ArticlePDF Available

Landsat based distribution mapping of high-altitude peatlands in Hindu Kush Himalayas — a case study of Broghil Valley, Pakistan


Abstract and Figures

In the alpine regions of Hindu Kush, Himalayas and Karakorum, climatic and topographic conditions can support the formation of peat, important for the livelihood of the local communities, and ecological services alike. These peatlands are a source of fuel for the local community, habitat for nesting birds, and water regulation at source for rivers. Ground-based surveys of high-altitude peatlands are not only difficult, but also expensive and time consuming. Therefore, a method using cost-effective remote sensing technology is required. In this article we assessed the distribution and extent of high-altitude peatlands in a 2000 ha area of Broghil Valley using Landsat 8 data. The composite image was trained using a priori knowledge of the area, and classified into peatland and non-peatland land covers using a supervised decision tree algorithm. The Landsat-based classification map was compared with field data collected with a differential GPS. This comparison suggests 82% overall accuracy, which is fairly high for high altitude areas. The method was successfully applied and has the potential to be replicated for other areas in Pakistan and the high-altitude regions of the neighbouring Asian countries.
Content may be subject to copyright.
J. Mt. Sci. (2020) 17(1): 42-49 e-mail:
Abstract: In the alpine regions of Hindu Kush,
Himalayas and Karakorum, climatic and topographic
conditions can support the formation of peat,
important for the livelihood of the local communities,
and ecological services alike. These peatlands are a
source of fuel for the local community, habitat for
nesting birds, and water regulation at source for rivers.
Ground-based surveys of high-altitude peatlands are
not only difficult, but also expensive and time
consuming. Therefore, a method using cost-effective
remote sensing technology is required. In this article
we assessed the distribution and extent of high-
altitude peatlands in a 2000 ha area of Broghil Valley
using Landsat 8 data. The composite image was
trained using a priori knowledge of the area, and
classified into peatland and non-peatland land covers
using a supervised decision tree algorithm. The
Landsat-based classification map was compared with
field data collected with a differential GPS. This
comparison suggests 82% overall accuracy, which is
fairly high for high altitude areas. The method was
successfully applied and has the potential to be
replicated for other areas in Pakistan and the high-
altitude regions of the neighbouring Asian countries.
Keywords: Peatland distribution; Chitral;
Qurumbar; Wakhi; Hindu Kush; Yarkhun
Peatlands are characterized by the
accumulation of organic matter (peat), which is
derived from dead and decaying plant material
under conditions of permanent water saturation
(Parish et al. 2008). Peatlands worldwide are
major carbon storage sites (Joosten et al. 2012;
Biancalani and Avagyan 2014). The formation of
peatlands in regions such as North America is
traced to the exposure of land after deglaciation
and is estimated at 13,000 Cyr BP (Calendar Years
before present) in the south and 7000 Cyr BP and
Landsat based distribution mapping of high-altitude
peatlands in Hindu Kush Himalayas – a case study of
Broghil Valley, Pakistan
Ahmad KHAN1*; e-mail:
Ahmad SAID2; e-mail:
Imran ULLAH3; e-mail:
*Corresponding author
1 Department of Geographical Sciences, University of Maryland, College Park, MD, 20770, USA
2 True Green Office, 20 Raffaele Road, Plymouth, MA 02360, USA
3 Departments of Paediatrics and Biochemistry, University of Texas, Southwestern Medical Center, Dallas, Texas 75235, USA
Citation: Khan A, Said A, Ullah I (2020) Landsat based distribution mapping of high-altitude peatlands in Hindu Kush
Himalayas – a case study of Broghil Valley, Pakistan. Journal of Mountain Science 17(1).
© Science Press, Institute of Mountain Hazards and En vironment, CAS a nd Springer-Verlag GmbH Ger many, part of Springer Nature 2020
1st Revision: 25-Apr-2019
2nd Revision: 21-Aug-2019
Accepted: 11-Oct-2019
J. Mt. Sci. (2020) 17(1): 42-49
4000 Cyr BP in the north (Prest 1970).
In Hindu Kush – Himalayas (HKH), naturally
occurring high-altitude pastures above the tree line
are characterized by wet conditions and low
temperatures that result in incomplete degradation
of organic material. The partially degraded organic
material accumulates over time (Khan and Said
2012; Khan et al. 2010; Lafleur et al. 2005; Laiho
2006) to create high-altitude peatlands (Ranhotra
and Kar 2011; Kumaran et al. 2016). These
peatlands might date back to the deglaciation of the
region after collision of the Indian Sub-continent
with Asia and the consequent Himalayan orogeny,
50 million years ago (Kumaran et al. 2016; Frisch
Meshed and Blakey 2011; UNDP 2002;
Christopherson 2000).
Various disciplines define peat differently
(Bond 1986). The American Geological Institute
defines peat as “an unconsolidated early stage of
coal deposit of semi-carbonized plant remains of a
water saturated environment such as a bog or fen
and of persistently at least 75% moisture” (Gary et.
al 1974). Peat has less than 25% ash content, which
distinguishes it from other organic soil materials. It
has higher rank phytogenic material, and has lower
British thermal unit (BTU) value on water
saturated basis (Bond 1986).
Bond (1986) has described soil consisting of
greater than 75% organic matter in dry state as the
fuel grade peatlands in Florida. In order for a peat
deposit to be classified as fuel grade, the deposit
must be at least four feet thick (or deep), and not
less than 80 contiguous acres per square mile and
yield not less than 8,000 BTU per pound (moisture
Carbonization of peat results in carbon
enrichment, which is essential for making it
desirable fuel. This is also a basis for peat
classification against wood (Hreibljan et al. 2015).
According to one study, one pound of wood
contains 20% fixed carbon and provides 9,300
BTU energy when is moisture/mineral free, while
one pound of peat contains 28% fixed carbon and
provides 10,600 BTU energy. Lignite contains 47%
fixed carbon and provides 12,400 BTU energy (U.S.
Department of Energy 1979). Thus, peat is a better
fuel than wood and approaches lignite.
The fuel usefulness of peat can be determined
from its fixed carbon, ash content, moisture, and
volatile contents. Peat has higher volatile content
than coal, a positive attribute that makes peat a
suitable fuel. The fixed carbon content in peat is
responsible for its combustion energy (U.S.
Department of Energy 1979; Clark 2008). Volatile
and fixed carbon contents are important attributes
that can determine utility and suitability of a peat,
and can categorize it for use. Based on the
definitions applied, “peat in Broghil Valley could be
defined as the organic soils created by impartial
decomposition of plant material under humid cold
environment, where thickness of peat is influenced
by the terrain and geographic process taken place
thousands of years ago”. The definitions don’t
touch the incremental process in peat, which under
cold environments is extremely slow.
Ullah and Khan (2010) analysed 30 samples of
Peatlands from Broghil and found that organic
matter ranged from 32% to 93% with average of 47%
organic matter in peatlands samples collected
above the water-level and 59% organic matter in
peatlands samples collected from below the water
levels. In addition to key ecological functions in
furnishing nesting habitat for birds and regulating
the hydrological processes of high-altitude
meandering streams (Joosten et al. 2012), the
Broghil Valley peatlands provide grazing grounds
for around 20,000 goats, sheep, cows, oxen,
donkeys, horses and yaks.
The 1600 people living in Broghil valley
learned to use peat as a fuel from a migrant family
that arrived from China in the 1940s. Peat use
varies seasonally: in summer (May–September),
when household energy is partly derived from
fuelwood, animal dung or agriculture residues,
daily consumption of peat per household is 40–50
kg. In winter, when people are mostly confined to
their houses by heavy snowfall and cold weather,
this rate jumps to almost 80–100 kg per day per
household (discussions with local community
during Broghil Valley visit in 2008, and subsequent
surveys by Pakistan Wetlands Program 2008–
2012). Local people are well aware of the
importance of peatlands for their livelihoods and,
accordingly, place high value on them. However,
they are now obliged to mine peat because they
have no access to alternative fuels such as wood,
kerosene oil or liquid propane gas, and have
become dependent on peat fuel for their heating
and cooking needs (personal observations 2008).
The resulting degradation of peatlands will be
J. Mt. Sci. (2020) 17(1): 42-49
detrimental for local communities in the long run.
The distribution of high-altitude peatlands in
the Pakistani Himalayas has not yet been studied
in detail. Examples in Broghil Valley were noted by
the author in July 2008 (PWP 2008) and, since
then, have become a focus of research and
conservation for concerned organisations in the
region (Khan et al. 2010; Ning 2012; Rafae 2015;
Shah and Ahmad 2013). Based on ecological value,
landscape features, ecological significance and
biodiversity of the area in addition to socio-
economic significance of the peatlands, Broghil was
declared a National Park in 2010 (Government of
Khyber Pakhtunkhwa 2010). However, there is no
inventory of the natural resources including high
altitude peatlands of the valley to form a basis for
conservation planning in the valley.
In view of the importance of the Broghil
peatlands for carbon storage, their role in
supporting the local flora and fauna, their function
as water buffers (Joosten et al. 2012; Biancalani
and Avagyan 2014), and the continuing use of peat
by the local community, it is important to know the
distribution of these peatlands for monitoring and
conservation. Ground-based studies of the high-
altitude peatlands in Broghil valley are expensive
and difficult because of the harsh climatic
conditions. These difficulties make remote sensing
solutions highly desirable (Shrestha and Zinck
2001). Thus, the mapping of peatlands using
Landsat data (Arvidson et al. 2001) would be an
important step towards assessing the peatland
resources of the valley, the impact of peat
extraction on the ecosystem, and the livelihood of
the local community. Once developed, the remote
sensing method could also be applied to determine
the extent of other high-altitude peatlands across
the mountain ranges of Pakistan and neighbouring
countries. It could also be used to analyse Landsat
time-series data to estimate the rate of any changes
due to natural phenomena such as global warming
or due to extraction by the local community.
In this study, we developed and validated a
technique for mapping high-altitude peatlands
using a single date Landsat 8 image from northern
Pakistan. The Landsat data are available from
USGS at 30 metre spatial resolution for the whole
world. We aim to demonstrate the utility of remote
sensing data applications to map high altitude
peatlands, as a key ecosystem feature. This
technique provides a quick snap shot of target
resources, such as high-altitude peatlands, that
otherwise could only be studied with labor
intensive field techniques requiring time and
logistic resources.
1 Materials and Methods
1.1 Study area
The study area is bounded by the coordinates
36°54'43.5'' N, 73°03'47.5'' E in the northwest and
36°41'21.28'' N, 73°53'18.86'' E in the southeast.
Broghil valley lies at the northern edge of the
Chitral District of Khyber Pakhtunkhwa Province
in Pakistan, at altitudes ranging from 3,100 m
(10,200 ft) at Kishmanjah village to 4,300 m
(14,000 ft) at Kurambar Lake. Broghil Valley has
borders with Afghanistan to the north and west,
where a narrow strip of Afghan territory (Wakhan)
separates it from Tajikistan; with the Gilgit
Baltistan province of Pakistan to the east; and with
Yarkhoon Valley (also part of Chitral District) to
the south (Figure 1).
The climate of northern Pakistan is very cold
and moist due to the presence of high mountain
systems (Khan and Said 2012; Shah and Ahmad
2013). The eastern part of the country lies within
the moist temperate zone of the western Himalayas,
while to the north-west the Karakorum and Hindu
Kush ranges present a much drier environment
(Ives and Messerli 1989). In these areas,
temperatures drop below freezing during winter.
The maximum July/August temperature in the
valleys is 20°C-25°C, while the minimum
temperature in January drops to -10°C (Dijk and
Hussein 1994).
Figure 1 The study area, Broghil Valley in Khyber
Pakhtunkhwa, Pakistan. (Landsat 8 image of July 28,
2013 (source:
J. Mt. Sci. (2020) 17(1): 42-49
1.2 Landsat image analysis
Application of low- and medium-resolution
remote sensing data in high altitude landscapes is
challenging due to snow cover, cloud cover and
shadows. However, the high temporal resolution of
these data is valuable and results in opportunities
for image capture during periods of low clouds and
snow cover. The data used in this research are
derived from a single 185 km × 170 km scene from
Landsat 8 OLI (30 m × 30 m spatial resolution),
identified in the Landsat Worldwide Reference
System 2 (WRS-2) as path/row 150/035. The
peatlands of Broghil Valley, being snowbound for
most of the year (November to June), could only be
detected in images with minimum snow cover. We
used the scene captured on 28 July 2013.
We stacked all the bands excluding the
thermal band (Band 8) and pre-processed the
imagery for radiometric and geometric corrections
(Arvidson et al. 2001; Chander et al. 2009). For
visualization in PCI Geomatics Focus we used three
bands, band 6 (SWIR), band 5 (NIR), and band 4
(Red)—which enabled better visual interpretation
of the land covers. The reflectance of peatlands,
due to higher moisture content in comparison to
other land covers in the valley, enabled
identification of peatlands vs. non-peatlands land
covers. Analyses were performed using the Easi
programming language and Focus application of
PCI Geomatics. Image analysis to separate
peatland and non-peatland land covers was done
using supervised classification with a bagged
decision tree algorithm developed by the Global
Land Analysis and Discovery Team at the
University of Maryland (Breiman 1996; Breiman et
al. 1984; Hansen et al. 1996; Khan et al. 2016).
To classify the area of interest in the image,
the Broghil valley was marked with a polygon using
Google Earth kml file, which was converted to a
shapefile in ArcGIS10.2. The area falling outside of
the Broghil region was masked as no-data. Using
visual interpretation of the Landsat data, hand-
drawn training samples for peatlands and non-
peatlands (Figure 2) were created (Tokola et al.
1999). The Landsat-composite was related to
training data via a bagged classification tree
algorithm (Breiman 1996; Hansen et al. 1996;
Potapov et al. 2012; Bwangoy et al. 2013). A
classification tree software was developed by our
research group following the algorithm described
by Ripley (1996). Classification trees employ an
Figure 2 Peatlands and non-peatlands land covers training of Landsat image (July 28, 2013). Area marked in red
box is projected above as inset.
J. Mt. Sci. (2020) 17(1): 42-49
entropy measure, referred to as deviance, to split
multidimensional space of dependent variables
into successively more homogeneous hyper
volumes, called nodes. The best univariate split was
sought from all independent variables, and the
process was repeated until a perfect tree was fit or
preset conditions for termination of tree growth
were met. We terminated each classification tree
when additional splits decreased model deviance
by less than 0.001 of the deviance of the total
training set population (Bwangoy et al. 2010, 2013).
To further avoid overfitting, we used a set of seven
bagged tree models each derived from a 20%
random sample of training pixels. Each tree
reporting a per-pixel probability of wheat cover
class membership; the per-pixel median of the
seven model outputs was taken as the result
(Potapov et al. 2012). Peat was categorized if this
median value was equal to or greater than 50%.
The area of peatlands in the Broghil valley was
calculated with pixel counts in PCI Geomatica
The classification script was run in perl to
classify the image. The results were compared with
high resolution images from Google Earth for
corrections. The training data were revised
accordingly and the classification script was run
again. The iteration of classification was repeated
several times to achieve a result that compared
satisfactorily with a visual analysis of Google Earth
The validity of the classification was calculated
with ground truth point samples, which were taken
with a Differential GPS in the field. The GPS points
were converted to KML format to overlay on
Google Earth and then to a shape file in ArcMap
10.1. The shape file was overlaid on the
classification results to calculate the Confusion
Matrix for the accuracy assessment.
1.3 Ground-based survey
In 2010, during a survey of land use in the
valley, 34 ground-surveyed points were recorded
with differential GPS. These consisted of 26 points
from peatlands and eight points from non-peatland
areas (personal work with Pakistan Wetlands
Program). These GPS points were used to validate
the peatlands classification using Landsat 8 data.
The GPS points were overlaid as a kml data layer
on the Google Earth image to identify points falling
in peatland and non-peatland areas. The kml data
layer was converted into an ESRI shape file with
ArcMap 10.1. The shape file was overlaid on the
peatlands classification map to calculate the
Confusion Matrix (Stehman and Czaplewski 1998;
Xie et al. 2008) for estimating overall, user and
producer accuracies of the Landsat based peatland
map of the valley.
2 Results
The supervised classification map shows
peatlands distributed along the Broghil river from
Garum Chashma (lower valley) to near Qurumbar
Lake (higher altitude). The total area covered by
peatlands is about 1,933 ha, forming about 1.66%
of the valley’s total area (~118,500 ha). The total
area of the valley, calculated with measuring tool in
Google Earth, was about 1200 hectares (Figure 3).
The reference data were composed of 38
points, of which 26 (68%) were from peatlands,
while 12 (32%) were from non-peatland area.
According to the confusion matrix created from the
38 GPS points (Figure 3), 20 peatland points
corresponded to peatland map, while six of the
peatland points were mapped as non-peatland area.
Eleven of the reference points from the non-
peatland area correspond to non-peatland map,
while one mapped as peatland (Table 1).
The ground reference points were recorded
with differential GPS and overlaid on the Landsat-
based map. This comparison resulted in user
accuracy of 77% and a producer accuracy of 95%
for peatlands, with user and producer accuracies of
92% and 65% respectively for non-peatland areas.
The overall accuracy result was 82%, which
demonstrates a highly reliable map product of the
high altitude peatlands using Landsat data.
3 Discussion
Peat is defined in different ways in accordance
to the local ecological setup. Peatlands in high
altitude mountainous areas of Asia, such as Broghil
in Pakistan, might have characteristic
resemblances with the northern peatlands that
occur in cold to cool areas of mountains, arctic and
J. Mt. Sci. (2020) 17(1): 42-49
sub-arctic in Russia. A detailed phyto-sociological
analysis has not been performed; however, based
on observations of physical characteristics of the
peatlands in Broghil Valley, these peatlands are
located at high elevation above the tree line, have
ground flora, have high moisture content and are
subject to cold temperatures. While high altitude
peatlands such as those of the Broghil valley have
no tree cover, since they occur above the tree line,
other types including boreal, prairie, temperate,
and oceanic peatlands, have either dense or sparse
tree cover (Tarnocai and Stolbovoy 2006).
The Broghil Valley is a typical high-altitude
valley, where ground-based surveys are hindered
by financial constraints, as well as accessibility
constraints caused by the difficult and steep terrain.
Application of the remotely sensed medium-
resolution data presents a low-cost opportunity to
investigate land covers. However, these approaches
are generally limited by high cloud intensity and
snow cover. Clear-sky satellite data such as 30
meter Landsat are scarce for the desired July-
September period, when the Broghil valley is green,
with only permanent snow cover left. In this study
we searched the Landsat archive maintained by
USGS and found images with minimum snow and
cloud cover in the peak vegetation season. These
images provided an opportunity to map peatlands,
which otherwise are hard to access and study.
We found that classification of the Landsat 8
data successfully characterizes the occurrence and
distribution of high-altitude peatlands, particularly
when sample points from ground-based (such as
differential GPS) surveys can be used to validate
the results. Supervised classification was applied
primarily on the basis that adequate a priori
knowledge about the research area exists, and it is
a small area covered in part of one scene (Cihlar
2000). Although we tested the application of
Landsat data to map peatlands in a small area, this
technique can be applied at a larger geographic
scale. Our method, using remotely sensed data,
could potentially be applied to large areas such as
peatlands in Hindu Kush Himalayas, to easily
distinguish vegetation cover in arid and barren
high-altitude mountain landscapes for
identification and monitoring purposes (Hansen &
Loveland 2012). Supervised classification resulted
in lower user accuracy compared to producer
accuracy because some non-peatland areas, such as
Figure 3 Landsat-based map of high-altitude peatlands of Broghil Valley, Pakistan. Peatlands are mapped as green,
while gray is non-peatlands.
Table 1 Accuracy assessment of Broghil peatland map using field-based GPS reference points.
Total User accuracy
Peatlands Not Peatlands
Map Peatlands 20 6
No Peatlands 01 11
Total 21 1
Over all accuracy 82%
Producer accurac
95% 65%
J. Mt. Sci. (2020) 17(1): 42-49
agricultural lands in the vicinity of peatlands, were
mis-classified as peatlands. This mis-classification
probably occurred because most of the agricultural
lands are converted peatlands and contain high
moisture content. This mis-classification could
likely be reduced with further refinement of the
training dataset. The method used for Broghil can
be applied to wider mountainous areas for
assessment of the extent of high-altitude peatlands
in Pakistan and the HKH region. Additionally,
there are other significant protected areas in the
HKH region in Pakistan including Khunjerab,
Central Karakorum, Qurumbar, Shandur-Handrap
and Deosai National Parks. These protected areas
have peatland resources which need inventory and
The high-altitude mountainous areas of Hindu
Kush, Karakorum and Himalayas are not only
difficult to access but logistically expensive for
ground-based surveys. Application of the Landsat
data, available free of cost and with a 16-day revisit
cycle, can be a cost-effective method to assess
distribution of the high altitude peatlands in the
region. The Landsat time series data can also be
applied to study changes in high-altitude peatlands
such as loss over time due to harvesting by the local
population, degradation due to livestock grazing,
and loss due to natural disasters such as glacial
lake outbursts, avalanches and landslides.
4 Conclusion
The method developed and applied in this
paper can be extended to map high altitude
peatlands in adjacent protected areas to
understand their distribution, biodiversity value,
and conservation needs, and also to map changes
to peatlands over time. The Landsat data cover the
high-altitude region of Pakistan from 1988 until
present. This time series can be analysed to map
and quantify land use changes in high-altitude
peatlands, either as a natural or an anthropogenic/
human-caused process. Such a time series could
also be effectively analysed for climate impacts on
high altitude peatlands, which may be subject to
warming in the region. Well-designed reference
data will be required for peatland map validation.
Finally, pixel-based random sampling to
synchronize mapping unit for validation is
recommended for any future studies.
Arvidson T, Gasch J, Goward SN (2001) Landsat 7's long-term
acquisition plan - an innovative approach to building a global
imagery archive. Remote Sensing of Environment 78: 13-26.
Bond P, Campbell KM, Scott TM (1986) An overview of peat in
Florida and related issues. State of Florida, Department of
Natural Resource Management and Bureau of Geology;
Special Publication No. 27: 151.
Biancalani R, Avagyan A (2014) Towards climate-responsible
peatlands management. Food and Agriculture Organization of
the United Nations. p 100.
Breiman L (1996) Bagging predictors. Machine Learning 24:
Breiman L, Friedman JH, Olshen RA, Stone CJ (1984)
Classification and Regression Trees. Chapman and Hall/CRC,
Boca Raton, FL. p 368.
Bwangoy J-RB, Hansen MC, Potapov P, et al. (2013) Identifying
nascent wetland forest conversion in the Democratic Republic
of the Congo. Wetlands Ecology and Management 21(1): 29-
Chander G, Markham BL, Helder DL (2009) Summary of
current radiometric calibration coefficients for Landsat MSS,
TM, ETM+, and EO-1 ALI sensors. Remote Sensing of
Environment 113: 893-903.
Christopherson RW (2000) Geosystems: An Introduction to
Physical Geography. 4th Edition, Prentice-Hall, Inc., Canada.
p 626.
Cihlar J (2000) Land cover mapping of large areas from
satellites: status and research priorities. International Journal
of Remote Sensing 21 (6-7): 1093-1114.
Clark D (2008) The origins and development of peat industry in
Ireland. Peatlands International 1: 12-15.
Cooper DJ, Wolf EC, Colson C, et al. (2018 ) Alpine peatlands of
the Andes, Cajamarca, Peru. Arctic, Antarctic and Alpine
Research 42 (01): 19-33.
Dijk AV, Hussein MH (1994). Environmental profile of North-
West Frontier Province Pakistan. The Netherlands Ministry of
Foreign Affairs and IUCN Pakistan, Islamabad.
Frisch W, Mesched M, Blakey T (2011) Plate Tectonics:
Continental Drift and Mountain Building. Heidelberg:
Springer Berlin Heidelberg. p 207.
Gary MR, McAfee Jr, Wolf CL (eds.) (1974) Glossary of Geology.
American Geological Institute, Alexandria, VA. p 805.
Hansen MC, Loveland TR (2012) A review of large area
monitoring of land cover change using Landsat data. Remote
Sensing of Environment 122: 66-74.
Hansen MC, Dubayah R, DeFries R (1996) Classification trees:
An alternative to traditional land cover classifiers.
International Journal of Remote Sensing 17(5): 1075-1081
Hreibljan JA, Cooper DJ, Sueltenfuss J, et al. (2015) Carbon
storage and long-term rate of accumulation in high-altitude
Andean peatlands of Bolivia. Mires and Peat 15: 1-14.
J. Mt. Sci. (2020) 17(1): 42-49
Ives JD, Messerli B (1989) The Himalayan Dilemma:
Reconciling development and conservation. Routledge, New
York. p 324.
Joosten H, Tapio-Bistrom M, Tol S (2012) peatlands – guidance
for climate change mitigation through conservation,
rehabilitation and sustainable use (2nd ed.). Food and
Agriculture Organization of the United Nations. p100.
Khan A, Said A (2012) Wetlands Management Plan for Broghil
National Park. Pakistan Wetlands Program, Islamabad. p 68.
Khan A, Hansen CM, Potapov P, et al. (2016) Landsat-based
wheat mapping in the heterogeneous cropping system of
Punjab, Pakistan. International Journal of Remote Sensing 37:
Khan A, Said A, Hamid A (2010) Socio-economic profile of
selected communities in the CIWC and NAWC Pakistan.
Pakistan Wetlands Program and GE4DE, ILO, Islamabad.
Kumaran NK, Padmalal D, Limaye RB, et al. (2016) Tropical
Peat and Peatland Development in the Floodplains of the
Greater Pamba Basin, South-Western India during the
Holocene. PLoS ONE 11(5): 1-21.
Lafleur PM, Moore TR, Roulet NT, Frokling S (2005) Ecosystem
respiration in a cool temperate bog depends on peat
temperature but not water table. Ecosystems: 619-629.
Laiho R (2006) Decomposition in peatlands: Reconciling
seemingly contrasting results on the impacts of lowered water
levels. Soil Biology & Biochemistry 38: 2011-2024.
Ning W (2012) Conservation of High Altitude Peatlands in the
Greater Himalayas. In, Conference presentation at Ramsar
COP 11 (p.11). Bucharest, July 10, 2012,: Ramsar Secretariate.
Avialable online at:
Parish F, Sirin A, Charman D, et al. (2008) Assessment on
peatlands, biodiversity and climate change. Global
Environment Center, Kuala Lumpur and Wetlands
International, Wageningen. p 179.
Potapov PV, Turubanova SA, Hansen MC, et al. (2012)
Quantifying Forest cover loss in Democratic Republic of the
Congo 2000-2010 with Landsat ETM Plus data. Remote
Sensing of Environment 122: 106-116.
Prest VK (1970) Quaternary Geology of Canada. In: Douglas
RJW (ed.), Geology and Economic Minerals of Canada,
Geological Survey of Canada. pp 675-764.
Rafae U (2015) An overview of Broghil Valley CCA Project
Jurisdiction. In: IADC Inc presentation online at:
Ranhotra PS and Kar R (2011) Palynological study of glacio-
geomorphic features and its relevance to Quaternary
palaeoclimate and glacial history. Current Science 100(5):
Ripley BD (1996) Pattern Recognition and Neural Networlds.
Cambridge University Press: New York, NY, USA.
Shah KA, Ahmad H (2013) Peatlands of Broghil national park,
Pakistan: human use and management strategy. In: Ning W,
et al. (eds.), High-Altitude Rangelands and their Interfaces in
the Hindu Kush Himalayas. Special Publication, Kathmandu,
Nepal: International Center for Integrated Mountain
Development (ICIMOD). pp 137-145.
Shrestha DP, Zinck JA (2001) Land use classification in
mountainous areas: integration of image processing, digital
elevation data and field knowledge (application to Nepal).
International Journal of Applied Earth Observation and
Geoinformation 3: 78-85.
Stehman SV, Czaplewski RL (1998) Design and analysis for
thematic map accuracy assessment: Fundamental principles.
Remote Sensing of Environment 64: 331-344.
Tarnocai C. Stolbovoy V (2006) Northern Peatlands: their
characteristics, development and sensitivity to climate change.
In: Martini IP, Cortizas A, Chesworth M (eds.) Peatlands:
Evolution and Records of Environmental and Climate
Changes. Elsevier, Amsterdam. pp 17-51.
Tokola T, Lofman S, Erkkila A (1999) Relative calibration of
multitemporal landsat data for forest cover change detection:
Remote Sensing of Environment 68: 1-11.
Ullah I, Khan A (2010) Organic matter contents in selected
peatland and wetlands of Pakistan. The role of water level.
UNDP Pakistan (2002) Indus dolphin. Documentary by UNDP
Pakistan, Islamabad and WWF Pakistan.
U.S. Department of Energy 1979 Peat-Prospectus: United States
Department of Energy, Division of Fossil Fuel Processing,
Washington D.C. p 79.
Xie YC, Sha ZY, Yu M (2008) Remote sensing imagery in
vegetation mapping: a review. Journal of Plant Ecology 1: 9-
ResearchGate has not been able to resolve any citations for this publication.
Full-text available
The aim of this guidebook is to support the reduction of GHG emissions from managed peatlands and present guidance for responsible management practices that can maintain peatlands ecosystem services while sustaining and improving local livelihoods. This guidebook also provides an overview of the present knowledge on peatlands, including their geographic distribution, ecological characteristics and socio-economic importance. This publication considers the environmental and pedological issues associated with peatland use and management before entering into the details of technical options for climate-responsible peatland use.
Full-text available
Holocene sequences in the humid tropical region of Kerala, South-western (SW) India have preserved abundance of organic-rich sediments in the form of peat and its rapid development in a narrow time frame towards Middle Holocene has been found to be significant. The sub-coastal areas and flood plains of the Greater Pamba Basin have provided palaeorecords of peat indicating that the deposits are essentially formed within freshwater. The combination of factors like stabilized sea level and its subsequent fall since the Middle Holocene, topographic relief and climatic conditions led to rapid peat accumulation across the coastal lowlands. The high rainfall and massive floods coupled with a rising sea level must have inundated > 75% of the coastal plain land converting it into a veritable lagoon-lake system that eventually led to abrupt termination of the forest ecosystem and also converted the floodplains into peatland where accumulation of peat almost to 2.0-3.0 m thickness in coastal lowlands and river basins during the shorter interval in the Middle Holocene. Vast areas of the coastal plains of Kerala have been converted into carbon rich peatland during the Middle Holocene and transforming the entire coastal stretch and associated landforms as one of the relatively youngest peatlands in the extreme southern tip of India. Unlike the uninterrupted formation of peatlands of considerable extent during the Holocene in Southeast Asia, the south Peninsular Indian region has restricted and short intervals of peatlands in the floodplains and coastal lowlands. Such a scenario is attributed to the topographic relief of the terrain and the prevailing hydrological regimes and environmental conditions as a consequence of monsoon variability since Middle Holocene in SW India. Considering the tropical coastal lowlands and associated peatlands are excellent repositories of carbon, they are very important for regional carbon cycling and habitat diversity. The alarming rate of land modification and development is destabilizing these carbon pools resulting in large scale carbon emissions to the atmosphere and loss of low-latitude peat palaeorecords. Therefore, these palaeorecords are to be conserved and addressed for better understanding and utilizing the carbon pool for effective climate change adaptation. This communication is the first attempt of addressing the peat formation and peatland development during the Holocene from the tropical region of Peninsular India.
Full-text available
(1) The high-altitude (4,500+ m) Andean mountain range of north-western Bolivia contains many peatlands. Despite heavy grazing pressure and potential damage from climate change, little is known about these peatlands. Our objective was to quantify carbon pools, basal ages and long-term peat accumulation rates in peatlands in two areas of the arid puna ecoregion of Bolivia: near the village of Manasaya in the Sajama National Park (Cordillera Occidentale), and in the Tuni Condoriri National Park (Cordillera Real). (2) We cored to 5 m depth in the Manasaya peatland, whose age at 5 m was ca. 3,675 yr. BP with a LARCA of 47 g m-2 yr-1. However, probing indicated that the maximum depth was 7–10 m with a total estimated (by extrapolation) carbon stock of 1,040 Mg ha-1. The Tuni peat body was 5.5 m thick and initiated ca. 2,560 cal. yr. BP. The peatland carbon stock was 572 Mg ha-1 with a long-term rate of carbon accumulation (LARCA) of 37 g m-2 yr-1. (3) Despite the dry environment of the Bolivian puna, the region contains numerous peatlands with high carbon stocks and rapid carbon accumulation rates. These peatlands are heavily used for llama and alpaca grazing.
Full-text available
Before being used in scientific investigations and policy decisions, thematic maps constructed from remotely sensed data should be subjected to a statistically rigorous accuracy assessment. The three basic components of an accuracy assessment are: 1) the sampling design used to select the reference sample; 2) the response design used to obtain the reference land-cover classification for each sampling unit; and 3) the estimation and analysis procedures. We discuss options available for each of these components. A statistically rigorous assessment requires both a probability sampling design and statistically consistent estimators of accuracy parameters, along with a response design determined in accordance with features of the mapping and classification process such as the land-cover classification scheme, minimum mapping unit, and spatial scale of the mapping.
The methodology used to construct tree structured rules is the focus of this monograph. Unlike many other statistical procedures, which moved from pencil and paper to calculators, this text's use of trees was unthinkable before computers. Both the practical and theoretical sides have been developed in the authors' study of tree methods. Classification and Regression Trees reflects these two sides, covering the use of trees as a data analysis method, and in a more mathematical framework, proving some of their fundamental properties.
Wheat is a staple food of Pakistan’s population of more than 180 million. The average annual harvest of 24 million tonnes places Pakistan eighth in the world in wheat production. Roughly, 76% of wheat is produced by Punjab province, the breadbasket of Pakistan. The current wheat area and yield reporting system operated by the Punjab Crop Reporting Service (CRS) delivers estimates several months after harvest. The delayed production data cannot contribute to in-season decision support system, such as timely production estimation. However, freely available satellite imagery at medium spatial resolution offers a data source for addressing this limitation. Using Landsat data from the 2013 to 2014 winter wheat growing season and a supervised bagged classification tree approach, we estimated the area of wheat cultivated in Punjab. Validation consisted of a two-stage probability-based cluster field sample. Field data from 110 sample points distributed among twelve 20 km × 20 km blocks were collected. Overall accuracy of the Landsat-based wheat map was 76% with wheat commission error of 38% and wheat omission error of 30%. The Landsat-based map totalled 6.13 Mha of wheat area, and the field-based sample estimate 6.96 ± 2.15 Mha. Both figures are comparable to the official Punjab CRS figure of 6.77 Mha. The study presents a robust approach for early wheat area estimation providing information relevant to decision support systems concerning food production and trade.
How are mountains formed? Why are there old and young mountains? Why do the shapes of South America and Africa fit so well together? Why is the Pacific surrounded by a ring of volcanoes and earthquake-prone areas while the edges of the Atlantic are relatively peaceful? Frisch and Meschede and Blakey answer all these questions and more through the presentation and explanation of the geodynamic processes upon which the theory of continental drift is based and which have led to the concept of plate tectonics.