Mae Chaem, Northern Thailand
“Best bet” Land-use Systems
Thematic reports
Impact of different land uses on biodiversity
Biodiversity and Productivity Assessment for Sustainable
Agroforest Ecosystems
Unique id: IDA1AINC
Source file: D:\Projects\ASB\ASB Country and Thematic
reports\Above ground biodiversity assessmet WG\PART D.xml
Authors: A.N. Gillison, N.Liswanti
--------------------------------------------------------------------------
Period: 4 May to 5 June 1999
Funding agency: ACIAR
(Australia)
CIFOR code: R-BIO-16-1-ICR02
Summary
An ecoregional survey was conducted in the Mae Chaem
watershed of Northern Thailand during 7 May – 3 June
1999. The objectives were: first, to
evaluate a method of rapid biodiversity survey already developed as part of the
ASB system-wide initiative; and second, to use this method to determine how
indicators of biodiversity respond to land-use management under varying natural
resource conditions and by different ethnic groups. A final objective was to
assess how biodiversity varies with profitability and how such information can
be used to develop alternative options to slash and burn and/or regressive,
intensive cropping systems. Knowledge of this kind is needed to develop appropriate
tradeoffs between sustainable biodiversity and economic returns. A range of
land-use types and natural resource gradients under management by Hmong, Karen
and lowland Thai ethnic groups was selected. These traversed mainly the central
and southern regions of the watershed, from 2330 to 500m elevation,including a
wide range of landforms and geological strata.
Building on earlier reconnaissance, 28 representative sites were chosen,
in which 40 x 5m plots were used to record all vascular plant species, plant
functional types (PFTs), vegetation structure, plant uses and site physical
features, including soil physico-chemical variables. The survey procedures were
those applied in other ASB ecoregional sites. Due to logistic difficulties,
birds were censused in only 26 of the 28 sites.
Results indicate that the socioeconomic and biophysical
complexity of the watershed, in particular the overlay of differing historical
land use patterns, is such that it was under-represented by the selected sites.
Although broadly representative of the biophysical aspects of the watershed,
these sites did not account for sufficient variation in land management due to
ethnic differences and recent immigration patterns. Sufficient data to
establish linkages between biodiversity and profitability could not be obtained
in the time available. Despite these limitations, for biodiversity we found
significant correlations between PFTs, soils and birds. As with the Jambi
survey, these indicate that PFTs, rather than species, can be used to
characterise soil nutrient availability.
This is important in assessing agricultural productivity and related
profitability. Plant-bird correlations indicate that PFTs are potentially more
efficient in predicting bird distribution than plant species, but that
vegetation structure is also useful. The most productive dryland agriculture
appears to be associated with intensive, permanent cropping systems with the
lowest recorded biodiversity. In such systems, intensive use of pesticides,
herbicides and fertilizers, together with the loss of key biota and
environmental services, suggests that high profitability may be relatively
transient. For this reason, aspects of integrated pest management, the cultural
and material values attached to non-timber forest products, and environmental
services provided by forested lands must be considered when seeking tradeoffs
between biodiversity and profitability for planned land management. By itself,
species richness is an inappropriate indicator of potential profitability.
Upland, dry, deciduous, Dipterocarp/Oak, open forests and woodlands were found
to be unusually rich in plant species, functional types and birds. When
converted to forestry (Pinus kesiya) plantations, with moderate tending by fire and
slashing, such areas retained most of the plant species and PFTs. Nonetheless,
significant changes in vegetation structure and, thus, animal habitat result in
a 50% reduction in bird species in such plantations. Conversion of such areas
to intensive agricultural cropping systems results in dramatic losses in
biodiversity.
While surveys of this kind can play an important role in
providing baseline information on biodiversity and productivity, the challenge
is to convey this information to planners and managers so that an effective
value can be placed on biodiversity. This is necessary in order to arrive at
acceptable tradeoffs between sustaining biodiversity and to ensure an
acceptable economic return. The greatest challenge will be to communicate this
to recent immigrants in the Mae Chaem region who are unfamiliar with the
existing natural resource base and who are as a result, inclined to value it
least. Significant outputs of the survey are: a refined and readily
transferable survey method and software package for data entry, storage and
analysis; key data for a policy analysis matrix; new and significant additions
to baseline biodiversity data for Thailand; preliminary statistical models for
forecasting impacts of land use on biodiversity; linkages between plant functional
types, soil nutrient availability and bird species richness; effective
‘training of trainers’ and a means of establishing indicators of biodiversity
and agricultural productivity. The methods applied in this survey and the
information acquired are relevant to biodiversity and productivity assessment
within similar montane landscapes in mainland South Asia.
Part D: Mae
Chaem List of Tables, Figures and Annexes
Tables
Table 1 Site
locations and descriptions surveyed within the Mae Chaem
Watershed
Table 2a Site physical environmental features
with symbols used in analyses
Table 2b Summary data, diversity indices and
complexity measures for plant
species and PFTs
Table 2c Site vegetation structural data
Table 3 Presence – absence data for birds (26
plots)
Table 4 Soil physico-chemical data for all
plots
Table 5 Soil correlates with first two MDS
vector scores from each data set
Table 6 Correlations between key plant and soil
attributes for all sites
Table 7 Correlation table for MDS eigenvector scores
for all data sets
Table 8 Plant-based correlates with bird and soil
vectors
Table 9 Correlation of plant diversity indices with
birds and soil
Figures
Fig. 1 Mae Chaem Watershed showing topography and plot
locations
Fig.
2 Classification of all
sites according to presence-absence of all vascular plant species (ref: Table 2a for interpretation of symbols)
Fig.
3 Multi-Dimensional Scaling
(MDS) of all sites according to presence-absence of all vascular plant species
Fig. 4 Classification
of all sites according to species-weighted Plant Functional
Attributes (PFAs)
Fig. 5 MDS
of all sites according to species-weighted Plant Functional Attributes
(PFAs)
Fig.
6 Classification of all
sites according to species-weighted Plant Functional Types (PFTs)
Fig. 7 MDS
of all sites according to species-weighted Plant Functional Types
(PFTs)
Fig. 8 Classification
of 26 sites according to presence-absence of bird species
Fig. 9 MDS of 26 sites according to
presence-absence of bird species
Fig. 10 Classification of all sites
according to soil physico-chemical data
Fig. 11 Classification of all sites
according to soil physico-chemical data
Fig. 12 Correlation between soil pH (H20)
and the first MDS vector for PFTs
(ref: Table 5)
Fig. 13 Correlation between soil Nitrogen
and the first MDS vector for PFTs
(ref: Table 5)
Fig. 14 Relationship between species
richness and PFT richness across all Land
Use Types.
Annexes
Annex I List
of plant families, genera and species together with PFTs and plant
uses, arranged according to
sites. Mae Chaem watershed
Annex II List of proposed sites following initial
reconnaissance of Mae Chaem
watershed
Annex III List of participants at meeting with staff
from Thailand Royal Forestry
Department
Annex IV List of participants in survey of Mae Chaem
watershed
Annex V Itinerary of A.N. Gillison and N. Liswanti
Annex VI Profiles
of species, PFT and spp/PFT richness /area curves for Mae Chaem
Plots showing sample
representativeness and response
characteristics for
each LUT
Annex VII Land use in the Mae Chaem wateshed
Cool, moist, upland, evergreen myrtaceous forest
(2330m) Doi Inthanon
(site 21)
(b) Dry, deciduous,
Dipterocarp/ Oak open forest (site 4).
Coppicing Dipterocarpus
tuberculatus; leaves commonly used for roofing
thatch
Deciduous geophytic liane , edible yam (Dioscorea sp.) in upland, dry,
deciduous, Dipterocarp/Oak open
forest (site 23)
Intensive, upland, permanent cropping systems, Ban Mae
Tho (near site
23)
(f) Mushrooms are an important non-timber forest product in Mae
Chaem
1. Purpose
To undertake a baseline study of biodiversity and
associated profitability under different land use conditions within the Mae
Chaem watershed in Northern Thailand, as part of the
CIFOR commitment to the ICRAF-led, CGIAR system-wide project: Alternatives to Slash and Burn, Phase II.
(See Objectives and Box 1)
To present preliminary results at a methodology
workshop on ‘Environmental Services and
Land Use Change: Bridging the Gap Between Policy and Research ‘ (at Chiang
Mai).
2. Background
The survey was originally proposed and accepted by the ASB
Global Steering Group and ACIAR as being parallel and complementary to the
previously completed, intensive biodiversity baseline study in a lowland,
tropical forested landscape, in Jambi, Central Sumatra.
It forms part of an overall project that has its origins in Phases I and II of
the ICRAF-led consortium on Alternatives to Slash and Burn, a CGIAR System-wide
Global Ecoregional initiative. CIFOR’s contribution as a consortium member has
been to coordinate the biodiversity assessment component (above- and
below-ground), in particular focusing its activities on above-ground
vegetation. Sites were established in three ecoregional benchmark areas: The
Western Amazon basin in Brazil
(Rôndonia and Acre) and Perú (Pucallpa and Yurimaguas); Indonesia
(JambiProvince)
and Cameroon
(Mbalmayo-Makam III). Regional teams were trained in field methods using the
PFA proforma (Gillison, 1988, Gillison and Carpenter, 1997), and data were
collected across a range of different land-use gradients. There was close
correspondence among patterns of richness in plant functional types (PFTs) and
vascular plant species and similar land-use types across regions. Although the
numbers of plots from each benchmark site were relatively few (108 in total),
the observed trends were fairly constant across continents. An intensive survey
in the Jambi lowlands of Sumatra (Gillison and Liswanti,
1999) revealed close correlations between certain components of vegetation
structure and PFAs with certain animal groups. These outcomes provided a
theoretical basis for generating hypotheses about the value of biodiversity
indicators that are being tested in additional sites in the respective study
areas in Indonesia.
They also formed a basis for undertaking a parallel study in mainland Asia, in
the present case in Mae Chaem, Northern Thailand (Fig.
1a), where the methods could be refined, and where we could acquire new and useful biodiversity baseline
data.
In the Jambi survey, a calibrational study for biodiversity
indicators included multiple plant and animal taxa, as well as soils and other
site physical variables. That survey revealed a significant correlation between
plant attributes (vegetation structure, vascular species and Plant Functional
Types or PFTs), several insect groups, and birds. Although the insect taxa constitute a very
important element of biodiversity, their collection was logistically demanding
and subsequent identification costly.
Because birds tend to be conspicuous, usually well known and relatively
easy to document, it was decided that for the Mae Chaem survey only plants and
birds would be surveyed, together with soils and other key physical
environmental variables. It was assumed
the combined information from both
plants and birds would adequately represent the broad biodiversity
pattern and provide a basis for examining linkages with profitability. Site reconnaissance of the Mae Chaem
watershed was completed with a team from
ICRAF in August 1998. From this, 22 potential sites were identified (Annex II)
across a broad range of land use types (LUTs) and natural environments. These
formed the basis for a subsequent survey in May 1999, in which 28 sites were
selected and documented (Table 1).
|
3. Objectives The survey goal was to provide data to
help:
Develop a
generic method of rapid biodiversity assessment that can be used to
characterise and quantify impacts on biodiversity and productivity along
definable land use intensification gradients at the landscape level.
Identify
key biophysical determinants of above-ground biodiversity and productivity
for human needs.
Provide a
means of identifying readily measureable components of biodiversity that can
be used to build acceptable management models by maximising profitability
while sustaining biodiversity, i.e. best bet alternatives for slash and burn.
Define
indicators of thresholds of sustainablility that can be used by managers to
detect conditions that may lead to irreversible decline in resource capacity.
Provide a
means of ready communication and transfer of methods to prospective clients
and beneficiaries (e.g. training, computer and other media).
4. Hypothesis to be tested
Above-ground
biodiversity, as measured by the PFA vegetation assessment proforma, and
avifauna diversity vary with profitability, as measured in dollars per
hectare.
5. Outputs
Field
manual and CD-ROM package for rapid survey proforma using PFAs, for use in
English and Thai.
Transfer of
technology by ‘training the trainers’ in Thailand.
Spatial
models of representative sub-sets of biophysical, ecoregional gradients in
the Mae Chaem watershed in northern Thailand.
Correlative
models of key biophysical features for extraction of biodiversity and total
factor productivity (profitability) indicators.
Means for
coupling biodiversity and productivity with other ASB-related co-located
studies of GHGs and carbon stocks.
Standardised
database for the above.
Capacity
for generating and testing thematic maps of key groups of biota and related
productivity potential under different land use conditions.
Output to
systems and linkages models for ASB.
|
6. Methods
A reconnaissance (Annex II) of the Mae Chaem watershed
during a visit in September, 1998 provided a useful basis for the subsequent
selection of sites in May, 1999 (Table 1, Annexes IV,VII). Meetings with
representatives of the Royal Forestry Department (Annex III) and personnel from
Chiang Mai University (CMU) (socio-economists and soil specialists) provided
useful background for site selection. Criteria for site selection therefore
involved a tradeoff between sampling a representative range of land use types
(LUTs) and capturing the different socioeconomic conditions in the time
available. Using the gradsect site selection procedure (Gillison and
Brewer,1985; Wessels et al.,1998), a primary climate gradient was
selected according to elevation (thermal) and known rainfall pattern. The ICRAF
office at Chiang Mai provided a comprehensive Digital Elevation Model (DEM)
(Fig.1b) that was used to help locate elevational gradients, mapped parent rock types and land cover. Sites
were selected to cover a wide range of LUTs, fallow systems and forest
successional types. Against this biophysical platform, 28 sites were selected
to represent, as far as possible, different ethnic group management approaches
to land use (Hmong, Karen, and lowland Thai).
Their location is shown in Fig 1b.
While these provided a useful coverage of land use conditions in the
watershed they constituted only a very minimal stratification within a highly
complex biophysical and socioeconomic matrix.
Within each site, a 40 x 5m transect was marked out and
vegetation sampled according to the method described in Part C, Section 2 for
the Jambi survey, using the PFA rapid survey proforma. As in earlier surveys,
voucher specimens were taken for every species in every plot and later
identified at the herbarium at the Royal Forestry Department in Bangkok
(Annex I). Unlike the Jambi survey, a new software package (PFAPro) (precursor
to VegClass) was used to help compile field data at the conclusion of each
day. Using local informants, known uses
by the Karen for each species were also recorded in Thai and cross-referenced
(Annex VIII).
Box 1 Activities, outputs and potential application
|
Activities
Outputs
Potential application
1. Site location Baseline digital database; bio- Biodiversity and co-located
physical environmental data. socioeconomic studies
2. Baseline survey Geo-referenced
data Basis for potential mapping
for plant and animal of
key taxa and functional
distribution; link with groups;
spatial models of
GHGs, soil, carbon stocks, biodiversity
and productivity
productivity
data per land for human needs.
use type.
3. Predictive
modelling Indicators of biodiversity Thematic
maps of
and productivity. biodiversity and related
productivity.
Identification of thresholds Use in monitoring resource
of sustainability of conditions
and in forecasting
biodiversity
and productivity. impacts of land
use. Use bio-indicators to forecast productivity potential; Means of identifying ‘best-bet’ alternatives to
Slash and Burn.
4. Field testing
of models Refined models. Same
as 3 above.
5. Comparisons of Generic
method with Cost-efficient basis for
ecoregional
data and broad
ecoregional RBA
and for broad-scale
spatial
models. application. geographic
and physical environmental
comparisons
|
These data are expected to complement ongoing profitability
studies by ICRAF staff in Chiang Mai. Soil samples were collected at 0-15 cm
depth, at three equally spaced locations in each transect, and bulked. These
were stored in double thickness nylon mesh bags and transported to CMU as
vehicles became available. Soil
physico-chemical elements were analysed as indicated in Table 4.
Unlike the Jambi survey, the long distances between sites
created logistic difficulties for sampling avifauna, which meant the
ornithologist worked solo on many occasions. This also limited the time
available at each site, although a dawn and an evening census were conducted
for 26 plots (Table 3), using a similar
species-asymptote approach to that described in Part C, Section 4 of the Jambi
report (Jepson and Djarwadi, 1999). Some
uncertainties in identification limited subsequent data analysis. Data were
analysed using the same methods described for the Jambi survey. This involved
simple linear and polynomial regression combined with standard techniques of
exploratory data analysis. In addition to these, new diversity indices
developed by Gillison et al., (1999) were used to explore correlations
between avifauna and soil features. These included Shannon-Wiener and Simpson
diversity indices for both plant species and PFTs (Table 2b), as well as
Fisher’s alpha index, which is less sensitive to plot size. A ‘complexity’
measure (also described by Gillison et al.,1999 and Gillison, Carpenter
and Thomas, unpubl.) is derived from computing a minimum spanning tree distance
between unique PFTs in each plot. The
present study provided an opportunity to evaluate the utility of each of these
measures for developing biodiversity indicators albeit within a very complex
environment.
7. Results
Records of a total of 1590 plant species (665 unique
species) were obtained from the 28 sites, together with 418 Plant Functional
Types (Annex I). For each record, local
Karen or Thai names were recorded wherever possible, together with uses by the
Karen people. These ‘utility’ data are one component of profitability data that
are being recorded as part of an ongoing project by staff from ICRAF and CMU.
At the completion of the present survey there were insufficient socioeconomic
data available to identify useful connections between biodiversity and
profitability. Nevertheless, variation
in soil nutrients correspond closely with agricultural productivity, the
highest nutrient levels occurring in permanent cropping systems with high
fertilizer input. Exploratory data analysis of the plant, bird, and soil data
sets revealed the following patterns:
Plant species
Classification (Fig. 2) of presence-absence data using a
Bray-Curtis (default Czechanowski) coefficient (Belbin, 1992) produced three
identifiable groups. Two contained both high elevation closed and open forests,
each containing Pinus kesiya and
dipterocarps, but with otherwise marked species differences. These are
represented by a cluster from Doi Inthanon (sites 1,2,3,4,5,6) and another from
Ban Mae Tho and Ban Mae Aep in the West of the watershed (13,23,24,25). The
third group represents a mixed group broadly divided into upland and lowland
fallows (refer symbols in Table 2a). This pattern is broadly reflected in the
ordination (Fig. 3) produced from Multi-dimensional Scaling (MDS).
Plant functional
attributes (PFAs)
The classification of species-weighted, individual PFAs
(Fig. 4), reveals a primary pattern of four groups, represented by mixed upland
forests and open forests combining both Doi Inthanon and Ban Mae Tho and Ban
Mae Aep (1,2,9,13,23,24,25,26) with the next division, containing a mixed group
with Pinus kesiya and upland fallows
(3,6,7,11,12,14,27). The next division contains low elevation, dry, deciduous
Oak/ Dipterocarp woodland (4,5) and fallows derived from this vegetation type
(8,16,18), a montane evergreen forest (21), an agroforestry Coffee plantation
(15), and a mixed group of early to mid-stage upland fallows (10,17,28). The
final group contains three permanent cropping systems (19,20,22). This
relatively confused pattern is brought into clearer perspective in the
ordination (Fig. 5), where the permanent cropping systems (D) are clearly
isolated from the rest of the 28, sites as are the fallows from the
Dipterocarp/Oak woodlands, next to which are one-year, upland fallows. Later
stage, upland fallows containing Pinus
kesiya are shown further along a
gradient towards closed forests and their late-stage fallows. The bottom
left-hand corner of the graph shows two isolated upland open forests,
represented by a species and PFT-rich Dipterocarp/Oak site (23) and a
plantation, both with Pinus kesiya (24).
This ordination reflects a gradient of increasing functional complexity
along a series of fallows of increasing age and variable origin, towards
complex forests and woodlands. These are readily distinguished along a
seasonality/precipitation gradient separating evergreen from deciduous
vegetation types.
Plant functional types
(PFTs)
A classification of 485 unique PFTs (each PFT is a unique
combination of PFAs) in Fig. 6 shows a similar trend to that in Fig. 4. While
interpretation is aided by comparison with the ordination (Fig. 7) the intense
clustering makes detection of sub-clusters difficult. Outliers (including a
mixed, deciduous, community forest represented by an asterisk), climax evergreen forests and
permanent cropping systems are clearly differentiated from the central cluster.
Birds
Presence-absence data are listed in Table 3. As only 26 of
the 28 sites were recorded for birds, no direct comparison can be made with the
plant-based analyses. Nonetheless, the classification (Fig. 8) shows two clear
divisions; one dominated by closed forests and late-stage fallows and the other
by open woodlands and early fallows. Other than this, no other interpretable
pattern can be detected with respect to vegetation or land use type. The ordination (Fig. 9) shows more
interpretable outliers with respect to closed forests, upland dry, deciduous
forests with Pinus kesiya and a mixed
deciduous community forest. Somewhat surprisingly, two closed-secondary forests
(late fallows aged 10, and 20 years respectively) are at opposite ends of the
gradient from primary closed forests (bottom of graph).
Soils
A classification based on all physico-chemical data (Table
4) shows a primary division that accords generally with parent rock type (Fig
10). The first of these (sites 1,2,3,5,16,17,18,21,25,26,27) represent
quartzites, gneisses and amphibolites. The second division indicates soils with
higher fertility and higher pH, derived from quartzites, phyllites and granites
with limestone inliers. Certain permanent cropping or agroforestry systems with
high fertilizer input (15,22) are also included in this second group. Within
these two divisions but with a few exceptions, lower-level clustering reflects
differences in LUT. The ordination (Fig. 11) shows similar patterning, with a
strong fertility gradient expressed along vector 2 from low nutrient upland
forests on gneiss(1,21) to heavily fertilized, base-rich soils in permanent
cropping systems (22) and late fallows with high organic matter and closed
canopy of pioneer trees (9,13).
In seeking predictive connections between soil variables
with other data sets, it is useful to explore correlations between the ‘raw’
soil data and the eigenvector scores from plant and bird data sets. This is
appropriate, partly because of the sparse nature of some data (e.g.
presence-absence rather than abundance data), but also because most of the
variance in each data set can be accounted for in the first two MDS vectors.
For these reasons, it is logical to examine the potential value of correlations
between the ‘raw’ soil data and each of the vectors obtained from the plant and
bird data sets. The results are shown in Table 5, where each soil variable is
compared with the vectors from PFAs, PFTs (modi),
plant species and bird species. An inspection of the table reveals closest
correlation between soil pH, CEC, organic matter, N and the first vector from
the PFA and PFT analyses. Soil P is highly correlated with the first PFT
vector. There are no significant correlations between plant species and soils
or between soils and birds (bird data are missing for sites 20,22). Figure 12
indicates how pH (H20) varies directly with PFT vector scores along
what can be generally assumed to be an increasing gradient of nutrient availability.
Soil N, on the other hand, shows an inverse relationship with PFT vector scores
(Fig.13). While this may run counter to a perceived gradient, the highest N
values arise from deep litter accumulation in forests and late fallows that may
be supported on otherwise low nutrient soils (e.g. high elevation forests
(1,21) in the Doi Inthanon region). Both relationships can be related to
adaptive responses of plants to nutrient availability as expressed in PFAs and
PFTs.
Linear regression between soil attribute values and certain
key plant attributes such as species, PFT richness, and vegetation structural
variables (mean canopy height and basal area) reveals a significant correlation
only between P, PFT and species richness (Table 6). This correlation is heavily
influenced by outlying high P values in site 22 – an intensive, fertilised,
cropping system. On the other hand, very
high correlations can be observed between pH, CEC, organic carbon (OM),
N and vegetation structure. Significant
correlations also exist between vegetation structure and the first of two MDS
soil vectors (Table 7). Soil attributes were poorly correlated with avifauna.
Plant attributes and land use type
Richness in vascular plant species is highly correlated
(R-Sq 86.6%) with richness in PFTs (Fig. 14). This, in turn, corresponds with
gradients of land use intensity. It is reflected along a gradient of richness
extremes from a depauperate, permanent cropping systems, through systems in
which herbicides and organic pesticides are used, to late fallows, to closed
upland forests, to increasingly species-rich and PFT-rich upland dry,
deciduous, open, Dipterocarp/ Oak forests and plantations with Pinus kesiya .
When cumulative species/ area, PFT/area and spp/PFT /area
curves are constructed for each 40x5m plot in 5x5m increments, variation in the
slope of the curves tends to correspond with LUT. This is generally visible across all
cumulative values, but is particularly reflected in the spp/PFT ratios. Annex
VI provides graphic profiles for all LUTs and illustrates a number of cases
where, for example, a species and PFT-poor, high elevation (2330m) evergreen
montane rain forest can be distinguished from a lower elevation (1590m) rain
forest and secondary forest (20 year fallow, 1214m) by the cumulative
species/PFT richness ratio values. In closed canopy, mature forests where
ecological niches are apparently reduced due to the heavily buffered light
regime, more species are contained in fewer PFTs than in more disturbed or
open, patchier habitats, such as dry deciduous woodlands and open forests (e.g.
mixed, deciduous forest (6) and dry, deciduous, Oak woodland (5) which is both
burned and grazed). In species-rich conditions, where habitat is modified by
plantation tending, the high ratio values tend to plateau early (2,13,15,24).
In one year, species-poor fallows in permanent cropping systems, the ratio
values are typically low and the curve tends to asymptote early or else
continues to drop (19, 22). Dry,
intermediate fallows in flooded paddy rice terraces, on the other hand, provide
a unique environment that supports a moderately high number of species and
PFTs, with consistently high ratios (20).
8. Discussion
The highly complex socioeconomic and biophysical environment
in the Mae Chaem watershed generated sampling problems that were not
encountered in surveys in other ecoregional studies (e.g. Cameroon,
Western Amazon basin and Jambi, Central
Sumatra). Although most of the key LUTs were sampled, differences
in land use management due to ethnic background – in some cases by recent
immigrants, could not be accommodated in the time available. In addition, the
considerable distances between sites and the need for dawn and dusk bird
censuses also resulted in truncated bird inventories, with 26 of the 28 plots
sampled by the RFD ornithologist. To add to the field problems, wet weather
limited road access to a number of sites, especially in the Ban Mae Yot area. A representative set of
photographic records of the 28 plots has been scanned for follow-up reporting
and publication and for use by collaborating institutions (ICRAF, RFD, CMU). A
sub-set of these is presented in Annex VII (a,b,c,d,e,f). The complete biophysical
data set (plants, birds, soils) will be made available in electronic media
format to each of the collaborating institutions in accordance with ASB/CIFOR
policy. A backup copy is presently available at CIFOR (N. Liswanti) and with
the primary author at the Center for Biodiversity Management (Yungaburra,
Queensland). Because all site data are
spatially-referenced, it is now possible to generate thematic maps of key
patterns of biodiversity and plant and bird assemblages. Such output will help
identify significant gaps in the knowledge base and indicate locations where
additional samples need to be acquired in order to obtain an adequate data set
for modelling purposes. While not all the intended outputs were achieved, the
following outcomes can be matched against the original objectives:
Develop a
generic method of rapid biodiversity assessment that can be used to
characterise and quantify impacts on biodiversity and productivity along
definable land use intensification gradients at the landscape level.
The gradient-based survey and rapid vegetation recording
method proved efficient in previous multi-taxa surveys, especially in the
multi-taxa survey in Jambi. Subsequent evaluation during the Mae Chaem survey
was limited, due to restricted numbers of samples and limited avifaunal data.
For these reasons, the present survey does not represent an adequate test of
the method. Despite sampling limitations, certain aspects of the survey method
are encouraging. Exploratory data analysis has shown potentially useful
correspondences between plant species, PFAs, PFTs and LUTs. In particular, the
PFT and PFA data sets are more predictive of LUT soil nutrient availability
than are plant species (Tables 5,6), and only the PFAs and PFTs are
significantly correlated with birds (Table 6). Vegetation structure (mean
canopy height and basal area) corresponds more closely with certain soil
variables than do species, PFAs or PFTs (Table 6). While this may be useful
within a region (or watershed), indicators based solely on structure must be
used with caution as structural equivalence between different regions may be
otherwise ecologically distinct. Generic
plant functional attributes and types, on the other hand, can provide a more
sensitive means of ecological comparison when used both within and among
regions (Gillison and Thomas, 1999).
Identify the
key biophysical determinants of above-ground biodiversity and productivity for
human needs.
Analyses to date indicate plant
functional characteristics are most closely associated with avifauna
distribution and exhibit the most significant correlations with certain soil
attributes (pH, CEC, OM, N) as well as with overall soil
information, as expressed by multi-dimensional gradient analysis. While no
direct harvest data were available, discussions with landowners indicated that
base-rich soils and fertilised, permanent cropping systems were most productive
in dryland agriculture. Because LUTs could be characterised by both richness
and composition of functional attributes and functional types, it is logical to
assume these may be used to derive indicators of productivity for human needs.
This is further supported by the correspondence between plant functional
indicators and numbers of non-agricultural plants used by local people for
medicine, food, building and other purposes. In establishing predictive
correlates between LUTs, biodiversity (richness and composition of species and
PFTs and vegetation structure) care must be taken to use indicators that can be
used to discriminate between species depauperate (potentially agriculturally
unproductive) habitats and highly productive intensive, permanent cropping
systems as both are associated with low biodiversity (see next).
Provide a
means of identifying readily measureable components of biodiversity that can be
used to help build acceptable management models by maximising profitability
while sustaining biodiversity, i.e. for ‘best bet’ alternatives to slash and
burn.
There seems little doubt the most
agriculturally productive systems (fertilised, permanent cropping systems) are
the poorest in biodiversity. While this suggests productivity/profitability
potential may be contra-indicated by increasing species and PFT richness it
does not take into account inceasing soil acidity and overall loss of
environmental services that are difficult to quantify but can be inferred from
empirical observations. Lessons learnt in other tropical countries in similar
environments suggest intensive systems of this kind inevitably lead to a
decline in crop production. Outside these intensive systems, simple measures of
richness in species-weighted PFTs, PFAs, mean canopy height and basal area can
be used in a policy analysis matrix framework (in progress by ICRAF, Chiang
Mai) to help indicate acceptable tradeoffs between biodiversity and
profitability. As a first pass in
resource asessment, vegetation structure alone may give a reasonable indication
of both biodiversity and soil nutrient availability in the Mae Chaem watershed.
This can be refined where needed, by adding species-weighted PFAs and PFTs. The
high value placed on non-timber forest products by most ethnic groups suggests
this should be carefully considered when resolving tradeoffs with other
economic benefits gained by land-clearing and fertilising. The only
‘agroforest’ examined in the survey was a coffee plantation containing other
mixed crops (site 15). The owner of this site preferred the economic returns
from this management system to those obtained from more labour-intensive
rice-cropping, citing as part of his reason, the improved ecosystem services
arising from agroforestry.
Define
indicators of thresholds of sustainablility that can be used by managers to
detect conditions that may lead to irreversible regression in resource
capacity.
The survey found no indicators of thresholds of sustainability. This is
partly because complex lag effects in landscape dynamics prevent early
detection of such thresholds. By the time indicators are apparent, it will
almost certainly be too late for remedial action. Nonetheless, the framework of
land-use types within which the study was conducted provides guidelines that can
be used to maintain production while sustaining biodiversity. The most serious
threat to biodiversity loss is wholesale land clearing, with its concomitant
erosion, in-filling of streams,
intensive use of pesticides and herbicides, uncontrolled burning and
hunting. At an absolute minimum, reservoirs of living propagules are needed to
ensure that viable populations of biota are maintained and available to
reoccupy rehabilitated landscapes. Such reservoirs are either severely depleted
or are totally absent in widespread, intensive cropping systems. Provided
sufficient forested patches and representative water catchments can be
maintained within an agricultural mosaic, this will help to maintain
biodiversity and ensure continued access to culturally and economically
important non-timber forest products. There is no simple answer to the question
of what a ‘minimal area’ for conservation management should be. The key to
sustainable management will be a function of the value that landowners place on
a piece of land in terms of its potential for economic return and for
sustaining forest-based livelihood. While indigenous landowners are generally
well aware of the value of forested lands, many immigrants are not. It is the
latter group of land managers who require access to information that will
allow them to more effectively value the
natural resource in order to achieve sustainable biodiversity and
profitability. The present study has provided only a very limited baseline
against which land-use planning can take place and only then for certain areas.
Provide a
means of ready communication and transfer of methods to prospective clients and
beneficiaries (training, computer and other media).
While more sampling is clearly
needed, the results suggest the survey method can be readily applied by persons
with limited technical experience to acquire a basic knowledge of the natural
resource. The training course that preceded the survey was highly successful,
with several participants subsequently using the method for research associated
with land management. The use of a rapid survey proforma has clear potential,
and when coupled with the CIFOR-developed, user-friendly, ‘Vegclass’ software
(previously PFAPro), data entry, data management and limited analysis of
metadata can be undertaken by people with limited training. A training manual
on rapid vegetation classification and survey is in preparation. Its completion
in early 2000 will enable this package to be transferred to collaborating
agencies. It should be made clear the methodology described in this report is
aimed not at the average land owner. It is directed at mid and upper level
management, for whom the objective is to rapidly assess the natural resource
and produce information that can serve as a decision support to facilitate
management adaptation to changing physcial environments and economic climates.
Key data from surveys can be used to assist in evaluating the living resource
base so that appropriate policy interventions can be developed.
Hypotheses tested
Above-ground
biodiversity, as measured by the PFA vegetation assessment proforma, and
avifauna diversity vary with profitability, as measured in dollars per hectare.
In the present study it economic data were insufficient to
evaluate profitability although this study is ongoing. In terms of resource
off-take, it seems that highest profitability in dollar terms will come from
the most intensive, permanent cropping systems where biodiversity is lowest. In
testing this hypothesis, one must consider both long- and short-term
biophysical and economic outcomes. Empirical evidence suggests that in very
intensive and widespread cropping systems, long-term buildup of pesticide and
herbicide residues, soil acidification and/or salinisation, together with
breakdown in soil structure, will ultimately lead to crop decline and an
increase in crop pathogens and pests. This will almost certainly be associated
with a loss in environmental services and contribute to a decline in health and
economic status in the local human populace. It is, therefore, possible to
speculate that, whereas short-term economic gains may offset biodiversity loss
(i.e. an inverse relationship between profitability and biodiversity), in the
long term, there must be a better balance in land management planning at the
outset to ensure the subsequent maintenance of both biodiversity and
environmental services, as well as an acceptable economic return. Under such
conditions, biodiversity may tend to vary directly, rather than inversely, with
profitability.
9.Outcomes
As a result of this survey, a field manual for rapid
vegetation assessment has been modifed and software re-designed to match the
field proforma format. Field trials suggest both the ‘VegClass’ software
(formerly PFApro) and the proforma can be used with relatively little training
by persons with limited background in botany, ecology or computing
applications. A training manual that includes this package is planned for
completion in early 2000, and it is intended this will be made available to all
collaborating agencies in CD-ROM and hard-copy, together with a translation
into Thai.
Because all surveys are spatially referenced, it is now
possible to generate spatial distribution models of key biota for certain areas
of the Mae Chaem watershed. The recent availability of a very comprehensive DEM
with roading infrastructure and other data overlays from ICRAF Chiang Mai now makes exploratory
modeling possible. This database can be used in conjunction with the DOMAIN
spatial modelling software to identify key gaps in the sampling framework for
further survey, especially in the north of the watershed. It can also be used
to help model the impact of specific land management practices in a limited
number of areas.
Correlative models suggest that plant functional
characteristics are potentially useful as indicators of both agricultural and
agroforestry productivity potential (cf. Vanclay et al.,1997) as
such as species richness in birds (cf.Lawtonet
al.,1998). While these models are consistent with outcomes from the
Sumatran baseline study (Gillison and Liswanti, 1999), they require more
rigorous testing across a wider array of land use and tenurial conditions in
the Mae Chaem wateshed.
As at the other ASB, ecoregional sites, the survey has
provided a framework for the co-location of sites for studies of GHGs and
carbon stocks. The regression models
derived from the Jambi survey (see also Hairiah and van Noordwijk, 1999), using
mean canopy height and basal area, can be used to generate a first estimate of
above-ground carbon for the sites investigated in the Mae Chaem survey.
The program ‘VegClass’ has been used to generate a
standardised database for actual and derived meta-data for all 28 sites. The
software can be used to export data to industry-standard Excel and MS-Access
formats, thereby making it readily accessible for statistical analysis. These
and other data sets included in the present report will be made available to
all collaborating agencies. These procedures will facilitate data and
information transfer to ASB systems and linkages models.
Data acquired from the Mae Chaem survey have been added
to those from earlier ecoregional surveys. They will form an integral part of a
system-wide database now being developed and examined for congruent patterns of
biodiversity response to land use impact. This ‘global’ analysis will help
evaluate the generic capacity of the survey method as well as providing
information relevant to other research areas within ASB and elsewhere.
10. Conclusions
The survey identified key biodiversity elements across a
wide range of biophysical and socioeconomic conditions. One such element
consists of the upland dry, deciduous woodlands and open forests. These are
extraordinarily rich in plant species and functional types, and are also
important indigenous reservoirs of non-timber forest products. Conversion of
these to Pinus kesiya
forestry plantations with a moderate tending and fire-hazard reduction program
and limited grazing appears to have only a limited impact on many
naturally-occurring biota. This is partly because of the numerous cryptic
(below-ground storage organs) and other strategies employed by many plants for
fire and drought avoidance and tolerance. The conversion of such resources to
widespread, intensive, permanent agricultural cropping systems increases
productivity but dramatically reduces biological diversity. Empirical evidence suggests this relationship
cannot be sustained in the long term, due to potential crop decline under
increasing soil acidity and pesticide resistance as well as increasing loss in
environmental services and key biota.
Such biota are potentially important players in integrated pest
management, as well as significant contributors to forest-based resources that
are important to the long-term survival of local people. While there were
insufficient data to indicate predictive relationships between profitability
and biodiversity, significant correlations were found between plant functional
types, avifauna and soil nutrient availability. These suggest that plant-based
features may be useful in identifying and monitoring biodiversity and in
deriving indicators of biodiversity and related productivity for human needs.
The extent to which these reflect
profitability needs further exploration. Regionally specific conditions must be
carefully evaluated before any generic conclusions can be reached; models of
profitability developed in Jambi Province of Central Sumatra may not apply in
Mae Chaem, due to socioeconomic and cultural differences. The survey provided a
useful, preliminary framework for identifying important trends in biodiversity
and related impact from varying land management regimes. The complexity of the Mae Chaem watershed,
with its elevational, geological and biological extremes, coupled with
historical overlays of variable land-use practices by different ethnic groups,
requires more comprehensive sampling before any realistic models of land-use
impact on biodiversity can be constructed and tested. The Mae Chaem watershed
has many features in common with the extensive, montane, forested landscapes of
the South-Asia mainland, where some ethnic groups range from the Burmese border
to Northern Vietnam. The results from this survey
suggest that although linkages between profitability and biodiversity were not
identified, the biophysical outcome is sufficient to indicate that the generic
methodology may be applied across this montane South-Asian region. A carefully designed and coordinated survey
using the methodology applied in the Mae Chaem and other ASB ecoregional
surveys would help provide important basic knowledge to the region as a whole,
while focusing on areas of specific developmental interest.
11.
Acknowledgements
The close cooperation and assistance of the Royal Forestry
Department, the ICRAF office and field staff in Chiang Mai and the staff of ChiangMaiUniversity
is gratefully acknowledged. The survey was made possible through funding
supplied by ACIAR.
12. References
Belbin, L.
(1992). PATN Pattern Analysis Package: Technical Reference. CSIRO Div.
Wildlife and Ecology, Canberra.
Carpenter, G.,
Gillison, A.N. , Winter, J. (1993). DOMAIN: a flexible modelling procedure for
mapping potential distributions of plants and animals. Biod. Cons. 2,
667-680.
Gillison,
A.N. (1988). A Plant Functional
Attribute Proforma for Dynamic Vegetation Studies and Natural Resource Surveys.
Tech. Mem. 88/3, Commonwealth Scientific and Industrial Research organization,
Division of water Resources, Canberra.
Gillison, A.N.,
Brewer, K.R.W. (1985). The use of gradient directed transects or gradsects in
natural resource surveys. J. Environ. Manage. 20: 103-127.
Gillison, A.N.,
Carpenter G. (1997). A generic plant functional attribute set and grammar for
dynamic vegetation description and analysis. Funct. Ecol.11:
775-783.
Gillison, A.N.,
Carpenter, G., Thomas, M.R. Plant
functional diversity and complexity: two new complementary measures of species diversity. (Unpubl. 1999)
Gillison, A.N.,
Liswanti, N.L., (Eds.), (1999). An intensive biodiversity baseline study in
Jambi province, Central Sumatra, Indonesia.
In: Gillison, A.N. (coordinator), Above-ground biodiversity assessment working
group summary report 1996-99. Impact on biodiversity of different land uses.
Alternatives to slash and burn project. ICRAF, Nairobi,
pp. 41-53.
Gillison, A.N.,
Liswanti, N., 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., van
Noordwijk, M. (1999). Soil properties
and carbon stocks. In: Gillison, A.N., Liswanti, N.L., (Eds.), An intensive
biodiversity baseline study in Jambi province, Central Sumatra,
Indonesia. In:
Gillison, A.N. (coordinator), Above-ground biodiversity assessment working
group summary report 1996-99. Impact on biodiversity of different land uses.
Alternatives to slash and burn project. ICRAF, Nairobi,
pp. 143-154.
Jepson, P,
Djarwadi (1999). Birds. In: Gillison, A.N., Liswanti, N.L.,
(Eds.), An intensive biodiversity baseline study in Jambi province, Central Sumatra,
Indonesia. In:
Gillison, A.N. (coordinator), Above-ground biodiversity assessment working
group summary report 1996-99. Impact on biodiversity of different land uses.
Alternatives to slash and burn project. ICRAF, Nairobi,
pp. 41-53.
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.
Magurran,
A.E., (1988). Ecological Diversity
and its Measurement. Croom Helm, Lond.
Vanclay, J.K.,
Gillison, A.N., Keenan, R.J. (1996).
Using plant functional attributes to quantify site productivity and
growth patterns in mixed forests. For. Ecol. Manage. 94,
149-163.
Wessels, K.J.,
Van Jaarsveld, A.S., Grimbeek, J.D., Van der Linde, M.J. (1998). An evaluation
of the gradsect biological survey method. Biod. Cons. 7,
1093-1121.
Observers: AG: Andy Gillison; NL: Nining Liswanti; RP: Rachun Pooma; JP: Jim Peters; KS:
Kitiya Suriya; W: Wanaree; PN: Prathuang Narintrangkol; P: Prasit Wang. S.Dpt = soil depth (cm); Ltr = litter depth (cm). Symbols: HP = High-elevation >1000m
incl. Pinus , L = Low elevation <1000m, Af = agroforest, = closed forest, = upland early fallow system , = intensive cropping system, early fallow
from dry deciduous forest, = dry,
deciduous forest frequently with Pinus
kesiya; numerals indicate fallow period in years.