Content uploaded by Arif Budiman
Author content
All content in this area was uploaded by Arif Budiman on Oct 08, 2014
Content may be subject to copyright.
ORIGINAL ARTICLE
Planning hydrological restoration of peatlands
in Indonesia to mitigate carbon dioxide emissions
Julia Jaenicke &Henk Wösten &Arif Budiman &
Florian Siegert
Received: 18 September 2009 /Accepted: 5 January 2010 /
Published online: 3 February 2010
#The Author(s) 2010. This article is published with open access at Springerlink.com
Abstract Extensive degradation of Indonesian peatlands by deforestation, drainage and
recurrent fires causes release of huge amounts of peat soil carbon to the atmosphere.
Construction of drainage canals is associated with conversion to other land uses, especially
plantations of oil palm and pulpwood trees, and with widespread illegal logging to facilitate
timber transport. A lowering of the groundwater level leads to an increase in oxidation and
subsidence of peat. Therefore, the groundwater level is the main control on carbon dioxide
emissions from peatlands. Restoring the peatland hydrology is the only way to prevent peat
oxidation and mitigate CO
2
emissions. In this study we present a strategy for improved
planning of rewetting measures by dam constructions. The study area is a vast peatland
with limited accessibility in Central Kalimantan, Indonesia. Field inventory and remote
sensing data are used to generate a detailed 3D model of the peat dome and a hydrological
model predicts the rise in groundwater levels once dams have been constructed. Successful
rewetting of a 590 km² large area of drained peat swamp forest could result in mitigated
emissions of 1.4–1.6 Mt CO
2
yearly. This equates to 6% of the carbon dioxide emissions by
civil aviation in the European Union in 2006 and can be achieved with relatively small
efforts and at low costs. The proposed methodology allows a detailed planning of
hydrological restoration of peatlands with interesting impacts on carbon trading for the
voluntary carbon market.
Keywords Dam construction .Drainage canal .Groundwater level rise .
Hydrological modelling .Illegal logging
Mitig Adapt Strateg Glob Change (2010) 15:223–239
DOI 10.1007/s11027-010-9214-5
J. Jaenicke (*):F. Siegert
GeoBio Center, Ludwig-Maximilians-University Munich & Remote Sensing Solutions GmbH,
Wörthstrasse 48, 81667 München, Germany
e-mail: jaenicke@rssgmbh.de
H. Wösten
Alterra, Wageningen University and Research Centre, P.O. Box 47, 6700 AA Wageningen,
The Netherlands
A. Budiman
WWF-Indonesia, Kantor Taman A9, Unit A-1, Kawasan Mega Kuningan, Jakarta 12950, Indonesia
1 Introduction
Of the tropical peatlands worldwide 70% are located in Southeast Asia, 22 million ha of
these in coastal and sub-coastal regions on the islands of Sumatra, Borneo and West Papua
in Indonesia (Page and Banks 2007). Tropical peat is an accumulation of partially decayed
organic matter which has been formed over thousands of years in waterlogged environ-
ments that lack oxygen. In Indonesia peat deposits with up to 20 m in thickness store huge
amounts of carbon (Whitten et al. 1987; Sorensen 1993; Jaenicke et al. 2008). Under
undisturbed conditions, tropical peatlands are covered with peat swamp forests which
comprise ecosystems with many endemic species and high biodiversity. Since the 1980s the
Indonesian peatlands have been extensively logged, drained and converted to plantation
estates as a result of economic development (Curran et al. 2004; Rieley and Page 2005;
Hansen et al. 2009). In Southeast Asia 12 million ha of peatlands are currently deforested
and drained, including over 1.5 million ha of tropical peat swamp forests in the Indonesian
province of Central Kalimantan (Hooijer et al. 2006). Canals and ditches are not only built
to control and lower the groundwater level for plantation operations and small-scale
agriculture but also to facilitate access to peat swamp forests and to extract timber logs. The
extent of these diverse canals and thus the impact on drainage depth varies. For example,
the drainage depth of oil palm plantations in Sarawak, Malaysia, is −60 cm (Melling et al.
2005) whereas it is about −30 cm in farm fields in Central Kalimantan, Indonesia
(Jauhiainen et al. 2004).
Once peat is drained, it oxidises due to microbial activity and releases stored carbon to
the atmosphere as carbon dioxide. This ongoing rapid peat decomposition leads to the
irreversible process of peatland subsidence. In developed peat, drainage depth is related to
peat organic matter oxidation rates and peat subsidence (Wösten et al. 1997; Furukawa et
al. 2005). On average 60% of peat subsidence is caused by oxidation and 40% by
irreversible drying or shrinkage of the peat (Wösten et al. 1997). Lowering the groundwater
level which naturally is close to the peat surface throughout the year while fluctuating with
the intensity and frequency of rainfall, results in an increase in CO
2
emissions. In a recent
review it is estimated that an increase of drainage depth by 10 cm results in the emission of
about 9 t CO
2
ha
−1
a
−1
(Couwenberg et al. 2009).
Another severe consequence of drainage is the occurrence of peat fires. Under natural
circumstances peat consists of 90% water and 10% plant matter and hardly ever burns.
However, if the groundwater level falls below a critical threshold of −40 cm, the dry peat
surface becomes susceptible to fire (Takahashi et al. 2003;Usupetal.2004; Wösten et al.
2008). Fires are most severe during El Niño events, as in 1997/98 when about 2.4–6.8 million
ha of peatlands burnt in Indonesia releasing huge amounts of the greenhouse gas CO
2
(Page
et al. 2002; Van der Werf et al. 2008). With a groundwater level at about −100 cm the burn
depth was estimated to be 51 cm on average releasing up to 9.4 Gt of carbon dioxide in
Indonesia (Page et al. 2002). The failed Mega Rice Project, a resettlement project initiated in
1995 in Central Kalimantan, contributed largely to this ecological devastation. Drainage
canals, up to 30 m wide and 10 m deep, with a combined length of 4,500 km disrupted the
peatland ecosystem over an area of more than 1 million ha. There exists a positive feedback
of recurrent fires which leads to progressive forest degradation and continuous release of CO
2
with regional and global consequences for the environment and climate (Siegert et al. 2001;
Cochrane 2003; Langner et al. 2007).
Complete rewetting is the only way to prevent fires and peat oxidation by microbial
decomposition. Due to its high permeability peat acts as a sponge, i.e. it shrinks when dried
and swells when rewetted, unless water contents fall below a threshold value at which
224 Mitig Adapt Strateg Glob Change (2010) 15:223–239
irreversible drying occurs (Wösten et al. 2008). Therefore, one of the most important peatland
restoration measures is blocking of drainage canals by dams and thus raising the groundwater
level of the surrounding peatland. Damming activities performed in the former Mega Rice
Project area, in Sebangau National Park and in Merang peatland of South Sumatra have
shown that the water retention upstream of dams could be increased thereby decreasing peat
desiccation during the dry season (Suryadiputra et al. 2005;CKPP2008;Jauhiainenetal.
2008). Few rehabilitation attempts have been undertaken in the past (Page et al. 2008),
however within the context of ongoing discussions concerning climate change tropical
peatlands have now been recognised as major sources of greenhouse gas emissions (Rieley
and Page 2005; Hooijer et al. 2006;Uryuetal.2008). The carbon content of the peat soils in
Indonesia is about 18 times higher than that of pristine peat swamp forest (Jaenicke et al.
2008). Therefore, peatland rehabilitation projects are of high interest for carbon trading on the
voluntary carbon market. While peat oxidation causes continuous release of carbon dioxide,
peat fires are the source of huge amounts of CO
2
emissions in short time. These emissions
can be mitigated if peatland rewetting measures are implemented.
The objective of this study was the development of an efficient and cost-effective
methodology to plan hydrological restoration of disturbed tropical peatlands. The study was
conducted in the Sebangau catchment in Central Kalimantan under supervision of the
World Wildlife Fund (WWF) aiming at mitigation of carbon dioxide emissions. The surface
of tropical peat shows little slope; with gradients of only 0.2–1 m per kilometre in the centre
they appear virtually flat (Page et al. 1999; Rieley and Page 2005). In addition, the
Sebangau peat dome is covered with dense vegetation which makes an in situ assessment of
the entire hydrology impossible. The proposed restoration programme comprises several
steps: 1) planning: selection of locations best suited for effective restoration measures and
dam construction, 2) hydrological modelling: predicting the effect of dams, 3) implemen-
tation: dam construction, 4) monitoring: monitoring the performance of dams in time. The
methodology presented here for steps 1) and 2) builds on a combined approach of field
inventory, remote sensing, geospatial analysis and 3D peat dome topography assessment as
well as sophisticated hydrological modelling. Steps 3) and 4) are briefly discussed in
Section 4and will remain as a future research topic.
2 Study area, materials and methods
2.1 Study area
The hydrological restoration project will be carried out in a 1,480 km² area of the Sebangau
catchment which is located in the Indonesian province of Central Kalimantan on the island
of Borneo (Fig. 1). The catchment is part of a 7,347 km² large peat dome which contains
the largest remaining continuous area of dense peat swamp forest in Borneo and stores
about 2.3 Gt of peat soil carbon (Jaenicke et al. 2008). The extent of the study area is
defined by natural, hydrological borders, i.e. the Sebangau River to the east, tributary
streams to the southwest and north and the highest elevation of the peat dome to the
northwest. As most Indonesian peatlands the Sebangau peat dome is ombrogenous, i.e.
rainfall is the only source of water and nutrients. Organic matter accumulation started
around 26,000 years ago (Page et al. 2004). The climate of Central Kalimantan is
determined by a dry season which usually begins in May and lasts until October and a wet
season from November until April. Annual rainfall varies between 2,000 and 4,000 mm and
is influenced by periodic El Niño events which cause a prolonged dry season. During the
Mitig Adapt Strateg Glob Change (2010) 15:223–239 225
dry season the groundwater level in the peat drops as precipitation decreases. The Sebangau
ecosystem is renowned for its high conservation value and important natural resource
functions. Consequently, the Sebangau catchment was designated as National Park in 2004,
also to protect the largest population in the world of the endangered Bornean orang-utan.
Nevertheless, the Sebangau peat dome is suffering from serious drainage in recent years
due to the construction of hundreds of canals by illegal loggers. Until 1997 timber
concessions constructed thousands of kilometres of simple railway tracks to transport felled
timber to the Sebangau River (Boehm and Siegert 2004). The concession companies
removed their infrastructure equipment but illegal loggers excavated canals along the
former railway tracks to enable timber transport (Fig. 2). Difficult access restricts the
knowledge of the total number of canals in Sebangau peat dome to estimations by local
fisherman and environmental organisations. In this study, field surveys were conducted to
map all canals within two specific areas located in the eastern part of the peatland. Burn
scars occurring on Landsat satellite imagery since 1997 as well as fire hotspots yearly
detected by the MODIS satellite sensors (FIRMS 2009) demonstrate the negative impacts
of canal drainage on the Sebangau peatland.
The eastern part of the Sebangau catchment was selected for hydrological restoration due
to its vicinity to the city of Palangka Raya and its relative easy access via the Sebangau
Fig. 1 Landsat ETM+ satellite image from August 2007 showing the study area located in Central
Kalimantan on the island of Borneo, Indonesia. Dark green: peat swamp forest, red: fire scars in the year
2006
226 Mitig Adapt Strateg Glob Change (2010) 15:223–239
River and tributary streams. Two water sub-catchments, named after their main outlet rivers
Bakung and Bangah, were identified for the project (Fig. 1). Outlet rivers give loggers
access to the forest and thus most drainage canals start there. On the basis of a Digital
Terrain Model (DTM) the two catchments were delineated comprising a total area of
590 km². It is assumed that if all canals actually draining the peat within a specific
catchment are blocked, it will be possible to permanently raise groundwater levels to the
original situation in which groundwater levels are normally at or close to land surface.
2.2 Remote sensing
Difficult access of tropical peat swamp forests and limited project funds, require the use of
remote sensing data and modelling techniques in combination with field surveys of canal
attributes. Optical satellite imagery from Landsat ETM+, SPOT HRVIR and ALOS AVNIR
sensors, radar satellite data from the Shuttle Radar Topography Mission (SRTM) and high
resolution airborne laser scanning data (LIDAR) were used to: 1) generate a Digital Terrain
Model (DTM) of the peat surface and determine peat thickness, and 2) localise drainage
canals for hydrological modelling of groundwater levels. Hydrological modelling allows
identification of areas with good restoration potential and helps to optimise the number and
location of dams required for rewetting a specific area. Canal location, length, width, depth
and slope as well as peat bulk density, hydraulic conductivity and the stratification by peat
thickness are required parameters for the modelling.
LIDAR (LIght Detection And Ranging) measurements were acquired in August 2007 for
the northern part of the study area along a 34 km long and 0.4 km wide flight stripe running
from west to east. LIDAR systems are active, airborne remote sensing systems which
radiate pulses of laser light to the terrain and measure the time delay between transmission
of the pulse and measurement of the reflected signal by the sensor. The three dimensional
clouds of points were differentiated into ground points and non-ground points reflected
from vegetation. To extract ground points from vegetation points the terrain-adaptive bare
Fig. 2 Typical drainage canal in the Sebangau catchment used to transport timber
Mitig Adapt Strateg Glob Change (2010) 15:223–239 227
earth filtering algorithm from Cloud Peak software was applied (Ballhorn et al. 2009).
LIDAR measurements allow assessing the terrain height beneath forests with unrivalled
accuracy. The ground surface generated by airborne Laser data has a spatial resolution of
1 m. LIDAR data were used to assess the peat dome topography across the Sebangau
catchment and to validate the DTM generated for the study area.
The elevation of the DTM was calculated from SRTM imagery acquired in February
2000. Kriging interpolation in ArcGIS was used to generate a dome shaped peat surface
model as indicated by the LIDAR and SRTM data. For this surface grid points at 500–
1,000 m intervals extracted from the SRTM data, were interpolated. SRTM data represent in
deforested peat areas a Digital Terrain Model (DTM), i.e. bare-earth model. However, in
forested areas they display a so called Digital Surface Model (DSM) because the SRTM C-
band radar sensor does not penetrate the dense peat swamp forest cover. The tree canopy
height was estimated by means of deforested patches, like burn scars, rivers and canals.
Different peat swamp forest types were identified by analysing their texture variations in the
radar imagery in combination with spectral information from a Landsat ETM+ image also
acquired in February 2000. The terrain model, together with peat drilling data, formed the
basis for modelling peat thickness. Peat thickness drillings using manually operated peat
corers are laborious and expensive. The limited terrain accessibility restricts these drillings
usually to sites adjacent to drainage canals and along logging railway tracks. A total of 129
drilling measurements were available for the study area but not evenly distributed to
directly apply spatial interpolation. Therefore, correlation was used to provide missing peat
thickness information (Jaenicke et al. 2008). The correlation function makes use of a
biconvex shape model typically for ombrogenous, tropical peatlands (Rieley and Page
2005; Jaenicke et al. 2008). A strong correlation coefficient of r=0.87 was obtained
between peat surface and peat thickness.
2.3 Hydrological modelling
For hydrological modelling, the physically-based SIMGRO (SIMulation of GROundwater
flow and surface water levels) model was used to simulate water flow in the saturated zone,
unsaturated zone, river channels and over the peat surface (Querner et al. 2008; Querner
and Povilaitis 2009). Using the DTM and the watercourses map, delineations of the project
area were determined with the hydrology extension in the GIS package ArcView. Saturated
groundwater flow was modelled using the finite element method for which the model area
was subdivided into triangular segments. The top of the mineral layer was set as aquifer
bottom. Hydraulic conductivity of the peat is an essential element of hydrological
modelling. In turn, the hydraulic conductivity and also the moisture retention relationship
of the peat is strongly influenced by the degree of humification of the peat. Based on
hydraulic conductivity measurements using the pumping test method as reported by Ong
and Yogeswaran (1992) and by Takahashi and Yonetani (1997) the peat profile in this study
is schematised in a two layer system consisting of a fibric to hemic peat top layer (0–1m)
with an average hydraulic transmissivity (cumulative thickness multiplied by conductivity)
of 30 m
2
d
−1
and a deeper, sapric peat layer with an average hydraulic transmissivity of
2.2 m
2
d
−1
. While using these average values it should be realised that the relatively few
measurements available for tropical peatlands show a considerable range. In addition, a peat
water storage coefficient is required as a model input parameter. This coefficient was not
measured directly but obtained in the model calibration process and set to 0.5 (Wösten et al.
2006). Groundwater levels calculated using both the original and calibrated model for the
test site directly south of Palangka Raya (Fig. 1) are shown in Fig. 3a. The correlation
228 Mitig Adapt Strateg Glob Change (2010) 15:223–239
coefficient (R
2
), the root mean square error (RMSE) and the mean square error (MSE) for
the calibrated model are 0.74, 5.22 and 7.79 respectively. After calibration the model was
validated and the results are shown in Fig. 3b. The calibrated and validated model
represents groundwater levels measured in a dip well at the test site with acceptable
accuracy (within 0.10 m).
3 Results
3.1 Peat dome 3D topography
The 3D topography of the peat layer is an essential input for hydrological modelling of
groundwater levels. The DTM of the peat dome surface was used for slope calculations to
identify water sub-catchments and to determine the number and location of dams for
hydrological restoration. LIDAR data analysis showed that the surface of the Sebangau peat
dome towards the centre is elevated by a maximum of 13 m above its margins with an
average gradient of 0.7 m per kilometre which appears flat when in the field (Fig. 4). The
SRTM derived peat dome surface correlates very well with the LIDAR measurements; the
average discrepancy is only 0.35 m (Fig. 4). The LIDAR as well as SRTM DSM reveal
Fig. 3 Measured and calculated groundwater levels relative to land surface at the test site (Lat = 2.323S, Lon =
113.903 E) versus time. aModel calibration, bModel validation
Mitig Adapt Strateg Glob Change (2010) 15:223–239 229
different peat swamp forest types (low, medium, tall pole), which in accordance with field
investigations have different maximum canopy heights depending on local substrate
conditions (Page et al. 1999). Biomass data, i.e. breast height diameter, tree height and tree
species, were collected in October 2007 and 2008 along the transect shown in Fig. 4and
these data confirm the results. Even across large distances with little relief it is possible to
derive the DTM from the SRTM DSM using spatial interpolation between deforested
patches. The result was a detailed DTM of the Sebangau peat dome and its sub-catchments
with 30 m spatial resolution. Figure 5shows the fine topography along cross sections in the
middle of Bakung and Bangah catchments. The slope of the southern part of Bakung
catchment appears relatively steep but the gradient is only 1 m per kilometre at maximum.
Besides detailed peat dome topography, hydrological modelling requires peat thickness and
bedrock data. The result of the thickness modelling reveals an average peat thickness of
5.4±0.95 m within the study area and a maximum depth of approximately 10.7 m in the
centre of the Sebangau peat dome. The margin of error results from comparison of the peat
thickness model with in situ measurements. The large deviations result probably from
bedrock unconformity, which is not taken into account in the model. About half of the in
Fig. 4 The LIDAR DTM and the peat surface derived from SRTM data (Model) agree very well. The SRTM
DSM data reveal relative canopy heights of various peat swamp forest types
Fig. 5 DTM cross sections in the middle of the Bakung and Bangah catchment (from north to south)
230 Mitig Adapt Strateg Glob Change (2010) 15:223–239
situ thickness values are larger than the model result, while the other half are smaller. This
suggests that discontinuities in the mineral ground topography are balanced by spatial
Kriging interpolation and thus the modelled volume results are close to reality (Jaenicke et
al. 2008).
3.2 Canal delineation
During field surveys in the Bakung and Bangah catchments the origin of 65 drainage canals
was recorded. Eventually all these canals need to be blocked to rewet the surrounding
peatland. The field team also recorded direction, length, width and depth of all canals as
well as water depth, water flow, mud sedimentation or weed growth. With an average depth
of 0.7 m and an average width of 2.4 m the canals are relatively small in terms of their
cross-sectional dimensions, but they are closely spaced with an average distance of about
200 m in the Bakung and of about 800 m in Bangah catchment and they extent for distances
up to 13 km. All information was stored in a geodatabase and a ranking was assigned
indicating the priority of a canal to be closed. Long, wide and deep canals with a high water
level and flow were assigned a high priority, whereas canals filled with mud and weeds
were categorised as low priority. Twenty-two canals showed a high or medium need for
closure. Canal lengths were estimated by consulting local people since access to the canals
is very laborious and because GPS recordings are inaccurate due to dense forest cover
hampering the GPS receiver. Narrow canals were invisible even from high resolution
satellite images (SPOT and ALOS AVNIR, both at 10 m spatial resolution) because the tree
canopy covers the streams (Fig. 6). However, knowing the outlet of the canal, the direction
and approximate length it was possible to delineate most canals.
Fig. 6 SPOT satellite image
from May 2004 showing the
course of canals and railway
tracks in the Bangah catchment as
bright green lines as well as sites
of illegal logging (pink and bright
green “dots”). The origin of
drainage canals recorded during
field work is superimposed as
yellow dots
Mitig Adapt Strateg Glob Change (2010) 15:223–239 231
3.3 Identification of locations for dam construction
Dams act as flow barriers but they cannot store water for long periods as water will
eventually seep through the surrounding peat. As dams restrict water flow rather than stop
all water movement, they do not have to be watertight and thus construction can be
relatively simple. To determine the optimal number and location of dams required for
efficient drainage reduction, the surface slope was determined along each canal selected to
be closed. Hydrological model simulations revealed that a cascade of closely spaced dams
is most effective for water control (Wösten and Ritzema 2001). The steeper the slope, the
more dams are needed to reduce drainage. Figure 7shows the slope of a medium priority
canal in the Bangah catchment (length 10 km, width 3 m, depth 1 m). The absolute
elevation difference of the canal from its origin at the top of the peat dome to its outlet into
Bangah river is 3.1 m. Because the slope of the canal is not constant over its total length it
was subdivided into two sections: an upper, relatively flat section (Fig. 7, Slope1) and a
lower, steep section (Fig. 7, Slope2). The distance between dams required to reduce
drainage is determined by the hydraulic head difference, i.e. difference between upstream
and downstream canal water level across a dam. Field experiments showed that for small
canals the water level over each dam should be limited to about 25 cm to reduce seepage
and to prevent erosion. Thus, the canal in Fig. 7requires a series of 13 dams to overcome
the 3.1 m elevation difference
1
. In the upper section of the canal a spacing of 975 m
between dams is sufficient to keep water level differences low, while in the steeper section
the spacing needs to be reduced to 320 m. The Bakung catchment requires the construction
of 141 dams to efficiently reduce drainage. For the Bangah catchment 84 dams are
needed in addition to 30 dams previously constructed. Figure 8shows the location of
dams planned and already built, as well as the priority status of the canals superimposed on
the DTM. The Bakung catchment is smaller than Bangah catchment but requires more
dams because of the steeper topography and higher density of canals to be closed. Figure 9
shows an example of a relatively simple dam in the Bangah catchment mainly made of
locally available material.
3.4 Prediction of groundwater level rise
The effect of dams on groundwater levels is predicted by hydrological modelling comparing the
situation before and after dam construction. Figure 3shows that in wet years calculated
groundwater levels are at or close to land surface whereas in dry years they drop to about 1 m
below land surface. On average the groundwater level at the undisturbed test site is −16 cm.
This value provides an indication of the intended long-term average groundwater level after
successful blocking of drainage canals in the Bakung catchment. The calibrated and validated
hydrological model was applied to the whole of the Bakung and Bangah catchment for the 25
November 1997, an extremely dry period. Figure 10a shows that dams can raise groundwater
levels up to 50–70 cm under these very dry weather and peat conditions. For larger areas the
1
H slope1ðÞ=0:25 þH slope2ðÞ=0:25 þ...þH slopen
ðÞ=0:25 ¼N damsðÞ
D slopen
ðÞ
=N dams
ðÞ
¼S dams
ðÞ
H maximum elevation difference of the canal within each “slope section”
N optimum number of dams (rounded up to be on the save side)
D distance of each “slope section”
S spacing between dams
232 Mitig Adapt Strateg Glob Change (2010) 15:223–239
rise is approximately 10–30 cm. Rise in groundwater levels is presented in classes rather than
as absolute values to reflect the uncertainty in the calculated results. The areas affected by
rewetting are strongly influenced by the slope of the peatland area surrounding the canal as this
determines the catchment area draining to the canal. Figure 10b shows surface water levels in a
12 km long canal. Compared to the situation without dams, the result is a rise of the canal water
level of up to 35 cm in the upstream part of the canal. The resulting rewetting of the peatland
area surrounding this canal is up to 50 cm. Hydrological modelling of the rise of groundwater
levels on a daily base for the years 2006, 2007 and 2008 shows that on average this rise is
20 cm during the dry season. As a consequence, construction of dams considerably increases
Fig. 7 Slope of the peat surface
next to a canal in Bangah catch-
ment as measured in the modelled
DTM (0 marks the most upstream
part of the canal). 13 dams are
required to reduce large scale
drainage
Fig. 8 Location of dams to be
constructed for an efficient re-
duction of drainage in the
Bakung and Bangah catchments.
Only canals ranked as medium
and high priority should be
closed. Data are superimposed on
the peat surface DTM
Mitig Adapt Strateg Glob Change (2010) 15:223–239 233
the water retention capacity of the blocked areas thereby creating favourable wet conditions for
vegetation re-growth and eventually peatland restoration.
3.5 Mitigation of carbon dioxide emissions
Rewetting of drained tropical peatlands will potentially lead to large mitigations of carbon
dioxide emissions (Couwenberg et al. 2009). Quantifying the rise in groundwater levels of
hydrological restoration projects in peatlands together with an estimation of the mitigation
in CO
2
emissions caused by this rise, is important information to make greenhouse gas
emission mitigations tradable under the voluntary carbon market or REDD (Reducing
Emissions from Deforestation and Degradation) mechanism. Continuous, long-term
groundwater level measurements in tropical peat swamp forests are rare. The only available
12 year average groundwater level recorded at the relatively intact test site is −16 cm,
whereas this level in an adjacent, drainage affected, selectively logged forest is −47 cm for
the years 2004 and 2005 with normal precipitation (Jauhiainen et al. 2008). Preliminary
groundwater level measurements in the drainage affected Bangah catchment indicate an
average level of −49 cm. Consequently, an average annual groundwater level of −50 cm
was assumed to be a baseline level for the project area before hydrological restoration
started. After construction of all dams, hydrological modelling indicates a rise of annual
average groundwater levels of 20 cm. With a reported emission mitigation of approximately
0.8–0.9 t CO
2
ha
−1
a
−1
per centimetre groundwater level rise (Couwenberg et al. 2009;
Hooijer et al. 2006), rewetting of the 590 km
2
area of the combined Bakung and Bangah
catchments results in an estimated mitigated emission of 1.4–1.6 Million tons CO
2
annually. This estimated emission mitigation will not be achieved in the first year after all
dams have been constructed because only with time sedimentation of organic and mineral
material upstream of the dams makes them fully effective. Higher emissions are expected
during El Niño years, such as in 1997, 2002, 2006 and 2009 due to very low groundwater
Fig. 9 Simple dam in the Bangah catchment made of locally available material (3 m long, 1 m wide and
2.5 m deep)
234 Mitig Adapt Strateg Glob Change (2010) 15:223–239
levels in addition to drainage. In the project area, long-term measurements of groundwater
levels (before and after dam construction) as well as subsidence and gas flux emissions are
needed to confirm these preliminary results. In this study, conservative estimates were used of
both the reduced CO
2
emission rate per centimetre groundwater level rise (Couwenberg et al.
2009; Hooijer et al. 2006) as well as of the magnitude of the groundwater level rise itself.
Results are reported as a class to reflect the uncertainty in the calculations. Other greenhouse
gases such as methane (CH
4
) and nitrous oxide (N
2
O) are not taken into account because they
are relatively unimportant in tropical peatlands (Furukawa et al. 2005; Strack 2008).
4 Discussion
Canals constructed for drainage and illegal logging have destroyed the hydrological integrity of
many tropical peatland ecosystems (e.g. Giesen 2004;Wöstenetal.2006; Hoekman 2007;
CKPP 2008). The only way to prevent soil subsidence, peat decomposition, peat fires and
Fig. 10 Hydrological modelling
applied to the Bangah catchment
for very dry conditions on 25
November 1997. aGroundwater
level rise in the whole area after
construction of 114 small dams b
Rise of the surface water level
(swl) in a single canal after dam
construction
Mitig Adapt Strateg Glob Change (2010) 15:223–239 235
associated carbon dioxide emissions is the restoration of the hydrological integrity by raising
groundwater levels and thus rewetting the peat to its original situation. Many studies have
shown that groundwater levels control greenhouse gas emissions from tropical peatlands (e.g.
Furukawa et al. 2005; Hooijer et al. 2006; Hirano et al. 2008; Jauhiainen et al. 2008;
Couwenberg et al. 2009). However, very few practical hydrological restoration measures of
degraded tropical peatlands have been reported (Jauhiainen et al. 2008;Pageetal.2008). The
aim of this study was to develop a detailed plan to rewet a 590 km² large area of highly
inaccessible peat swamp forest drained by a dense network of small canals that are used by
illegal loggers. The case as such is typical for many tropical peatlands in Indonesia and the
proposed methodology is transferable to other drained tropical peatlands thereby increasing the
knowledge base for future hydrological restoration activities. A detailed 3D peat dome model
generated using remote sensing data, together with identified dam construction sites, provided
input for hydrological modelling to quantify the effects of dams on raising groundwater levels.
To verify the calculated groundwater levels a monitoring programme is under construction
aiming at measurement of these levels in wells installed at a dam along two transects left and
right, and perpendicular to the canal at 5, 25, 50, 150 and 300 m distances from the canal. Also
water discharges will be measured in both blocked and unblocked canals. In this study wider
canals were clearly visible in high resolution satellite imagery, while hardly visible, smaller
canals were determined as follows: 1) canals do not run parallel to the river or cross each other
because they are constructed to facilitate extraction of timber logs from the forest, 2) while in
reality the course of the canals might be not completely straight, small meanders do not have
any impact on the number of dams required for rewetting. Dams need to be adapted to the
characteristic high hydraulic conductivity (Wösten and Ritzema 2001) and low load bearing
capacity (Salmah 1992) of tropical peat. Reduced water flow in the canals allows sedimentation
of organic and mineral material upstream of the dam which in turn facilitates the re-growing of
vegetation. Eventually, original peat forming vegetation will fill in the canal thereby restoring
the resistance to water flow in the peat swamp forest to its original value of approximately
30 m/day. To keep subsidence of the area surrounding the dam low, dam construction should
not be too heavy. Materials like gelam timber poles and peat are suitable for dam construction
and they are locally available. Blocking of a canal can be regarded successful if the blocked
canal sections continue to hold water during the dry season. Since some drainage canals are
used for navigation and transportation by local people, ownership of each canal should be
considered and consensus should be reached before dam construction starts. Failure to do so
can result in damage to the dam structures as has happened frequently in the past. After
construction, monitoring and maintenance of the dams is very important, especially in the first
years (CKPP 2008). Previous work in the Bangah catchment demonstrated that a field team can
build 30 dams in 7 days, i.e. 53 days are required to construct all 225 dams required for the
Bakung and Bangah catchments together. Labour costs for one dam (transport and material
costs excluded) are approximately 150,000 IDR which is equivalent to about 10 Euro. An
annual emission mitigation of 1.5 Mt CO
2
from restored tropical peatlands is a significant
amount corresponding to 6% of the carbon dioxide emissions by civil aviation in the European
Union in 2006 (UNFCCC 2009), and therefore of interest for carbon crediting on the voluntary
carbon market. This mitigation can be achieved with relatively small efforts and at low costs by
focusing on construction and maintenance of simple dams made of locally available material. In
case oxidation by drainage is limited to the top 50 cm of an active peat layer the total carbon at
stake is 2.1 times higher than that of the aboveground biomass
2
. This total amount of carbon at
2
A carbon content of 140.5 t/ha for peat swamp forest (Uryu et al. 2008) and of 58 kg/m³ for peat soils
(Neuzil 1997; Shimada et al. 2001; Supardi et al. 1993) is assumed.
236 Mitig Adapt Strateg Glob Change (2010) 15:223–239
stake increases to 22 times the aboveground biomass if no hydrological restoration measures
were implemented and continuous oxidation of the whole 5.4 m thick peat layer was allowed to
take place. Increased awareness of the large amounts of carbon at risk due to tropical peatland
drainage and fires promotes interest in alternative funding mechanisms such as REDD and
carbon credits to safeguard these carbon stocks. Canal blocking in tropical peatlands is not only
a technical but also a social challenge. Illegal logging was, besides gold mining, a main source
of income for people in Central Kalimantan. Now that funding through REDD and carbon
credits becomes a realistic alternative it should also be used to improve livelihoods of local
people. Restoration can only be successful if local communities are actively involved in
planning and implementation of restoration measures as demonstrated in this study by WWF.
Acknowledgments The authors would like to thank Guenola Kahlert, WWF Germany, for financial
support. Special thanks to the WWF Indonesian field team for collecting canal data and to Prof. Hidenori
Takahashi, University of Hokkaido, for the long-term measurements of rainfall and groundwater level at the
test site. We gratefully acknowledge the Global Land Cover Facility (GLCF) for providing SRTM data
without expense, and the US Geological Survey (USGS) for providing Landsat ETM+ imagery.
Open Access This article is distributed under the terms of the Creative Commons Attribution
Noncommercial License which permits any noncommercial use, distribution, and reproduction in any
medium, provided the original author(s) and source are credited.
References
Ballhorn U, Siegert F, Mason M, Limin S (2009) Derivation of burn scar depth and carbon emissions with
LIDAR in Indonesian peatlands. PNAS www.pnas.org/cgi/doi/10.1073/pnas.0906457106
Boehm H-DV, Siegert F (2004) The impact of logging on land use change in Central Kalimantan, Indonesia.
Int Peat J 12:3–10
CKPP (2008) Provisional report of the Central Kalimantan Peatland Project. CKPP Consortium, Nov 2008,
Palangka Raya, Indonesia, p 72. http://www.ckpp.org. Cited 20 June 2009
Cochrane MA (2003) Fire science for rainforests. Nature 421:913–919
Couwenberg J, Dommain R, Joosten H (2009) Greenhouse gas fluxes from tropical peatlands in Southeast
Asia. Global Change Biol. doi:10.1111/j.365-2486.2009.02016.x
Curran LM, Trigg SN, McDonald AK, Astiani D, Hardiono YM, Siregar P, Caniago I, Kasischke I (2004)
Lowland forest loss in protected areas of Indonesian Borneo. Science 303:1000–1003
FIRMS (2009) Fire Information for Resource Management System. University of Maryland. http://maps.
geog.umd.edu/firms/
Furukawa Y, Inubushi K, Ali M, Itang AM, Tsuruta H (2005) Effect of changing groundwater levels caused
by land-use changes on greenhouse gas fluxes from tropical peat lands. Nutr Cycl Agroecosyst 71:81–91
Giesen W (2004) Causes of peat swamp forest degradation in Berbak NP, Indonesia, and recommendations for
restoration. Water for Food and Ecosystems Programme, Arcadis. http://www.waterfoodecosystems.nl
Hansen MC, Stehman, SV, Potapov PV, Arunarwati B, Stolle F, Pittmann K (2009) Quantifying changes in
rates of forest clearing in Indonesia from 1990 to 2005 using remotely sensed data sets. Environ Res Lett
4:034001(12pp)
Hirano T, Jauhiainen J, Inoue T, Hidenori T (2008) Controls on the carbon balance of tropical peatlands.
Ecosystems. doi:10.1007/s10021-008-9209-1
Hoekman DH (2007) Satellite radar observation of tropical peat swamp forest as a tool for hydrological
modelling and environmental protection. Aquatic Conserv: Mar Freshw Ecosyst 17:265–275
Hooijer A, Silvius M, Wösten H, Page SE (2006) PEAT-CO2, assessment of CO2 emissions from drained
peatlands in SE Asia. Delft Hydraulics report Q3943
Jaenicke J, Rieley JO, Mott C, Kimman P, Siegert F (2008) Determination of the amount of carbon stored in
Indonesian peatlands. Geoderma 147:151–158
Jauhiainen J, Jaya A, Inoue T, Heikkinen PJ, Martikainen P, Vasander H (2004) Carbon balance in managed
tropical peat in Central Kalimantan. In: Päivänen J (ed) Proceedings of the 12th International Peat
Congress. International Peat Society, Tampere, pp 653–659
Mitig Adapt Strateg Glob Change (2010) 15:223–239 237
Jauhiainen J, Limin SH, Silvennoinen H, Vasander H (2008) Carbon dioxide and methane fluxes in drained
tropical peat before and after hydrological restoration. Ecology 89(12):3503–3514
Langner A, Miettinen J, Siegert F (2007) Land cover change 2002–2005 and the role of fire derived from
MODIS imagery. Global Change Biol 13:1–12
Melling L, Hatano R, Goh KJ (2005) Soil CO2 flux from three ecosystems in tropical peatland of Sarawak,
Malaysia. Tellus 57B:1–11
Neuzil SG (1997) Onset and rate of peat and carbon accumulation in four domed ombrogenous peat deposits,
Indonesia. In: Rieley JO, Page SE (eds) Biodiversity and sustainability of tropical peatlands. Samara,
Cardigan, pp 55–72
Ong BY, Yogeswaran M (1992) Peatland as a resource for water supply in Sarawak. In: Aminuddin BY, Tan
SL, Aziz B, Samy J, Salmah Z, Siti Petimah, Choo ST (eds) Proceedings of the International Symposium
on Tropical Peatland, Kuching, Sarawak, May 1991. Ministry of Agriculture, MARDI, pp 255–268
Page SE, Banks C (2007) Tropical peatlands: distribution, extent and carbon storage—uncertainties and
knowledge gaps. Peatl Int 2:26–27
Page SE, Rieley JO, Shotyk W, Weiss D (1999) Interdependence of peat and vegetation in a tropical peat
swamp forest. Phil Trans R Soc Lond B 354:1885–1897
Page SE, Siegert F, Rieley JO, Boehm H-DV, Jaya A, Limin S (2002) The amount of carbon released from
peat and forest fires in Indonesia during 1997. Nature 420:61–65
Page SE, Wüst RA, Weiss D, Rieley JO, Shotyk W, Limin SH (2004) A record of Late Pleistocene and
Holocene carbon accumulation and climate change from an equatorial peat bog (Kalimantan, Indonesia):
implications for past, present and future carbon dynamics. J Quaternary Sci 19:625–635
Page SE, Hoscilo A, Wösten JHM, Jauhiainen J, Silvius M, Rieley JO, Ritzema H, Tansey K, Graham L,
Vasander H, Limin S (2008) Restoration ecology of lowland tropical peatlands in Southeast Asia: current
knowledge and future research directions. Ecosystems. doi:10.1007/s10021-008-9216-2
Querner EP, Povilaitis A (2009) Hydrological effects of water management measures in the DovinėRiver
basin, Lithuania. Hydrol Sci 54:363–374
Querner EP, Jansen PC, Kwakernaak C (2008) Effects of water level strategies in Dutch peatlands: a scenario
study for the polder Zegveld. In: Farrell C, Feehan J (eds) Proceedings of the 13th International Peat
Congress. International Peat Society, Tullamore, pp 620–623
Rieley JO, Page SE (eds) (2005) Wise use of tropical peatlands: focus on Southeast Asia. Alterra,
Netherlands. www.restorpeat.alterra.wur.nl
Salmah Z (1992) Water management in deep peat soils in Malaysia. Dissertation, Silsoe College, Cranfield
Institute of Technology, UK
Shimada S, Takahashi H, Haraguchi A, Kaneko M (2001) The carbon content characteristics of tropical peats
in Central Kalimantan, Indonesia: estimating their spatial variability. Biochemistry 53:249–267
Siegert F, Ruecker G, Hinrichs A, Hoffmann AA (2001) Increased damage from fires in logged forests
during droughts caused by El Niño. Nature 414:437–440
Sorensen KW (1993) Indonesian peat swamp forests and their role as carbon sink. Chemosphere 27:1065–
1082
Strack M (ed) (2008) Peatlands and climate change. International Peat Society, Jyväskylä
Supardi, Subekty AD, Neuzil SG (1993) General geology and peat resources of the Siak Kanan and
Bengkalis Island peat deposits, Sumatra, Indonesia. In: Cobb JC, Cecil CB (eds) Modern and acient
coal-forming environments, Boulder, Colorado, Geological Society of America Special Paper 286
Suryadiputra INN, Dohong A, Waspodo RSB, Muslihat L, Lubis IR, Hasudungan F, Wibisono ITC (2005) A
guide to blocking of canals and ditches in conjunction with the community. Wetlands International—
Indonesia Programme, Bogor
Takahashi H, Yonetani Y (1997) Studies on microclimate and hydrology of peat swamp forest in Central
Kalimantan, Indonesia. In: Rieley JO, Page SE (eds) Biodiversity and sustainability of tropical peatlands.
Samara, Cardigan, pp 179–187
Takahashi H, Usup A, Hayasaka H, Limin SH (2003) Estimation of ground water levels in a peat swamp
forest as an index of peat/forest fire. Proceedings of the International Symposium on Land Management
and Biodiversity in Southeast Asia, Bali, Indonesia, 17–20 September 2002
UNFCCC (2009) GHG data. http://unfccc.int/ghg_data/ghg_data_unfccc/items/4146.php. Cited 25 Aug 2009
Uryu Y, Mott C, Foead N, Yulianto K, Budiman A, Setiabudi, Takakai F, Nursamsu, Sunarto, Purastuti E,
Fadhli N, Hutajulu CMB, Jaenicke J, Hatano R, Siegert F, Stüwe M (2008) Deforestation, forest
degradation, biodiversity loss and CO
2
emissions in Riau, Sumatra, Indonesia. WWF Indonesia
Technical Report, Jakarta, Indonesia
Usup A, Hashimoto Y, Takahashi H, Hayasaka H (2004) Combustion and thermal characteristics of peat fire
in tropical peatland in Central Kalimantan, Indonesia. Tropics 14:1–19
238 Mitig Adapt Strateg Glob Change (2010) 15:223–239
Van der Werf GR, Randerson JT, Giglio L, Gobron N, Dolman AJ (2008) Climate controls on the variability
of fires in the tropics and subtropics Global Biogeochem Cycles 22, GB3028, doi:10.1029/
2007GB003122
Whitten AJ, Damanik SJ, Anwar J, Hisam N (1987) The ecology of Sumatra. Gadjah Mada University Press,
Yogyakarta
Wösten JHM, Ritzema HP (2001) Land and water management options for peatland development in
Sarawak, Malaysia. Int Peat J 11:59–66
Wösten JHM, Ismail AB, Van Wijk ALM (1997) Peat subsidence and its practical implications: a case study
in Malaysia. Geoderma 78:25–36
Wösten JHM, Hooijer A, Siderius C, Satriadi Rais D, Idris A, Rieley JO (2006) Tropical peatland water
management modelling of the Air Hitam Laut catchment in Indonesia. Int J River Basin Manag 4:233–
244
Wösten JHM, Clymans E, Page SE, Rieley JO, Limin SH (2008) Peat—water interrelationships in a tropical
peatland ecosystem in Southeast Asia. Catena 73:212–224
Mitig Adapt Strateg Glob Change (2010) 15:223–239 239
