sumatra, Indonesia. In:
A.N. Gillison (coordinator) Alternatives to Slasha nd Burn Project: Phase II.
Above-ground biodiversity assessment working group summary report 1996-99.
ICRAF, Nairobi
Gillison, A.N., Liswanti, N. and Arief-Rachman, I.
(1996). Rapid Ecological Assessment, KerinciSeblatNational Park Buffer Zone, Central Sumatra: Report for
Plant Ecology. CIFOR Working Paper No. 14., Bogor,
Indonesia.
Hairiah, K., and van Noordwijk, M. (1999). Soil properties
and carbon stocks. Section 10 in: Gillison, A.N. and Liswanti, N. (compilers).
An intensive biodiversity baseline study in Jambi province, Central Sumatra,
Indonesia. In: A.N. Gillison (coordinator) Alternatives to Slasha nd Burn
Project: Phase II. Above-ground biodiversity assessment working group summary
report 1996-99. ICRAF, Nairobi.
Lawton, J.H., Bignell, D.E., Bolton,
B., Bloemers, G.F., Eggleton, P., Hammond,
P.M., Hodda, M., Holt, R.D., Larsen, T.B., Mawdsley, N.A., Stork., Srivastiva,
D.S. , Watt, A.D., (1998). Biodiversity
inventories, indicator taxa and effects of habitat modification in tropical
forest. Nature, 391, 72-76.
Vanclay, J.K.,
Gillison, A.N. and Keenan, R.J. (1997). Using Plant Functional Attributes to
quantify site productivity and growth patterns in mixed forests. Forest Ecology and Management94,
149-163
Wessels, K.J.,
Van Jaarsveld, A.S., Grimbeek, J.D. and Van der Linde, M.J. (1998). An
evaluation of the gradsect biological survey method. Biodiversity and Conservation7:
1093-1121.
