Applied Geography 31 (2011) 1166e1172
Contents lists available at ScienceDirect
Applied Geography
journal homepage: www.elsevier.com/locate/apgeog
Historical spatiotemporal analysis of land-use/land-cover changes and carbon
budget in a temperate peatland (Turkey) using remotely sensed data
Fatih Evrendilek a, *, Suha Berberoglu b, Nusret Karakaya a, Ahmet Cilek b, Guler Aslan a, Kerem Gungor a
a
b
Department of Environmental Engineering, Abant Izzet Baysal University, Golkoy Campus, Bolu, 14280, Turkey
Department of Landscape Architecture, Cukurova University, Balcalı-Adana, Turkey
a b s t r a c t
Keywords:
Carbon cycle
Change detection
Environmental monitoring
Peatlands
Risk assessment
Rate and amount of carbon (C) storage in peatlands play a vital role in global biogeochemical cycles despite
their globally small spatial extent. Carbon dioxide (CO2) emissions from drained peatlands is a function of
drainage intensity and extent of peatlands, peat thickness, land use and land cover (LULC) type to which
peatlands are converted, and management practices. Historical LULC change-detection of Yenicaga peatlands and wet flats yielded an estimated conservative decrease in C pool from 2.43 0.15 million metric
tons (Mt) in 1944 to 0.65 0.04 Mt in 2009 due mostly to drainage by agricultural and afforestation
activities. Drainage-induced C emission rate was estimated to range from 12.5 to 32.5 t C ha 1 yr 1. Our
extrapolation of the C emission rate to the entire peatland area of 240 km2 in Turkey resulted in CO2
emissions of 0.30e0.78 Mt in 2009, equivalent to 0.01%e0.02% of the global LULC-related CO2 emissions of
3230 Mt in 2009.
Ó 2011 Elsevier Ltd. All rights reserved.
Introduction
Local land use/land cover (LULC) changes alter global biogeochemical cycles and include quantitative and qualitative changes
such as the rate, amount, and direction of land conversions, land
fragmentation, land degradation, and loss of habitats and biodiversity (Baldocchi, 2008; Wali et al., 1999). Cumulative changes in
LULC are one of the most important human-induced processes that
alter global environment and climate (Dickinson, 1995; Hall, Tian,
Qi, Pontius, & Cornell, 1995). Peatlands are one of the most significant hotspots subjected to intensive anthropogenic alterations
across the world and in Turkey such as peat mining, drainage and
clearance for agriculture, overgrazing, and afforestation. Though
spatially small (3% of the global land surface), peatlands play
a globally vital role in global climate change as a sink of carbon
dioxide (CO2) from the atmosphere and a source of methane (CH4)
to the atmosphere (Limpens et al., 2008; Tomlinson, 2010; Wilson
et al., 2009).
A variety of change detection techniques have been developed
and implemented to quantify the extent to which LULC changes
affect ecosystem structure and function at multiple spatiotemporal
scales (e.g., local-to-global and intra-annual-to-interannual scales)
(Mouat, Mahin, & Lancaster, 1993; Coppin & Bauer, 1996; Serpico &
* Corresponding author. Tel.: þ90 374 254 1000; fax: þ90 374 253 4558.
E-mail address: fevrendilek@ibu.edu.tr (F. Evrendilek).
0143-6228/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.apgeog.2011.03.007
Bruzzone, 1999; Singh, 1989). Airborne remote sensing has the
ability to provide high spectral and spatial resolution imagery as
well as to monitor near-surface soil water content across peatland
land cover, while spaceborne remote sensing provides more
temporally and spatially continuous and rapid coverage at relatively low cost (Yu, Zang, Wu, Liu, & Na, 2011). Integrated use of
aerial photographs and high spatial resolution satellite sensors
(e.g., GeoEye, IKONOS and Quickbird) can bridge the gap between
what is observed in situ and what is recorded by remote sensing,
thus overcoming the limitations of medium-resolution sensors
(e.g., Landsat, SPOT, and SAR systems) due to the heterogeneous
nature of peatland vegetation communities.
In Turkey, peatlands are rare ecosystems with their current
estimated area of 240 km2 (Dengiz, Ozaytekin, Cayci, & Baran, 2009)
and the only habitats where very uncommon plant species, and
important migratory waterfowl and shorebirds exist. Despite the
globally and locally important role and ability of peatlands in
biodiversity, and CO2 and CH4 exchanges along the soil-vegetationatmosphere continuum, to the authors’ knowledge, there have been
no studies quantifying LULC changes in a peatland ecosystem and
their associated carbon (C) fluxes in Turkey based on remotely
sensed data. The objective of this study was to (1) quantify LULC
changes in the Yenicaga peatland of northwestern Turkey between
1944 and 2009 using GeoEye-1 and aerial photograph and (2)
explore implications of environmental monitoring and risk assessment for LULC change- and drainage-induced C fluxes.
F. Evrendilek et al. / Applied Geography 31 (2011) 1166e1172
Materials and methods
Study area
The Yenicaga peatland is located about 38 km east of the city of
Bolu (40 47N0 , 321’E) in the northwestern Black Sea region of
Turkey (Fig. 1). The climate in the Yenicaga region is classified as
a cool temperate zone, with mean annual temperature and
precipitation values of 10.2 C and 538 mm, respectively, and with
a prevailing wind direction from north (Dengiz et al., 2009;
Evrendilek, Karakaya, Aslan, & Ertekin, 2011). About 60% of the
mean annual precipitation falls during the winter and spring
(TSMS, 2005). The Yenicaga basin as a part of the Alpine-Himalayan
belt is a strike-slip fault wedge basin covered by Quaternary
accumulated and water-saturated sediments upon which Yenicaga
peatlands developed (Arca, 2004). Main dominant natural vegetation of Yenicaga peatlands and wet flats is in the Euro-Siberian
origin phytogeographically and includes Phragmites australis and
1167
Carex pseudocyperus in reedbed; Alopecurus arundinaceus, Juncus
compressus, Bidens tripartite, Chenopodium album, and Equisetum
palustre on overgrazed lands; and Cirsium canum, Orchis laxiflora,
Euphorbia palustris, Senecio paludosus, and Teucrium scordium as fen
vegetation on peatlands and wet flats drained for agriculture and
grasslands (Sumer, 2002).
Processing of remotely sensed data
The remotely sensed data used for this study consist of aerial
photography at the scale of 1. 35,000 in 1944 and GeoEye-1 image
acquired on 1 October 2009. The GeoEye-1 image consists of four
spectral bands: three bands in the visible part of the spectrum
located at blue (450e520 nm), green (520e600 nm) and red
(625e695 nm), and one band in near infrared (760e900 nm), with
a spatial resolution of 0.41 m and 1.65 m for panchromatic and
multispectral images, respectively. In accordance with the use of
pan-sharpened multispectral GeoEye-1 imagery at a 1-m spatial
Fig. 1. Location map of Yenicaga peatlands.
1168
F. Evrendilek et al. / Applied Geography 31 (2011) 1166e1172
resolution, panchromatic aerial photograph was also resampled to
a 1-m resolution. In combination with the remotely sensed data
and field observations, ancillary data including topographic and
agricultural maps at the scale of 1:25000 were used to improve
classification and mapping of LULC. However, high spatial resolution remotely sensed data have high within-field spatial variation
which in turn may lead to an inaccurate per-pixel classification
(Berberoglu, Lloyd, Atkinson, & Curran, 2000). In this study, the
object-oriented classification approach involved the performance
of image segmentations and classifications based on multiresolution segmentation algorithm and standard nearest neighbor
classifier, implemented in Definiens Developer. Multiresolution
segmentation is a region growing method for repetitive segmentations using randomly selected seed pixels across a scene to create
image objects (Definiens, 2004). The segmentation was operated as
a heuristic optimization procedure which minimized the average
heterogeneity of image objects at a given spatial resolution for the
whole scene (Bian & Walsh, 1992).
The entire process can be described sequentially as follows: (1)
The image was segmented into pattern cells (one or more pixels); (2)
each segment was compared with its neighbors to determine
similarity, and values of spectral (or color) and shape heterogeneities
were computed and utilized to determine whether to merge the
segment pairs. The spectral heterogeneity was measured by
standard deviation of spectral values in each layer for the image
segments, while the shape heterogeneity referred to the compactness and smoothness of the boundary of the merged pair of adjacent
image segments; (3) the segment continued to grow as compared
with all the neighbors until there is no remaining joinable region at
which point the segment was labeled as a completed region; (4) this
process was repeated for next uncompleted cells until all the cells
were labeled. A pair of image segments was merged when the
spectral and shape heterogeneities did not exceed a user-defined
scale parameter. The scale parameter plays a major role in terminating, and thus, controlling the size of image objects (Definiens,
2004); (5) a supervised classification method with the nearest
neighbor algorithm was carried out using the segments instead of
the traditional pixels. Three wavebands for GeoEye-1, and one
waveband for aerial photograph were utilized during the image
segmentation; and (6) finally, classification results were crosstabulated in order to derive the ‘frometo’ change detection needed
to estimate ‘class-by-class’ changes.
Estimation of SOC pools and CO2 emissions
A total of 26 disturbed and undisturbed soil samples were
randomly taken from peatlands for a depth of 0e30 cm in order to
estimate bulk density (BD, g cm 3 or t m 3), soil organic carbon
Fig. 2. Object-oriented supervised classification of air photo for Yenicaga peatlands in 1944.
F. Evrendilek et al. / Applied Geography 31 (2011) 1166e1172
(SOC, %), soil nitrogen (SN, %), and soil C/N ratios. Soil samples were
dried at 65 C for 48 h and weighed (1.5e2.0 mg) into tin capsules
for analysis of C and N contents with an elemental CHN analyzer
(Euro EA3000, EuroVector, Milano, Italy). Soil organic carbon stock
was estimated for peatlands and wet flats using the following
equation (Evrendilek & Wali, 2001):
SOC stock t C ha
1
¼ peatland thickness ðmÞ * BD t m
* SOC ðC%Þ * 10; 000 m2 ha 1
3
The amount of change in SOC storage was estimated as
a product of changes in peatland and wet flat area (ha) and SOC
stock determined in this study. Mean annual rate of drainageinduced CO2 emissions in Yenicaga was quantified using a linear
relationship between CO2 emissions and groundwater depth found
across Southeast Asia peatlands by Wosten and Ritzema (2001),
and Hooijer et al. (2010).
Results and discussion
Quantification of LULC changes
The images were geometrically corrected and geocoded to the
Universal Transverse Mercator (UTM) coordinate system using
1169
a reference image and regularly distributed ground control points
(GCPs) selected from the remotely sensed image and air photo. The
images were enhanced using linear contrast stretching and histogram equalization to improve the image quality to help identify
GCPs in rectification. Second order polynomial function with
a cubic convolution algorithm was used for the air photo. The
transformation had root mean square errors (RMSE) ranging from
0.3 to 0.5 which indicated that the images were accurate being well
within one pixel.
Mapping LULC was carried out using object-oriented supervised classification of the air photo and the GeoEye-1 image,
via Definiens software (Definiens Imaging GmbH, München,
Germany). Utilizing ground truth information from field survey
records, topographic maps, and detailed field reports coincided
with the image acquisition date, LULC classification map of the
study area was corrected manually. In addition, expert knowledge
was also utilized for the 1944 air photo classification. Objectoriented supervised classification discriminated among 11 classes:
(1) peatlands, (2) settlements, (3 and 4) agricultural areas with and
without crops, (5) shadows of trees and buildings, (6) water
bodies, (7) bareground, (8) wet flats, (9) ditches, (10) woodland,
and (11) roads. The detailed LULC classes were aggregated into
eight categories: (1) peatlands, (2) agriculture, (3) roads, (4)
woodland, (5) water bodies, (6) settlements, (7) bareground, and
(8) wet flats (Figs. 2 and 3).
Fig. 3. Object-oriented supervised classification of GeoEye-1 image for Yenicaga peatlands in 2009.
1170
F. Evrendilek et al. / Applied Geography 31 (2011) 1166e1172
Table 1
Cross-tabulation of land-use/land-cover (LULC) changes (ha) in Yenicaga peatlands between 1944 and 2009.
LULC
in 2009 (ha)
Peatland
Agriculture
Road
Woodland
Water body
Settlement
Bareground
Total
Amount of change (ha)
Percent change (%)
Rate of change (ha yr-1)
Peatland
221.4
202.5
1.7
58.3
20.8
0.1
1.3
506.1
185.8
37
2.9
Agriculture
Road
in 1944 (ha)
Woodland
Water body
0.7
704.5
47.7
89.0
0
11.0
267.7
1121.0
552.8
49
8.5
1.6
23.3
5.1
4.4
0
1.4
5.1
40.9
74.1
181
1.1
0.1
20.7
1.0
58.1
0
0.6
25.8
106.4
243.9
229
3.8
Water bodies and ditches were difficult to separate spectrally
and grouped into the water bodies class in order to prevent the
issue of seasonality effect during the change detection process.
Croplands planted to crops and fallow areas had different spectral
characteristics, but were merged into a single class of agriculture.
The settlement class also included bareground associated with
villages, while the woodland class included solitaire and clustered
trees and shrubs. Conversion to agriculture, or agricultural encroachment towards peatlands and wet flats as clearly evident in its
spatiotemporal distribution (Figs. 2 and 3) appeared to be the most
remarkable driving force behind LULC changes in the Yenicaga
peatland over the period of 66 years.
55.4
1.4
1.4
17.5
220.2
0.1
0
296.0
52.2
18
0.8
Settlement
0.1
13.0
7.4
12.1
0
5.8
0.5
38.9
13.6
35
0.2
Bareground
0.3
211.9
26.2
50.7
0.5
1.9
160.0
451.4
63.8
14
1.0
Wet flats
40.7
496.1
24.5
60.1
2.3
4.2
54.7
682.9
682.9
100
10.5
Total
320.3
1673.8
115.0
350.3
243.8
25.3
515.2
4191.1
According to our cross-tabulation results for the period of
1944e2009, in decreasing order of areal extent (ha), Yenicaga peatland ecosystem underwent increases in agriculture, woodland, roads,
and bareground, but reductions in wet flats, peatlands, water body,
and settlement (Table 1) (Fig. 4). The percent increase detected for the
study period ranged from 14% in bareground to 229 in woodland,
while the percent decrease ranged from 18% in water body to
a complete loss of wet flats. Net annual rates of increase ranged from
1.0 ha yr 1 for bareground to 8.5 ha yr 1 for agriculture, while net
annual rates of decrease ranged from 0.2 ha yr 1 for settlement to
10.5 ha yr 1 for wet flats. The 13.6-ha decrease in the settlement is
most likely to result from population dynamics for the study period
Fig. 4. Detection of land-use/land-cover (LULC) changes in Yenicaga peatlands between 1944 and 2009.
1171
F. Evrendilek et al. / Applied Geography 31 (2011) 1166e1172
during which the census data of 1935e2009 for Yenicaga town
showed that population increased from 701 in 1935 to 6364 in 2000
and decreased to 5175 in 2009 (Dogan & Kizilkaya, 2010; TSI, 2011).
Also, urbanization and subsequent concentration of the population in
concrete apartment buildings in Yenicaga town may have contributed to the decrease in the settlement area.
Yenicaga peatlands decreased by 37% at a rate of 2.9 ha yr 1 with
a total degradation of 185.8 ha (Table 1). When taken into account
together, the degradation of the peatlands and wet flats occurred at
an annual rate of 13.4 ha yr 1. Spatial distribution of LULC changes
(Fig. 4) reveals that peatlands and wet flats were degraded in the
total amount of 828 ha mostly to agriculture and woodland after
taking into account the conversion of the 40.7-ha wet flats to
peatlands (Table 1). Agricultural and afforestation activities appear
to be the main drivers for the historical degradation of 37% of the
original peatland area between 1944 and 2009. The root causes
behind the adoption of intensive agriculture and afforestation on
Yenicaga peatlands and wet flats can be attributed to low environmental awareness and short-term profit maximization by
cropland production, overgrazing, peat extraction, and tourism and
recreational activities.
Another striking change revealed by the LULC change detection
was an 18% decline in the total area of water bodies at a rate of
0.8 ha yr 1. The surface area of Lake Yenicaga declined from
280.4 ha in 1944 to 219.5 ha in 2009 by 22% at an annual rate of
0.94 ha yr 1. The decline in the surface area of the lake may be
attributed to two main drivers: (1) the attempt made in 1955 and
ceased in 1990 to drain the lake as a precaution to the 1955 malaria
outbreak in Bolu and environs (Dogan & Kizilkaya, 2010; SaygiBasbug & Demirkalp, 2004), and (2) increased temperature and
evapotranspiration due to climate change. Mean annual air
temperature in Bolu was 9.73 C between 1941 and 1945 and
11.03 C between 2001 and 2005, with a long-term mean of
10.28 0.26 C for the period of 1931e2005 (TSMS, 2005). Fitting
a linear regression line to air temperature time-series data
smoothed by a 5-year moving average revealed a temperature
increase for Bolu at annual rates of 0.008 C yr 1 between 1931 and
2005 and 0.012 C yr 1 between 1941 and 2005.
Quantification of SOC pools and CO2 emissions
In situ measurements in Yenicaga peatlands showed that peat
thickness ranged from 0.5 m to 12 m, with mean minimum and
maximum values of 1.7 0.9 m and 4.6 1.7 m, respectively
(n ¼ 37) (GTZ, 2010). A mean peat thickness reported for various
countries such as Malaysia, Brunei, Papua New Guinea, and Europe
varied between 1.5 m and 3 m in the related literature (Bradley
et al., 2005; Byrne et al., 2004; Gorham, 1991; Hooijer et al., 2010;
Hoper, 2002). Our in situ soil samples to a depth of 0.3 m indicated mean SOC, SN, BD and C/N values of 24 7%, 2 0.3%,
0.5 0.2 t m 3, and 14 2, respectively, thus leading to a mean SOC
density of 120 kg m-3 for Yenicaga peatlands (n ¼ 18). Peatland C
densities were reported to range from 24 kg Cm 3e95 kg Cm 3
(Shimada, Takahashi, Haraguchi, & Kaneko, 2001; Page et al., 2002;
Wetlands International, 2003, 2004; Wosten & Ritzema, 2001).
Also, BD values were reported to range from 0.09 t m 3 to 0.2 t m 3
for peatlands across the world (Byrne et al., 2004; Bradley et al.,
2005; Gorham, 1991; Hoper, 2002) and ranged from 0.09 t m 3 to
0.78 t m 3 for Yenicaga peatlands and wet flats (Dengiz et al., 2009).
Our estimate of unit area SOC stock in the Yenicaga peatlands and
wet flats ranged from 2040 126 t C ha 1 to 5520 238 t C ha 1
depending on the mean minimum and maximum values of peat
thickness (worst- and best-case scenarios), respectively. When the
SOC stock was extrapolated to Yenicaga peatlands and wet flats
given the LULC changes between 1944 and 2009, a total historical
degradation of SOC pool was estimated to vary between 0.15 0.08
million metric tons (Mt) C (at an annual rate of 3 t C ha 1 yr 1) in the
best-case scenario and 10.43 3.85 Mt C (at an annual rate of
194 t C ha 1 yr 1) in the worst-case scenario (Table 2). The range of
estimated C storage in Yenicaga peatlands and wet flats was between
2.43 0.15 and 6.56 0.28 Mt C in 1944 and between 0.65 0.04
and 1.77 0.08 Mt C in 2009.
Drainage of Yenicaga peatlands and wet flats for agriculture and
afforestation dominated the human-induced disturbances of the
peatland C budget in the region. Present and future C emissions
from drained peatlands and wet flats were quantified using data
concerning LULC changes, peat thickness, and groundwater depthinduced C emissions (Table 2). The range of groundwater depth of
0.5 me1.3 m was assumed to represent the most common water
table depths below the peatlands and wet flats of the study region
(Dengiz et al., 2009). Wosten and Ritzema (2001), and Hooijer et al.
(2010) found a linear relationship between C emissions and
groundwater depth at a mean annual C emission rate of
2.5 t C ha-1 yr-1 (9.1 t CO2 ha 1 yr 1) with every 10 cm water table
drawdown across Southeast Asia peatlands (R2 ¼ 0.71; n ¼ 8). Based
on the range values of water table in this study, net annual rate of C
emissions was estimated at a minimum of 12.5 t C ha 1 yr 1 and
a maximum of 32.5 t C ha-1 yr 1.
Byrne et al. (2004) reported median values of C emissions from
ombrotrophic and minerotrophic peatlands drained across Europe
as 2.3 and 4.1 t C ha 1 yr 1 for grasslands and 4.4 and 4.1 t C ha 1 yr 1
Table 2
Site specific and literature-related values used to estimate C and N fluxes for Yenicaga peatlands.
Yenicaga peat thicknessb
Bulk density in Yenicaga peatlands
and wet flats
Soil organic carbon in Yenicaga
peatlands and wet flatsb
Soil nitrogen in Yenicaga peatlands
and wet flatsb
C/N ratio in Yenicaga peatlands
and wet flatsb
Groundwater depth below Yenicaga
peatlands and wet flats
Emission factor
Peatland SOC contentd
SOC content in Yenicaga peatlands
and wet flats
Total SOC degradation of
Yenicaga peatlandsb
Total SOC degradation of
Yenicaga wet flatsb
Net annual degradation rate
of Yenicaga peatlands
and wet flatsb
Water table change-induced net
annual C emission rate from
drainage of Yenicaga
peatlands and wet flatsc
Water table change-induced
total CO2 emissions to atmosphere
from Yenicaga peatlands
and wet flatsc
Water table change-induced total CO2
emissions to atmosphere from
Turkish peatlandsa,c
Unit
Min
m
tm
1.7
0.09a
3
Max
4.6
0.78a
3.0
0.5b
%
12.1
%
1.0
2.2
2.0
8.6
21.1
14.0
m
0.5a
1.3a
0.9b
t C ha-1 yr-1
kg C m-3
kg C m-3
1.6c
24
11a
27.2c
95
274a
120b
Mt C
0.03
2.34
0.67
0.12
8.09
2.31
3
194
55
12.5
32.5
22.5
Mt C
t C ha
-1
yr
-1
t C ha-1 yr-1
35.1
Mean
24.0
Mt CO2 yr
1
0.010
0.026
0.018
Mt CO2 yr
1
0.30
0.78
0.54
Multiply numbers above by 3.67 for conversion of C to CO2 emissions. C: carbon; N:
nitrogen; SOC: soil organic carbon.
a
according to Dengiz et al. (2009).
b
according to this study.
c
according to Hooijer et al. (2010).
d
according to Shimada et al. (2001); Wosten and Ritzema (2001); Page et al.
(2002); Wetlands International (2003, 2004).
1172
F. Evrendilek et al. / Applied Geography 31 (2011) 1166e1172
for croplands, respectively (multiply by 3.67 for conversion of C to
CO2). The range of unit area C emissions calculated for Southeast Asia
peatlands was between 1.6 and 27.2 t C ha 1 yr 1 (Hooijer et al.,
2010). IPCC (2006) reported mean annual emissions from cultivated organic soils as 5 t C ha 1 yr 1 for boreal/cool temperate,
10 t C ha 1 yr 1 for warm temperate, and 20 t C ha 1 yr 1 for tropical/
sub-tropical climate regimes.
Total CO2 emissions to the atmosphere from Yenicaga peatlands
and wet flats in 2009 were in the range of 0.010e0.026 Mt CO2 yr 1
(Table 2). If the historical rate of peatland and wet flat degradation
continues at the estimated rate of 13.4 ha yr 1, the remaining
320.3-ha peatland is expected to be completely degraded by 2033.
A steady decline in the degradation rates is most likely to occur by
virtue of increases in environmental awareness, and preventive and
mitigative measures being taken as the remaining peatlands
become increasingly scarce. The extrapolation of net annual C
emission rates to the entire peatland area of 240 km2 across Turkey
leads to a national estimation range of 0.3e0.78 Mt CO2 yr 1. The
national estimate of CO2 emissions from peatlands accounts for
0.01e0.02% of the global CO2 emissions of 3230 Mt CO2 in 2009
caused by LULC changes (Friedlingstein et al., 2010).
Conclusions
Environmental monitoring of LULC changes may serve as an
early warning signal and a risk assessment tool for the loss and
depreciation of natural capital. In this way, preventive and mitigative feedbacks can be tailored in a timely fashion to secure
ecosystem and human well-being and minimize adverse environmental impacts. Integration of field observations, and high-resolution airborne and spaceborne remote sensing is a promising
approach for spatiotemporal analysis of LULC change-induced C
fluxes. Proximal sensing and sampling of spectral information in
combination with flux towers, and remote sensing is in need of
future research in order to better differentiate among contributions
of peatland plant species or plant functional types to biogeochemical cycles, a capability that remote sensing techniques may
lack, thus assisting in a spatiotemporal scaling of ecosystem
processes and rehabilitation of damaged peatland ecosystems.
Acknowledgments
The research project (COST-CAYDAG-109Y186) was funded by
the Scientific and Technological Research Council of Turkey. We are
grateful to Mr. Aydın Gülmez, Yenicaga Municipality, and German
Society for Technical Cooperation (GTZ) in collaboration with the
Turkish Ministry of Environment and Forestry for their logistic
support with field measurements. We would also like to thank
anonymous reviewers for their constructive comments on an
earlier version of the manuscript.
References
a basin, Bolu-Turkey.
Arca, M. S. (2004). Neotectonics and evolution of the Yeniçag
MSc Thesis in the Department of Geological Engineering submitted to the
Graduate School of Natural and Applied Sciences of the Middle East Technical
University, Ankara, Turkey.
Baldocchi, D. (2008). “Breathing” of the terrestrial biosphere: lessons learned from
a global network of carbon dioxide flux measurement systems. Australian
Journal of Botany, 56, 1e26.
Berberoglu, S., Lloyd, C. D., Atkinson, P. M., & Curran, P. J. (2000). The integration of
spectral and textural information using neural networks for land cover
mapping in the Mediterranean. Computers & Geosciences, 26, 385e396.
Bian, L., & Walsh, S. J. (1992). Scale dependencies of vegetation and topography in
a mountainous environment of Montana. Professional Geographer, 45, 1e11.
Bradley, R. I., Milne, R., Bell, J., Lilly, A., Jordan, C., & Higgins, A. (2005). A soil carbon
and land use database for the United Kingdom. Soil Use and Management, 21,
363e369.
Byrne, K. A., Chojnicki, B., Christensen, T. R., Drosler, M., Freibauer, A., Frolking, S.,
et al. (2004). EU Peatlands: Current carbon stocks and trace gas fluxes, Carboeurope-GHG, concerted action synthesis of the European greenhouse gas budget.
Viterbo, Italy: Department of Forest Science and Environment.
Coppin, P. R., & Bauer, M. E. (1996). Digital change detection in forest ecosystems
with remote sensing imagery. Remote Sensing Reviews, 13, 207e234.
Definiens. (2004). eCognition. München, Germany: Definiens Imaging.
Dengiz, O., Ozaytekin, H., Cayci, G., & Baran, A. (2009). Characteristics, genesis and
classification of a basin peat soil under negative human impact in Turkey.
Environmental Geology, 56(6), 1057e1063.
Dickinson, R. (1995). Land processes in climate models. Remote Sensing of Environment, 51, 27e38.
Dogan, M., & Kizilkaya, B. (2010). A preliminary assessment of the water and air
a coast. Journal of Black Sea/
quality parameters in and around the Lake Yeniçag
Mediterranean Environment, 16(2), 167e194.
Evrendilek, F., & Wali, M. K. (2001). Modelling long-term C dynamics in croplands in
the context of climate change: a case study from Ohio. Environmental Modelling
& Software, 16(4), 361e375.
Evrendilek, F., Karakaya, N., Aslan, R. G., & Ertekin, C. (2011). Using eddy covariance
sensors to quantify carbon metabolism of peatlands: a case study in Turkey.
Sensors, 11(1), 522e538.
Friedlingstein, P., Houghton, R. A., Marland, G., Hackler, J., Boden, T. A., Conway, T. J.,
et al. (2010). Update on CO2 emissions. Nature Geoscience, 3, 811e812.
Gorham, E. (1991). Northern peatlands: role in the carbon cycle and probable
responses to climatic warming. Ecological Applications, 1, 182e195.
GTZ (German Society for Technical Cooperation). (2010). Stratigraphic assessment of
a Lake, Bolu-Turkey. Wetlands and climate change Project,
the mire of Yeniçag
Ankara.
Hall, C. A. S., Tian, H., Qi, Y., Pontius, G., & Cornell, J. (1995). Modeling spatial and
temporal patterns of tropical land use change. Journal of Biogeography, 22,
753e757.
Hooijer, A., Page, S., Canadell, J. G., Silvius, M., Kwadijk, J., Wosten, H., et al. (2010).
Current and future CO2 emissions from drained peatlands in Southeast Asia.
Biogeosciences, 7, 1505e1514.
Hoper, H. (2002). Carbon and nitrogen mineralization rates in German agriculturally used fenlands. In G. Broll, W. Merbach, & E. M. Pfeiffer (Eds.), Wetlands in
Central Europe. Soil organisms, soil Ecological processes, and trace gas emissions
(pp. 149e164). Berlin: Springer.
IPCC. (2006). IPCC Guidelines for National Greenhouse Gas Inventories. The Intergovernmental Panel on Climate Change (IPCC).
Limpens, J., Berendse, F., Blodau, C., Canadell, J. G., Freeman, C., Holden, J., et al.
(2008). Peatlands and the carbon cycle: from local processes to global implications e a synthesis. Biogeosciences Discussions, 5, 1379e1419.
Mouat, D. A., Mahin, G. C., & Lancaster, J. S. (1993). Remote sensing techniques in the
analysis of change detection. Geocarto International, 2, 39e50.
Page, S. E., Siegert, F., Rieley, J. O., Boehm, H. V., Jaya, A., & Limin, S. (2002). The
amount of carbon released from peat and forest fires in Indonesia during 1997.
Nature, 420, 61e65.
a Lake
Saygi-Basbug, Y., & Demirkalp, F. Y. (2004). Trophic status of shallow Yeniçag
(Bolu, Turkey) in relation to physical and chemical environment. Fresenius
Environmental Bulletin, 13, 385e393.
Serpico, S. B., & Bruzzone, L. (1999). Change detection. In C. H. Chen (Ed.), Information processing for remote sensing (pp. 319e336). Singapore: World Scientific
Publishing.
Shimada, S., Takahashi, H., Haraguchi, A., & Kaneko, M. (2001). The carbon content
characteristics of tropical peats in Central Kalimantan, Indonesia: estimating
their spatial variability in density. Biogeochemistry, 53, 249e267.
Singh, A. (1989). Digital change detection techniques using remotely sensed data.
International Journal of Remote Sensing, 10, 989e1003.
a. M.Sc. Thesis, Bolu, Turkey: Graduate School of
Sumer, N. (2002). Flora of Yeniçag
Abant Izzet Baysal University, Department of Biology.
Tomlinson, R. W. (2010). Changes in the extent of peat extraction in Northern
Ireland 1990e2008 and associated changes in carbon loss. Applied Geography,
30, 294e301.
TSI (Turkish Statistics Institute). (2011). Demographic statistics. Available at: http://
www.turkstat.gov.tr/Start.do last accessed on February of 2011.
TSMS (Turkish state Meteorological Service). (2005). Turkish meteorological data
archive system. Available at: http://tumas.dmi.gov.tr/wps/portal last accessed on
December of 2006.
Wali, M. K., Evrendilek, F., West, T., Watts, S., Pant, D., Gibbs, H., et al. (1999).
Assessing terrestrial ecosystem productivity: usefulness of regional carbon and
nitrogen models. Nature & Resources, 35(4), 21e33.
Wetlands International. (2003). Maps of peatland distribution and carbon content in
Sumatera, 1990e2002.
Wetlands International. (2004). Maps of peatland distribution and carbon content in
Kalimantan, 2000e2002.
Wilson, D., Alm, J., Laine, J., Byrne, K. A., Farrell, E. P., & Tuittial, E. S. (2009).
Rewetting of cutaway peatlands: are we recreating hot spots of methane
emissions? Society for Ecological Restoration International, 17(6), 796e806.
Wosten, J. H. M., & Ritzema, H. P. (2001). Land and water management options for
peatland development in Sarawak, Malaysia. International Peat Journal, 11,
59e66.
Yu, W., Zang, S., Wu, C., Liu, W., & Na, X. (2011). Analyzing and modeling land use
land cover change (LUCC) in the Daqing City, China. Applied Geography, 31(2),
600e608.