Content uploaded by Ben DeVries
Author content
All content in this area was uploaded by Ben DeVries on Jan 20, 2018
Content may be subject to copyright.
Monitoring the Effects of Hydrological Restoration
Efforts in Degraded Tropical Peatlands
Central Kalimantan, Indonesia
MSc Thesis by Ben DeVries
19 February, 2010
Monitoring the Effects of Hydrological Restoration
Efforts in Degraded Tropical Peatlands
Master thesis Land Degradation and Development Group submitted in partial
fulfillment of the degree of Master of Science in International Land and Water
Management at Wageningen University, the Netherlands
Study program:
MSc International Land and Water Management (MIL)
Student registration number:
810801-180-020
LDD 80336
Supervisor(s):
Dr.ir. H. Wösten
Prof.dr.ir. L. Stroosnijder
Examinator:
Prof.dr.ir. L. Stroosnijder
Date: 19 February, 2010
Wageningen University, Land Degradation and Development Group
Alterra, Soil Science Centre
World Wide Fund for Nature (WWF Indonesia; WWF Germany)
Preface
This research project is actually a culmination of work that has been taking place in the
Sebangau National Park for a number of years. As such, this project was obviously not an
individual endeavor, and there are a number of people to thank for their contributions. First I
want to thank my supervisors, Henk and Leo, for introducing me to this thesis topic with such
enthusiasm, and also for your continued support throughout the project, both in Indonesia and
in the Netherlands. I also appreciated the technical assistance from a number of people
throughout the project, including Christian (Alterra) for helping me to understand issues
relating to hydrological modeling, and Julia (RSS) for assisting with remote sensing data. On
the side of WWF, I would like to thank Guénola (WWF Jakarta) and Arif (WWF Indonesia) for
welcoming me to the WWF team. Even before arriving in Indonesia, I was assisted with
logistic matters by a large team of very competent staff at WWF-Indonesia headquarters in
Jakarta. Thanks especially to Lely for assisting me with my visas and all other government
permits and documents during my 4-month stay in Indonesia.
I arrived in Palangkaraya, Central Kalimantan, to find a team of dedicated conservationalists
working on some very complex problems. I can safely say this is one of my favourite groups
of people with which I have had the pleasure to work. Their dedication to their task of
environmental protection and restoration was an inspiration for me, and I learned a great deal
from our time together. I would like to thank Ibu Sendy for providing strong overall leadership
to the WWF-Central Kalimantan team, and especially to Panda for leadership on the
Sebangau project. I am extremely grateful to everyone in the WWF office in Palangkaraya
and especially to the members of the Sebangau team: to Okta for sharing your expertise with
me, and staying up late working through the dataset; to Makmun for introducing me to the
communities around Paduran (and finally for teaching me to ride a motorcycle!); to Rosidi for
sharing your GIS data and helping me with some of the GIS analysis; to Priyono and the rest
of the Sebangau team for your dedication to the work in the field, which was often quite
challenging – both physically and mentally. Thanks also to the staff of Teman Nasional
Sebangau (Sebangau National Park) for your continued presence and contribution to the field
work. Thanks to all of those in the WWF-Central Kalimantan office in Palangkaraya who were
always available to help with logistics, advice, or anything else – especially Ibu Raya and Pak
Ian. I also had the privilege to meet a diverse group of people along the way, many of whom
made small but significant contributions to this work. Thanks to Pak Baharan and the whole
Bangah community for graciously hosting us on a number of occasions, and for providing
much needed assistance in our field work in Sungai Bangah. Thanks to Pak and Ibu Selowadi
for your assistance, your humour, and for even naming a canal after me! Finally, I want to
give a very big „terima kasih‟ to the wonderful Untung family – Pak and Ibu Untung, Daniel,
and Willy – for inviting me to live with them for over three and half months. I don‟t think I‟ve
ever experienced hospitality as warm as yours. Thank you for taking such good care of me
and teaching me so much about Palangkaraya and Indonesia.
I want to give a final thank you to my family in Canada. I can only consider myself incredibly
lucky to have such a big, caring family. Even from such a long distance away, your continued
support has still been immensely important to me. It‟s because of you that I‟ve had the
courage and confidence to take chances, travel the world, and continue developing my career
and myself.
Terima kasih banyak! | Hartstikke bedankt! | Thank you very much!
Summary
In the recent past, rapid destruction of sensitive tropical peatland ecosystems has
necessitated the formulation of comprehensive strategies aiming towards restoration and
sustainable environmental management. In cooperation with the World Wide Fund for Nature
of Indonesia (WWF-Indonesia), research was undertaken in the Sebangau National Park in
the province of Central Kalimantan, Indonesia, to assess the efficacy of hydrological
restoration of the Sebangau peat swamp forest. In this study, dams were constructed across
drainage canals in the Sebangau National Park, and groundwater and surface water levels in
monitoring transects near the dams were monitored on a monthly basis and analyzed using
ArcGIS9.3 and a radar satellite-derived Digital Elevation Model (DEM). Groundwater
dynamics showed that artificial drainage by constructed canals disrupts regional subsurface
drainage patterns by increasing drawdown on a local scale, while regional flows remain intact.
The scale of this drawdown was found to be dependent on the direction of regional
subsurface flows, suggesting that the extent of drainage by artificial canals is orientation-
dependent. Dams constructed by WWF-Indonesia were found to function by reducing local
drawdown, retarding subsurface drainage during the dry season, and retaining floodwater
during the wet season. Water retention and associated rewetting was found not only to be
dependent on local precipitation, but was also found to progressively increase over time
despite varying levels of rainfall, suggesting a role for deposited organic debris in dam
performance. The implications of these results on peatland ecosystem health and sustainable
management, and the potential for hydrological modeling in the Sebangau peatlands are also
discussed in this report.
Table of Contents
Preface ...................................................................................................................................... iv
Summary ................................................................................................................................... vi
Table of Contents ..................................................................................................................... vii
Table of Figures ....................................................................................................................... viii
1. Introduction .............................................................................................................................1
1.1 Background ...................................................................................................................... 1
1.2 Problem Description ......................................................................................................... 2
1.3 Research Objectives ........................................................................................................ 3
2. Materials and Methods ...........................................................................................................4
2.1 Site Description ................................................................................................................ 4
2.1.1 Location ..................................................................................................................... 4
2.1.2 Topography ............................................................................................................... 5
2.1.3 Soils and Geology ..................................................................................................... 5
2.1.4 Climate ...................................................................................................................... 6
2.1.5 Hydrology .................................................................................................................. 7
2.1.6 Land Cover and Land Use ........................................................................................ 9
2.2 Ground Truthing, Dam Planning, and Monitoring Site Selection ................................... 11
2.3 Groundwater Monitoring and Analysis ........................................................................... 12
2.4 Canal Discharge Measurements .................................................................................... 14
3. Results and Discussion ....................................................................................................... 16
3.1 Differences between Tropical and Temperate Peatlands .............................................. 16
3.2 Formation and Current State of the Sebangau Peatlands ............................................. 17
3.3 Tools for Hydrological Restoration Planning .................................................................. 19
3.4 Seasonal Groundwater Trends in the Sebangau Peatlands ......................................... 21
3.5 Effects of Artificial Drainage on Peatland Hydrology ..................................................... 25
3.6 Efficacy of Water Retention by Dams ............................................................................ 27
3.7 Variance in Dam Performance over Time ...................................................................... 31
3.8 Interrelationships between Groundwater Hydrology and Peat Fire Risk ....................... 36
3.9 Potential for Hydrological Modeling ............................................................................... 40
4. Conclusions ......................................................................................................................... 43
References .............................................................................................................................. 44
Appendix 1 – Ground Truth Data ............................................................................................. 47
Appendix 2 – SSI Dam Description ......................................................................................... 49
Appendix 3 – Tubewell Hydrographs....................................................................................... 51
Appendix 4 – Diver Datalogger Data ....................................................................................... 58
Table of Figures
Figure 1 – Extent of peatlands across Indonesia.. .................................................................... 1
Figure 2 – Sebangau National Park boundary and western part of Block C. ............................ 4
Figure 3 – Digital Elevation Model (DEM) of eastern Sebangau peatlands. ............................. 5
Figure 4 – Yearly rainfall between 1997 and 2006 and monthly rainfall and evapotranspiration
between January and September 2009. .............................................................................. 7
Figure 5 – Main hydrological processes within the peat column. ............................................. 8
Figure 6 – LandSat-derived land cover classes for SSI canal and the Bangah sub-catchment.
.............................................................................................................................................. 9
Figure 7 – Tubewell construction and groundwater measurement in the WWF Sebangau
project. ................................................................................................................................ 12
Figure 8 – Tubewell transect design for canal 21. ................................................................... 13
Figure 9 – Diver datalogger setup ........................................................................................... 14
Figure 10 – Schematic of canal discharge measurements.. ................................................... 15
Figure 11 – Ground truth measurements and DEM-derived slope of Canal 22. ..................... 20
Figure 12 – Groundwater trends at the SSI monitoring grid betwen September 2006 and
September 2009.. ............................................................................................................... 22
Figure 13 – Monthly groundwater trends for the fifth tubewell of the third left-hand transect
and precipitation and evapotranspiration data measured at the SSI dam from September
2008 to November 2009. .................................................................................................... 23
Figure 14 – Hourly water table depth at canal 21 as measured by diver datalogger between
July and October 2009. ...................................................................................................... 24
Figure 15 – Groundwater monitoring transects and groundwater levels at canal 21. ............. 26
Figure 16 – SSI groundwater trends from November 2008 to November 2009 for a transect
adjacent to the large dam and a transect 3.5km upstream of the dam. ............................. 29
Figure 17 – View of the SSI dam looking upstream during the dry season. ........................... 29
Figure 18 – Interpolated water table contour maps with respect to soil surface and mean sea
level .................................................................................................................................... 30
Figure 19 – Interpolated groundwater levels at the SSI monitoring grid relative to soil surface
and normalized against DEM-derived soil surface elevation averaged for each year from
September 2006 to September 2009.. ............................................................................... 33
Figure 20 – SSI dam spillway and natural bypass to on the south side of dam. ..................... 34
Figure 21 – Interpolated monthly groundwater levels at the SSI monitoring grid relative to the
soil surface.. ....................................................................................................................... 35
Figure 22 – Repairs on the first SSI dam on 27 August. ......................................................... 36
Figure 23 – Hot spots within the Sebangau National Park as detected by satellite imagery.. 37
Figure 24 – Permanent dam and small dam in Canal 21. ....................................................... 38
Figure 25 – Head difference (cm) across each dam in canal 21 expressed as relative water
levels.. ................................................................................................................................. 39
Figure 26 – Groundwater levels relative to soil during July 2009 at the three canal 21
monitoring transects. .......................................................................................................... 39
Figure 27 – Downstream and upstream of the first dam at canal 21 after a fire at the end of
August 2009. ...................................................................................................................... 39
Figure 28 – Hydraulic head differences between surface water bodies and the groundwater
zone included in SIMGRO drainage calculations ............................................................... 41
Figure 29 – Predicted changes in groundwater levels using the SIMGRO hydrological model
after construction of dams in the Bangah sub-catchment. ................................................. 41
Table 1 - Forest cover and associated peat depth within the Sebangau National Park ......... 10
Table 2 - Comparison of dam spacing calculations using DEM-derived slopes and manually
measured (Ground Truth) slopes. ...................................................................................... 21
1
1. Introduction
1.1 Background
Indonesian peatlands represent a significant proportion of the world‟s tropical peatlands, with
some estimates suggesting they constitute half of global tropical peatland coverage. Within
Indonesia, nearly all of the country‟s peatlands occur on the islands of Sumatra, Kalimantan
(Indonesian Borneo), and Papua (formerly Irian Jaya). Several studies estimate that
Kalimantan contains between 3.5 to 7 million hectares of peatland, representing up to 40% of
the country‟s total peatland area (Joosten, 2004). Though the province of Central Kalimantan
covers under 30% of the total surface area of Kalimantan, it contains the bulk of the island‟s
peatlands with over 2.2 million hectares (Boehm and Siegert, 1999). After large tracts of this
peatland were targeted for agricultural development, some of the last remaining native peat
swamp forests in Central Kalimantan are found in the Sebangau National Park.
Figure 1 – Extent of peatlands across Indonesia. Study Area denoted by square.
Despite their relatively limited extent, tropical peatlands constitute a significant reservoir of
water, carbon, and biodiversity on earth. Among growing recognition of the severity of tropical
peatland degradation around the world, a wide range of water management solutions are
being proposed to restore these fragile environments. One such measure being proposed in
the peat swamps of Central Kalimantan is the canal blocking method, in which small dams
are constructed across decommissioned logging canals to retard surface water flow in canals
and eventually restore the upstream groundwater level. Conservation groups such as the
World Wide Fund for Nature (WWF) have been supporting efforts to rewet peat swamps in
the Sebangau peatlands through the construction of such canal blocking structures. These
efforts are linked not only to growing concerns over biodiversity loss in these forests, but also
to concerns regarding increased carbon emissions from drained, fire-prone peat soils and
associated climate change.
Alterra and Wageningen University in the Netherlands have been involved in peatland
research and restoration in the tropics and beyond for a number years and under a number of
different frameworks involving a variety of international partners
2
(http://www.restorpeat.alterra.wur.nl). Ending in 2007, the RESTORPEAT project, a
continuation of the STRAPEAT project, was an EU-funded initiative aimed at the restoration
of tropical peatlands and the sustainable use of natural resources they offer. The
CARBOPEAT project is another partnership involving Alterra and a number of international
partners investigating carbon-human-climate relationships in tropical peatlands
(http://www.carbopeat.org). Such research programmes aim to strengthen efforts to restore
and sustainably manage tropical peatlands, and to curb the detrimental effects of peatland
degradation in the tropics.
Lacking in many of these restoration programmes, however, are data to quantitatively assess
the efficacy of these methods in environmental restoration. Such data are especially useful
when peat swamp water management plans are connected to carbon accounting schemes,
where restoration of so-called 'carbon sinks' contributes to a reduction in carbon emissions.
To this end, Alterra and other groups have been actively involved in establishing quantitative
relationships between peatland hydrology, carbon emissions, and ecological restoration
potential. The thesis research described herein took place within a partnership between
Alterra-WUR and the German and Indonesian World Wide Fund for Nature (WWF) as part of
an effort to strengthen the quantitative monitoring of peatland restoration effort in the
Sebangau National Park in Central Kalimantan.
1.2 Problem Description
Changing land use practices have created considerable pressure on the delicate peat
swamps of Kalimantan. In particular, illegal and concession logging in protected peat forests
has resulted in lowering of the groundwater table and peat subsidence. Peat subsidence is
due largely to a combination of deforestation and drainage through the extensive network of
artificial channels that have been dug throughout peat swamps in Kalimantan to aid in the
evacuation of harvested timber (Kool et al., 2006). In addition to degrading the peat-based
ecosystems in this region, loss of groundwater is leading to increased susceptibility to forest
fires which leads to increased carbon emissions from peat soils and higher susceptibility to
flooding during wet seasons (Wösten et al., 2008). Among the methods being proposed to
rehabilitate degraded peat swamps is the construction of small dams across inactive logging
channels to retard canal water flow and trap deposited organic debris, encouraging
groundwater recharge upstream. The problem that this research attempts to address thus
relates to the uncertainty surrounding the effects of small dams in rehabilitating degraded
peat swamps. The problem of accurate hydrological predictions is of interest not only to local
communities and policy makers in Kalimantan, but also to such international actors as the
World Wide Fund for Nature (WWF), who promote peat rehabilitation (and thus increased
carbon sequestration in peat soils) as an activity which can contribute to 'carbon neutrality'.
While preliminary modelling of the regional hydrology in the Sebangau has been performed
using the integrated SIMGRO model (Jaenicke et al., 2009), these efforts suffer from a lack of
3
field data, necessitating the excessive use of assumptions on canal morphology, dam
placements, and other vital parameters (Siderius C., Alterra, personal communication, 2009).
While such computer simulations are important to predicting large-scale patterns in the
Sebangau peatlands, insufficient quantitative data renders model predictions unreliable.
There is therefore a need for more robust data collection from within the Sebangau National
Park, both for ongoing monitoring of activities by WWF and the National Park authorities and
for future hydrological modelling.
1.3 Research Objectives
The ultimate aim of this research is to gain an understanding of the hydrological functions of
tropical peatlands and the effects of water management – including exploitation, degradation,
and restoration – on these functions. Specifically, this research concerns the effect of the
construction of small dams across canals within the Sebangau National Park forest on the
regional groundwater hydrology in the peat swamps of the Sebangau watershed. To address
this question, the following sub-questions were directly addressed:
1) According to the literature, to what extent are the characteristics and functions
(physical, hydrological, and others) of tropical peatlands understood? How can this
understanding of tropical peatland contribute to planning for peatland restoration?
2) What theoretical and practical considerations are being taken in planning hydrological
restoration of the Sebangau peatlands? Are the tools currently available for
hydrological restoration planning valid for the Sebangau National Park?
3) How do artificial drainage canals affect sub-surface flow regimes in the Sebangau
peatlands?
4) What are the effects of canal blocking structures on local hydrology in the vicinity of
the structure over time?
5) What is the current state of the art of hydrological modelling for Central Kalimantan
peatlands? What barriers exist to the implementation of the SIMGRO model for the
Sebangau National Park?
4
2. Materials and Methods
2.1 Site Description
2.1.1 Location
The research described in this report took place within the Taman Nasional Sebangau
(Sebangau National Park). The National Park was officially created on 19 October 2004
(Ministerial Decree SK423/Menhut-II/2004, Indonesia) and covers approximately 568,700 ha
in Central Kalimantan Province. Situated southwest of Palangkaraya, the capital of Central
Kalimantan, the park constitutes part of Katingan District, Pulang Pisau District, and
Palangkaraya Municipality, and is bordered by Sungai (S.; “River”) Katingan to the west and
S. Sebangau to the east. The National Park is flanked by areas of intensive peatland
conversion and degradation, the most widely known of which is the former Mega Rice Project
(MRP) area, which was slated for a failed intensive rice production project in the 1990‟s
The specific study sites were chosen based on previous work being done by WWF-Indonesia
in collaboration with Alterra-WUR Netherlands and Remote Sensing Solutions (RSS GmbH)
Germany, in which three areas of interest (AOI) were delineated (Jaenicke et al., 2009),
roughly covering the S. Rasau, S. Bakung, and S. Bangah sub-catchments (Figure 3). The
data from which this thesis is derived was collected in two study canals as shown in Figure 3:
the Sanintra Sebangau Indah (SSI) Canal and Canal 21 south of the S. Rasau sub-
catchment. Additional data was collected in Canal 23 in the S. Bangah sub-catchment.
Figure 2 – Sebangau National Park boundary and western part of Block C (taken from
http://www.ckpp.org).
5
Figure 3 - Digital Elevation Model (DEM) of eastern Sebangau peatlands indicating the three major
tributaries and three study areas (red squares).
2.1.2 Topography
The Sebangau peatlands are situated on the flat lowland delta of Central Kalimantan. As
such, slopes within the study areas are very small. A Digital Elevation Model (DEM) for the
eastern region of the Sebangau National Park derived by RSS GmbH (Jaenicke et al., 2009;
Figure 3) from radar satellite data reveals the characteristic „dome‟ shape formed by the
ombrogenous Sebangau peatlands. The peat dome rises from a minimum altitude of
approximately 5m above mean sea level (ASL) at the banks of S. Sebangau to a maximum of
approximately 20m ASL in the deep interior of the park. Beyond the extent of the DEM, the
peatlands slope downwards towards the west to S. Katingan to complete the dome shape.
Resulting from an accumulation of peat material over time, this dome shape gives rise to
varying depths of peat throughout the area, such that higher altitude regions are situated on
deeper columns of peat soil than that of lower altitude areas.
2.1.3 Soils and Geology
Given their substantial makeup of soil organic matter (SOM), peatland soils are classified as
histosols, which are comprised of at least 30% SOM (FAO-UNESCO, 1990). The Sebangau
peatlands are mainly composed of two subgroups of histosols: fibrists and hemists
(RePPProT, 1987). Fibrists and hemists differ in the degree to which their organic material
has decomposed. Fibrist soils have a remarkably fibrous texture and contain a large
proportion of identifiable preserved plant material, while hemist soils are moderately
decomposed and contain a lesser amount of identifiable plant material (Wüst et al., 2003).
Extensive oxidation of peat in the top layers in response to water table depression results in
6
progressive decay of fibric and hemic matter. In extreme cases, the resulting black soils could
be classified as saprists, which are more extensively degraded than fibrists or hemists (Moore
et al., 1996).
High rainfall in the Kalimantan lowlands results in reduced ability of aerobic oxidation of
organic material and buildup of peat material. The high water holding capacity of peat in turn
reduces peatland drainage, thus magnifying the reduction in aerobic decomposition of peat
(Hirano et al., 2009). The persistent presence of preserved organic matter gives peatlands
their unique chemical characteristics compared to dryland soils. Release of organic material
into soil pore water generates humic and fulvic acids which result in low pH in the peat soils
and associated watercourses (Weiss et al., 2002). In combination with a low pH, the peat
domes of Central Kalimantan are ombrogenous, which results in low soil nutrient availability
(Lähteenoja et al., 2009). While most of the Sebangau peatlands are situated on quartz sand
deposits (Boehm, 1999) and a sandy mineral sub-layer is evident under shallow peat near the
riverbanks, some of the southern part of the Sebangau catchment towards the Java Sea were
established on underlying pyrite clay soils, which release sulphuric acid and further reduce
soil and water pH when excessively drained (RePPProT, 1987; Haraguchi, 2007).
2.1.4 Climate
According to the Köppen classification system, the island of Borneo is classified as an Af
Tropical Rainforest (Peel et al., 2007). The associated high rainfall and constant amounts of
sun exposure characteristic of equatorial rainforests have given rise to vast extents of
rainforest ecosystems throughout Kalimantan. Annual rainfall data were acquired from the
Palangkaraya airport and the CIMTROP Sebangau research station (Takahashi, 2006,
unpublished data) and are shown in Figure 4. From these data it is evident that in the past 10
years, precipitation in Central Kalimantan has ranged from 1800mm to 3200mm, consistent
with the classification of the region as a wet tropical climate. Also apparent in Figure 4 are the
effects of recurring years (1997, 2002, 2006) in which the El Niño Southern Oscillation
(ENSO) is active. These years are associated with longer dry seasons (and hence lower
annual rainfall) in Kalimantan, which have important consequences on peatland hydrology.
Monthly rainfall data are shown for 2009 in Figure 4 as measured at the WWF meteorological
station at SSI canal (WWF Indonesia, 2009). Distinct wet and dry seasons are evident from
these data, with the wet season extending roughly from October to May, and the dry season
extending from June to September. The rainfall measured during 2009 indicates that this year
was again associated with an ENSO event. The hydrological consequences of the extended
dry season are discussed in section 2.1.5.
7
0
500
1000
1500
2000
2500
3000
3500
1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
Rainfall (mm)
0
50
100
150
200
250
January
February
March
April
May
June
July
August
September
P, ET (mm)
Rainfall (mm)
ETo(mm)
Figure 4 - Yearly rainfall between 1997 and 2006 at the CIMTROP Setia Alam research station in the
Sebangau National Park (top; Takahashi (2006)) and monthly rainfall and evapotranspiration at the
WWF-Indonesia SSI Field Station between January and September 2009 (bottom; WWF-Indonesia).
2.1.5 Hydrology
For the purpose of discussing peatland hydrology, the following water balance is used:
SRETP
where P is precipitation, ET is evapotranspiration, R is surface flow and interflow (including
runoff, canal drainage, and sub-surface flow from the system), ΔS is the change in combined
unsaturated soil water and saturated groundwater storage. As a result of high hydraulic
conductivity (discussed below), precipitation in tropical ombrogenous peatlands is nearly
completely absorbed by the soil until flooding occurs, at which point surface flow through
natural rivers and drainage canals becomes the predominate drainage route. Despite the high
amounts of rainfall in Central Kalimantan peatlands, dry seasons still have the potential to
create a precipitation deficit. The extended dry season of 2009 (Figure 4) resulted in a water
deficit between June and September owing to the fact that evapotranspiration (ETO)
exceeded precipitation during these months. Rearranging the above water balance to yield:
SRETP )(
demonstrates the consequence of such a deficit. Since ombrogenous peatlands are
exclusively rain-fed, R only includes outflow from the system and is thus greater than zero. A
8
deficit (where (P – ET) < 0) therefore results in a negative change in combined unsaturated
and saturated storage (where ΔS < 0), implying a drop in the water table. Longer periods of
water deficit can thus result in a dramatic drop in the water table, a phenomenon reported in
Central Kalimantan during especially dry ENSO years. Forecasted changes in precipitation
patterns across tropical peatlands therefore have serious implications on the ability of
peatlands to store water (Li et al., 2007). The above water balance also implies that increased
drainage (R) through the construction of drainage canals results in a decrease in combined
storage (ΔS), as discussed in section 2.1.6.
The region of the peat column in which the water table experiences seasonal fluctuations is
commonly known as the acrotelm. Below the acrotelm is the catotelm, which is situated below
the groundwater table and is permanently saturated (Quinty and Rochefort, 2003). Figure 5
demonstrates the roles of the different hydrological processes within an arbitrary unit of the
peat dome. Relatively high lateral flow through the acrotelm (interflow) compared to the flow
through the catotelm is a result of differences in soil structure in these two regions. In the
fibric and hemic soils of the acrotelm, large pore size and high water content give rise to very
high infiltrability and saturated hydraulic conductivity (kS). Given the high flow rates in these
soils, kS is indeed difficult to accurately measure in the field or in the laboratory. Macropores
in the soil column result in preferential flows which have a significant influence on kS of peat
soils. In water-logged fibric and hemic soils, preserved root and woody plant matter maintain
these preferential flow routes and give rise to anisotropy between the vertical and horizontal
components of kS (Kruse et al., 2009). Excessive drainage of peatlands alters the structure of
these soils and affects the preferential flow and resulting anisotropy of kS.
Figure 5 - Main hydrological processes within the peat column (photo courtesy of H. Wösten, Alterra-
WUR). Size of arrows indicates relative importance of each process.
9
2.1.6 Land Cover and Land Use
Since the study area is located in a National Park and most exploitive activities are
consequently banned by law, most of the area is covered by natural forest and can be
classified as lowland tropical peat swamp forest. However, some exploitation of the Sebangau
peat forests occurred before creation of the national park and left patches of cleared areas
and an extensive network of artificial drainage canals. Further to direct clearing of forests,
forest fires during especially dry ENSO years have left burn scars scattered throughout the
study area. Figure 6 outlines the main land cover classes in the study areas, as well as the
main study canals: SSI, 21, and 23. Riverine forest or sedge swamp is commonly seen at the
banks of S. Sebangau and its tributaries (including S. Bangah). Burn areas can also be found
in the vicinity of the rivers and canals, the most significant of which is a result of fires in 2002
and 2006 and can be seen around the mouth of canal 21. Most of the open areas unaffected
by fires are a direct result of clear-felling logging activities, in which all trees in a given area
are cut down. The SSI canal is a remnant of concession logging in the 1990‟s by the Sanintra
Sebangau Indah company, and is thus a site of clear-felling, as is evident in Figure 6.
Figure 6 - LandSat-derived land cover classes for SSI canal and the Bangah sub-catchment. (GIS data
courtesy of M. Rosidi, WWF Indonesia).
The land cover classes presented in Figure 6 were derived from LandSat satellite imagery,
but do not provide a detailed description of the different types of forest found throughout the
Sebangau National Park. Studies have uncovered changes in vegetation in northern and
southern locations within what is now the Sebangau National Park (Page et al., 1999).
According to the study of Page et al. (1999), biodiversity and structure of the Sebangau peat
swamp forest is dependent on peat dome structure and nutrient availability in the soils.
Changes in forest zones seem to follow the shape of the peat dome, suggesting a role for
proximity to river flood levels and peat thickness on these forest growth-determining factors.
Table 1 below summarizes the types of forests identified in the Sebangau catchment.
10
Distance from
River
Peat
Depth
Forest Type
0 to 1.5km
up to 1.5m
Riverine forest up to a canopy height of 35m, influenced by
seasonal riverine flooding; now largely destroyed and
replaced by low-canopy sedge swamp.
1.0 - 1.5km
up to 2m
Transitional riverine-mixed swamp forest, influenced mainly
by catchment outflow.
up to 4.0km
up to 6m
Mixed swamp forest with a large diversity of tree species,
and several canopy layers up to 35m high.
4 - 6km
up to 6m
Transitional mixed swamp-low pole forest with a maximum
canopy height of 30m
6 - 11km
7 - 10m
Low pole forest on uneven forest floor with 2 canopies of
maximum height 20m and 15m respectively
11 - 25km
maximum
Tall interior forest on permanently non-flooded peat with at
least 3 canopy layers of maximum height 45m, 25m, and
15m respectively
various
maximum
Very low canopy forest covering a small area at maximum
elevation with a permanently high water table and a canopy
of maximum height 1.5m
Table 1 - Forest cover and associated peat depth within the Sebangau National Park (Page et al.,
1999).
Prior to the creation of the Sebangau National Park, concession and illegal logging operations
alike posed a significant threat to the health of peatland ecosystems throughout Central
Kalimantan. While the protected Sebangau peatlands contain some of the only remaining
pristine peat swamp forests in Kalimantan, logging has left a legacy of degradation within the
National Park. With a lack of infrastructure throughout the region, companies and individuals
excavated drainage canals throughout the forest to aid in the transport of harvested timber
out of the forest. Even after the banning of logging activities in the forest, these canals
continue to accelerate drainage of the Sebangau peat swamps.
While there are no permanent human settlements within the boundaries of the Sebangau
National Park, some temporary settlements exist, mainly for fishing communities and
secondary service sectors. These communities mainly engage in fishing in S. Sebangau and
its tributaries and to a smaller degree in the artificial canals. Aside from fishing, harvesting of
non-timber forest products, including latex and aloe, takes place within the National Park
boundaries. On the eastern banks of S. Sebangau (outside the National Park boundaries),
some land is cultivated for smallholder agriculture as well as large-scale palm plantations.
Located south of the National Park are a cluster of transmigrasi settlements. A result of
government sponsored transmigration policies moving residents from Java to other less
11
populated islands, these communities have added to the impact on the environment around
the Sebangau National Park. Currently, most of these transmigrasi communities are engaged
in small-scale agriculture, agroforestry, fishing, and other economic activities (personal
communication, WWF Indonesia staff).
2.2 Ground Truthing, Dam Planning, and Monitoring Site Selection
Characterization of canals within the study area was undertaken between 2005 and 2009 by
the WWF Sebangau team for the purposes of planning restoration measures and to provide a
baseline for future hydrological monitoring and modeling. For every accessible canal, the
coordinates of the canal mouth, the direction of canal flow, and the width and depth of the
canal incision (regardless of water depth) were measured and recorded. Where possible,
surface water discharge was measured as described in section 2.4. Other qualitative
information such as economic uses of the canal and priority for blocking were documented.
For several select canals, canal location, direction, and bank slope were measured to ground
truth the DEM provided by RSS GmbH. Slopes were measured at approximately 40-50m
intervals up to 2 – 4km inland by using a 50m waterpass and measuring elevations along the
canal banks relative to the canal mouth. Measured slope profiles were compared to derived
elevations from the RSS GmbH DEM in the ArcGIS9.2 and Microsoft Excel 97 software
packages. In the case of small dams, measured or DEM-derived slopes were used to
calculate optimal spacing of the dams within a cascade. To achieve a given head difference
across the dams, spacing of dams within a particular section of a dam cascade was
calculated using the following formula:
m
h
d
where d is the spacing between dams (in m),
h
is the desired head difference across the
dams (in cm), and m is the average slope of the planned cascade (in cm/m). The above
calculation was performed for two sections of the canal transect, dividing the peat dome
shape into a steeper area closer to the riverbank and a relatively flat area extending away
from the river.
Dams were built in small canals using gelam wood imported from outside the National Park
according to the above dam placement calculations (with some exceptions). After surveying
the canal of interest, the location of a large „permanent‟ dam was negotiated with local fishing
communities based on the economic functions of the canal. After building the permanent
dam, cascades of simple small dams were installed according to optimized dam locations.
The regions in the immediate neighbourhood of certain dams were selected for hydrological
monitoring. As shown in Figure 6 (section 2.1.6 Land Cover and Land Use), three locations
were monitored for the purpose of this study: SSI (downstream dam), canal 21 (permanent
dam and two small dams), and canal 23 (permanent dam).
12
2.3 Groundwater Monitoring and Analysis
Groundwater levels were monitored on a monthly basis between September 2006 and
November 2009 for the SSI study site and between June and November 2009 for the canal
21 and canal 23 study sites. To monitor water table levels, simple tubewells were constructed
from perforated PVC pipes of approximately 3m in length, covered with a fine nylon mesh to
minimize sediment infiltration into the tubewell (Figure 7). The bottom of the tubewell was
fitted with a pointed wood block to allow penetration into the soil. At the SSI study site, five
tubewell transects were installed on each side of the canal immediately upstream of the
downstream dam, separated by 50m, creating a 250mX500m grid of 50mX50m cell size
consisting of 25 tubewells on each side of the canal. An additional three transects (total 15
tubewells) were installed and monitored 3.5km upstream of the dam during 2009. 20
tubewells were similarly installed in two transects before and after the first permanent dam in
canal 23 to create a 50mX500m grid of 50mX50m cell size. The transects at canal 23
replaced five tubewells installed along the canal banks prior to dam construction. Tubewell
transects were installed in canal 21 following a slightly different pattern than that of SSI and
canal 23. Out of eight dams in total, left and right transects were installed at dam number 1, 5,
and 8 (counting from downstream). Each transect consisted of five tubewells at a distance of
5m, 25m, 50m, 150m, and 300m from the canal, respectively. Groundwater levels were
measured from the tubewell transects on a monthly basis by measuring the depth from the
top of the tubewell to the groundwater level, and subtracting it from the height of the tubewell
above the soil surface. Positive groundwater levels thus represent flooding conditions, while
negative groundwater levels are situated below the soil surface.
Figure 7 - Tubewell construction (left) and groundwater measurement in the WWF Sebangau project.
13
Groundwater trends were analysed per tubewell in Microsoft Excel 97, and spatial
groundwater patterns were calculated using the Inverse Distance Weighted (IDW) method in
the ArcGIS9.2 software package. The IDW interpolations were carried out according to the
simple Shepard model (Shepard, 1968). Given known values uk (k=0,1,…N) at various
measuring points xk, the value u at each point x within a 2-dimensional monitoring grid was
estimated based on the formula:
where Dk is a function inversely related to the distance d between the known point xk and the
interpolated point x, according to the formula:
The exponent p defines the sensitivity of the interpolation to the distance between the two
points, where large values of p give greater influence to closer distances and increase
sensitivity to increasing distances. Conversely, small values of p give less influence on closer
points, thereby smoothing peaks in the interpolation. For the analyses described in this report,
a default value of 2 was assigned to p for all interpolations.
In some cases, groundwater levels relative to the soil surface were normalized against soil
surface elevation above mean sea level (ASL) by adding water table depths to DEM-derived
soil elevations. Alternatively, local soil surface elevation was measured with respect to canal
elevation with a water-pass, and local water tables were expressed using these elevations.
Here, tubewells were installed such that the top of each PVC pipe was at the same altitude
above a common datum as all others in the transect (Figure 8). Subtracting the measured
height of the tubewell above the soil surface from the reference height H thus gives the
elevation of the soil surface (En) above the datum.
Figure 8 - Tubewell transect design for canal 21. Tubewells were installed using a waterpass such that
the tops were at equal elevation. The groundwater readings for each tubewell (Tn) were subtracted from
the reference tubewell height (H) to obtain relative groundwater level. Likewise, the height of each
tubewell (En) was subtracted from the reference height (H) to obtain relative soil surface elevation.
H
Tn
En
datum
14
To analyze groundwater dynamics at higher temporal resolution, two diver dataloggers (along
with one barometric diver) were installed near canal 21 between August and October 2009.
The divers consist of a pressure cell which measures and logs vertical water pressure on an
hourly basis, while the barometric diver logs barometric pressure at the same interval. For
each diver, a tubewell was constructed to accommodate the diver below the groundwater
table. The tubewell was installed and the height above the soil surface was measured
(represented by B in Figure 9) and the diver was suspended by a string of known length (A).
Since the diver logged the pressure above its pressure cell (C) the depth of the water table
relative to the soil surface (h) was determined by the following formula: h = B – A + C.
Figure 9 - Diver datalogger setup. The diver was suspended from a string from the top of a tubewell, and
water pressure readings (C) were converted to water table depth (h) by the formula h = B - A + C.
2.4 Canal Discharge Measurements
Drainage from canals was estimated by taking the product of flow velocity and cross-sectional
area according to the continuity equation:
dAvQ
where Q is discharge through a cross section of area A, and v is the velocity of the water flow
at each point dA integrated over cross section A. Cross sectional area was estimated by
assuming a rectangular cross sectional area of the artificial drainage canals as shown in
Figure 10 (top). The average depth (hav) was calculated by measuring the depth (h) at three
points separated by a distance of w/4. The average cross-sectional area (A) was calculated
by averaging the cross-sectional areas at the start-point and end-point of the float
measurement (Figure 10, bottom, shown as striped areas AI and AII). Surface water velocity
(vs) was measured with the float method using pieces of waterlogged wood approximately
10cm long and 2-3cm in diameter. The floats were released approximately 1 meter upstream
of the start point (0m) to allow for acceleration to maximum velocity, and timing began at 0m.
The float travel time was recorded for 3 separate trials and averaged. Surface velocity (vs)
was calculated by dividing the length (L) by the average travel time.
15
Figure 10 - Schematic of canal discharge measurements. Average cross sectional areas for the start
and end points (striped areas) were calculated by taking the average depth and canal width (top), and
discharge estimated by taking surface velocity and average cross sectional area for a stretch (bottom).
Discharge (Q) was then estimated by multiplying average velocity (v) by the average cross-
section area and a correction factor (k) according to the following approximation of the
continuity equation:
AkvQS
where:
85.0)( y
k
fk S
This approximation assumes a uniform horizontal velocity profile perpendicular to the
direction of flow. To account for the non-linear vertical velocity profile, measured surface
velocity (vS) was multiplied by a correction factor k. The correction factor k is a function of the
ratio of canal surface roughness (kS) to canal depth (y) as shown in the above approximation,
and was taken to be 0.85 in all cases.
Canal discharge measurements in canal 21 were compared to direct discharge
measurements from the spillway of the permanent downstream dam. Spillway discharge was
measured by setting a calibrated 40L basin underneath the dam and recording the time
needed to fill the basin over three trials. Discharge from the dam spillway was calculated by
dividing the volume of the basin by the average time required to fill it.
16
3. Results and Discussion
3.1 Differences between Tropical and Temperate Peatlands
A discussion on the functions of peatlands necessitates a definition of the term, as mineral-
based wetlands and peatlands can have very different biophysical characteristics. While the
Ramsar Convention provides a very broad definition of wetlands, they are commonly defined
as areas whose soils are permanently or seasonally saturated with water (Neue et al., 1997).
Peatlands form a subset of this category, with the key distinction being that peatlands contain
dead organic material formed and deposited on location. This organic material is defined as
peat and differs from other classes of dead organic material deposited off-site from their
source. Peatlands on which peat is continuously accumulating are known as mires, in
contrast to other peatlands where peat accumulation has stopped (Schuman and Joosten,
2008).
Peatlands can be further characterized as ombotrophic or minerotrophic based on their
primary nutrient source. Ombotrophic peatlands receive nearly all their minerals from rain
water, while minerotrophic peatlands receive their minerals from runoff originating from
upstream mineral soils or from underlying groundwater (Lähteenoja et al., 2009). The
concentration of such common minerals as calcium and magnesium is thus a convenient
indicator of the type of peatlands in question. If calcium to magnesium concentration ratios
mirror that of rainwater, a peatland can be said to be ombotrophic. Minerotrophic peatlands
typically form in depressions and floodplains, where contact with mineral soil and mineral-
containing runoff gives the peat a mineral content comparable to that of its mineral source.
Ombotrophic peatlands, on the other hand, form on flat terrain as convex peat „domes‟,
rendering the deposition of minerals from runoff impossible. While convex peat domes are
nearly always ombrotrophic, they may receive relatively small amounts of mineral nutrients
from groundwater and underlying mineral soil via capillary rise in regions where the peat is
very young and shallow (Lähteenoja et al., 2009). Nutrient concentrations thus decrease with
increasing peat depth, as underlying mineral sources becoming increasingly inaccessible
(Morley, 1981). Similarly, mires can be classified into (raised) bogs and fens, the former being
ombotrophic and generally more acidic than the latter, owing to the influence of precipitation
and lack of input from mineral soil (Sjörs, 1950).
Despite a significant body of knowledge on temperate and boreal peatlands, relatively little is
known about tropical peatlands. Key differences between functional characteristics of
temperate and tropical peatland include vegetation-determining and climate-determining
factors. The vegetative source of peat is an important determinant for many of the physical
properties of peat soils. In temperate peatlands, grasses and shrubs provide the bulk of
organic material to the soil, with sphagnum mosses being the prime contributor in temperate
ombotrophic peatlands. Tropical peatlands, in contrast, consist of a higher number of tree
17
species, which in turn contribute litter and dead matter to the peat soil. Owing to the woody
nature of dead tree matter, tropical peat soils tend to have a markedly different structure than
that of temperate peat soils. Notably, the accumulation of preserved organic material from
trees results in macropore formation in tropical peat soils, with consequences on soil water
transport (Jauhiainen et al., 2005). Hydraulic conductivity in tropical peatlands is discussed in
more detail below.
The second major difference between temperate and tropical peatlands arises from climatic
differences. The hydrological, chemical, and biological processes which facilitate the
formation and decomposition of peat are all sensitive to climatic parameters, which are
considerably different between temperate and tropical regions. For example, after lowering of
the water table, aerobic decomposition occurs at an accelerated rate in the unsaturated zone
of peat soils as compared to that of temperate peatlands. Notably, respiration from tropical
peatlands has been shown in most cases to be temperature-dependent, with warmer
peatlands releasing CO2 at a higher rate than their cooler counterparts (Chimner, 2009).
Additionally, constant temperatures in the tropics throughout the year provide starkly different
environmental conditions than in temperate peatlands, which experience seasonal
fluctuations in temperature. Despite similarities in characteristics between temperate and
tropical peatlands, the differences noted above warrant consideration in peatland
management strategies.
3.2 Formation and Current State of the Sebangau Peatlands
Radiocarbon dating in sites near the Sebangau catchment suggest that peatlands began to
develop over topogenous freshwater swamps in the region starting in the mid-Holocene era
(about 5000-6000 years BP) at approximately the same time that post-glacial sea levels were
at their maxima. Poor drainage in the catchment inhibited decomposition of organic matter,
causing the formation of an ombrogenous peat dome. Progressive deepening of peat layers
towards the centre of the dome gave rise to abrupt changes in vegetation and the
development of different forest zones (Morley, 1981). Today, forest vegetation in the
Sebangau watershed varies in species composition and canopy height with distance from the
Sebangau River as described in section 2.1.6 Land Cover and Land Use.
In the past 20 years, forested peatlands have come under serious threat from increasing
human activities in and around the forests. Fluctuating policies and unsustainable practices
surrounding forest management in Indonesia have resulted in the destruction of vast areas of
rainforest, particularly in Kalimantan (Casson and Obidzinski, 2002). More recently, rapid
conversion of native peat swamps into agricultural land has revealed the fragility of these
ecosystems (Wösten et al., 2008). The recent land use changes in the Central Kalimantan
peatlands has been closely associated with government-sponsored transmigrasi
programmes, where communities from other more populated islands (particularly Java) were
18
moved to sparsely populated areas in the hopes that native forest soils can be easily
converted to high-yielding agricultural land. However, in many cases, these policies failed to
take into account differing biophysical, cultural, and socioeconomic considerations. Indeed,
the Indonesian government's concurrent Mega Rice Project has been named a failure due to
the fact that planners did not account for the unique physical and hydrological characteristics
of the peat swamps, and rapid conversion of peatlands has significantly impaired their
ecological function (Ludang et al., 2007). With global attention now turning to the use of
biofuels, the potential for tropical countries like Indonesia to produce the necessary palm
products constitutes an additional threat to these tropical peat swamps. These potentially
destructive factors illustrate the great need for further understanding of peat swamp dynamics
and sustainable management options.
In response to the issues described above, the plight of Kalimantan rainforests has been
increasingly featured in public environmental discussions in recent years. Of special concern
have been the Sebangau peatlands, as they are among the last remaining habitats of the
endangered orangutan (Morrogh-Bernard et al., 2003). Additionally, there has been growing
recognition of the role of peat swamps and other wetland biomes in sequestering carbon and
mitigating against climate change. Notably, the recognition of the role of terrestrial carbon
sinks in combating climate change in the Kyoto Protocol and subsequent agreements may
provide the necessary incentive for policy makers to formulate more sustainable peatland
management policies (Hecker, 2005).
Some of the programmes working towards restoration and sustainable management of
peatlands in the tropics are described in section 1.1 Background. The research described in
this report follows the completion of both the RESTORPEAT and CARBOPEAT, and was
carried out under a partnership between Alterra, the World Wide Fund for Nature‟s (WWF)
Indonesian and German chapters, and Remote Sensing Solutions (RSS GmbH), a
consultancy agency based in Germany. WWF-Indonesia is among several actors in
Kalimantan who have begun to work with the concept of canal blocking as a peat swamp
restoration technique under the auspices of the Central Kalimantan Peatlands Project
(CKPP). CKPP is a consortium project funded by the government of the Netherlands, and
aims to restore large areas of degraded peatlands in Central Kalimantan through hydrological
and ecological restoration (CKPP; http://ckpp.wetlands.org). The work of WWF Indonesia
within the CKPP mainly focuses on hydrological restoration in the Sebangau National Park
through canal blocking and reforesting in degraded areas. Large scale drainage and
conversion of areas surrounding the national park has provided evidence for the detrimental
impact of extensive peatland drainage on ecosystem health. Previous studies in the Block C
area of the former Mega Rice Project area east of the National Park have shown that areas of
peatland where the water table is allowed to drop to 40cm below the soil surface are at a
significantly increased risk of fire (Wösten et al., 2008). WWF‟s canal blocking programme is
therefore aimed at preventing such a drop in groundwater levels within the Sebangau forest,
19
and to facilitate the ecological restoration of already degraded parts of the National Park.
Furthermore, raising the water table within the Sebangau peatlands will slow the
decomposition of organic carbon and associated emissions of greenhouse gases.
3.3 Tools for Hydrological Restoration Planning
One of the key activities comprising peatland restoration efforts by WWF-Indonesia in Central
Kalimantan is the blocking of drainage canals with dams. Due to high void space, high water
content, and nearly negligible load bearing capacity, peat soils present a significant challenge
to the construction of such load-bearing structures (Islam and Hashim, 2009). As such, the
stability of dams built for peat rewetting depends on their ability to withstand water pressure
from upstream. In some cases, dams built across large canals in the ex-MRP area failed
under water pressure during the high flows of the wet season (S. Limin, 2009, personal
communication). Effects of water pressure on dam integrity were also observed during
September 2009 at the 1st downstream dam of the SSI canal, where peat erosion at the sides
of the dam leads to tunneling. Especially for dams in smaller canals, accumulated litter
deposits in the canal and natural revegetation surrounding the canal are expected to
strengthen the dam against upstream water pressure over time. The ability of the small dams
to withstand canal water pressure is therefore of most concern during the first wet season
following the dam‟s construction, where high water flows are most likely to have an impact on
the dam‟s integrity.
The parameter that is used to monitor the pressure exerted on a particular dam is the surface
water level (head) difference across the dam. For the purpose of small dams in the WWF
study area, a maximum head difference of 20cm to 30cm is recommended to prevent dam
collapse and minimize erosion and tunneling through surrounding peat soil (Jaenicke et al.,
2009), which is accomplished by calculating optimal dam-spacing based on canal slope. To
this end, a Digital Elevation Model (DEM) was generated by Remote Sensing Solutions
GmbH from radar and LandSat imagery from February 2000 (Jaenicke et al., 2009). First, a
Digital Surface Model (DSM) was generated from radar imagery, which included forest
canopy. Tree canopy height was then subtracted from the DSM in areas of known forest
types (described in section 2.1.6 Land Cover and Land Use) as seen in LandSat images. The
resulting DEM was verified by measuring slopes at selected canals using a waterpass, taking
elevation measurements at 50m intervals. The results from one of the ground-truth missions
in the southern area of the National Park are shown below in Figure 11.
20
Canal 22 Slope
0
50
100
150
200
250
0500 1,000 1,500 2,000 2,500 3,000 3,500 4,000 4,500
Distance (m)
Relative elevation (cm)
Elevation (cm)
Elevation DEM (cm)
Figure 11 - Ground truth measurements and DEM-derived slope of Canal 22.
Due to the abundance of litter and the fibrous nature of peat soil, uncertainties of the exact
location of the soil surface gave rise to manual measurement error, evident in the frequent
variation in the manual elevation measurements in Figure 11. The DEM elevations appear to
consistently underestimate the elevation, which could be indicative of a systematic error in
DEM calculation. This trend was also observed in ground truth results for other canals (shown
in Appendix 1 – Ground Truth Data). In the case of most of the ground truth missions, the
systematic error of approximately 50cm arises during the initial rise moving inland away from
the canal mouth. From 1.5km onwards, the DEM and ground truth elevations seem to rise
uniformly, indicating that the error is a result of the DEM‟s failure to detect the higher slopes
near the edge of the peat dome. Although predicted variations in tree canopy height were
considered in the DEM (Jaenicke et al., 2009), neglecting small-scale changes in tree canopy
height from the river banks towards the inland forest can lead to an over-compensation at the
river banks. Since the DEM and waterpass elevations are expressed relative to each other in
Figure 11, such an over-compensation inevitably gives rise to a error as the canopy height
increases moving inland. Systematic error notwithstanding, the DEM seems to give a good
estimation of the characteristic peat dome shape, which is especially evident in ground-truth
comparisons done further than 2km upstream from the river. The DEM is therefore a useful
tool in the planning of dam placements along the course of a canal.
After verifying modeled slopes along several canals, optimal dam locations were calculated
by dividing the desired head difference (25cm) by the average slope in the region. Differences
between calculations using the DEM and those using ground-truth slope profiles were
21
minimized by separating canal transects into two stretches based on the peat dome shape. In
the case of Figure 11, “slope 1” extends from the mouth to 1.5km upstream, while “slope 2”
extends beyond the 1.5km marker. Following these classifications, dam spacing is calculated
as shown in Table 2. According to these results, using the DEM for dam planning, five dams
spaced approximately 300m from each other would be installed in the downstream region
(Slope 1), with dams upstream of the 1.5km mark (Slope 2) being installed at 1.5km intervals.
The differences between DEM and ground truth calculations reflect the systematic error in the
DEM for relatively higher slopes (slope 1), with the error diminishing for flatter areas (slope 2).
DEM
Ground
Truth
Average Slope 1
(x = 0-1.5km)
0.075 cm/m
0.11 cm/m
Average Slope 2
(x > 1.5km)
0.016 cm/m
0.017 cm/m
Dam Spacing (m) -
Slope 1
333
227
Dam Spacing (m) -
Slope 2
1563
1471
Table 2 - Comparison of dam spacing calculations using DEM-derived slopes and manually measured
(Ground Truth) slopes.
3.4 Seasonal Groundwater Trends in the Sebangau Peatlands
To assess the effects of restoration measures on groundwater levels in the Sebangau
peatlands, it is first necessary to discuss the seasonal trends within the peatlands. In the large
SSI canal, WWF-Indonesia has built a long cascade of five large dams, each separated by
several kilometers, where monitoring of the groundwater table has taken place at the dam
closest to the canal mouth since the dam‟s completion in 2005. A detailed description of the
groundwater monitoring grid at the first SSI dam is provided in Appendix 3 – Tubewell
Hydrographs. Groundwater trends since September 2006 at one of the tubewells
approximately 150m downstream of the dam along the canal bank are shown in Figure 12.
Within this time frame, the groundwater at this tubewell fluctuates between -80cm and +30cm
(where negative water table depths are below the soil surface and positive are above the soil
surface). The groundwater fluctuations at this monitoring point show a clear seasonal trend,
where a drying period occurs between May and October, followed by a rise in the water table.
Especially dry ENSO years (2006 and 2009) are associated with dramatic drops in
groundwater levels, where water tables fall below the level at which risk of peat fires have
been found to increase significantly (Wösten et al., 2008). According to these data, dry ENSO
years are either directly preceded or followed by a significant flooding in the area, which is
likely a result of redistribution of rainfall during these years.
22
-100
-80
-60
-40
-20
0
20
40
Sep
06 Nov
06 Jan
07 Mar
07 May
07 Jul
07 Sep
07 Nov
07 Jan
08 Mar
08 May
08 Jul
08 Sep
08 Nov
08 Jan
09 Mar
09 May
09 Jul
09 Sep
09
Depth (cm)
SCL1
Critic al GW limit
Figure 12 - Groundwater trends at the SSI monitoring grid betwen September 2006 and September
2009. The threshold for fire risk is shown as a dotted red line.
To further analyze the groundwater dynamics, water table depths at a tubewell located 250m
away from the SSI canal were superimposed on locally collected precipitation data for some
of the year 2009. Figure 13 demonstrates groundwater fluctuations in response to
precipitation at this monitoring point over several months. From the precipitation and
groundwater trends, there appears to be a lag in groundwater response to precipitation
changes. Where the water table is situated above the soil surface as is the case in February
and March, this lag could be indicative of a surplus in water, despite the relative decrease in
rainfall. Subsequent decrease in rainfall as seen from April onwards is associated with a rapid
decline in water table depths for this monitoring point. A lag in groundwater response has
been observed in a similar study where groundwater recharge was modeled on a monthly
basis in response varying precipitation (Wösten et al., 2008). However, the data shown in
Figure 13 suggest that this lag only occurs during periods of groundwater discharge, and not
during recharge phases, contrasting the observations of Wösten et al. Similarly, hourly diver
datalogger and rainfall data (described below and shown in Figure 14) do not show a lag in
groundwater recharge following precipitation, but rather an instantaneous rise in the water
table, followed by the resumed steady discharge from the saturated zone. The observation
that recharge into the soil column occurs nearly instantaneously with precipitation, while
discharge from the soil is delayed during drying periods may indicate a difference between
rates of vertical infiltration into the soil and lateral drainage from the unsaturated soil column.
Further precipitation-groundwater monitoring is needed to shed light on these trends.
23
Figure 13 – Monthly groundwater trends for the fifth tubewell of the third left-hand transect and
precipitation and evapotranspiration data measured at the SSI dam from September 2008 to November
2009.
Since monthly groundwater measurements are rather limited in their temporal resolution,
diver dataloggers were installed in canal 21 to provide further insight into the fluctuations of
groundwater hydrology on a smaller timescale. After downloading the three months of diver
data, barometric pressure from the baro-diver was subtracted from each of the divers‟ hourly
data to give a corrected value. The results of the 3-month installment of both divers are
shown in Figure 14. Diver 18330 (blue) was installed north of canal 21, near the first dam,
while diver 17927 (red) was installed on the opposite (south) side of the canal. Hourly rainfall
data from the SSI meteorological station is shown superimposed on the diver graph (in
green). Given the close proximity of canal 21 to SSI, there is a near-instantaneous response
of the groundwater to precipitation in the region, confirming that rainfall data from SSI may in
most cases be applicable to most of the research area. During times of drought, as in the
latter part of August and early September, a prolonged lack of rainfall results in the steady
decline in the groundwater levels. However, the reasons for the small differences in discharge
rates between the two diver locations, as observed directly after heavy rainfall events, are not
clear from these results.
24
26/07/09 05/08/09 15/08/09 25/08/09 04/09/09 14/09/09 24/09/09 04/10/09 14/10/09 24/10/09 03/11/09
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
18330 (cm)
17927 (cm)
Rainfall (cm)
Date
Depth from Soil Surface (cm)
-18
-17
-16
-15
-14
-13
-12
-11
-10
24-08-2009
12:00 25-08-2009
0:00 25-08-2009
12:00 26-08-2009
0:00 26-08-2009
12:00 27-08-2009
0:00 27-08-2009
12:00 28-08-2009
0:00 28-08-2009
12:00
Depth from Soil Surface (cm)
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
ETo (cm)
18330 (cm)
Eto (cm)
Figure 14 - Hourly water table depth at canal 21 as measured by diver datalogger (red and blue) and
precipitation data at the SSI field station (green) between July and October 2009. Water table depth
and evapotranspiration for a dry three day period are shown in the bottom panel.
In Figure 14, there is also evidence for diurnal variation in groundwater levels. To examine
this phenomenon more closely, groundwater from three days without rain (25 – 27 August,
2009) were superimposed on hourly evapotranspiration (ETO), as in Figure 14. Though it is
tempting to link groundwater fluctuations with periodic fluctuation in evapotranspiration such a
direct relationship would require a source of water for recharge on a daily basis. A more likely
cause for the observed diurnal fluctuations in groundwater levels could be from the incursion
of tidal waters into the coastal regions of the Sebangau watershed. Research related to peat
water chemistry in and near the Sebangau National Park points to tidal effects from the Java
Sea to the south reaching up to 80km inland, extending northward beyond the SSI canal
(Haraguchi, 2007). To confirm this hypothesis, diurnal water table fluctuations should
correlate with surface water fluctuations in nearby canals, in which case flow resistance within
Rainfall (cm)
25
the peat subsoil would cause a damping effect on the range of fluctuation. This phenomenon
was observed in canal 21 when surface water discharge was measured at the same stretch
on two occasions. While the first measurement was taken to be 0.019m3/s, measurements
taken at the same stretch 28 days later showed a higher average depth (and thus higher
average cross-sectional area), but a markedly lower discharge (0.0097m3/s). This reduction in
discharge, despite an increase in canal cross-sectional area, could possibly be explained by
friction from incoming tides. Regular discharge measurements at a canal stretch located in
the vicinity of a diver datalogger may assist in testing the hypothesis of tidal influence on
groundwater fluctuations.
While useful in discussing small-scale temporal fluctuations in groundwater levels, caution
must be taken in interpreting diver datalogger data, as the diver measurements seem to be
underestimates according to periodic manual measurements. It is unlikely that water retention
in unsaturated peat is the cause of any underestimates, as the air entry value of fibrist peat is
negligible, and tubewell water levels should closely match those of the true water table. It is
possible, however, that a systematic error is the result of a consistent underestimate of
barometric pressure by the baro-diver. Perhaps more accurate measurements of barometric
pressure directly above the groundwater column can be obtained by installing the baro diver
inside the tubewell above the water table.
3.5 Effects of Artificial Drainage on Peatland Hydrology
An effective restoration strategy for artificially drained peatlands requires an understanding of
the effects of drainage canals on peatland hydrology. The effects of drainage canals on local
hydrology were monitored in the area around canal 21, where a cascade of eight dams and
three monitoring transects were installed. Here, transects were designed perpendicular to the
canal such that tubewells were placed in close proximity to each other near the canal bank,
and farther apart moving away from the bank. Specifically, tubewells in each transects were
installed 5m, 10m, 50m, 150m, and 300m away from the canal. Such an arrangement allows
for a more detailed observation of the influence of the drainage canal on sub-surface flows in
the vicinity of the canal. Of the three transects shown in Figure 15 (top), the first transect
from the mouth of the canal (right-hand side) showed the most clear effect of the canal on the
surrounding area. As shown in the middle panel of Figure 15, there is a marked difference in
groundwater depths from the soil surface between the right-hand and left-hand transects.
Extending to the left of the canal, the groundwater level was situated well below the soil
surface for the duration of the dry season. In contrast, the groundwater level to the right of the
canal seems to follow the soil surface closely. While the groundwater levels relative to the soil
surface are useful in discussing temporal trends as in section 3.4 Seasonal Groundwater
Trends in the Sebangau Peatlands, analyzing groundwater flows requires a common datum
for all groundwater depths.
26
Figure 15 – Top: Groundwater monitoring transects (solid lines) at canal 21 with respect to dams (X)
and diver dataloggers (triangles), with left (L) and right (R) directions indicated. The predicted location of
the canal is shown as a solid line, and the true location of the dam is shown as a broken line. Middle and
bottom: Groundwater levels near dam 1 (right-hand transect) at canal 21 from July 2009 are shown with
respect to the soil surface (middle) and an arbitrary datum (bottom).
Soil Surface
Water Table
Predicted
Actual
L
R
L
R
27
To observe the groundwater patters in comparison to local topography around the canal, the
tubewells in canal 21 were installed such that the tops of the tubes in each transect were at
the same elevation as each other (Figure 8, section 2.3 Groundwater Monitoring and
Analysis). With such an arrangement, it is possible to measure the water table depths relative
to an arbitrary datum, regardless of the change in soil surface elevation. As shown in Figure
15 (bottom), the water table in the right-hand (southern) transect is situated consistently at the
soil surface and follows the soil surface as it descends away from the canal, with some minor
drawdown adjacent to the canal. The left-hand (northern) transect, on the other hand, does
not follow the soil surface upwards as it does in the right-hand transect. It is conceivable that
before the construction of canal 21, the water table followed the soil surface very closely in
this area, as seen in the right-hand transect. Given the local topography, Darcian flow would
imply an approximate north to south direction of sub-surface flow. Construction of the canal
then disrupted this sub-surface flow, causing a higher rate of drawdown on the northern
(upslope) side than that of the southern (downslope) side. These results suggest that
groundwater drawdown from drainage canals is dependent on the original flow direction of the
groundwater. When canals are dug along the contour lines of the water table (and thus
perpendicular to subsurface flow), as in canals 21 and SSI, drawdown next to the canal is
higher on the upslope side of the canal, where gravity-driven subsurface flow directs
groundwater into the canal. Aside from a smaller degree of drawdown on the downslope side
of the canal, the original subsurface flow is generally uninterrupted by the canal.
In addition to topographical effects on peat drainage, oxidative decomposition of the peat after
drainage can also be a contributing factor to reduced conductivity within the top layers of the
soil column. First, peat decomposition can lead to reduced anisotropy in hydraulic conductivity
due to loss of preferred pore orientation (Kruse et al., 2009). A second factor to be considered
is the ability of decomposed peat to retain water. In many cases, drained, deforested, or
burned peatlands exhibit lower capacity to hold water, often leading to extensive flooding
during the wet season (Page et al., 2009). This scenario likely holds true for the region
surrounding canal 21, which is still recovering from large fires in 2002 and 2006. Peat
drainage thus gives rise to a cascade of consequences within the soil column. First, drainage
leads to peat decomposition, subsidence, and increased fire risk, which in turn change the
structure of the soil matrix and thus the water retention qualities of the soil. Given these
contributing factors, an important question to be answered relates to the relative importance
of surface runoff in instances of reduced infiltrability and conductivity resulting from drainage-
associated peat decomposition and compaction.
3.6 Efficacy of Water Retention by Dams
Regularly monitoring the impact of canal blocking on the groundwater levels around the canal
is essential to a hydrological restoration programme involving canal blocking. Though time-
scale hydrological data is limited for the peatlands of Central Kalimantan, the large dam built
28
at the SSI canal in 2004 (Figure 17) provides a good opportunity for a preliminary assessment
of dam performance in peatlands rewetting. The structural details of the SSI dam discussed in
this report are described in Appendix 2 – SSI Dam Description. The effect of the SSI dam on
the groundwater in the surrounding area was assessed through regular groundwater
measurements at a 250mX500m monitoring grid (described in detail in Appendix 3 – Tubewell
Hydrographs). In addition to this monitoring grid, water table depths were measured on a
monthly basis in three transects situated 3.5km away from the first SSI dam.
Figure 16 demonstrates the difference between one transect next to the large dam (SAL1-5;
dotted lines) and another transect 3.5km upstream of the dam (SFL1-5; solid lines) between
November 2008 and November 2009. While groundwater levels in the SAL transect rise well
above the soil surface between December and May, SFL water levels remain comparatively
close to the soil surface. Similarly, water levels near the dam remain considerably higher than
those upstream in the dry season, confirming the role of the dam in retaining water during the
dry season. In Figure 16 there appears to be little difference between groundwater levels
between the two regions during the early wet season (November to January). At the peak of
the wet season, however, floodwater is retained behind the dam in the SAL transect, while it
drains as surface flow in the SFL transect. Though more data is needed to confirm this trend,
this observation could indicate that the dam serves as a drainage barrier in this large canal by
performing two functions: retarding subsurface drainage during the dry season, and retaining
floodwater during the wet season.
In the SAL transect, the effect of the dam is most obvious between the months of February
and May. During this time, the tubewell closest to the canal maintains the highest water table,
with the effect decreasing as distance from the canal increases. In contrast, groundwater
levels in the distant transect are nearly equal, and do not show any trends with relation to
distance from the canal. This difference indicates that in transects near the large SSI dam,
groundwater levels are elevated to a degree dependent on the distance from the dam, an
effect which disappears in areas further upstream of the dam, as in the SFL transect.
The effects of the dam on drainage can also be seen by observing differences in groundwater
recharge rates between the two monitoring transects. In the distant SFL transect,
groundwater levels in all of the five tubewells seem to be nearly uniform throughout the
months of November 2008 to August 2009. At the beginning of the 2009 wet season,
however, there is a divergence between the tubewell groundwater levels, owing to different
groundwater recharge rates. During these months, the groundwater levels at the SFL1
tubewell (closest to the canal) increased at a lower rate than groundwater levels at he SFL5
tubewell (furthest from the canal). This seemingly inverse relationship between the distance of
a monitoring point to the canal and the recharge rate is likely due to the fact that close
proximity to a drainage canal and associated groundwater drawdown causes a reduction in
29
groundwater recharge in response to precipitation. Interestingly, this inverse relationship is
not observed directly adjacent to the dam in transect SAL. In this transect, there is no
discernable divergence in groundwater levels, and recharge seems to occur at comparable
rates, regardless of the distance from the canal. Since the presence of a large dam restricts
drainage from the region directly upstream, there is no significant drawdown in at the SAL1
tubewell (near the canal and the dam) and thus no significant difference in groundwater
recharge rates between SAL1 and other tubewells in the SAL transect. Further monitoring
and comparing of these transects over several years will help in confirming these
observations and in further elucidating the mechanism of canal blocking by these dams.
-150
-100
-50
0
50
100
Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov
Depth SS (cm)
SFL1
SFL2
SFL3
SFL4
SFL5
SAL1
SAL2
SAL3
SAL4
SAL5
Figure 16 - SSI groundwater trends from November 2008 to November 2009 for a transect adjacent to
the large dam (dotted lines) and a transect 3.5km upstream of the dam (solid lines).
Figure 17 - View of the SSI dam looking upstream during the dry season.
30
To gain further insight into the efficacy of peatland rewetting by the canal blocking method,
spatial groundwater patterns were mapped based on data from the monitoring transects. To
this end, average groundwater levels over the entire 250mX500m SSI monitoring grid were
interpolated using the Inverse Distance Weighted (IDW) method in ArcMap9.2 with the
ArcHydro extension to generate water table contour maps within the monitoring grid. The IDW
algorithm estimates the value of a pixel based on the known values of neighbouring pixels.
The weight (or priority) a pixel is given is dependent on the proximity of that pixel to the pixel
of interest (described in more detail in 2.3 Groundwater Monitoring and Analysis). The
interpolated groundwater levels relative to the soil surfaced averaged between 2006 and 2009
are shown in Figure 18. In this figure, blue regions indicate flooding conditions while brown
regions indicate regions where the groundwater table is on average situated below the soil
surface. The approximate location of the canal is represented by a blue line, and water flows
in the canal from west to east. The large dam is situated in line with the eastern most
transects (transects SAL and SAR, extending north and south, respectively).
Figure 18 - Interpolated water table contour maps with respect to soil surface (left) and mean sea level
(right). Flow direction in the canal is indicated by a blue arrow. A possible path for preferential rewetting
is indicated as a large green arrow in the right-hand figure.
According to the interpolated groundwater levels expressed relative to the soil surface (Figure
18, left), flooding conditions prevailed in the transects adjacent to the dam, while the effect of
the dam rapidly declines moving away from the dam. Interestingly, the higher average
groundwater levels to the south of the canal suggest that the dam rewets the upstream area
in an asymmetric fashion. To ascertain whether this observation is due to actual flow of water
or an effect of surface topography, water table depths were normalized against soil surface
elevations above mean sea level (ASL) according to the DEM. The normalized average
31
groundwater levels between 2006 and 2009 are shown in Figure 18 (right). In this figure, the
rise of the water table with the natural peat dome topography from southeast to northwest is
apparent, with some other distinctive features. Notably, the elevated water table south of the
canal observed in the left panel of figure 18 is still evident. Just as the orientation of the canal
affects the degree to which drawdown occurs on each side as discussed in section 3.6, water
retention behind the large SSI dam also seems to be affected by the orientation of the canal
with respect to the water table contour. In this case, the general direction of sub-surface flow
is to the south-east, resulting in the preferential flow of water upstream of the dam to the
southern side of the canal (shown as green arrows in Figure 18, right).
3.7 Variance in Dam Performance over Time
While the 3-year averages described in section 3.6 Efficacy of Water Retention by Dams give
an indication of the spatial effects of the SSI dam on the upstream area, the annual variability
of the water table within the monitoring grid is not clear, given the variability in precipitation
from year to year (see section 2.1.4 Climate). Average groundwater levels were thus
calculated for each tubewell in the SSI monitoring grid on a yearly basis from September of
each year to the August of the following year. The interpolated yearly average water table
depths are shown relative to the soil surface and normalized against soil surface elevation in
Figure 19 (bottom). The total precipitation in the region was calculated for the same time
period (September to August) and is shown superimposed on Figure 19. Interestingly, the
average yearly groundwater levels do not seem to follow the total measured precipitation, as
would be expected if no other factors were at play in the groundwater dynamics. Between
2007/2008 and 2008/2009, there seems to be an increase in the overall average groundwater
levels, with an associated decrease in total precipitation. The implications of this observation
may be understood by rearranging the water balance in section 2.1.5 Hydrology as follows:
outin RETRPS )(
where a term for inflow from upslope areas (Rin) has been introduced, since only a small sub-
region of the Sebangau peat dome is being considered in this case. Assuming that Rin is
dependent on total precipitation and that total yearly ET remains relatively constant from year
to year, a decrease in P (and thus a decrease in
)( in
RP
) and an increase in
S
implies a
significant decrease in Rout over time, as shown by taking the derivative of the above equation
over time:
0
)(
0)(
)(:0)(
)(:0)(
)()(
out
outin
outin
in
outin
dR
dRRPd
dRRPd
observedSd
observedRPd
dRRPdSd
In other words, it seems from these results that since completion of the large downstream SSI
32
dam in 2005, there has been a progressive reduction in drainage from the area directly
upstream of the dam, despite yearly variation in precipitation in the area. The apparent
increase in the dam‟s canal blocking function over time could possibly be a result of a
progressive build-up of peat sediment behind the dam, increasing the dam‟s ability to retain
surface water and promote infiltration into the surrounding peat soil. To confirm the validity of
this conclusion, further monitoring of groundwater levels in the SSI grid should be coupled to
surface water discharge regularly measured from the SSI canal. Secondly, to test the role of
sedimentation behind the dam in reducing drainage over time, sediment stage (defined as the
height of the canal bed above an arbitrary datum below the bed) should also be monitored
over time.
33
Figure 19 - Interpolated groundwater levels at the SSI monitoring grid relative to soil surface (previous
page) and normalized against DEM-derived soil surface elevation averaged for each year from
September 2006 to September 2009. Rainfall on each figure was measured during the same time period
at the SSI field station and Palangkaraya airport.
In Figure 19, there is also evidence for uneven water retention between the northern and
southern transects. This difference is most obvious in the eastern-most transects SAL and
SAR (north and south of the canal, respectively). In all of the three years, there seems to be a
higher degree of retention from the dam. To gain further insight into this pattern, groundwater
levels for each month from February 2009 to September 2009 were interpolated for all points
within the SSI monitoring grid as described in section 2.3 Groundwater Monitoring and
Analysis. These interpolated monthly groundwater levels are shown relative to the soil surface
in Figure 21. The elevated water table in the SAR transect compared to that of its northern
34
counterpart was most obvious during the wet season, and especially moving into the dry
season during the month of May as water levels are falling. During the dry season between
June and September, the water table within the entire monitoring grid fell below the soil
surface, and the pattern between transects SAL and SAR were no longer evident. As shown
in Figure 20, elevated canal stage during the wet season has resulted in the formation of a
bypass channel to on the southern side of the SSI dam (in the direction of the SAR transect).
The area south of the canal just downstream of the SSI dam acts as a retention area during
the wet season. Thus, groundwater levels south of the dam are significantly elevated during
the wet season as a result of free surface water flow. Since the peat dome slopes in a south-
easterly direction, this bypass was likely formed as a direct result of preferential gravity flow
within the canal. When the water levels in the canal are lowered during the dry season, the
water table in the area around the bypass retreats below the soil surface, surface water flow
is again restricted to the main canal, and surface water discharge within the bypass is
reduced to zero. Establishing an additional groundwater monitoring transect downstream of
the dam and monitoring surface water discharge from both the dam and the bypass on a
monthly basis would confirm the relationship between bypass flow and groundwater retention
to the south of the dam.
Figure 20 - SSI dam spillway (left) and natural bypass to on the south side of dam (right). (Photo
courtesy of Henk Wösten, February 2009).
35
Figure 21 - Interpolated monthly groundwater levels at the SSI monitoring grid relative to the soil
surface. The broken black line in the September panel indicates the distances from the canal until which
the groundwater levels increase in response to dam structural repairs between August and September
measurement dates. The shaded rectangle in the September panel indicates a lack of data for that
transect.
36
In addition to uneven retention and the formation of a bypass, the monthly interpolated
groundwater levels in figure 21 show an unexpected increase in groundwater levels in some
tubewells between the months of August and September. Given the fact that September
experienced very little precipitation (Figure 4, section 2.1.4 Climate), groundwater levels are
expected to decline from August to September as a result of continued surface and sub-
surface drainage. On the contrary, there is an increase in groundwater levels between August
and September in the tubewells closest to the canal in either direction. This seeming anomaly
can be explained by the fact that the large dam underwent repairs (Figure 22) between the
August and September measurement dates to slow water leakages around the periphery of
the dam. The increase in groundwater levels within the SSI monitoring after these repairs
provide a rough estimate of the area around the canal that is most sensitive to changes in
dam functionality. By determining which tubewells experienced an increase in groundwater
levels from August to September, the area that responds to such changes in dam functionality
was estimated (full dataset shown in Appendix 3 – Tubewell Hydrographs). This area (shown
as a black line in the September panel of appendix 21) covers more of the right-hand (south)
side of the canal than the left-hand side, again suggesting an asymmetric effect of the dam in
rewetting. Taken together with the observed bypass to the south of the dam, these results
suggest that dam-induced rewetting, like canal-induced drawdown, acts in an asymmetric
manner depending on the orientation of the canal with respect to the water table contours.
Figure 22 - Repairs on the first SSI dam on 27 August.
3.8 Interrelationships between Groundwater Hydrology and Peat Fire Risk
In addition to reducing carbon emissions from peatlands due to oxidative decomposition in
drained areas, the objectives of canal blocking and peatland restoration can be extended to
the prevention of peat fires. Between the years of 1997 and 2007, 21% of Borneo‟s land mass
37
was affected by fires, the majority of which constitutes Kalimantan‟s extensive peatlands
(Langner and Siegert, 2009). The fact that these fire events were more widespread and
severe during the extended dry seasons of ENSO years exemplifies the role of canal blocking
and peatland rewetting in the mitigation against peat fires. The Sebangau National Park was
affected by an unseasonably high number of fire events during the extended dry season of
2009, as shown in Figure 23. Here, „hot spots‟ are defined as fire events as interpreted from
satellite data. There is a clear connection to the observed hot spots and the precipitation,
where the number of hot spots dramatically dropped with the beginning of the October rains.
While it is clear that heavy precipitation prevents ignition and spread of fires during the wet
season, the height of the water table also plays a critical role in stemming spread of fires
during the dry season. Serious fires during 1997 in Central Kalimantan were estimated to
have extended up to 70cm below the soil surface (Boehm et al., 2001), while other studies
have also shown the effects of subsurface peat fires on wetland processes (Ellery et al.,
1989). A hydropedological modeling study in Central Kalimantan found a correlation between
hot spots in 1997 and regions where the water table fell below 40cm below the soil surface
(Wösten et al., 2008). This depth is therefore used as a threshold for fire risk assessment as
shown in Figure 12 (section 3.4 Seasonal Groundwater Trends in the Sebangau Peatlands),
and provides a useful tool in the restoration and management of peatlands.
Figure 23 - Hot spots within the Sebangau National Park as detected by satellite imagery. (Data
courtesy of M. Rosidi, WWF-Indonesia, 2009).
As is evident in Figure 12 (section 3.4 Seasonal Groundwater Trends in the Sebangau
Peatlands), groundwater levels around the SSI monitoring grid fell below the -40cm risk
threshold on two occasions within the time of monitoring. As an ENSO-associated year, the
2009 dry season saw dramatic drops in groundwater levels in both SSI and canal 21. In
Figure 26, groundwater levels in the monitoring transect downstream of the dam cascade in
canal 21 remained below the threshold for all of the dry season extending into the following
wet season. In contrast, groundwater levels in the two monitoring transects within the dam
38
cascade remained far above the -40cm threshold for the entire dry season as a result of water
retention by the dam cascade. Since the dry season can be considered to be the time period
carrying the highest risk of peat fires, this trend has important consequences for the function
of dams in the Sebangau peatlands. The area surrounding first downstream monitoring
transect (downstream of the dam cascade) was in fact affected by a fire at the end of August
2009. In assessing the damage from the fire, it was noted that the above-ground fire damage
extended from the mouth of the canal to the first permanent dam. Interestingly, the area
upstream of this dam was unaffected by fire, as shown in Figure 27. Taken together with the
observation that groundwater levels remained elevated within the dam cascade, but not
downstream of the cascade, this fire event demonstrates the positive role of the dam in
preventing the spread of subsurface peat fires.
Just as groundwater levels throughout the year play a role in the spread of peat fires, the
history of fires in turn affects the distribution of water throughout the peat column. While the
flooding conditions shown in Figure 26 are certainly a result of the water retention by the dam
cascade, the land cover surrounding this part of canal 21 likely plays a role as well. Areas of
peatlands scarred by fires are known to be more susceptible to flooding due to decreased
uptake of water by vegetation and reduced soil water retention capacity (Page et al., 2009),
which could possibly explain the flooding conditions surrounding canal 21. Thus, the dam
cascade in canal 21 acts together with the land characteristics in influencing the hydrological
conditions in the area. First, flooding conditions prevail in the upstream reaches of the burn
scar area. In addition to these flooding conditions, the dam cascade in the downstream
reaches of the canal induces the retention of water in the upstream reaches.
Figure 24 - Permanent dam (left) and small dam (right) in Canal 21.
39
Figure 25 - Head difference (cm) across each dam in canal 21 expressed as relative water levels.
Locations of groundwater monitoring transects are indicated by arrows.
Figure 26 - Groundwater levels relative to soil during July 2009 at the three canal 21 monitoring
transects. Letters correspond to transect locations as shown in figure 20.
Figure 27 - Downstream (left) and upstream (right) of the first dam at canal 21 after a fire at the end of
August 2009. The fire spread from the mouth of the river to the 1st dam.
A
B
C
A
C
B
40
3.9 Potential for Hydrological Modeling
While the results of this study provide a useful analysis of the local effects of dams on
drainage and peatland hydrology, it fails to provide an understanding of the large-scale effects
of canal blocking on peatland hydrology. Scaling up hydrological analyses is especially
important for large-scale programmes and policies focussed on peatland ecosystem health.
For example, accounting for the prevention of carbon emissions through peatland restoration
and reforestation requires detailed and accurate hydrological monitoring methods. For such
an analysis, an integrated hydrological model is needed to simulate above and below-ground
processes. Though such modeling was not undertaken in this study, a brief description of the
current state of the art for one such hydrological model is provided, preliminary model outputs
for the Sebangau peatlands are discussed in light of the results of this thesis, and a
discussion regarding future modeling are given in this section of the report.
SIMGRO (SIMulation of GROundwater) is a dynamic integrated model which simulates soil-
water-atmosphere interactions within Soil-Vegetation-Atmosphere Transfer (SVAT) units
(Walsum et al., 2006). Paired with MODFLOW, the model's outputs are extended to describe
hydrological flows both locally and regionally (Harbaugh et al., 2000). The regional aspect of
the modeling is especially useful in planning water management interventions in a specific
region, as in the case of the Sebangau peatlands. The capability of the SIMGRO model to
scale up hydrological model outputs to the regional scale stems from its ability to link
processes within all of the relevant zones, including atmosphere, the soil column,
groundwater aquifers, and surface water bodies. Of particular interest to this study is the
integration of groundwater and surface water processes as groundwater discharge or
recharge. In general, water flows through the soil column are calculate in three dimensions
according to Darcy`s Law (Walsum et al., 2006):
zyx
Kq ,,
where q is the discharge flux (md-1) and is a product of Kx,y,z, the 3-dimensional anisotropic
conductivity tensor, and the 3-dimensional gradient in pressure,
, which is determined by
hydraulic head, h. As shown in Figure 28 below, the total hydraulic head difference between
surface water and groundwater zones is a sum of several hydraulic head differences
contributing to horizontal flow, vertical flow, and radial flow towards the canal. The effects of
slope (gravity) on canal-associated groundwater drawdown noted in section 3.5 Effects of
Artificial Drainage on Peatland Hydrology would be expected to factor into these sub-surface
transport processes. Effectively, the difference in canal surface water head and groundwater
head in the vicinity of the canal defines the rate at which water discharges into the canal (or
recharges into the groundwater zone). Elevation of the surface water head by a dam cascade
as in Figure 25 (section 3.8 Interrelationships between Groundwater Hydrology and Peat Fire
Risk) should therefore reduce the discharge, q, in the above equation by reducing the total
head difference, and in turn the pressure gradient.
41
Figure 28 - Hydraulic head differences between surface water bodies and the groundwater zone
included in SIMGRO drainage calculations (Walsum et al., 2006).
Calibration and validation of the SIMGRO model have been carried in the Block C area
adjacent to the Sebangau National Park in studying peat swamps restoration strategies
(Wösten et al., 2008). Following the calibration for the Central Kalimantan case, preliminary
modelling was done under the Alterra-WWF partnership, using 1997 climatic data simulating
the effects of hypothetical dams within the Sebangau National Park (Jaenicke et al., 2009).
The results of this preliminary modeling are shown in Figure 29 below for the area around the
Bangah sub-watershed, where changes in water levels after installation of all hypothetical
dams are estimated based on altered surface water resistance within the canals.
Figure 29 - Predicted changes in groundwater levels using the SIMGRO hydrological model after
construction of dams in the Bangah sub-catchment. (Figure courtesy of C.Siderius, Alterra-WUR, 2009).
The preliminary modeling results shown in Figure 29 suggest that dam cascades have the
potential to rewet areas in varying degrees, depending on the location of the canal. Notably,
canal 21 (indicated by an arrow) seems to have the potential to rewet a significant area north
42
of the canal, with minimal influence on groundwater levels to the south of the canal. This
prediction is consistent with local-scale observations around canal 21, where canal-induced
drawdown was only evident on the north-side of the canal due to gravity flow towards the
canal. Elevation of the water table thus only takes place on the north side of the canal, while
subsurface flow continues uninterrupted away from the canal to the south.
Despite the fact that the SIMGRO predictions are in some ways consistent with observations
from this study, attempts to model the hydrology within the Sebangau peatlands have faced a
number constraints. Ground-truth missions in the Sebangau National Park in 2007-2009
illustrate the inaccuracies of the currently used watercourse map regarding a number of
details (data shown in Appendix 1). Notably, the locations of a number of the small canals
were estimated based on data collected at the mouth of the canal, and other estimated or
assumed data. Since small canals are very difficult to detect using satellite imagery, canal
courses are mostly based on estimates by WWF staff or the local communities. Further
ground truthing and canal characterization is required to improve the watercourse map, which
in turn will improve drainage predictions by SIMGRO.
43
4. Conclusions
The aim of this study was to determine the effects of hydrological restoration efforts on
peatland function. More specifically, the effects of drainage canals on peat hydrology were
observed, and the efficacy of canal blocking methods in mitigating against drainage-induced
peat degradation was assessed through groundwater and surface water monitoring. Canals
were found to induce local drawdown in a slope-dependent manner. The results of this study
show that the side of the canal from which water is flowing towards the canal experienced a
greater degree of drawdown than the side from which water is flowing away, according to the
original water table contours. Thus, canals dug parallel to the water table contours should
exhibit the greatest asymmetry in groundwater drawdown, as sub-surface flow would be
expected to perpendicular to the canal. These observations were consistent with preliminary
modeling predictions (Jaenicke et al., 2009), where water retention was predicted to take
place in blocked canals in an asymmetric manner, depending on the slope of the water table
on either side of the canal. This observation has implications on restoration planning in the
Sebangau peatlands, as the prioritization of canals to be blocked could take the severity of
hydrological drawdown into consideration.
The presence of dams was found to affect the drainage and local hydrology in and around the
canals in a number of ways. Namely, dams function by reducing local drawdown, retarding
subsurface drainage during the dry season, and retaining floodwater during the wet season.
The function of a large dam over a period of three years was found to improve, whereby
drainage was progressively reduced despite varying water inputs to the system, suggesting a
positive role of post-dam sedimentation in dam performance over time. While these results
require more extensive monitoring to confirm their validity over time, these observations
provide preliminary affirmation into the efficacy of hydrological restoration through canal
blocking.
Several constraints were faced in this research which should be addressed in future studies.
First, groundwater monitoring transects should be established in canals lacking dams to
determine the hydrological patterns in continually drained peat. Second, further control data
could be collected in native peatlands, away from drainage canals, to provide a comparison
between native, drained, and restored peat within the same study area. Such controls may
also be important in the future if the current hydrological restoration programme is to be
integrated with Forest Carbon or other carbon trading schemes, where accurate and detailed
baseline and project data are required. Hydrological modeling using such integrated regional
models as SIMGRO will without a doubt be instrumental in the scaling-up of this study. To this
end, refinements in input data from the Sebangau peatlands are required, including more
detailed data on canal courses and characteristics.
44
References
Boehm, H.D.V., Siegert, F., Rieley, J.O., Page, S.E., Jauhiainen, J., Vasander, H. and Jaya,
A. 2001. Fire impacts and carbon release on tropical peatlands in Central Kalimantan,
Indonesia. Proceedings of the 22nd Asian Conference on Remote Sensing.
Boehm, H.-D.V. and Siegert, F., 1999. Application of Remote Sensing and GIS to survey and
evaluate tropical peat, International Conference and Workshop on Tropical Peat Swamps:
"Safeguarding a Global Natural Resource", Penang, Malaysia.
Casson, A. and Obidzinski, K., 2002. From New Order to Regional Autonomy: Shifting
Dynamics of "Illegal" Logging in Kalimantan, Indonesia. World Development, 30(12): 2133-
2151.
Chimner, R.A., 2009. Soil Respiration Rates of Tropical Peatlands in Micronesia and Hawaii.
Wetlands, 24(1): 51-56.
Ellery, W.N., Ellery, K., McCarthy, T.S., Cairncross, B. and Oelofse, R., 1989. A peat fire in
the Okavango Delta, Botswana, and its importance as an ecosystem process. African Journal
of Ecology, 27(1): 7-21.
FAO-UNESCO, 1990. Guidelines for soil description, Rome.
Haraguchi, A., 2007. Effect of sulfuric acid discharge on river water chemistry in peat swamp
forests in Central Kalimantan, Indonesia. Limnology, 8(2): 175-182.
Harbaugh, A.W., Banta, E.R., Hill, M.C. and McDonald, M.G., 2000. MODFLOW-2000, The
U.S. Geological Survey Modular Ground-Water Model - User Guide to Modularization
Concepts and the Ground-Water Flow Process. . United States Department of the Interior and
United States Geological Survey.
Hecker, J.H., 2005. Promoting Environmental Security and Poverty Alleviation in the Peat
Swamps of Central Kalimantan, Indonesia. Prototype EnviroSecurity Assessments; Part 1:
Case Study., Institute for Environmental Security.
Hirano, T., Jauhiainen, J., Inoue, T. and Takahashi, H., 2009. Controls on the Carbon
Balance of Tropical Peatlands. Ecosystems, 12(6): 873-887.
Islam, M.S. and Hashim, R., 2009. Bearing Capacity of Stabilised Tropical Peat by Deep
Mixing Method. Australian Journal of Basic and Applied Sciences, 3(2): 682-688.
Jaenicke, J., Rieley, J.O., Mott, C., Kimman, P. and Siegert, F., 2008. Determination of the
amount of carbon stored in Indonesian peatlands. Geoderma, 147(3-4): 151-158.
Jaenicke, J., Siegert, F. and Wösten, H., 2009. Planning hydrological restoration measures by
canal blocking in the Sebangau water catchment, Central Kalimantan. Remote Sensing
Solutions GmbH.
Jauhiainen, J., Takahashi, H., Heikkinen, J.E.P., Martikainen, P.J. and Vasander, H., 2005.
Carbon fluxes from a tropical peat swamp forest floor. Global Change Biology, 11: 1788-1797.
Joosten, H., 2004. The IMCG Global Peatland Database.
Kool, D.M., Buurman, P. and Hoekman, D.H., 2006. Oxidation and compaction of a collapsed
peat dome in Central Kalimantan. Geoderma, 137(1-2): 217-225.
Kruse, J., Lennartz, B. and Leinweber, P., 2009. A modified method for measuring saturated
hydraulic conductivity and anisotropy of fen peat samples. Wetlands, 28(2): 527-531.
45
Lähteenoja, O., Ruokolainen, K., Schulman, L. and Alvarez, J., 2009. Amazonian floodplains
harbour minerotrophic and ombrotrophic peatlands. CATENA, 79(2): 140-145.
Langner, A. and Siegert, F., 2009. Spatiotemporal fire occurrence in Borneo over a period of
10 years. Global Change Biology, 15(1): 48-62.
Li, W., Dickinson, R.E., Fu, R., Niu, G.-Y., Yang, Z.-L. and Canadell, J.G. 2007. Future
precipitation changes and their implications for tropical peatlands. Geophys. Res. Lett., 34.
Ludang, Y., Jaya, A. and Inoue, T., 2007. Geohydrological Conditions of the Developed
Peatland in Central Kalimantan. World Applied Sciences Journal, 2(3): 198-203.
Moore, T.A., Shearer, J.C. and Miller, S.L., 1996. Fungal origin of oxidised plant material in
the Palangkaraya peat deposit, Kalimantan Tengah, Indonesia: Implications for `inertinite'
formation in coal. International Journal of Coal. Geology, 30(1-2): 1-23.
Morley, R.J., 1981. Development and Vegetation Dynamics of a Lowland Ombrogenous Peat
Swamp in Kalimantan Tengah, Indonesia. Journal of Biogeography, 8(5): 383-404.
Morrogh-Bernard, H., Husson, S., Page, S.E. and Rieley, J.O., 2003. Population status of the
Bornean orang-utan (Pongo pygmaeus) in the Sebangau peat swamp forest, Central
Kalimantan, Indonesia. Biological Conservation, 110(1): 141-152.
Neue, H.U., Gaunt, J.L., Wang, Z.P., Becker-Heidmann, P. and Quijano, C., 1997. Carbon in
tropical wetlands. Geoderma, 79(1-4): 163-185.
Page, S., Hosciło, A., Wösten, H., Jauhiainen, J., Silvius, M., Rieley, J., Ritzema, H., Tansey,
K., Graham, L., Vasander, H. and Limin, S. 2009. Restoration Ecology of Lowland Tropical
Peatlands in Southeast Asia: Current Knowledge and Future Research Directions.
Ecosystems, 12(6): 888-905.
Page, S.E., Rieley, J.O., Shotyk, O.W. and Weiss, D., 1999. Interdependence of Peat and
Vegetation in a Tropical Peat Swamp Forest. Philosophical Transactions: Biological Sciences,
354(1391): 1885-1897.
Peel, M.C., Finlayson, B.L. and McMahon, T.A., 2007. Updated world map of the Köppen-
Geiger climate classification. Hydrology and Earth System Science, 11: 1633-1644.
Quinty, F. and Rochefort, L., 2003. Peatland restoration guide. Canadian Sphagnum Peat
Moss Association et New Brunswick Department of Natural Resources and Energy., Québec,
Québec, 106 pp.
RePPProT, 1987. Review of phase I, East and South Kalimantan, Regional Planning
Programme for Transmigration (RePPProT). Volume 1. Land Resources Department/Bina
Program, Jakarta.
Schuman, M. and Joosten, H., 2008. Global Peatland Restoration Manual, Institute of Botany
and Landscape Ecology, Greifswald University, Germany.
Shepard, D., 1968. A two-dimensional interpolation function for irregularly-spaced data,
Proceedings of the 1968 23rd ACM national conference. ACM.
Sjörs, H., 1950. On the Relation between Vegetation and Electrolytes in North Swedish Mire
Waters. Oikos, 2(2): 241-258.
Walsum, P.v., Velhiuzen, A., van Bakel, P., van der Bolt, F., Dik, P., Groenendijk, P.,
Ouerner, E. and Smit, M. 2006. SIMGRO 6.0.3 Theory and model implementation (User
Manual). WUR-Alterra, Wageningen, The Netherlands.
46
Weiss, D. Shotyk, W., Rieley, J., Page, S.E., Gloor, M., Reese, S. and Martinez-Cortizas, A.
2002. The geochemistry of major and selected trace elements in a forested peat bog,
Kalimantan, SE Asia, and its implications for past atmospheric dust deposition. Geochimica et
Cosmochimica Acta, 66(13): 2307-2323.
Wetlands-International, Central Kalimantan Peatlands Project (CKPP).
http://ckpp.wetlands.org.
Wösten, J.H.M., Clymans, E., Page, S.E., Rieley, J.O. and Limin, S.H., 2008. Peat-water
interrelationships in a tropical peatland ecosystem in Southeast Asia. CATENA, 73(2): 212-
224.
Wüst, R.A.J., Bustin, R.M. and Lavkulich, L.M., 2003. New classification systems for tropical
organic-rich deposits based on studies of the Tasek Bera Basin, Malaysia. CATENA, 53(2):
133-163.
47
Appendix 1 – Ground Truth Data
Examples of two ground truth missions are shown in the data below, including actual canal
locations and relative elevations as measured with a waterpass.
Figure A1-1 - Ground Truth measurements for Bangah canal 'XY'
S.Bangah
48
Figure A1-2 - Ground Truth data for Bangah canal 10.
S.Bangah
49
Appendix 2 – SSI Dam Description
The following diagrams summarize the structural features of the SSI dam discussed in this
thesis report. The figures were adapted and translated from a report from the Central
Kalimantan Peatlands project (Usup et al., 2007).
Aerial View
Width-wise
cross-section
Length-wise
cross-section
Flow
Flow
50
Reference:
Usup, A., Hendri, S.T. and Bambang, A.Md. 2007. Laporan Kegiatan – Studi Konstruksi Tabat
Pada Kawasan Taman Nasional Sebangau. Departemen Pendidikan Nasional Universitas
Palangka Raya. Central Kalimantan Peatlands Project.
Turbulent flow occurs due to swift
currents in the water, causing a
deepening of the canal.
Energy and weight of water are
distributed equally in all directions
to allow for infiltration of water at
the bottom of the structure.
Increase in flow energy due to
narrowing of the flow base.
Flow
Directs the flow of water into
the spillway.
51
Appendix 3 – Tubewell Hydrographs
All available hydrographs from all tubewells installed to date by WWF-Indonesia are included
in this appendix. Hydrographs were prepared for monitoring transects in the SSI canal, canal
21, and canal 23. The naming convention for the SSI tubewells follows the grid shown in figure
A3-1. All transects within SSI canal are preceded by an „S‟. Transects are labelled
alphabetically starting from the downstream-most transect. Transects A to E are located
immediately downstream as shown in figure A3-1. Transects F to H are located similarly
starting 3.5km upstream of the dam, but are not shown here. Transects are labelled „left‟ (L)
and „right‟ (R) relative to the direction of canal flow. Finally, tubewells within each transect are
numbered in order of their proximity to the canal (the closest tubewell being assigned the
number 1). The first transect adjacent to the dam moving in a northern (left) direction is thus
named SAL1, the second SAL2, and so on according to the grid labels in figure A3-1.
Figure A3-1 - SSI Tubewell grid naming system
Figure A3-2 - Location of dam (arrow) at canal 23
S.Bangah
S.Sebangau
52
53
54
55
56
Monitoring data for the first dam in canal 23 (constructed in August 2009) are shown below. The
location of the dam at the canal outlet is indicated in figure A3-2. Four transects were installed near
the dam as follows: two left and right transects 50m upstream of the dam, and two left and right
transects downstream of the dam, each consisting of five tubewells each.
57
The hydrographs below contain all available data from canal 21. The letters correspond to
transect locations as outlined in figure 24, section 3.8.
-140
-120
-100
-80
-60
-40
-20
0
20
-400 -300 -200 -100 0 100 200 300 400
Distance from Canal (m)
Depth SS (cm)
jun-09
jul-09
aug-09
sep-09
oct-09
nov-09
-140
-120
-100
-80
-60
-40
-20
0
20
-400 -300 -200 -100 0 100 200 300 400
Distance from Canal (m)
Depth SS (cm)
jun-09
jul-09
aug-09
sep-09
oct-09
nov-09
-140
-120
-100
-80
-60
-40
-20
0
20
40
-400 -300 -200 -100 0 100 200 300 400
Distance from Canal (m)
Depth SS (cm)
jun-09
jul-09
aug-09
sep-09
oct-09
nov-09
A
B
C
58
Appendix 4 – Diver Datalogger Data
The raw data retrieved from the diver datalogger are shown below. All divers were placed in
the vicinity of canal 21 (locations shown in figure 15, section 3.5). Divers 18330 and 17927
were installed in a tubewell (figure A4-1 top; also shown in figure 9, section 2.3) such that
they were constantly submerged below the water table. The barometric (baro) diver was
suspended from a small tree (figure A4-1 bottom) between the two diver-tubewells, where it
measured atmospheric pressure on an hourly basis.
Figure A4-1 - Installation of the diver dataloggers (top) and the barometric diver (bottom left and right)
near canal 21.
59
The divers were installed at canal 21 on 31 July 2009, and were taken out on 23 October
2009 for a total of nearly three months of continuous data. The raw measurements (in cm
water) are shown below for all three divers.
