Content uploaded by Neha Midha
Author content
All content in this area was uploaded by Neha Midha on May 09, 2016
Content may be subject to copyright.
J. Indian Soc. Remote Sens. (September 2010 : Special issue on Biodiversity and Landscape Ecology) 38 : 487-500
Assessment of Forest Fragmentation in the Conservation Priority
Dudhwa Landscape, India using FRAGSTATS Computed Class
Level Metrics
N. Midha • P.K. Mathur
Received: 7 April 2010 / Accepted: 2 June 2010
Keywords Conservation priority • Land use/cover • Fragmentation metrics • Dudhwa National Park •
Katerniaghat wildlife sanctuary
N. Midha. P.K. Mathur ()
Wildlife Institute of India
Post Box #18, Chandrabani
Dehra Dun, 248 001, Uttarakhand, India
email : mathurpk@wii.gov.in, mathurpk@yahoo.com
Photonirvachak
RESEARCH ARTICLE
Abstract The Dudhwa landscape, a priority
conservation area representing Terai ecosystem
(woodland-grassland-wetland complex) has witnessed
a sea change in past 150 years or so on account of
long history of forest management, changes in land
use, and rapid economic development. We assessed
fragmentation in two constituent protected areas
(Dudhwa National Park-DNP and Katerniaghat
Wildlife Sanctuary-KAT) of the landscape due to
forest management activities (clear cutting,
development of rail and road network, and
plantations) and compared the magnitude among
them using select metrics at the forest class level. We
applied FRAGSTATS spatial pattern analysis software
(ver.3.3) on different forest classes deciphered by land
use/ cover maps generated using IRS P6 LISS IV
digital data. Study amply revealed that the forests in
DNP are less fragmented and of better habitat quality
than forests of KAT. The set of seven metrics (patch
density, mean patch size, edge density, mean shape
index, mean core area, mean nearest neighbour, and
interspersion and juxtaposition index) at the class level
quantified in the present study are simple and proved
useful for quantifying complex spatial processes and
can be used as an effective means of monitoring in
Dudhwa landscape.
Introduction
World over, habitat loss and fragmentation has been
recognized as a key issue facing the conservation of
488 J. Indian Soc. Remote Sens. (September 2010 : Special issue on Biodiversity and Landscape Ecology) 38 : 487-500
biological diversity. Human activities have modified
the environment to the extent that most common
landscape patterns portray mosaic of human
settlements, agricultural land, and scattered fragments
of natural ecosystems. Most conservation reserves,
even large reserves, are becoming increasingly
surrounded by intensively modified environment and
in the long term appear destined to function as
isolated natural ecosystems (Bennett, 2003).
Fragmentation is a dynamic process that results in
marked changes to the pattern of habitat in a landscape
through time. The term ‘fragmentation’ has been defined
as simultaneous reduction of forest area, increase in
forest edge, and subdivision of large forest areas into
smaller non-contiguous fragments (Laurance, 2000). The
consequences of fragmentation include habitat loss for
some plant and animal species, habitat creation for others,
decreased connectivity of the remaining vegetation,
decreased patch size, increased distance between
patches, and an increased in edge at the expense of
interior habitat (Reed et al., 1996).
The Dudhwa landscape in Uttar Pradesh state of
India is one such example where spatial and temporal
changes of land use have played a major role in
promoting forest fragmentation. The landscape
witnessed a sea change during past 150 years or so
on account of long history of forest management,
changes in land use, and rapid economic development.
Forest reserves were carved out as early as 1880s and
were subjected to extensive working which included
extensive clear felling of natural forests and raising
of monoculture plantations of exotics viz., teak
(Tectona grandis) and eucalyptus (Eucalyptus
citridora). This fostered the establishment of a
massive rail and road infrastructure in Dudhwa
landscape. The landscape severely got influenced
during the country’s post independence era of 1950s
due to abrupt changes in land use policy, settlement
of refugees, expansion of agriculture and large scale
conversion of grassland and swamp habitats into
agriculture land (De, 2001; Mathur and Midha, 2008).
All these factors worked against conservation,
resulting once extensive wilderness into smaller forest
fragments. Today, the Dudhwa landscape harbours
three large forest fragments, now each serving as a
protected area i.e. Dudhwa National Park (DNP),
Katerniaghat Wildlife Sanctuary (KAT) and Kishanpur
Wildlife Sanctuary besides several small, scattered
forest fragments separated spatially amidst human
dominated matrix (Fig. 1).
Over recent decades, many authors have raised
concern about conservation implications of such large
scale alterations on wildlife flora and fauna. Incessant
increase in anthropogenic pressure has resulted into
local extinction of some species such as Indian great
one-horned rhinoceros (Rhinoceros unicornis),
swamp deer (Cervus duvauceli duvauceli) and hog
deer (Axis porcinus) from these areas (Holloway, 1973;
Qureshi et al., 2004; De, 2001; Kumar et al., 2002;
Jonhsingh et al., 2004; Midha and Mathur, 2010). Islam
and Rahmani (2004) have also drawn attention to the
problem of large-scale conversion of natural
grasslands into croplands and plantations outside as
well as inside the protected areas and has worriedly
voiced that it is impeding the survival of key and
threatened grassland bird species like swamp francolin
(Francolinus gularis) and Bengal florican
(Houbaropsis bengalensis).
Recognition of above alterations leading to
fragmentation in Dudhwa landscape has driven land
managers to consider and adopt a landscape
perspective for management. Pre-requisit to this is an
adequate quantitative understanding of how the
landscape has been transformed by fragmentation and
how it will affect landscape characteristics in the future
(Reed et al., 1996). The degree of fragmentation has
been described as a function of the varying size,
shape, spatial distribution, and density of patches
(Jorge and Garcia, 1997). Scientists have been using
metrics for assessing fragmentation and its impact
(Mladenoff et al., 1993; Reed et al., 1996; Lele et al.,
2008). The present study specifically aimed to
understand and compare the magnitude of forest
fragmentation in two protected areas of the Dudhwa
landscape i.e., DNP and KAT due to the influence of
forest management and changes in land use.
489
J. Indian Soc. Remote Sens. (September 2010 : Special issue on Biodiversity and Landscape Ecology) 38 : 487-500
Fig.1 Location of the study area - Dudhwa landscape which includes Dudhwa National Park and Katerniaghat Wildlife
Sanctuary
Study Area
The study area ‘Dudhwa landscape’ stretches across
Lakhimpur Kheri and Bahraich districts of state Uttar
Pradesh in India. It lies between latitude N 280 15'
44.7" and 280 17' 13.4" and longitude E 810 10' 38.7"
and 810 15' 17.2" (Fig.1). It includes DNP and KAT.
The DNP was established in 1977 and covers an area
of 680.3 km2 while KAT was designated as a wildlife
sanctuary in 1976 and encompasses an area of 400.6
km2. The tract experiences very gentle slopes towards
the south-east. The average elevation is 160 m above
mean sea level. The soil consists of Gangetic alluvial
formations. The moist deciduous forests in the tract
dominated by valuable Sal (Shorea robusta) have well
interspersed tall grasslands and swamps (Jha, 2000;
De, 2001; Mathur et al., 2010). Unique complex of
woodland-grassland-wetland ecosystem harbours a
variety of floral and faunal life, including several
charismatic and obligate species viz., tiger (Panthera
tigris), Asian elephant (Elephas maximum), Indian
great one-horned rhinoceros, swamp deer, and pygmy
hog (Sus salvanicus) (Mathur et al., 2010).
Undoubtedly, the landscape has undergone
irreparable changes due to past intensive forest
management, changes in land use, and developmental
activities inside as well as outside the two protected
areas under consideration.
Methodology
Landscapes often depict a mosaic of different patches.
Thus, landscapes are usually characterized by the
structure and composition of constituent patches
besides their spatial pattern or configuration.
Information on the structure, composition and
configuration of patches and spatial pattern of varied
490 J. Indian Soc. Remote Sens. (September 2010 : Special issue on Biodiversity and Landscape Ecology) 38 : 487-500
landscapes has been widely assessed using software
FRAGSTATS, a spatial pattern analysis programme
for quantifying landscape features (McGarigal and
Marks, 1994). Indices computed by FRAGSTATS
characterize each patch in the mosaic, each patch
class (class) in the mosaic, and the landscape mosaic
as a whole. McGarigal and Marks (1994) suggested
that class indices can be used as indicator for
fragmentation as these separately quantify the
amount and distribution of each class or forest class
and thus measure the fragmentation of particular
forest class. Details of dataset and software used for
computation of select metrics at the class level to
quantify fragmentation in two protected areas in the
present study are given below:
Dataset: We used land use/cover maps for DNP and
KAT generated during the Ministry of Environment
and Forests sponsored project ‘Mapping of National
Parks and Wildlife Sanctuaries’ using IRS P6 LISS IV
digital data (Mathur and Midha, 2008). The maps
included 16 land use/ cover (forest and non-forest)
categories in case of DNP (Fig. 2) and 20 for KAT
(Fig. 3). However, for the purpose of present study
on fragmentation, we considered, evaluated and
interpreted only 9 forest classes out of altogether 16
land use/cover classes in case of DNP and 12 forest
classes in KAT out of total 20 classes mapped for
explaining fragmentation. We converted polygon
layers (vector format) developed for preparing land
use/ cover maps into raster format with a pixel size of
100 m.
Computation of Metrics: We used raster version of
FRAGSTATS spatial pattern analysis software
(ver.3.3) developed by McGarigal et al. (2002) to
assess different metrics. As stated above, we selected
seven metrics at class level based on several similar
studies (Hargis et al., 1998; Tinker et al., 1998;
Botequilha and Ahern, 2002; Corry and Nassauer, 2005;
Miyamoto and Sano, 2008). An eight-neighbourhood
criterion for the definition of patches was adopted.
As adopted by Tinker et al. (1998), we selected those
metrics which were standardized per unit area to carry
out comparison of metrics between two PAs. We have
chosen 100 m edge influence to assess mean core
area (Laurance, 2000; Kumar et al., 2002). Table 1 lists
the set of seven metrics (patch density - PD, mean
patch size-MPS, edge density - ED, mean shape index-
MSI, mean core area-MCA, mean nearest-neighbor
distance-MNN, and interspersion and juxtaposition
index-IJI) chosen in the present study. Accordingly, a
forest class with greater density of patches indicate
more fragmentation. Smaller MPS of similar patches
in a forest class also indicates greater fragmentation.
Amount of edge is expected to increase in fragmented
forest class. Fragmentation results into decreased core
area available in patches of a forest class. Managed
forests are expected to have less geometrically
complex patches in terms of shape. The mean nearest-
neighbor distance is expected to decrease as patches
become smaller and more isolated, indicating greater
fragmentation. The percentage values of IJI will
indicate the adjacency of each patch with all other
forest classes.
Results
The results on the assessment of fragmentation in
two protected areas – DNP and KAT and their
indicators and magnitude are presented below.
Forest fragmentation
Patch density and mean patch size: The values of
MPS for 12 different forest classes in KAT ranged
from 9.5 ha to 360 ha while these values for nine forest
classes considered in case of DNP ranged from as
low as 6.1 ha to as high as 1,256.8 ha (Table 2).
Explicitly, with low patch density, open sal forest in
KAT obtained highest and significantly large average
patch size of 360 ha. Moderately dense sal and
Terminalia alata forest classes in KAT obtained next
highest values of MPS (Table 2). Thus, in terms of
fragmentation, open sal, moderately dense sal, and
Terminalia alata forest classes in KAT were found
to be least fragmented. Interestingly, with lower
values of patch density, all different sal forest classes
registered the higher range of average patch size
491
J. Indian Soc. Remote Sens. (September 2010 : Special issue on Biodiversity and Landscape Ecology) 38 : 487-500
Fig. 2 Land use/land cover of DNP developed from IRS P-6 LISS IV at the scale of 1:25,000
Fig. 3 Land use/land cover of KAT developed from IRS P-6 LISS IV at the scale of 1:25,000
492 J. Indian Soc. Remote Sens. (September 2010 : Special issue on Biodiversity and Landscape Ecology) 38 : 487-500
Table 1 Metrics used at class level to quantify fragmentation in KAT and DNP
Metrics and units Description
Patch Density- PD (patches/100 ha) The number of patches of corresponding forest type divided by total area*
(ha) multiplied by 100
¾ A class with greater density of patches indicate that it is subdivided into
many patches and thus could be considered more fragmented
Mean Patch Size – MPS (ha) The average size of the patches of the corresponding forest type
¾ Smaller mean patch size indicates more fragmented forest
Edge Density – ED (m/ha) Total length of edge involving the corresponding forest type divided by total
area (ha)
¾ Amount of edge relative to total area is expected to increase in initial
stages of fragmentation
Mean Shape Index – MSI (None) The average shape index of patches of corresponding forest type, adjusted
by a constant for a square standard (raster)
¾ Patches are expected to become less geometrically complex in managed
forest
Mean Core Area – MCA (ha) The average core area of the patches of the corresponding forest type
¾ One of the main effects of fragmentation is the conversion of interior
habitat to edge habitat. It is expected that the amount of core area will
decrease as a result of fragmentation
Mean Nearest- Neighbor Distance - The average distance between patches of corresponding forest type, based
MNN (m) on edge to edge distance
¾ It is expected to decrease as patches become smaller and more isolated. It
affects the movement and dispersal of species
Interspersion and The adjacency of each patch with all other forest types
Juxtaposition Index - IJI (%) ¾It is expected to increase
*Total area of respective study site (DNP or KAT)
(Table 2). Hence, they were found to be least
fragmented forests in DNP.
Edge density, mean shape index, and mean core area:
The values of edge density in case of 12 different
forest classes in KAT ranged from 0.6 m/ha to 12.1 m/
ha. The lowest value was registered in case of khair
and sissoo (Acacia catechu and Dalbergia sissoo)
forest while highest value was obtained by tropical
semi-evergreen forest. In DNP, highest value of edge
density, being 8 m/ha was recorded by moderately
dense sal forest whereas lowest value of just 0.2 m/ha
was registered by class of other plantations (Table 2).
The results for mean shape index revealed that
values for all the forest classes in KAT were greater
than one indicating that the average patch shape in
all forest classes in sanctuary area was irregular (Table 2).
In DNP, three highest values of mean shape index
obtained were 2.42, 2.38, and 2.37 by three sal forest
classes i.e. open sal forest, dense sal forest, and
moderately dense sal forest, respectively thus
pointing out towards their native character to the tract.
Interestingly, all forest classes in KAT recorded
MCA < 50% of MPS while in contrast almost all forest
classes in DNP recorded values of MCA >50% of
MPS. As core area is affected by shape, this clearly
indicated that effect of edge is relatively low in
different forest classes of DNP.
Mean nearest neighbour, and interspersion and
juxtaposition index: In case of KAT,the average
distance of similar patch for almost all forest classes
came out to be short while the overall interpretation
493
J. Indian Soc. Remote Sens. (September 2010 : Special issue on Biodiversity and Landscape Ecology) 38 : 487-500
revealed that patches of different forest classes in
DNP were located relatively far from each other (Table 2).
In KAT, values for IJT ranged from 36% to 86%,
amply indicating that sal forests were moderately
interspersed among all different forest classes (Table 2).
The values for IJI in DNP ranged from 58% to 81%.
All sal forest classes obtained value more than 70%
indicating that all sal forest classes in the national
park area were highly interspersed.
Comparison of KAT and DNP
A comparison of extent of fragmentation in two
constituent protected areas - KAT and DNP allowed
several interesting revelations (Figs. 4 and 5).
Lower values of patch density obtained by most
of the forest classes in DNP than those in KAT clearly
indicated that forest classes in the national park area
were relatively less fragmented, except tropical
seasonal swamp forest in DNP which was represented
by almost double number of patches as compared to
KAT.
A comparison of values of MPS between DNP
and KAT showed some striking differences (Table 2
and Fig. 4b). Revelation pointed out that patches of
forest classes in DNP were larger in sizes, particularly
in case of four classes of sal forest. In other forest
classes viz., mixed deciduous forest, tropical semi-
evergreen forest, teak plantation, and other
plantations,the value of MPS were almost comparable
Table 2 Fragmentation measurement metrics computed for KAT and DNP
Forest Type % PD MPS ED MSI MCA MNN IJI
area (patches (ha) (m/ha) (ha) (m) (%)
/100ha)
Katerniaghat Wildlife Sanctuary - KAT
Dense Sal 3. 1 0.05 59.9 2.8 1.8 24.2 365.8 56.8
Moderately Dense Sal 5.0 0.03 168.3 4.1 2.0 81.1 628.1 52.6
Sal Mixed 1.9 0.04 54.5 1.8 1.7 23.9 556.7 61.5
Open Sal 7. 9 0.02 359.5 5.4 2.3 196.6 215.4 68.8
Mixed Deciduous 11.4 0.16 80.0 10.0 1.7 40.8 567.7 85.9
Tropical Semi-Evergreen 12.4 0.14 98.5 12.1 1.8 45.5 262.3 69.0
Tropical Seasonal Swamp 1. 6 0.12 15.2 2.4 1.3 4.8 743.6 74.4
Khair and Sissoo 0. 7 0.01 75.0 0.6 1.7 35.7 3876.9 65.6
Aegle marmelos 0.7 0.05 16.7 1.2 1.4 3.3 239.3 36.0
Terminalia alata 3.3 0.03 120.7 2.9 2.0 59.0 307.9 39.8
Teak Plantation 9.5 0.13 76.9 7.5 1.7 40.6 323.1 79.9
Other Plantations 1.0 0.10 9.5 1.2 1.2 3.4 671.6 72.1
Dudhwa National Park - DNP
Dense Sal 11.4 0.01 974.2 3.3 2.3 781.0 2181.8 70.5
Moderately Dense Sal 23.5 0.01 1256.8 8.0 2.3 965.3 424.9 80.6
Sal Mixed 3.7 0.01 361.1 1.5 1.8 259.1 1492.9 70.9
Open Sal 17.7 0.02 449.5 5.8 2.4 307.6 978.1 75.8
Mixed Deciduous 8. 3 0.14 59.9 7.0 1.5 29.5 596.2 79.8
Tropical Semi-Evergreen 2 .1 0.02 95.7 2.0 2.0 38.1 864.8 58.5
Tropical Seasonal Swamp 3. 9 0.21 18.5 5.8 1.5 4.5 326.9 64.4
Teak Plantation 5.0 0.04 114.8 3.3 1.7 65.3 1031.6 78.4
Other Plantations 0.1 0.02 6.1 0.2 1.2 0.1 1106.6 64.2
494 J. Indian Soc. Remote Sens. (September 2010 : Special issue on Biodiversity and Landscape Ecology) 38 : 487-500
Fig. 4 Comparison of Values of Metrics (Patch Density, Mean Patch Size, Edge Density, and Mean Shape Index) for
Different Forest Classes in Dudhwa National Park (DNP) and Katernaighat Wildlife Sanctuary (KAT)
(DS: Dense Sal Forest; MDS: Moderately Dense Sal Forest; SM: Sal Mixed Forest; OS: Open Sal Forest; MDF: Mixed
Deciduous Forest; TSE: Tropical Semi-Evergreen Forest; TSSF: Tropical Seasonal Swamp Forest; TA: Terminalia alata
Forest; KSF: Khair and Sissoo Forest; AF: Aegle marmelos Forest; TP: Teak Plantation; OP: Other Plantations)
495
J. Indian Soc. Remote Sens. (September 2010 : Special issue on Biodiversity and Landscape Ecology) 38 : 487-500
Fig. 5 Comparison of Values of Metrics (Mean Core Area, Mean Nearest Neighbour, Interspersion and Juxtaposition
Index) for Different Forest Classes in Dudhwa National Park (DNP) and Katernaighat Wildlife Sanctuary (KAT)
(DS: Dense Sal Forest; MDS: Moderately Dense Sal Forest; SM: Sal Mixed Forest; OS: Open Sal Forest; MDF: Mixed
Deciduous Forest; TSE: Tropical Semi-Evergreen Forest; TSSF: Tropical Seasonal Swamp Forest; TA: Terminalia alata
Forest; KSF: Khair and Sissoo Forest; AF: Aegle marmelos Forest; TP: Teak Plantation; OP: Other Plantations)
496 J. Indian Soc. Remote Sens. (September 2010 : Special issue on Biodiversity and Landscape Ecology) 38 : 487-500
in two PAs (Fig.4b). However, lower values of patch
density in such forest classes suggested that these
forests were present in less fragmented state in DNP
(Fig. 4a).
Figure 4c compares values of edge density and
revealed that sal forest classes in KAT and DNP
registered almost same edge density except for
moderately dense sal forest. However, all other forest
classes recorded higher values of edge density in
KAT except tropical seasonal swamp forest. The
overall comparison indicated that the effect of edge
was less prominent in DNP. Comparison of MPS in
two PAs could not allow any striking differences (Fig.
4d). Only, dense sal forest and moderately dense sal
forest accounted for slightly higher values of MPS
indicating their irregularity in shape and pointed
towards being present in relatively natural state in
DNP.
Notably, the values of MCA in DNP were much
higher than those in KAT (Fig. 5a). This could be
attributed to less edge effect in DNP as highlighted
above. These results indicated towards better habitat
quality in DNP as compared to KAT. In addition, this
difference in MCA also gave the cue about the process
of fragmentation in KAT which encompass significant
impact due to changes in configuration rather than
due to forest loss.
The values of mean nearest neighbour were
found to be low in KAT for most forest classes as
compared to DNP (Fig. 5b). This indicates that
although forests were less fragmented in DNP but
patches were located far apart from each other. Lower
values of mean nearest neighbour in case of tropical
seasonal swamp forest in DNP and tropical semi-
evergreen forest in KAT indicated that these two
featured forest classes were highly localized in
respective PA.
The interpretation of values for IJI indicated that
most of the forest classes in both PAs reported almost
same level of interspersion (Fig. 5c). Only two
exceptions were dense sal and moderately dense sal
forest classes in DNP those obtained significantly
higher values indicating their high interspersion.
Discussion and conclusions
We examined magnitude of forest fragmentation in
two protected areas within the Dudhwa landscape as
well as compared them. This way we compared the
result of forest management activities (clear cutting,
development of rail and road network, and plantations)
among them, ultimately allowing us to determine which
one has been most severely impacted.
Reed et al. (1996) has pointed that the changes
in the shape, edge, density and diversity related
measures reflect the impact of forest management in
the landscape. Several authors have concluded that
the forest fragmentation tends to increase the number
of patches, decrease the MPS and interior forest
habitat (enhanced edge and reduced MCA), and
decrease the amount of old growth forests (Ripple et
al., 1991; Spies et al., 1994; Bennett, 2003; Holt and
Debinski, 2003; Miyamoto and Sano, 2008). Present
study distinctly supports findings of above authors
and also reflected the impact of prolonged phase of
active forest management followed by a relatively
short phase of passive management of forests in DNP
during past three decades or so after its designation
as a national park.
Both PAs have the presence of large extent of
areas under plantations especially of teak indicating
that the area was widely planted. In addition,
moderately large patch size, irregular shape, and high
interspersion of these plantation patches added that
during the present passive phase, teak has managed
to establish itself well in KAT as well as DNP despite
being an exotic species to the tract. Plantations
covered 5% area of DNP while the extent of teak and
other plantations in KAT was 10.5%.
In KAT, the impact of clear cutting along with
infrastructure development is quite apparent as only
18% of the area remained under different classes of
sal forest albeit it being the dominant forest class of
this tract. Furthermore, open sal forest registered the
largest extent among all classes of sal forest while
dense sal forest class which is indeed the matured or
old growth forests contributed less. In contrast,
different sal forests were found to be least fragmented
in DNP and in natural state as indicated by their actual
areal extent, patch density, MPS, and high shape index.
497
J. Indian Soc. Remote Sens. (September 2010 : Special issue on Biodiversity and Landscape Ecology) 38 : 487-500
In addition, all the forest classes in KAT acquired
the irregular shapes but recorded the MCA < 50% of
MPS which pointed that the shape of patches of these
forests had been affected by edges created during
road development and clear cutting. Similarly, in
Oregon USA, Spies et al. (1994) reported a decrease
in mean interior patch area from 160 to 62 ha/patch
following clear cutting. In an analogous study, Ripple
et al. (1991) found that patch area was reduced by
17% as a result of clear cutting. Shinneman (1996)
found that shelter wood cutting had essentially
depleted core old-growth forest area in the Black Hills
National Forest, USA. Roads in this study reduced
MPS and patch core sizes by > 70%. Interestingly,
almost all the forest classes in DNP recorded MCA
>50% of MPS thus indicated the low effect of edge.
Patches of different forest classes in DNP were found
to be located relatively at far distance from each other.
It is amply clear that biotic pressure has
contributed to fragmentation of the Dudhwa
landscape. The comparison appears worthwhile
between two protected areas which belong to same
Terai tract and were once connected to each other.
The high patch density and concurrent less MPS of
forest classes in KAT as compared to forest classes
in DNP indicated forests in DNP were relatively less
fragmented than different forests in KAT. This was
especially applicable in the case of sal forests. The
areal extent and MPS of old growth forests i.e. dense
sal and moderately dense sal forest classes were quite
large in DNP as compared to KAT. Distinctly, DNP
recorded significantly higher values of MCA
representing greater forest interior habitat as
compared to trend followed by forest classes in KAT.
This could be attributed to less edge effect on the
shape of the patches of forest classes in DNP. The
mean shape index could not allow any striking
differences except in the case of Dense sal forest and
moderately dense sal forest which pointed towards
being present in relatively natural state in DNP. This
was further endorsed by these two forest classes
being better interspersed as compared to KAT.
Counter intuitively, the inter patch differences were
found to be large in DNP which indicated that although
forests are less fragmented in DNP but patches are
located far apart from each other. Most of the forest
classes reported similar interspersion in both
protected areas.
In nutshell, the analysis of chosen metrics
definitely revealed that the forests in the national park
were less fragmented and of better habitat quality
than forests in sanctuary (KAT). The magnitude of
fragmentation was related to dominating forest
classes, land use, and level of protection. The
plausible reason for the difference could be benefit
of protection that the national park has enjoyed for
such a longer period than the sanctuary and
suspension of forestry operations much earlier (ca.
30 years) in the national park.
Implication for management and conservation
The process of landscape change as a result of
fragmentation caused primarily by forest management
activities (clear cut, development of rail and road
network, and plantations) has far-reaching
consequences for native plants, vertebrates and
invertebrates, particularly survival of threatened
species. Several studies have reported that timber
harvesting or clear cutting results in disproportionate
removal of late succession forests (Ruggiero et al.,
1994; Tinker et al., 1998). For birds and mammals those
depend on late-succession stands, this could have
resulted in qualitative and quantitative reduction of
their suitable habitat in case of KAT. Further, large
core areas are important habitat features for some
mammals, especially forest carnivores (Tinker et al.,
1998) and thus might be getting greatly affected in
KAT. It has been well established that an elusive
carnivore species like tiger requires adequately large
‘inviolate’ areas for breeding. Tiger being territorial
animal advertises its presence in an area and maintains
a territory. The tiger land tenure dynamics ensure
presence of prime adults in a habitat which acts as
‘source’ population, periodically replacing old males
by young adults from nearby forest areas. Gopal et
al. (2007) in ‘guidelines for preparation of tiger
conservation plan’ has highlighted the critical
requirement of an ‘inviolate’ space of 800-100 so as
498 J. Indian Soc. Remote Sens. (September 2010 : Special issue on Biodiversity and Landscape Ecology) 38 : 487-500
to maintain a viable population of 80-100 tigers. Tiger
being a ‘flagship’ species also ensure viable
population of co-predators and prey species besides
ecological integrity of the entire habitat. DNP and
KAT are two large remnant and important habitat
blocks not only for tiger but several other endangered
species of Terai.
It is worth mentioning that the tenuous
interconnectivity among patches of similar forest
classes in DNP could effect movement and dispersal
of faunal species. In addition, clear cuts and roads
block the movement of some species, resulting in
population fragmentation and increased competition
for resources in remaining forest resources (Lovejoy
et al., 1986; Noss, 1993). The effect of fragmentation
on the species was outside the scope of this study,
but a direct analysis of the impact is warranted and
should be a significant part of current management
planning.
The FRAGSTATS analysis applied in the present
study has been used in various regions of the world.
However, because of the different data source and
classification classes, it was difficult to compare
directly the index values from various regions. The
set of seven metrics quantified in this study are simple
and proved useful for quantifying complex spatial
processes and can be used as an effective means of
monitoring in Dudhwa landscape. The approach of
landscape level assessment and monitoring by select
metrics has been recommended and adopted by many
authors for different protected areas across the world
(Riitters et al., 1995; Botequilha. and Ahern, 2002;
Schindler et al., 2008). As both PAs - DNP and KAT
are in recovery phase after prolonged phase of active
forest management, changes in composition of forest
patches and configuration must be monitored and
effects of land use and management interventions on
landscape spatial pattern must be analysed. This
knowledge can then be used to assess progress in
conservation efforts and to improve management
decisions not only for Dudhwa landscape, but also in
other similar landscapes.
Acknowledgements The research was conducted at
the Wildlife Institute of India (WII) with funding
support from the Ministry of Environment and Forests
(MoEF), Government of India for the WII-MoEF-
National Natural Resource Management System
(NNRMS) Project. Thanks are due to P.R. Sinha,
Director, WII and V.B. Mathur, Dean, Faculty of
Wildlife Science, WII for their advice and support.
Special thanks are due to senior forest officials of
Uttar Pradesh Forest Department for their help in
various ways and frontline staff for field logistics.
References
Bennett AF (2003) Linkages in the landscape. The
role of corridors and connectivity in wildlife
conservation. IUCN Forest Conservation
Programme, Conserving Forest Ecosystem Series
No.1 (www.iucn.org/themes/fcp/publications/
files/linkages_in_the_landscape.pdf)
Botequilha LA and Ahern J (2002) Applying landscape
ecological concepts and metrics in sustainable
landscape planning. Landscape and Urban
Planning 59:65–93
Corry RC and Nassauer JI (2005) Limitations of using
landscape pattern indices to evaluate the
ecological consequences of alternative plans and
designs. Landscape and Urban Planning 72:
265–280
De R (2001) Management plan of Dudwa Tiger Reserve
(2000-2001 to 2009-2010), Wildlife Preservation
Organization, Forest Department, Uttar Pradesh
Gopal R, Sinha PR, Mathur VB, Jhala YV and Qureshi
Q (2007) Guidelines for preparation of Tiger
Conservation Plan. A technical document of the
National Tiger Conservation Authority, Ministry
of Environment and Forests, Government of
India. NTCA/01/07
Hargis CD, Bissonette JA and David JL (1998) The
behavior of landscape metrics commonly used
in the study of habitat fragmentation. Landscape
Ecology 13:167-186
499
J. Indian Soc. Remote Sens. (September 2010 : Special issue on Biodiversity and Landscape Ecology) 38 : 487-500
Holloway CW (1973) Swamp deer in Uttar Pradesh.
Oryx 12:41-48
Holt RD and Debinski DM (2003) Reflections on
landscape experiments and ecological theory:
tools for the study of habitat fragmentation. In:
How Landscapes Change. Human Disturbance
and Ecosystem Fragmentation in the Americas
(Eds.: G.A. Bradshaw and P.A. Marquet),
Springer-Verlag, New York, USA. pp 201-218
Islam MZ and Rahmani AR (2004) Important bird areas
in India: Priority sites for conservation. Indian
Bird Conservation Network, Bombay Natural
History Society and Birdlife International (U.K.).
pp XVIII + 1133
Jha RN (2000) Management plan of Katerniaghat
Wildlife Sanctuary (2000-2001 to 2009-2010).
Wildlife Preservation Organisation. Forest
Department, Uttar Pradesh
Johnsingh AJT, Ramesh K, Qureshi Q, David A, Goyal
SP, Rawat GS, Rajapandian K and Prasad S (2004)
Conservation status of tiger and associated
species in the Terai Arc Landscape, India. Wildlife
Institute of India Report. Dehradun
Jorge LABJ and Garcia GJ (1997) A study of habitat
fragmentation in Southeastern Brazil using
Remote Sensing and Geographic Information
Systems (GIS). Forest Ecology and Management
98: 35-47
Kumar H, Mathur PK, Lehmkuhl JF, Khati DVS, De R
and Longwah W (2002) Management of forests
in India for biological diversity and forest
productivity, A new perspective - volume VI: Terai
Conservation Area (TCA). WII-USDA Forest
Service Collaborative Project Report, Wildlife
Institute of India. Dehradun. pp xiii + 158
Laurance WF (2000). Do edge effects occur over large
spatial scales? Trends in Ecology and Evolution
15:134–135
Lele N, Joshi PK and Agarwal SP (2008) Assessing
forest fragmentation in northeastern region
(NER) of India using landscape metrices.
Ecological Indicators 8:657–663
Lovejoy TE, Bierregaard RO, Rylands AB , Malcolm
JR, Quintela CE, Harper LH, Brown KS, Powell
AH, Powell GVN, Schubart HOR and Hays MB
(1986) Edge and other effects of isolation on
Amazon forest fragments. In: Conservation
Biology: The Science of Scarcity and Diversity
(Ed.: M. E. Soulé.). Sinauer Associates, Inc.,
Sunderland. Massachusetts. pp. 257-285
Mathur PK and Midha N (2008) Mapping of National
Parks and Wildlife Sanctuaries, Dudhwa Tiger
Reserve, WII-NNRMS-MoEF Project, Final
Technical Report, Wildlife Institute of India,
Dehradun, India. pp. 216
McGarigal K and Marks BJ (1994) Spatial pattern
analysis program for quantifying landscape
structure. USDA Forest Service General Technical
Paper. PNW–351. (www.umass.edu/landeco/
pubs/Fragstats.pdf)
McGarigal K, Cushman SA, Neel MC and Ene E (2002)
FRAGSTATS: Spatial pattern analysis program
for categorical maps. Computer software program
produced at the University of Massachusetts,
Amherst. (www.umass.edu/landeco/research/
fragstats/fragstats.html)
Mathur PK, Kumar H, Lehmkuhl JF, Tripathi A,
Sawarkar VB and De R (2010) Mammal indicator
species for protected areas and managed forests
in a landscape conservation area of northern
India. Biodiversity and Conservation (DOI
10.1007/s10531-010-9851-8)
Miyamoto A and Sano M (2008) The influence of forest
management on landscape structure in the cool-
temperate forest region of central Japan.
Landscape and Urban Planning 86:248–256
Mladenoff DJ, White MA, Pastor J and Crow TR
(1993) Comparing spatial pattern in unaltered old-
growth and disturbed forest landscapes.
Ecological Application 3:294-306
Midha N and Mathur PK (2010) Conservation
implications of the channel changes in Sharda
River on endangered swamp deer population and
floodplain ecosystem in Kishanpur Wildlife
Sanctuary, Uttar Pradesh. Current Science 98:
(5) 665-672
Noss RF (1993) The wildlands project land
conservation strategy. In: Wild Earth: Special
issue on the Wildland Project. Canton. New York.
pp. 10-25
500 J. Indian Soc. Remote Sens. (September 2010 : Special issue on Biodiversity and Landscape Ecology) 38 : 487-500
Qureshi Q, Sawarkar VB, Rahmani AR and Mathur PK
(2004) Swamp deer or barasingha (Cervus
duvauceli Cuvier, 1823) In: Ungulates of India
(Eds.: K. Sankar and S.P. Goyal) Envis Bulletin:
Wildlife and Protected Areas, Vol.07, No. 1 Wildlife
Institute of India, Dehradun, India. pp 181-192
Reed RA, Johnson-Barnard J and Baker WL (1996)
Fragmentation of a forested rocky mountain
landscape, 1950-1993. Biological Conservation
75: 267-277
Riitters KH, O’Neill RV, Hunsacker CT, Wickham JD,
Yankee DH, Timmins SP, Jones KB and Jackson
BL (1995) A factor analyses of landscape pattern
and structure metrics. Landscape Ecol 10:23–39
Ripple WJ, Bradshaw GA and Spies TA (1991)
Measuring forest landscape patterns in the
Cascade Range of Oregon, USA. Biological
Conservation 57: 73–88
Ruggiero LF, Aubry KB, Buskirk SW, Lyon LJ and
Zielinski WJ (1994) The scientific basis for
conserving forest carnivores: American marten,
fisher, lynx, and wolverine in the western United
States. USDA Forest Service General Technical
Report RM-254, Rocky Mt. For. & Range Exp.
Sta., Fort Collins, CO, USA
Schindler S, Poirazidis K and Wrbka T (2008) Towards
a core set of landscape metrics for biodiversity
assessments: A case study from Dadia National
Park, Greece. Ecological Indicators 8:502-514
Shinneman DJ (1996) An analysis of range of natural
variability, roads, and timber harvesting in a Black
Hills ponderosa pine forest landscape. M.A.
Thesis, Univ. of Wyoming, Laramine, WY, USA
Spies TA, Ripple WJ and Bradshaw GA (1994).
Dynamics and pattern of a managed coniferous
forest landscape in Oregon. Ecological
Application 4 (3):555–568
Tinker DB, Resor CAC, Beauvais GP and Kipfmueller
KF (1998) Watershed analysis of forest
fragmentation by clear cuts and roads in a
Wyoming forest. Landscape Ecology 13:149-165
