Content uploaded by Rohit R. S. Jha
Author content
All content in this area was uploaded by Rohit R. S. Jha on Jul 19, 2022
Content may be subject to copyright.
Assessing patterns of butterfly communities in a multi-use
landscape and its implications for conservation
A Thesis
submitted to the
Tata Institute of Fundamental Research
for the degree of
Master of Science
in
Wildlife Biology and Conservation
by
Rohit Ravindra Samita Jha
2014
National Centre for Biological Sciences
Tata Institute of Fundamental Research
Bengaluru, India
DECLARATION
This thesis entitled “Assessing patterns of butterfly communities in a
multi-use landscape and its implications for conservation” is a
presentation of my original research work. Wherever contributions of
others are involved, every effort is made to indicate this clearly, with due
reference to the literature, and acknowledgement of collaborative
research and discussions.
The work was done under the guidance of Dr. Krushnamegh Kunte at the
National Centre for Biological Sciences – Tata Institute of Fundamental
Research, Bengaluru, India.
______________
Rohit Ravindra Samita Jha
Candidate, M.Sc Wildlife Biology and Conservation
NCBS-TIFR
Bengaluru 560 065, India
In my capacity as supervisor of the candidate’s thesis, I certify that the
above statements are true to the best of my knowledge.
________________
Guide:
Dr. Krushnamegh Kunte
National Centre for Biological Sciences
Tata Institute of Fundamental Research
GKVK, Bellary Road, Bengaluru 560 065, India
Certificate
I certify that the thesis entitled “Assessing patterns of butterfly
communities in a multi-use landscape and its implications for
conservation” comprises original research work carried out by Rohit Jha
at the National Centre for Biological Sciences, Tata Institute of
Fundamental Research, Bengaluru, under the supervision of Dr.
Krushnamegh Kunte during the period November 2013 to May 2014 for
the degree of Master of Science in Wildlife Biology and Conservation of
Tata Institute of Fundamental Research. The results presented in this
thesis have not been submitted previously to this or any other university
for a Master of Science degree, or any other degree.
__________________
Prof. Apurva Sarin
Head (Academics)
National Centre for Biological Sciences
Tata Institute for Fundamental Research
GKVK, Bellary Road, Bengaluru 560 065, India
i
Executive summary
One of the fundamental goals in ecology is to explain spatial and temporal patterns of
biological diversity. This is important as it helps in determining associations of
environmental variables with observed patterns of community similarity in order to
understand mechanisms that may influence beta diversity. Moreover, it is also crucial
in determining appropriate locations for nature reserves as well as in assessment of
management strategies with respect to exploitation of forest resources. Butterflies are
one of the better known invertebrate taxon in the world. With an estimated 20,000
species in the world, they occupy a prominent place in conservation efforts and
biodiversity assessments as they can serve as a valuable barometer of overall
community complexity, especially in tropical forests.
The Garo hills in Meghalaya, north-eastern India, is part of the Indo-Burma
biodiversity hotspot region. Vast swathes of this autonomous region are occupied by
tropical forests, the majority of which are community-managed. Covering a spatial
extent of approximately 605 sq.km., the Balpakram-Baghmara (BBL) landscape in the
South Garo hills district of the region encompasses both the largest protected area in
the state, the Balpakram national park, as well as smaller state-managed reserved
forests. In such a landscape where protected areas, community-managed forests,
monoculture plantations and shifting cultivation (locally called jhum) co-exist, the
main aim of this study was to assess and compare the community structure of
butterflies in each of these land-management categories in order to determine the
effect of rapid land-use change on butterfly communities.
I enumerated butterflies during time-constrained 30-minute counts in all four
treatment areas between March and May 2014. I surveyed different sites only once
during this period, but took a total of 298 counts covering varied vegetation types and
ii
land-management systems multiple times. During these counts I captured individual
butterflies for identification purposes if needed, and released them on the spot. I also
recorded environmental variables such as light, relative humidity and temperature
during all counts.
Comparison of diversity indices and extrapolated species richness for each treatment
revealed that the community-owned and managed forests were not significantly
different from the state-owned patches of forest. Both harboured a similar
composition of habitat-specialist species as well. Butterfly communities in
monoculture plantations and jhum landscape differed significantly from those found
in forests (state- or community-managed) and were highly depauperate in terms of
habitat specialist species. Climatic variables of light, temperature and relative
humidity did not seem to affect butterfly communities in a significant manner across
treatments, although they did exert an overall effect on butterfly species richness and
abundances.
The study highlights the importance of community forests in butterfly conservation,
and perhaps biodiversity conservation in general, in the Garo hills. It also suggests
that the rapidly changing land-use pattern being witnessed in these areas, with forests
making way for monoculture plantations, might pose a threat to such diverse
ecological assemblages. It might be worth considering innovative ways of involving
people with forest conservation activities in order to preserve the unique and varied
biota of the region.
iii
Acknowledgements
First and foremost, I am deeply grateful to all my field assistants without whom this work
would not have been possible at all. In my quest for the 'meplip' (= butterfly in Garo) they
were enthusiastic at all times! Of them all, I particularly enjoyed working with Tilar
Marak, Kebi Sangma, James Marak, Nuking Marak, Gogon Marak, Maibin Sangma and
Susillas Marak. I also express sincere gratitude to all my hosts in Garo hills who
accommodated me on various occasions for different lengths of time during the study
period. I also thank Shikha Srikant of Samrakshan Trust for logistical help and the initial
support that I very much needed to begin fieldwork. Kashmira Kakati is thanked for her
constant encouragement on field. I thank Kamal Medhi of Samrakshan Trust for
providing me with GIS layers of the landscape.
I would also like to thank a lot of people who extended their unconditional help at all
times, on most occasions at short notice. Amongst them are Mr. Plinder Marak, President,
Siju Ecotourism and Conservation Society (SECOS); Mr. Sengcheng Marak, member,
SECOS, Mrs. Parsist Marak, President, Karawani Eco-tourism and Conservation Society;
Mr. Maxbone Sangma, Assistant Conservator of Forest (ACF), Balpakram National Park
division; Mr. Witnes Marak, Forester, Chimitap independent beat; Mr. Iyerson Marak,
Forester, Rongara range and Mr. Debakar Marak, Forester, Sanbana sub-beat.
I would like to particularly thank Mr. Gonsing Sangma and his wife who were my hosts
in Karawani, the place which I used as a base more or less. I had my most wonderful time
and food in Garo hills at this place. I would also like to particularly thank the people of
Karawani vilage near Baghmara for extending their kindness and warmth to me, and also
letting me be a part of one of the best Christmas and new year celebrations of my life.
I thank my guide Dr. Krushnamegh Kunte for helping me formulate, develop and give
shape to a small idea that I had into a full-fledged dissertation project. I would also like to
thank my co-guide Dr. Jagdish Krishnaswamy for giving me useful inputs whenever I
needed them. I thank Shankar Raman, Divya Mudappa, Suhel Quader, Hari Sridhar and
Nachiket Kelkar for initial discussions regarding the project, as well as Rajat Nayak for
assistance in analysis. I thank all my classmates for their support and encouragement
throughout the course, particularly Sreedhar, Lakshminarayanan, Rakesh, Aravind,
Iravatee and Chetana. Lakshminarayanan also proof-read this manuscript and provided
certain crucial inputs from which it greatly benefitted.
Without the blessings, love and encouragement of a supportive family, all of this would
never have been a reality. I thank the Department of Science and Technology for funding
this project, and NCBS for providing a wonderful campus and great atmosphere! I also
the Department of Forest and Environment, Government of Meghalaya for permits and
field support. Last but not the least, I thank Dr. Ajith Kumar for being a great course
director, as well as Chandni Gurusrikar for co-ordinating the M.Sc. course at NCBS.
Contents
Page
Executive Summary…………………………………………………………………….
i
Acknowledgements……………………………………………………………………..
iii
General Introduction…………………………………………………………………….
1
Manuscript:
Assessing patterns of butterfly diversity in a multi-use landscape and its implications
for conservation
4
Abstract………………………………………………………………………………….
5
Introduction……………………………………………………………………………...
Materials and Methods
6
Study Area………………………………………………………………………..
10
Field methods……………………………………………………………….........
13
Statistical Analyses……………………………………………………………….
17
Results………………………………………………………………………………....
19
Discussion……………………………………………………………………………….
27
References……………………………………………………………………………….
35
1
General Introduction
Explaining the underlying mechanisms of observed spatial and temporal patterns in
biological communities is one of the fundamental objectives in community ecology.
Habitat degradation accompanied by rapidly changing land-use patterns has been
recognised as the biggest threat to biodiversity, especially in tropical areas (Dobson et
al 1997, Laurance 1999). The problem increases manifold here in the tropics since,
more often than not, areas of high deforestation overlap with areas of high
biodiversity and endemism (Ehrlich 1994). Protected areas (PAs), although a
worldwide conservation success story (Bruner et al 2001), have become islands of
biodiversity over the years surrounded by a matrix of high human-use or modified
landscapes (DeFries et al 2005). Consequently, there has been a significant increase in
the proportion of secondary forests in the tropics in conjunction with large scale
deforestation for plantation crops. In such a scenario, it has become extremely crucial
to recognise and determine the capacity of small forest remnants such as sacred
groves and other community-managed forests outside the PA network to support
biodiversity in order to achieve conservation goals (Brown and Lugo 1990).
Invertebrates in general have been recognised as appropriate indicators of ecosystem
integrity (Dufrene and Legendre 1997). As bioindicators they have certain advantages
because of being ubiquitous in a wide range of environments, and moderate in growth
rate and population turnover (Hodkinson and Jackson 2005). As one of the more
better known insect groups taxonomically, butterflies have been used as indicator taxa
for biodiversity assessment studies (Stork et al 2003) as well as towards monitoring
ecosystem responses to environmental perturbations (Howard et al 1998, Parmesan et
al 1999) all over the world.
2
In this scenario, I aimed to compare butterfly community structure in a multi-use
heterogenous landscape with a large protected area around smaller reserved forests
interspersed with monoculture plantation agroforestry landscapes, community-
managed forests and shifting cultivation in the South Garo hills district of Meghalaya
in north-eastern India. I investigated the similarity and dissimilarity between the
butterfly communities of community-managed forests and protected areas, in terms of
species richness and abundance. I also investigated the kind of butterfly community
that inhabited the monoculture agroforestry habitats and completely deforested jhum
habitats of recent cultivation (<5 years) and compared it with the butterfly community
found in PAs in order to understand the impact of such land-use conversion on
butterfly species composition, and by extension on biotic communities in general.
I sampled at various places in the study landscape, the Balpakram-Baghmara
landscape, under a stratified random sampling design wherein every place across
habitat types (evergreen, deciduous, monoculture, shrub) and land-management
system categories (PAs, community-forests, plantations, jhum) had an almost equal
probability of getting picked as a sample unit. I sampled across the landscape around
various places while trying to cover the inherent vegetational and environmental
heterogeneity that may have existed within the broad habitat and management
categories. I enumerated butterflies in 298 time-constrained 30-minute counts during
the summer months of March to May, 2014 for an effort of 43 days and 149 hours.
Enumeration was done during periods of peak butterfly activity (0845 to 1515 hours),
and each count or transect was assigned to an appropriate land-management and
habitat category. In order to decipher the relationship of certain microclimatic patch-
level variables on butterfly species diversity and abundance, I measured temperature
and relative humidity at the start and end of all transects alongwith light intensity
3
which was also measured at an interval of every five minutes along the 30-minute
count. I analysed the collected data for common metrics of describing any biological
community such as species richness, abundance, evenness and proportion of
generalists and specialists species for all the land-management systems investigated,
and compared their values with appropriate statistical tests for significance.
I provide an overview of rapidly changing land-use pattern and high deforestation
rates in the tropics, the insufficiency of PAs in conserving biological diversity and the
need for studying the communities of diverse taxa in order to assess the impact of
habitat modification, and the value of butterflies as indicator taxa in order to execute
rapid biodiversity assessments of degraded landscapes. I describe my field methods
along with the statistical methods for analyses employed. Later, I provide the results
of my analyses with text, figures and graphs, and finally discuss the gaps in current
research with respect to the utility of agro-ecosystems in achieving conservation-
oriented goals. I also discuss the shortcomings of the study as well as ways in which
further analyses could be done to reveal potentially more interesting patterns.
I present the results of my study in the form of a manuscript to be submitted to the
peer-reviewed journal Conservation Biology.
4
Assessing patterns of butterfly communities in a multi-use landscape and its
implications for conservation
Rohit Jha1,2
1Post-Graduate Programme in Wildlife Biology and Conservation, National Centre for
Biological Sciences, Bengaluru 560065, India
2Wildlife Conservation Society – India Program, Centre for Wildlife Studies,
Bengaluru 560070, India
Running Title: Patterns of butterfly communities in a multi-use landscape
Corresponding author
Rohit R.S. Jha
A/304, Gaurav Bhakti C.H.S.L.,
Beverly Park, Mira road east,
Thane 401107, Maharashtra, India.
Email: rohite_001@yahoo.co.in
5
Abstract
General spatial patterns of species richness of biotic communities can be useful in
determining conservation policy. Increasing intensity of human-use and habitat
modification around protected areas has resulted in them becoming islands of
biodiversity at most places with a matrix of secondary habitats, monoculture
plantations, settled agriculture and habitation surrounding them. Thus opportunities to
keep large tracts of native habitats intact are dwindling rapidly. In such a scenario
there is a need to study the biotic communities of remnant forest patches, secondary
forests and agro-forestry landscapes in order to assess their conservation value. I
surveyed a heterogeneous, multi-use and dynamic forested landscape in Garo hills,
northeast India to investigate the differences in butterfly species assemblages of
community-managed forests, plantations and shifting cultivation areas with that of the
neighbouring protected areas. I found that the butterfly community of community-
forests were similar and comparable with those of protected areas in terms of
expected species richness, diversity indices, generalists/specialists ratio and species-
rank abundance curves, but significantly depauperate in monoculture and shifting
cultivation areas. This indicates that conservation effort should be directed to such
areas as well where biotic communities have done well and establishment of protected
areas is not an option. Various novel means such as payments for ecosystem services
(PES) and community-based sustainable ecotourism could be encouraged in order to
ensure long-term benefits from such forests to the local people as well as to
biodiversity, thus meeting conservation goals. This study also highlights the negative
impact that changing land-use pattern in favour of monoculture plantations could have
on butterfly communities in particular and biodiversity in general.
Keywords: Protected Areas, community-forests, monoculture, agro-forestry, jhum,
species diversity, species richness, Garo hills, Meghalaya
6
Introduction
One of the fundamental goals in ecology is to explain spatial and temporal patterns of
biological diversity (Cleary and Genner 2006). The recognition of such patterns in
biological communities and seeking to explain their underlying mechanisms lie at the
heart of community ecology. Spatial pattern, in fact, is a theme that has been
acquiring interest in both population and community ecology (Spencer et al 2002) as
population-level effects ultimately exert an influence on the development of landscape
patterns in multi-species communities (Drake et al. 1993). One of the areas of current
concern has been the overall poor understanding of the effects of habitat degradation
and change in land-use patterns on natural communities (Laurance 1994). In fact the
rapid conversion of natural habitats into agricultural and industrial landscapes, and
ultimately into degraded land, poses the greatest conceivable threat to biodiversity
(Dobson et al 1997). This is especially true in the tropics where large scale
deforestation and habitat modification has been witnessed, especially in the last few
decades (Laurance 1999, Sodhi et al 2004). The extent of monoculture plantations and
agroforests meanwhile has increased concurrently (FAO 2010). The proportion of
secondary forests as a result have increased in the tropics, necessitating an assessment
of their conservation values through studies of the inhabiting biotic communities
(Brown and Lugo 1990). Moreover, degradation of tropical forests has been shown to
not only affect community structure but also behaviour and functional morphology of
species (Bonte and Van Dyck 2009), thus clearly depicting its far-reaching effects.
The establishment of a protected areas' (PA) network with state control along with
implementation of pro-wildlife laws in a prohibitory and regulatory framework in PAs
have provided a refugia of sorts for biodiversity and has maintained source
7
populations for several species (Mansourian and Dudley 2008). In doing so, the PA
network has demonstrated success in conserving wildlife habitats worldwide (Bruner
et al 2001, Geldmann et al 2013) and have thus been central to conservation
strategies (Bruner et al. 2001, Ervin 2003) so far. However, their increasing isolation
in tropical forested landscapes (DeFries et al 2005) coupled with the fact that many of
them are limited in size (Woodroffe et al. 2005) has meant that PAs have become
islands of biodiversity over the years within increasingly intense human-use or
modified landscapes (Rindfuss et al 2004). In many cases, the coverage of this
network itself is inadequate in terms of representing a meaningful proportion of all
biomes that may exist in a country (Brooks et al 2004). On top of that, not all parks
are equally effective in achieving their objectives as managerial and interventionist
activities such as enforcement of rules, boundary demarcation, and direct
compensation to local communities differ from place to place (Bruner et al. 2001)
influenced by local socio-eco-political conditions alongwith availability of adequate
funds (Bruner et al. 2004). This has given rise to the sentiment that while the PA-
centric approach towards conservation of biological diversity and sustenance of
biological populations cannot be substituted and must remain as the central theme,
there is an urgent need to investigate the efficacy of areas outside this PA network in
meeting these specified goals. Such areas include forest tracts protected due to
traditional belief systems (like sacred groves) or those set aside for sustainable use of
forest-derived resources (like community-managed forests, hereinafter CFs), as well
as modern agroforestry, the so called 'wildlife-friendly' landscapes, that could
potentially encourage the sustenance and persistence of biological communities over
large areas.
8
Another prominent and fairly common feature of hill forests throughout the tropics is
shifting cultivation, known in India as jhum cultivation. As an agricultural practice it
has had a long history (Cairns and Garrity 1999) and is quite a common practice of
subsistence cultivation for many rural hill tribe populations in the tropical belt
throughout the world (Craswell et al 1997). In the jhum practice, a plot of forestland is
identified, standing vegetation is brought down (but not necessarily uprooted), it is
left to dry for a few days, biomass is then burnt and later seeds of various types of
grain and a variety of vegetables like gourds, chillies etc. are sown. It is a completely
rain-fed system and thus is independent of irrigation needs. After the yield is
harvested, the plot is abandoned in search of a new plot for the cycle of slash and burn
to repeat. While much has been written about the deleterious effects of jhum
cultivation on biodiversity and soil fertility right from the FAO in 1957, and even
actively discouraged by various laws and policies of the Indian state (Indian Forest
Policy 1894, 1952 and 1988), it has only recently garnered recognition as a better
economic (Ramakrishnan 1992), ecologically sustainable (Toky and Ramakrishnan
1983, Raman et al 1998) and environmentaly-friendly practice of agriculture (Fox
2000, Mertz 2002, Ickowitz 2006), only if the fallow periods were long enough to
allow biodiversity to recuperate and repopulate such areas (Grogan et al 2012).
Butterflies (class Insecta, order Lepidoptera) are a hyper-speciose group of organisms
with an estimated 20,000 species in the world (Vane-Wright 1978). They are one of
the better known insect groups taxonomically with their biology well-defined (Griffis
et al 1999). Butterflies have thus occupied a prominent place in assessing
conservation progress and biodiversity assessments (Ulrich and Buszko 2003, Stork et
al 2003). Butterflies have also been identified as an important indicator taxa towards
monitoring ecosystem responses to environmental perturbations (Howard et al 1998,
9
Parmesan et al 1999, Cleary 2004), including climate change (Roy and Sparks 2000,
Stefanescu et al 2003), as well as serving as a barometer of the overall community
complexity in an ecosystem (Hutzinger 2003). They are also extremely sensitive to
changes in certain microclimatic conditions usually affected by habitat degradation
such as humidity, temperature and light-levels (Murphy et al. 1990). Although many
studies have tried to investigate the cause of spatial variation seen in butterflies at
various sites across the world, relatively scant attention has been given to
understanding the role of land-ownership, past history of landscape change and
current land-use regime in shaping butterfly communities in a heterogenous landscape
mosaic.
In such a scenario I attempted to ask if the community-managed forests (CFs) in Garo
hills outside the ambit of the PA network and codified laws were able to support
comparable butterfly diversity vis-a-vis PAs or not. I also tried to investigate the
difference in butterfly community structure seen in agroforestry landscapes (rubber,
areca palm, jackfruit, bay leaf etc. plantations) that had replaced forests at some point
in time, thereby trying to predict the outcomes of such land-use changes on biological
communities in the future. Furthermore, I also looked at the butterfly community
composition of recent jhum areas (recently cleared or <5 years in fallow) as an
extreme point of comparison with PAs, as it involves almost complete deforestation
of patches, and with agroforestry-based monoculture plantation areas. Finally, in
order to understand mechanisms that might influence diversity acting as structuring
agents of assemblages, I also investigated the associations of some important
microclimatic variables with certain measures of diversity.
10
Materials and Methods
Study area
The Garo hills is an autonomous region in the Indian state of Meghalaya situated in
the north-eastern part of the country. Comprising of five administrative districts, the
region covers approximately a third of the state's total area of 22,429 km2. A
substantial part of the north-eastern region of the country, and all of Meghalaya, also
form a part of the Indo-Burma biodiversity hotspot, one of only 35 such recognised
regions of high biodiversity around the world (Conservation International 2010). The
topography of Garo hills is characterised by an undulating terrain of short hills (20-
1,050m above msl) covered largely by moist tropical forests. The region is
characterised by high rainfall during the south-west monsoon period (>2,000mm
average, June-October) and extreme climate during other months (10-36 °C).
Figure1: Location of study landscape within Meghalaya state in north-eastern India
11
The South Garo hills district of Meghalaya (Figure 1) covers an area of 1,849 km2, of
which 1,648km2, or approximately 89% is forested (Forest Survey of India 2014). A
third of this is highly disturbed and the rest is moderately disturbed mainly owing to
the widespread practice of shifting or jhum cultivation practiced by indigenous
peoples in these areas along with other biotic pressures (Forest Survey of India 2014).
The district also harbours the largest protected area of the state in Balpakram National
Park (BNP) with an extent of 220 km2 managed by the Meghalaya Forest Department.
Surrounding BNP is a mosaic of other smaller state-managed patches of forests,
namely the Siju Bird Sanctuary (6km2), Rewak Reserved Forest (~4km2) and
Baghmara Reserved Forest (~44km2); large tracts of local community-managed
forests (CFs) in various akhings or clan lands, monoculture plantations, areas under
shifting cultivation and human habitation. My study was done mainly within this
larger heterogeneous, multi-use landscape in and around the BNP, referred to
hereinafter as the Balpakram-Baghmara landscape (BBL) (Figure 2).
Figure 2: Map of study landscape depicting government-owned forests and community
lands along with sampling locations
12
The BBL occupies an extent of approximately 605km2 of which almost one-half is
presently evergreen or secondary semi-evergreen forest situated in the four state-
managed forest preserves (approximately 276km2) and under 36 tribal community
lands (approximately 60km2) called akhings (Goswami et al. 2014). Majority of this
forest occur at low elevations (altitudinal range 150-875m) and may be classified
under the Cachar tropical evergreen forest category, originally dominated by
Palaquium spp., Diospyros topiosa, Dipterocarpus turbinus, Messua ferrea and other
large evergreen trees (Champion and Seth 1968).
Figure 3: Land cover map of the study landscape generated through unsupervised
classification based on calculated NDVI values from a LANDSAT 8 scene dated
March30,2014. The red triangles denote sampling camps for the present study.
13
The deciduous elements include trees such as Betula alnoides, Cedrela toona,
Engelhardtia spicata and Ficus roxburghii (Tynsong et al. 2013).
Apart from forests, the akhings also comprise of considerable tracts of monoculture
plantation areas, fallow jhum areas of various ages, areas under paddy cultivation near
large river banks, degraded secondary forests, scrubland and habitation (Figure 3). It
is worth noting that almost all the remnant evergreen or semi-evergreen forest patches
now have only few very large evergreen stands with variable proportion of deciduous
trees and bamboo depending upon the extent of human disturbance, such as the
intensity and extent of jhum cultivation along with other forest resource extractive
practices. Despite this, the remaining forest patches still harbour a rich assemblage of
tropical floral and faunal elements (Kunte et al. 2012), including a large number of
threatened organisms such as various plants of medicinal importance (Barik 2008),
many threatened bird species such as the Grey-sided Thrush Turdus feae and Pale-
capped Pigeon Columba punicea, a diverse assemblage of small carnivores
(Wanniang 2007), and charismatic large mammals such as the tiger Panthera tigris
and Asian elephant Elephas maximus. A number of streams, rivers and rivulets criss-
cross the hilly landscape, the catchment for a majority of which are located in the
Balpakram National Park.
Field methods
I sampled for butterfly abundance and diversity at various locations in BBL covering
its spatial extent, land management categories and all broad habitat types. Sampling
effort was distributed in these habitat and land-ownership types according to their
respective extents in the landscape (Figure 3). Broad habitat types sampled included
evergreen forests, deciduous forests, monoculture plantations, scrubland, grassland
14
and edge habitats. I categorised all counts done in areas under active jhum cultivation
as well as in fallow land with growth of short vegetation (usually 2 to 5 years fallow)
in the 'shrub' category, while counts done in hilltop evergreen forest patches or in
lowland patches of riverine forests along streams and rivers have been categorised
into the 'evergreen' habitat category. Counts done in areas under cashew (Anacardium
occidentale), rubber (Hevea brasiliensis), areca palm (Areca catechu), jackfruit
(Artocarpus heterophyllus) and bay leaf (Cinnamomom tamala) crops have been
categorized under the 'monoculture' habitat category, while counts done in 'edge'
category include those walked in transition or ecotone areas where categorisation
under either of the clear habitat categories was difficult.
As far as land management system is concerned, I categorised all counts undertaken
in the four government-managed forest reserves (two reserved forests and protected
areas each) under 'protected area'. Counts walked in monoculture plantation areas
have been included in the category 'plantation', whereas those done in community-
managed forests including both the evergreen and deciduous (even secondary
deciduous forests of jhum fallow greater than 5 years) habitat types have been clubbed
together under 'community forests'. I categorised all counts done in active jhum
cultivation areas as well as in fallow lands of less than 5 years under the category
'jhum'. For analyses with land-management system as the predictor variable, counts
done along edges of any two systems were excluded from analyses. Similarly, for
analyses with habitat type as the predictor variable, counts during which two or more
habitat types were covered (ecotone areas) were excluded. A summary of counts done
in different land-management systems and habitat types is given in Table 1.
15
Table 1: Summary of 30-minute counts done in various habitats and land-management
regimes
Land-
management
system /
Habitat type
Evergreen
Deciduous
Monoculture
Shrub
Grassland
Mixed
Total
Protected
Areas
68
45
0
1
6
0
120
Community
forests
39
32
0
6
0
6
83
Plantation
0
0
40
0
0
0
40
Jhum (0 to
5 years)
0
0
0
21
0
0
21
Edge
6
6
0
2
0
20
34
Total
83
113
40
30
6
26
298
In the time-constrained 30-minute counts, I enumerated butterflies that included all
individuals of the six families, namely Hesperiidae, Lycaenidae, Nymphalidae,
Papilionidae, Pieridae and Riodinidae. The time-constrained count method is
independent of distance covered during any such count. I recorded a total of 9,665
butterfly individuals belonging to at least 264 distinct species. Of the total 9,665
individuals enumerated during all counts, 8,748 individuals (90.5%) were identified
till the species level while the others were either too fast flying to be reliably
identified or consisted of complex groups whose accurate identification was possible
only upon their dissection. Such individuals were excluded from all analyses.
Individuals that could not be accurately identified on the wing or when stationery
were either photographed or captured by a hand-net, examined and released on the
spot. Species in the 'dirty sailer' group (genus Neptis) have been considered as a single
entity as their accurate identification in field is difficult. Similarly, individuals of
species belonging to the genus Potanthus (Darts) too were impossible to determine in
field and were thus enumerated as a single entity. However since there are a much
16
higher number of species under Potanthus than in the 'dirty sailer' group, they have
been excluded from all analyses.
I attributed each count to an appropriate land-use and habitat category (Table 1).
Sampling was restricted to hours of weather favourable for butterfly flight (sunny and
less windy) between 0845 and 1515 hours, thus maximizing detection in all land-use
as well as land-cover types. I made a majority of all observations, walking in front at
all times, while a field worker would supplement with additional observations
walking at the back. The pace of walk during all counts was slow, on an average
approximately 0.75km/hour. During these butterfly counts, surrounding vegetation
along the path walked was disturbed in order to flush butterflies that may have been
taking cover to enhance detection overall and serving to minimize differences in
detection between land-use and land-cover types. I also recorded local weather
conditions such as temperature, relative humidity and light for all transects. While
temperature and relative humidity readings were noted only at the start and end of
transects with a Kestrel 3000 weather meter, light intensity (or luminosity) was
additionally recorded at every 5 minute interval as well during the 30-minute counts
using an HTC luxmeter. I also recorded locations and track of all counts with a hand-
held GPS unit (Garmin 72H) along with additional habitat and field observations. In
order to do a generalists' and specialists' species comparison across land-management
types, I classified all documented butterfly species into these two categories based on
their habitat preference. Specialist species included those that are known to live in
forest habitats (previous literature and Kunte pers.comm.) and are thus more
important for conservation. Classification of butterflies is based on the system as used
on the 'Butterflies of India' website (www.ifoundbutterflies.com, Kunte et al 2014).
17
Statistical analyses
I categorized community diversity in terms of a) number of species accumulated
against sampling effort, b) expected number of species at the point where species
accumulation curve would have reached its asymptote, c) species evenness, d)
Shannon-Wiener index of diversity, e) Simpson's index of diversity, f) species rank-
abundance curves, and g) species richness, abundance and proportion of specialist and
generalist habit butterflies. For the sake of simplicity, I have assumed no spatial
autocorrelation between samples. In his review of some of the popular methods of
species estimation, Baltanas (1992) had opined that although all species estimation
methods had their inherent biases, communities with similar attributes could be
compared with each other. In any biological survey of mobile species, there will
always be an unavoidable proportion of species that will be missed and go unnoticed.
Thus in order to estimate the total number of species, or true species richness, of any
area under investigation, it is important to take this factor into consideration and apply
suitable measure to incorporate this unknown component into analyses. While there
are several statistical approaches available to community ecologists, I used species
accumulation- and rarefaction-based curves here as they have proven reasonably
better than other estimators thus far (Gotelli and Colwell 2001).
Species rank-abundance curves are an excellent way of graphically representing the
dominance of a few species, or the lack of it, in a community of organisms. These
curves rank species in decreasing order of their relative abundance (Whittaker 1965).
Their shapes, and mathematically the slope value of the curve, enable rapid
visualisation of the extent to which species are distributed in a community.
18
In order to test for differences between means and variances of butterfly abundance
and species richness in the three other land-management categories in comparison
with their corresponding values in PAs, I used unpaired t-test as values for abundance
and species richness at the count level were normally distributed. For comparison of
species diversity indices across land-management types and a few other analyses, I
used the non-parametric equivalent of student's t-test, the Mann-Whitney U-test (or
the Wilcoxon test) as the data points were on a continuous scale but not normally
distributed. All response variables were tested for normality graphically as well as
with the Kolmogorov-Smirnov goodness-of-fit test.
In order to investigate the response of butterfly abundance and species richness to the
microclimatic variables of luminosity, temperature and relative humidity, I
constructed generalized additive models (GAM) (Hastie and Tibshirani 1986, Hastie
and Tibshirani 1990). The strength of GAMs is their ability to deal with highly non-
linear and non-monotonic relationships between the response and the set of
explanatory variables. In the case of abundance and species richness, it was fitted
using the poisson distribution, as these are count data. The response variables
abundance and species richness were both modelled as a function of site-level
microclimatic covariates using a log-link function in GAM. The plot of the centred
response along with standard error bands and p-values were used to assess and
interpret the GAM results.
For most of biodiversity-related analyses, I used the package 'BiodiversityR' (Kindt
and Coe 2005) within the open-source analytical software R (version 3.1) (R Core
Team 2013).
19
Results
In all, including counts done in edge and mixed category habitats, I enumerated
butterflies in 298 thirty-minute counts to document 8,748 butterflies of 264 distinct
species. With the help of this set of data, I constructed a species accumulation-effort
curve (Figure 4) to test if the landscape indeed was sampled well or not in terms of its
potential full suite of species.
Figure 4: Species accumulation-effort curve for the overall landscape with 95% confidence
intervals (n=298, bootstrap=100)
From my results, it appears that the stratified-random sampling scheme adopted for
this study was reasonably successful in documenting a high proportion of species that
may actually be present in the landscape. This is evident from the species
accumulation curve that appears to reach its asymptote. The chao, jackknife 1st order
and jackknife 2nd order indices of extrapolated species richness return figures of 295,
305 and 319 respectively, assuming a closed community, for the entire landscape. The
20
estimates also match published butterfly diversity from the study landscape (Kunte et
al 2012). I generated 95% confidence intervals around the curve by the bootstrap
method (bootstrap=100).
When the species accumulation curve was plotted with respect to each of the four
land-management systems under investigation, the curves still tended to flatten off
(Figure 5) indicating adequacy of sampling. Table 2 shows the comparison of
observed and extrapolated species richness in each of the four land-management
systems in the landscape.
Figure 5: Species accumulation curve as per land-management system with 95% C.I.
(bootstrap=100). The estimated species richness figure is an average of the chao, jackknife
1st order and jackknife 2nd order richness estimates for the respective management
categories
21
Table 2: Observed and extrapolated species richness as per land-management type
Land-management
system
n
Species
richness
Extrapolated species
richness
chao
jack1
jack2
Protected areas (PA)
120
219
260
264
283
Community forests (CF)
83
200
238
251
267
Plantations (PL)
40
123
155
162
177
Jhum (JH)
21
78
108
107
121
I compared mean butterfly abundance in PAs (30.8 ± 9.73) with the corresponding
values in CFs, PLs and jhum areas using the parametric unpaired t-test. I found
statistically significant differences between PAs with all the other three management
categories. While abundances in community-forests (mean 26.4 ± 8.96, t=3.22,
df=201, p=0.01) and jhum areas (mean 23.6 ± 10.53, t=3.06, df=139, p=0.01) were
significantly lower than in PAs, those in plantations were significantly higher (mean
35.1 ± 10.49, t=2.42, df=158, p=0.01) (Table 3).
Table 3: Summary of abundance and species richness across land-use types and test
statistics (unpaired t-test in comparison with PA)
Land-
management
n
Total
abundance
(individuals)
Mean
abundance /
count ± sd
p-
value
Species
richness
(r)
Mean r /
count ±
sd
p-
value
Protected areas
120
3691
30.8 ± 9.73
-
219
17.16 ±
5.39
-
Community
forests
83
2193
26.4 ± 8.96
0.01
200
16.27 ±
4.89
0.22
Plantations
40
1406
35.1 ± 10.49
0.01
123
16.35 ±
4.42
0.38
Jhum ( 0 to 5
years)
21
496
23.6 ± 10.53
0.02
78
13.09 ±
4.59
0.001
I also compared mean species richness in PAs (17.16 ± 5.39) with those in CFs, PLs
and jhum areas using the parametric unpaired t-test. I found only jhum areas to have
significantly lower species richness (mean 13.09 ± 4.59, t=3.26, df=139, p=0.001)
than PAs while those in both CFs (mean 16.27 ± 4.89, t=1.22, df=201, p=0.22) and
did not show any significant differences (Table 3).
22
Comparing mean Shannon-Wiener diversity index per count value for PAs (2.57 ±
0.44) with that obtained for CFs (mean 2.57 ± 0.41) using the non-parametric Mann-
Whitney U-test showed no significant difference (u=4798, z=0.43, p=0.65), while
jhum areas showed a significant lower difference (mean 2.32 ± 0.36, u=745, z=2.97,
p=0.002). Plantation areas, as like CFs, showed a non-significant result (mean 2.49 ±
0.35, u=1963, z=2.97, p=0.085) for this comparison, although only marginally so.
Similarly, when the mean Simpson diversity/evenness index per count obtained for
PAs (0.889 ± 0.079) was compared with the corresponding value obtained for CFs
(0.896 ± 0.071), there was no statistically significant difference (u=4840, z=-0.33,
p=0.72) observed between the two categories. However, while comparison of mean
Simpson diversity index per count for plantations (mean 0.885 ± 0.061) with respect
to PAs was non-significant (u=1921, z=1.88, p=0.058), even though marginally so,
that for jhum areas it was significantly lower (mean 0.873 ± 0.05, u=783, z=2.75,
p=0.005) (Table 4).
Table 4: Summary of Shannon-Wiener and Simpson diversity indices across land-use types and test
statistics (non-parametric Mann-Whitney U-test in comparison with corresponding Protected areas'
values)
Land-
management
system
Overall
Shannon-
Wiener index
Mean S-W
index per
count
p-value
Overall
Simpson
index
Mean Simpson
index per
count
p-value
Protected areas
4.36
2.57 ± 0.44
-
0.976
0.889 ± 0.079
-
Community
forests
4.39
2.57 ± 0.41
0.65
0.978
0.896 ± 0.071
0.72
Plantations
3.65
2.49 ± 0.35
0.085
0.949
0.885 ± 0.061
0.058
Jhum
3.62
2.32 ± 0.36
0.002
0.957
0.873 ± 0.05
0.005
Non-parametric Mann-Whitney U-test for significance was used in order to compare
the means of specialist species' richness percent per count value in PAs (15.36 ±
10.15) against the corresponding values in CFs, plantation and jhum areas. Except for
that in CFs (mean 16.31 ± 11.96, u=4926, z=-0.13, p=0.89), whose difference with
23
PAs was statistically non-significant, the corresponding values for plantations (mean
6.29 ± 7.38, u=1135, z=4.98, p<0.05) and jhum areas (mean 6.37 ± 5.70, u=582.5,
z=3.92, p<0.05) were found to be significantly lower than those in PAs (Table 5).
Non-parametric Mann-Whitney U-test for comparison of means of specialist species'
abundance percent per count between PAs (12.39 ± 9.57) and CFs was non-significant
(mean 13.01 ± 10.58, u=4858, z=-0.29, p=0.76). However, difference in mean
specialist species' abundance percent per count between PAs and plantations (mean
4.08 ± 5.77, u=1052, z=5.30, p<0.05), and between PAs and jhum areas (mean 4.40 ±
4.46, u=605.5, z=-3.78, p<0.05) were significantly lower (Table 6).
Table 5: Summary of comparison of specialist (in habit & occurrence) species richness across land-
management types with that of PA with test statistics (non-paramteric Mann-Whitney U-test)
Land-
management
system
Species richness
Total
species
Specialist
species
%
mean specialist
species / count
mean percent
specialist species
/ count *
p-
value
PA
219
60
27.4
2.73 ± 1.96
15.36 ± 10.15
-
CF
200
54
27.00
2.48 ± 1.64
16.31 ± 11.96
0.89
PL
123
24
19.51
1.15 ± 1.36
6.29 ± 7.38
0.00
JH
78
11
14.10
0.90 ± 0.88
6.37 ± 5.70
0.00
* If I were to encounter 100 species of butterflies in each count, I could expect 15.36%,
16.31%, 6.29% and 6.37% species to be specialists in PAs, CFs, plantation and jhum areas
respectively
Table 6: Summary of comparison of specialist (in habit & occurrence) species' abundances across
land-management types with that of PA with test statistics (non-paramteric Mann-Whitney U-test)
Land-
management
system
Abundance
Total
Specialists
%
mean specialists
per count ± sd
mean percent
specialists / count * ±
sd
p-
value
PA
3691
437
11.84
3.64 ± 2.73
12.39 ± 9.57
-
CF
2193
263
11.99
3.16 ± 2.33
13.01 ± 10.58
0.76
PL
1406
59
0.042
1.47 ± 1.99
4.08 ± 5.77
0.00
JH
496
22
0.044
1.04 ± 1.02
4.40 ± 4.46
0.00
* if I were to encounter 100 individuals of butterflies in each count, I could expect 12.39%,
13.02%, 4.08% and 4.40% of them to belong to the specialists category in PAs, CMFs,
plantation and jhum areas respectively
24
The land-management system-based rank abundance curves (Figure 6) showed that
while the percent abundance of the most abundant species found in PAs was only
~6%, the corresponding figure for jhum areas was almost two times higher at ~11%.
The most uneven community of all investigated in this study was that which was
recorded from monoculture plantation areas. Here as much as ~25% of total
abundance space was contributed by only two of the most abundant species in the
community. The long tail present in the rank abundance curves of PAs and CFs can
partly be explained by higher sample size and partly by the inherently higher number
of uncommon and rare species that they supported.
Figure 6: Rank-abundance curve of species across land-management systems
25
Table 7: Percent abundances of each of the five most abundant species in each land-
management type
Rank
Protected Areas
Community forests
Species
Percent
abundance
Species
Percent
abundance
1
Junonia iphita iphita
6.66
Eurema hecabe hecabe
5.88
2
Melanitis leda leda
6.52
Junonia lemonias lemonias
4.92
3
Neptis hylas varmona
5.31
Anthene emolus emolus
4.65
4
Ypthima baldus baldus
4.79
Junonia iphita iphita
4.56
5
Junonia lemonias lemonias
3.9
Melanitis leda leda
4.37
Total
27.18
24.38
Rank
Plantations
Jhum
Species
Percent
abundance
Species
Percent
abundance
1
Ypthima baldus baldus
14.15
Catopsilia pomona pomona
11.08
2
Junonia lemonias lemonias
10.31
Junonia lemonias lemonias
8.26
3
Neptis hylas varmona
6.33
Neptis hylas varmona
6.65
4
Eurema hecabe hecabe
6.25
Eurema hecabe hecabe
6.25
5
Junonia iphita iphita
5.54
Junonia hierta hierta
6.25
Total
42.58
38.49
Finally, generalized additive models constructed to investigate the potential
community structuring effects of microclimatic variables such as temperature, relative
humidity and luminosity showed statistically significant relationships of all the
variables with both, species richness as well as butterfly abundance.
Butterfly species richness and abundance showed no significant response to
increasing relative humidity until around 42% and 68% respectively, after which there
was an observed negative response (Figure 7). Butterfly abundance and species
richness had a non-linear response to temperature. There was a positive increasing
response until around 31.5 degrees Celsius for both abundance as well as species
richness, after which there was no or a very weak negative response to temperature
(Figure 8). A similar non-linear trend was obtained with luminosity as well where
butterfly species richness and abundance both showed a positive increasing response
26
with light intensity until around 4000 lux and 7000 lux respectively, after which any
relationship ceased to exist (Figure 9).
Figure 7: GAM models depicting effect of relative humidity on butterfly species richness
(left) and abundance (right) (both statistically significant)
Figure 8: GAM models depicting effect of temperature on butterfly species richness (left)
and abundance (right) (both statistically significant)
Figure 9: GAM models depicting effect of luminosity on butterfly species richness (left) and
abundance (right) (both statistically significant)
27
Discussion
My data shows that the butterfly diversity of the region in general is extremely rich.
With only 43 days and 149 hours of sampling effort, I documented a total of 8,748
butterflies belonging to 264 species, including many new records for the Garo hills
region. The overall species accumulation curve did tend to reach saturation, but the
land-management type-based and habitat-based accumulation curves did not do so
fully, indicating the probability of finding more species in the landscape, perhaps after
sampling in the more inaccessible reaches of evergreen forests in both protected areas
and community forests. My results demonstrate the sensitivity of butterflies to local
patch-scale microclimatic conditions such as humidity, light and temperature, in
accordance with similar previous studies (Kremen 1992, Hill et al 2001, Hamer et al
2003 and others), but with region specific interesting tipping points at which butterfly
activity seems to decline with increasing temperature. This finding implies that any
habitat modification activities such as logging that opens up the canopy might
possibly promote an increase of widespread species with a corresponding decline in
understorey species.
My results also indicate that patches of community-owned and managed forests in the
study landscape act as buffer to butterfly populations, especially for the specialist
species that are usually found only in dense forested habitats. Butterfly communities
found in both community-forests and protected areas are strikingly similar, in terms of
species richness, evenness and the proportion of specialist species that they support.
The suite of species that I encountered in protected areas though is more diverse than
in any other land-management category investigated. The expected species richness in
PAs too was estimated to be higher than in community forests. Occurrence of most
rare species (total abundance <4) was in PAs and community-forests indicating their
28
inherent capacity to support a larger suite of locally uncommon and rare butterfly
species. These results thus point toward the subsidiary, and not necessarily
substitutive, role of community forests to PAs in providing space for biotic
communities to persist. This finding with a class of organism that has relatively low
space and resource requirements corroborates with a similar study done with a study
system such as the Asiatic elephant Elephas maximus in the same landscape
(Goswami et al 2014) which has, rather contrastingly, large space and resource
requirements. This implies the potential applicability of studying community structure
of smaller but taxonomically more diverse systems in understanding effects of habitat
disturbance, building upon similar studies done elsewhere (Vasconcelos et al 2006,
Yamaura et al 2008, Bossart and Opuni-Frimpong 2009, Humpden and Nathan 2010,
and others). It also reiterates the useful role of butterflies as indicators of habitat
modification and their practicability in rapid ecosystem assessments (Daily and
Ehrlich 1995). However, few other studies have cautioned against the merits of this
approach and have rather opined that the 'apparent' effects of disturbance on such
'indicator' biotic communities might actually be as a result of their sensitivity to
particular metrics used, the taxa group studied, and the functional group investigated
(Hayes et al 2009). Thus, under this scenario of conflicted opinion, it must be
emphasised that habitat-assessment studies could strive to be integrative in approach
by factoring in the effects of disturbance or degradation on a wider variety of taxa in
order to unravel possibly interesting patterns and decipher a clearer picture (Basset et
al 1998).
Field-evaluations of butterfly sampling techniques indicate that light-gap size greatly
affects sampling results, especially in tropical forests (Hill et al 2001), and should be
of 'critical' concern in site-selection and sampling design (Sparrow et al 1994). I
29
assumed such local-level habitat factors, along with forest phenology and other
vegetation to be an inherent component of any particular land-management category.
Moreover, I was more interested in uncovering patterns that operated at a larger scale
than the patch. By sampling from various sites in the landscape with differing
vegetation and habitat attributes encompassing the inherent general variation within
them, I tried to incorporate the heterogeneity within each land-management and
habitat type in order to draw conclusions for the landscape.
Although Pollard walks (Figure 10), a belt-transect count based method where the
observer records all butterflies seen within a fixed distance from the transect path
during favourable weather conditions (Pollard 1977), is a more popular and well-
established method for monitoring abundances of butterflies, it may not be very useful
in tropical landscapes. The method was developed in the temperate region and has
been widely adopted in butterfly monitoring schemes around the world (Pellet et al
2012). However, since habitats in temperate regions are homogenous for vast
stretches, unlike that in the tropics, and butterfly communities too are comparably
much less diverse, Pollard walks may not be suitable for tropical forested landscapes.
Instead, time-constrained methods could provide an ideal alternative to enumerate
butterflies enabling the recording of species that would be missed while walking
along belt transects in a straight line. This is important as the time-constrained
method, being independent of distance covered, enables the documentation of a more
complete suite of species for any particular treatment under investigation as the
observer is not bound to include butterflies only within a certain distance from the
path. This also enables spending time at mud-puddling spots where several species
could be spotted due to a high turnover of butterflies (Kunte et al 2012). In areas
where abundances are high,
30
Figure 10: Several half-hour counts(separated by dotted blue lines) done along a forested
valley in areas of high and low butterfly abundances (left), and a typical belt-transect
'Pollard walk' (right)
distance covered during any such time-constrained count will be less as time taken to
record all butterfly sightings with proper identification will take time. Nevertheless,
time spent either photographing or netting specimens is excluded from the time spent
enumerating butterflies. Although visibility is undoubtedly high in the more open
habitats of jhum when compared to closed canopy forests, the fact that a greater
proportion of butterflies found in forested tracts are comparatively larger in size, and
conversely a greater proportion of butterflies found in shrub habitats are smaller in
size, the actual range within which butterflies are seen and can be accurately
identified along transect routes of different land cover types is more or less
comparable (Kunte pers.comm.). Nevertheless, it is important to be aware of the fact
that there is an active ongoing debate over the utility of detection probabilities for
such abundance estimation exercises (Gross et al 2007, Pellet 2007, Pellet et al 2012).
The significantly less diverse butterfly communities found in monoculture plantations
and recent jhum areas indicate a negative effect of such land-use patterns from natural
31
cover (with its inherent heterogeneity) to managed landscapes (with their designed
homogeneity) on biotic communities. This result is in contrast with Ranganathan et al
(2008) where they documented a 90% subset of regional native forest's bird species in
areca palm cultivation areas. While there may be obvious differences in ways that bird
species respond to cultivated landscapes than butterflies, it is worth noting that the
underlying causes of such apparently conflicting results might be the way in which
these agro-ecosystems are managed, and the proximity and matrix of such cultivation
areas with natural forests.
Assessing the combined effect of climate change and habitat degradation on 46
species of butterflies in Britain, Warren et al (2001) concluded that, in future, the dual
forces of habitat modification and climate change will likely cause specialists to
decline, leaving biological communities increasingly depauperate and dominated by
widespread, abundant and mobile habitat generalists. The results of my study concur
with their findings as the proportion of specialist species that I encountered in the two
highly human-modified landscapes of monoculture plantations and jhum cultivation
was significantly lower than those in natural forests (including both PAs and CFs). It
is therefore clear that more research is needed in such agro-ecosystems in a variety of
landscapes with differing management regimes before any certain crop or an agro-
forestry landscape is deemed important, relevant or even a priority for conservation.
Although the butterfly communities in plantations and jhum areas were significantly
more uneven than those in PAs and CFs, there were still quite a few species here that
were sighted on one or two occasions only, as a result of which the species rank
abundance curves show a long tail. I suspect this to be as a result of a number of
samples for plantations and jhum areas being located closer to PAs and CFs with
32
much diverse butterfly communities. It could also be an effect seen as a result of
relatively low sample sizes for these categories when compared to PAs and CFs.
However it does give an insight into the role played by habitat matrix in structuring
patterns of diversity observed over a landscape, or even in sustaining populations of
certain species that are able to survive in moderately hospitable environments as a
result of 'old' edges (in this case older jhum fallow and less intensive agro-forestry),
and short distance of forest fragments (in this case patches of CFs spread across the
landscape) from continuous patches (Debinski 2006), as was found in a study of
Amazonian ant communities in fragmented and continuous forest patches by
Vasconcelos et al (2006). Perhaps this also points to the sustainability of jhum
practice in general, if fallow lands are given considerable time to recover their biotic
elements as they were in the past. This is because for all my analyses I have included
secondary mature deciduous forests outside PAs at an advanced stage of recovery (10-
20 years) in the 'community-forests' category, and yet found no statistically significant
difference between butterfly communities of CFs as a whole with that of PAs.
However to further substantiate this, I recommend an appropriate sampling scheme
tailored to investigate this with larger samples across the full breadth of jhum fallows
at various stages of vegetation regeneration.
Since butterflies have also been shown to exhibit dependence and sensitivity to
alterations in landscape composition (Dennis et al 2003, Dover and Settele 2009,
Bergerot et al 2010, Rossi and van Halder 2010), it would be a worthwhile exercise to
analyse the data from a landscape perspective asking questions like the effect of
distance from PAs on butterfly communities, or the influence of human habitation (as
a proxy of human disturbance to habitat) on them. Similarly, analysing data in order
to examine the within taxa patterns of occurence can be informative as families and
33
subfamilies of butterflies tend to have different habitat requirements and respond
differently to habitat perturbations. Perhaps an analysis by species biomass and
butterfly functional guilds, as these reflect the true partitioning of resources within the
community (Basset et al 1998) rather than just abundance or presence-absence data,
too might reveal important patterns that otherwise get hidden. Another useful way of
analysing my data could be a community-level multivariate and ordination analysis
which might be able to distinguish unique community assemblages within land-
management or habitat types.
Tracts of community forests are important to the inhabiting local Garo tribal
community as most of their food and shelter needs are still met by resources that these
forests provide, thereby providing economic benefits as well. Traditionally, clan or
akhing leaders had set aside considerable tracts of forest, especially in the watershed
and dense vegetation areas as village preserves. Usage of forest resources from such
preserves was restricted with common consent. However, with the large-scale
promotion of monoculture plantations coupled with a fast growing human population
(a decadal growth of 27% during 2001-2011, Census of India 2011) has meant that
these village preserves have either disappeared or shrunk considerably as demand for
jhum plots too have increased. This has also meant a shortening of fallow cycles from
10-20 years on an average to about 4-6 years now. Only a few akhings in today's
scenario have still managed to hold on to the same extent of forests as they did till a
few decades back.
All this further accentuates the need to conserve the remaining tracts of forests in
order that both biodiversity and local people could reap sustainable benefits in the
long run. This could be achieved by means of a participatory approach or by adoption
34
of novel economic tools such as payments for ecosystem services (PES) along with
encouraging community-based sustainable ecotourism activities. As far as jhum areas
are concerned, they cannot be wished away in this landscape as they form a major
component of the local people's sustenance. Rather, efforts could be channelled
towards finding ways of lengthening fallow periods as this is not only a biodiversity
concern but also a genuine concern with regard to soil erosion and agricultural yields.
However against the backdrop of the general poor economic condition of people of
this region and in general slow acceptance of ideas such as ecotourism along with
extensive planning and commitment that it requires to achieve any amount of success,
as compared with quick, efficient, subsidy-based and blanket government policies
favouring adoption of monoculture crops, the success and long-term viability of such
interventionist and participatory approaches remain to be seen.
35
References
Baltanas, A. “On the use of some methods for the estimation of species richness.”
Nordic Society Oikos, 1992: 484-492.
Barik, S.K. “Environmental Issues and Management of Natural Resources:
Community Participation and Government Intervention in Meghalaya.” In Meghalaya
Human Development Report, by Government of Meghalaya Planning Department,
215-244. Shillong: Government of Meghalaya, 2008.
Basset, Y., Novotny, V., Miller, S.E. and Springate, N.D. “Assessing the impacts of
forest disturbance on tropical invertebrates: some comments.” Journal of Applied
Ecology, 1998: 461-466.
Bergerot, B., Fontaine, B., Julliard, R. and Baguette, M. “Landscape variables impact
the structure and composition of butterfly assemblages along an urban gradient.”
Landscape Ecology, 2010.
Bonte, D. and Van Dyck, H. “Mate-locating behaviour, habitat-use, and flight
morphology relative to rainforest disturbance in an Afrotropical butterfly.” Biological
Journal of the Linnean Society, 2009: 830-839.
Bossart, J.L. and Opuni-Frimpong, E. “Distance from Edge Determines Fruit-Feeding
Butterfly Community Diversity in Afrotropical Forest Fragments.” Environmental
Entomology, 2009: 43-52.
Brooks, T.M., Bakarr, M.I., Boucher, T., da Fonseca, G.A.B., Hilton-Taylor, C.,
Hoekstra, J.M., Moritz, T., Olivieri, S., Parrish, J., Pressey, R.L., Rodrigues, A.S.L.,
Sechrest, W., Stattersfield, A., Strahm, W. and Stuart, S.N. “Coverage Provided by
the Global Protected-Area System: Is It Enough?” BioScience, 2004: 1081-1091.
Brown, S. and Lugo, A.E. “Tropical Secondary Forests.” Journal of Tropical
Ecology, 1990: 1-32.
Bruner, A.G., Gullison, R.E. and Balmford, A. “Financial Costs and Shortfalls of
Managing and Expanding Protected-Area Systems in Developing Countries.”
BioScience, 2004: 1119-1126.
Bruner, A.G., Gullison, R.E., Rice, R.E. and da Fonseca, G.A.B. “Effectiveness of
Parks in Protecting Tropical Biodiversity.” Science, 2001: 125-128.
Cairns, M. and Garrity, D.P. “Improving shifting cultivation in Southeast Asia by
building on indigenous fallow management strategies.” Agroforestry Systems, 1999:
37-48.
Champion, H.G. and Seth, S.K. A Revised Survey of the Forest Types of India. Delhi:
Manager of Publications, 1968.
36
Conservation International. Biodiversity Hotspots. Washington DC, USA:
Conservation International, 2010.
Cleary, D.F.R. and Genner, M.J. “Diversity patterns of Bornean butterfly
assemblages.” Biodiversity and Conservation, 2006: 517-538.
Cleary, D.F.R. “Assessing the Use of Butterflies as Indicators of Logging in Borneo
at Three Taxonomic Levels.” Journal of Economic Entomology, 2004: 429-435.
Craswell, E.T., Sajjapongse, A., Howlett, D.J.B. and Dowling, A.J. “Agroforestry in
the management of sloping lands in Asia and the Pacific.” Agroforestry Systems,
1997: 121-137.
Daily, G.C. and Ehrlich, P.R. “Preservation of biodiversity in small rainforest patches:
rapid evaluations using butterfly trapping.” Biodiversity and Conservation, 1995: 35-
55.
Debinski, D.M. “Forest fragmentation and matrix effects: the matrix does matter.”
Journal of Biogeography, 2006: 1791-1792.
DeFries, R., Hansen, A., Newton, A.C. and Hansen, M.C. “Increasing Isolation of
Protected Areas in Tropical Forests over the past Twenty Years.” Ecological
Applications, 2005: 19-26.
Dennis, R.L.H., Shreeve, T.G., Van Dyck, H. “Towards a functional resource-based
concept for habitat: a butterfly biology viewpoint.” Oikos, 2003: 417-426.
Dobson, A.P., Bradshaw, A.D. and Baker, A.J.M. “Hopes for the Future: Restoration
Ecology and Conservation Biology.” Science, 1997: 515-522.
Dover, J. and Settele, J. “The influences of landscape structure on butterfly
distribution and movement: a review.” Journal of Insect Conservation, 2009: 3-27.
Drake, J.A., Flum, T.E., Witteman, G.J., Voskuil, T., Hoylman, A.M., Creson, C.,
Kenny, D.A., Huxel, G.R., Larue, C.S. and Duncan, J.R. “The Construction and
Assembly of an Ecological Landscape.” Journal of Animal Ecology, 1993: 117-130.
Dufrene, M. and Legendre, P. “Species Assemblages and Indicator Species: The Need
for a Flexible Asymmetrical Approach.” Ecological Monographs, 1997: 345-366.
Ehrlich, P.R. “Energy Use and Biodiversity Loss.” Philosophical Transactions:
Biological Sciences, 1994: 99-104.
Ervin, J. “Protected Area Assessments in Perspective.” BioScience, 2003: 819-822.
FAO. Global Forest Resources Assessment 2010. Rome, Italy: Food and Agriculture
Organization of the United Nations, 2010.
37
FAO. Shifting cultivation. Rome, Italy: Food and Agriculture Organisation, United
Nations, 1957.
Fox, J. “How blaming 'slash and burn' farmers is deforesting mainland Southeast
Asia.” Asia Pacific Issues, 2000: 1-8.
FSI. India State of Forest Report 2013. Dehradun: Forest Survey of India, 2014.
Geldmann, J., Barnes, M., Coad, L., Craigie, I.D., Hockings, M. and Burgess, N.D.
“Effectiveness of terrestrial protected areas in reducing habitat loss and population
declines.” Biological Conservation, 2013: 230-238.
Goswami, V.R., Sridhara, S., Medhi, K., Williams, A.C., Chellam, R., Nichols, J.D.
and Oli, M.K. “Community-managed forests and wildlife-friendly agriculture play a
subsidiary but not substitutive role to protected areas for the endangered Asian
elephant.” Biological Conservation, 2014: 74-81.
Gotelli, N.J. and Colwell, R.K. “Quantifying biodiversity: procedures and pitfalls in
the measurement and comparison of species richness.” Ecology Letters, 2001: 379-
391.
Griffis, K.L., Mann, S.S. and Wagner, M.R. “The Suitability of Butterflies as
Indicators of Ecosystem Condition: A Comparison of Butterfly Diversity Across
Stand Treatments in Northern Arizona.” USGS Conference. USGS, 1999. 125-135.
Grogan, P., Lalnunmawia, F. and Tripathi, S.K. “Shifting cultivation in steeply sloped
regions: a review of management options and research priorities for Mizoram state,
Northeast India.” Agroforestry Systems, 2012: 163-177.
Gross, K., Kalendra, E.J., Hudges, B.R. and Haddad, N.M. “Robustness and
uncertainty in estimates of butterfly abundance from transect counts.” Population
Ecology, 2007: 191-200.
Hamer, K.C., Hill, J.K., Benedick, S., Mustaffa, N., Sherratt, T.N.,Maryati, M. and
Chey, V.K. “Ecology of butterflies in natural and selectively logged forests of
northern Borneo: the importance of habitat heterogeneity.” Journal of Applied
Ecology, 2003: 150-162.
Hastie, T.J. and Tibshirani, R.J. “Generalized Additive Models.” Statistical Science,
1986: 297-310.
Hastie, T.J. and Tibshirani, R.J. “Generalized Additive Models.” Boca Raton, Florida:
Chapman & Hall/CRC, 1990.
Hayes, L., Mann, D.J., Monastyrskii, A.L. and Lewis, O.T. “Rapid assessments of
tropical dung beetle and butterfly assemblages: contrasting trends along a forest
disturbance gradient.” Insect Conservation and Diversity, 2009: 194-203.
38
Hill, J.K., Hamer K.C., Tangah, J. and Dawood, M. “Ecology of tropical butterflies in
rainforest gaps.” Oecologia, 2001: 294-302.
Hodkinson, I.D. and Jackson, J.K. “Terrestrial and Aquatic Invertebrates as
Bioindicators for Environmental Monitoring, with Particular Reference to Mountain
Ecosystems.” Environmental Management, 2005: 649-666.
Howard, P.C., Viskanic, P., Davenport, T.R.B., Kigenyi, F.W., Baltzer, M.,
Dickinson, C.J., Lwanga, J.S., Matthews, R.A. and Balmford, A. “Complementarity
and the use of indicator groups for reserve selection in Uganda.” Nature, 1998: 472-
475.
Humpden, N.N. and Nathan, G.N. “Effects of plant structure on butterfly diversity in
Mt. Marasabit Forest-northern Kenya.” African Journal of Ecology, 2010: 304-312.
Hutzinger, M. “Effects of fire management practices on butterfly diversity in the
forested western United States.” Biological Conservation, 2003: 1-12.
Ickowitz, A. “Shifting cultivation and deforestation in Tropical Africa: critical
reflections.” Development and Change, 2006: 599-626.
Kindt, R. and Coe, R. Tree diversity analysis. A manual and software for common
statistical methods for ecological and biodiversity studies. Nairobi, Kenya: World
Agroforestry Centre (ICRAF), 2005.
Kremen, C. “Assessing the indicator properties of species assemblages for natural
areas monitoring.” Ecological Applications, 1992: 203-217.
Kunte, K., S. Kalesh and U. Kodandaramaiah (eds.). Butterflies of India. v. 2.10. 08
July 2014. http://www.ifoundbutterflies.org/ (accessed July 10, 2014).
Kunte, K., S. Sondhi, B.M. Sangma, R. Lovalekar, K. Tokekar & G. Agavekar.
“Butterflies of the Garo Hills of Meghalaya, northeastern India: their diversity and
conservation.” Journal of Threatened Taxa, 2012: 2933-2992.
Landsat 8-OLI, scene dated 30 March 2014, path 137, row 42, courtesy of the U.S.
Geological Survey. The USGS home page is http://www.usgs.gov
Laurance, W.F. “Rainforest fragmentation and the structure of small mammal
communities in tropical Queensland.” Biological Conservation, 1994: 23-32.
Laurance, W.F. “Reflections on the tropical deforestation crisis.” Biological
Conservation, 1999: 109-117.
Mansourian, S. and Dudley, N. Public Funds to Protected Areas. Gland, Switzerland:
WWF International, 2008.
Mertz, O. “The relationship between fallow length and crop yields in shifting
cultivation: a rethinking.” Agroforestry Systems, 2002: 149-159.
39
Murphy, D.D., Freas, K.E. and Weiss, S.B. “An Environment‐metapopulation
Approach to Population Viability Analysis for a Threatened Invertebrate.”
Conservation Biology, 1990: 41-51.
Office of the Registrar General and Census Commissioner, India. Census of India
2011. New Delhi: Ministry of Home Affairs, Government of India, 2011.
Parmesan, C., Ryrholm, N., Stefanescu, C., Hill, J.K., Thomas, C.D., Descimon, H.,
Huntley, B., Kaila, L., Kullberg, J., Tammaru, T., Tennent, W.J., Thomas, J.A. and
Warren, M. “Poleward shifts in geographical ranges of butterfly species associated
with regional warming.” Nature, 1999: 579-583.
Pellet, J. “Seasonal variation in detectability of butterflies surveyed with Pollard
walks.” Journal of Insect Conservation, 2007.
Pellet, J., Bried, J.T., Parietti, D., Gander, A. and Heer, P.O. “Monitoring Butterfly
Abundance: Beyond Pollard Walks.” PLoS ONE, 2012.
Pollard, E. “A method for assessing changes in the abundance of butterflies.”
Biological Conservation, 1977: 115-134.
Ramakrishnan, P.S. “Shifting agriculture and sustainable development: an
interdisciplinary study from north-eastern India.” Parthenon Publishing Group, 1992.
Raman, T.R.S., Rawat, G.S. and Johnsingh, A.J.T. “Recovery of tropical rainforest
avifauna in relation to vegetation succession following shifting cultivation in
Mizoram, north-east India.” Journal of Applied Ecology, 1998: 214-231.
Ranganathan, J., Daniels, R.J.R., Chandran, M.D.S., Ehrlich, P.R. and Daily, G.C.
“Sustaining biodiversity in ancient tropical countryside.” Proceedings of the National
Academy of Sciences, 2008: 17852-17854.
R Core Team. A language and environment for statistical computing. Vienna, Austria:
R Foundation for Statistical Computing, 2013.
Rindfuss, R.R., Turner II, B.L., Entwisle, B. and Walsh, S.J. “Land Change Science:
Observing, Monitoring and Understanding Trajectories of Change on the Earth's
Surface.” In Land Change Science, by G., Janetos, A.C., Justice, C.O., Moran, E.F.,
Mustard, J.F., Rindfuss, R.R., Skole, D., Turner II, B.L., Cochrane, M.A. (Eds.)
Gutman, 351-366. Online: Springer, 2004.
Rossi, J.-P. and van Halder, I. “Towards indicators of butterfly biodiversity based on a
multiscale landscape description.” Ecological Indicators, 2010: 452-458.
Roy, D. B. and Sparks, T. H. “Phenology of British butterflies and climate change.”
Global Change Biology, 2000: 407-416.
40
Sodhi, N.S., Koh, L.P., Brook, B.W. and Ng, P.K.L. “Southeast Asian biodiversity: an
impending disaster.” Trends in Ecology and Evolution, 2004: 654-660.
Sparrow, H.R., Sisk, T.D., Ehrlich, P.R. and Murphy, D.D. “Techniques and
Guidelines for Monitoring Neotropical Butterflies.” Conservation Biology, 1994: 800-
809.
Spencer, M., Schwartz, S.S. and Blaustein, L. “Are there fine-scale spatial patterns in
community similarity among temporary freshwater pools?” Global Ecology &
Biogeography, 2002: 71-78.
Stefanescu, C., Peñuelas, J. and Filella, I. “Effects of climatic change on the
phenology of butterflies in the northwest Mediterranean Basin.” Global Change
Biology, 2003: 1494-1506.
Stork, N.E., Srivastava, D.S., Watt, A.D. and Larsen, T.B. “Butterfly diversity and
silvicultural practice in lowland rainforests of Cameroon.” Biodiversity and
Conservation, 2003: 387-410.
Toky, O.P. and Ramakrishnan, P.S. “Secondary succession following slash and burn
agriculture in north-eastern India: I. biomass, litterfall and productivity.” The Journal
of Ecology, 1983: 735-745.
Tynsong, H., Dkhar, M. and Tiwari, B.K. “Domestication, Conservation, and
Livelihoods: A Case Study of Piper peepuloides Roxb.—An Important Nontimber
Forest Product in South Meghalaya, Northeast India.” International Journal of
Biodiversity, 2013: 7pages.
Ulrich, W. and Buszko, J. “Species – area relationships of butterflies in Europe and
species richness forecasting.” Ecography, 2003: 365-373.
Vane-Wright, R.I. “Ecological and behavioural origins of biodiversity in butterflies.”
Symposium of Royal Entomological Society of London. London, UK, 1978. 56-70.
Vasconcelos, H.L., Vilhena, J.M.S., Magnusson, W.E. and Albernaz, A.L.K.M.
“Long-term effects of forest fragmentation on Amazonian ant communities.” Journal
of Biogeography, 2006: 1348-1356.
Wanniang, T. Management Plan of Balpakram National Park. Plan document,
Shillong: Meghalaya Forest Department, 2007.
Warren, M.S., Hill, J.K., Thomas, J.A., Asher, J., Fox, R., Huntley, B., Roy, D.B.,
Telfer, M.G., Jeffcoate, S., Harding, P., Jeffcoate, G., Willis, S.G., Greatorax-Davies,
J.N., Moss, D. and Thomas, C.D. “Rapid responses of British butterflies to opposing
forces of climate and habitat change.” Nature, 2001: 65-69.
41
Whittaker, R.H. “Dominance and Diversity in Land Plant Communities: Numerical
relations of species express the importance of competition in community function and
evolution.” Science, 1965: 250-260.
Woodroffe, R., Thirgood, S. and Rabinowitz, A. People and Wildlife: Conflict or
Coexistence? Cambridge, United Kingdom: Cambridge University Press, 2005.
Yamaura, Y., Kawahara, T., Iida, S. and Ozaki, K. “Relative Importance of the Area
and Shape of Patches to the Diversity of Multiple Taxa.” Conservation Biology, 2008:
1513-1522.
