Natural and human impact on the land use
Journal of Environmental Management 138 (2014) 15e23
Contents lists available at ScienceDirect
Journal of Environmental Management
journal homepage: www.elsevier.com/locate/jenvman
Natural and human impact on the land use and soil properties
of the Sikkim Himalayas piedmont in India
P. Prokop*, D. P1oskonka
Department of Geoenvironmental Research, Institute of Geography and Spatial Organization, Polish Academy of Sciences, Jana 22, 31-018 Krakow, Poland
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 4 December 2012
Received in revised form
6 December 2013
Accepted 22 January 2014
Available online 20 February 2014
Natural and human causes of change in land use and soil properties were studied in the Sikkim Himalayas piedmont over the last 150 years, with a special emphasis on the period 1930e2010. Analysis of
historical reports, combined with the visual interpretation of topographic maps and satellite images,
indicates that the land reforms related to the location of tea gardens caused rapid deforestation of the
higher elevated terraces in the late 19th century. Continuous population growth between 1930 and 2010
caused a shift in the major land use changes from the terraces to the floodplains. As a consequence, a
gradual extension of tea plantation and forestry development helped in stabilizing the land use of the
terraces, while the parallel deforestation of mountain catchments and floodplains for rice cultivation
intensified fluvial activity. The enlargement of river-channel area by about 42% between 1930 and 2010
excluded a large part of the floodplains from cultivation and increased risk of soil degradation. The
replacement of natural forest by monocultural tea and rice cultivation influenced the physical and
chemical properties of the soil. Statistically significant changes were observed only in some chemical
properties of the topsoil. Tea cultivation reduced the total carbon content by 26% and total nitrogen
content by 33% in the surface soil horizon. The influence of rice tillage on the soil properties is masked by
the fluvial activity. The combined effect of flooding and rice cultivation is reflected in the lower content of
total carbon and nitrogen in the surface of the soil, namely, 76% and 77% respectively. Taking into account
the long-term nature of the plantation, the soil still has the capability to support tea production. The
productivity of rice depends partly on fertilization levels and partly on the natural deposition of fresh
sediment eroded from mountains.
Ó 2014 Elsevier Ltd. All rights reserved.
Keywords:
Deforestation
Land degradation
Fluvial activity
Tea plantation
1. Introduction
The humaneenvironment interactions are characterized by
great diversity, with much feeding back, and many nonlinear processes, thresholds and time lags (Rudel et al., 2005; DeFries, 2008;
Lambin and Meyfroidt, 2010; An, 2012). Human impacts are most
frequently related to changing patterns of land use and are of
special importance in environmental studies (Dale et al., 1998;
Geist and Lambin, 2002; Carr et al., 2006). While landforms and
soils are subject to formation, change and even destruction by
natural forces over a geological time scale, land use and land cover
(LULC) changes resulting from human activity usually occur more
rapidly and have a strong impact on vegetation, water resources
and soil (Lambin et al., 2003; Vanacker et al., 2003; Geist and
Lambin, 2006; Davidar et al., 2010). On a smaller scale, there is a
wide variety of humaneenvironment relationships, with different
* Corresponding author. Tel./fax: þ48 124224085.
E-mail address: [email protected] (P. Prokop).
0301-4797/$ e see front matter Ó 2014 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jenvman.2014.01.034
patterns across regions and nations. The densely settled piedmont
zones of young mountains constitute an example of areas where
human activity is superimposed on changes induced by natural
forces (Tiwari, 2000; Liebault and Piegay, 2002; Starkel et al., 2008).
The tectonically active Sikkim Himalayas, with lithology which
is prone to mass movement, receive the highest annual rainfall
along the whole Himalayan front (Starkel, 1972; Dhar and Nandargi,
2000; Soja and Starkel, 2007). Their piedmont, as a transitional
zone down to the lowland plains, is under the strong influence of
the adjacent mountains. The nature and extent of the Himalayas
impact on their piedmont is largely a product of adjustments in
fluxes of water and sediment (Starkel and Basu, 2000; Grujic et al.,
2006; Ghosh and Carranza, 2010). Both are frequently accelerated
by various forms of human activity such as agriculture, logging,
mineral extraction or road building at the mountain margin
(Froehlich and Starkel, 1993; Tiwari, 2000). Increased fluvial sediment transport increases the flood risk, river bank erosion and
shifting of their braided courses, leading to direct changes in the
LULC system of the piedmont (Froehlich and Walling, 2007; Sarkar,
16
P. Prokop, D. Płoskonka / Journal of Environmental Management 138 (2014) 15e23
2008; Starkel et al., 2008). Therefore piedmont of the Sikkim
Himalayas creates specific environmental conditions either
favourable or restrictive for certain forms of human activity.
The present-day economy of the piedmont depends mainly on
tea estates established in the late 19th century (Ray, 2002). Cultivation of other crops (e.g. rice, millet, areca nut) is largely for local
consumption. A significant part of the piedmont area constitutes
reserved forest, national parks and wildlife sanctuaries. Forestry
based on commercially valuable trees and tourism is an important
contributor to the economy, as well as being an employer of large
numbers of people (Government of West Bengal, 2008e2009;
Madhusudan, 2011).
The presented paper attempts to determine the relative
importance of human activity as compared to natural processes in
the transformation of the environment of the Sikkim Himalayas’
piedmont over the last 150 years, with special emphasis on the
period 1930e2010. The natural and human impact is analyzed by
looking at land use change, and quantifying these changes by
comparing the soil properties under the mononocultural cultivation of tea and rice against the properties of the natural forest soil.
2. The study area
The study area is located in the piedmont of the Sikkim
Himalayas in the Jalpaiguri District of the West Bengal State in
India (Fig. 1). A hydrologic and geomorphic approach was
employed to delineate 176.6 km2 at the outlet of the Tista River
from the Himalayas. The area represents a system of Quaternary
fans decreasing in elevation from 200e300 m a.s.l. at the base of
the mountain to about 100 m a.s.l. over a distance of 15 km
(Nakata, 1972; Chakraborty and Ghosh, 2010). The fan surface,
fragmented by the wide braided channels of the Tista and its
tributaries, creates two major landforms, terrace and floodplain,
occupying an area of 93.3 km2 (53%) and 83.3 km2 (47%) respectively. The relative height of the terrace surface increases above
the present river bed, from 6e10 m at the front of the mountain to
25e30 m downstream.
The climate is subtropical monsoonal, with the warm rainy
season spanning from June to October and the dry cooler season
from November to May. The mean annual air temperature reaches
23 C, and fluctuates between 16 C in winter (January) and 28 C in
summer (August). The frontal zone of the mountains and piedmont
receives about 4000e6000 mm rainfall annually, more than 80% of
which falls during the summer monsoon. During this period, 10e15
days of rainfall, of up to 100 mm per day, is observed every year
(Soja and Starkel, 2007; Bookhagen, 2010).
The Tista river drainage area of 8640 km2 covers the whole of
the Sikkim Himalayas. In the piedmont, it receives water from the
Lish and Gish, draining 51 km2 and 157 km2 respectively of the
Himalayan margin. Every year, at the mountain’s edge, the Tista
discharge may reach 3000e5000 m3 s 1. The peak discharge in the
devastating flood of 1968 was calculated at 18,745 m3 s 1, about
45 km downstream of the mountain (Murray and Bochin, 1973).
The 1952 records show a maximum discharge of 255 m3 s 1 for the
Lish and 630 m3 s 1 for the Gish (Dutt, 1966). Several days of
continuous rainfall cause regional floods every 20e30 years, while
the high frequency of severe hourly rainfall causes flash floods
almost every year (Starkel and Basu, 2000; Sarkar, 2008). During
such events, the water level rises in the river channels up to 5e6 m
and a heavy sediment load is deposited on the floodplains (Dhar
and Nandargi, 2000). In the past three decades, the river beds
along the margins of the Himalayas have experienced aggradations
of about 6 cm year 1 in the case of the larger trans-Himalayan Tista,
and between 9e14 cm year 1 in the case of the smaller Lish and
Gish (Sarkar, 2008). Thick alluvial cover creates good conditions for
water infiltration and during the winter season the braided channels are totally dry.
Soils from relatively stable terraces and active floodplains
revealed two distinct stages of pedogenic development, A-(Bw)-C
and A-C, with the dominant soil classified as Inceptisols (Dystrudepts) and Entisols (Fluvaquents and Endoaquents) respectively
(Soil Survey Staff, 2010). Good drainage conditions of Inceptisols
result in rapid leaching and enhancement of the weathering process, giving rise to a humus horizon, thereby lowering the pH (4e5)
and development structural Bw (Cambic) horizon. The Entisols of
active floodplains, imperfectly drained, without any diagnostic
horizon except ochric epipedon, show little alteration of the original deposits and higher pH (5e7), because they receive new
Fig. 1. Location of the study area in part of the Sikkim Himalaya and piedmont, with a schematic longitudinal profile. White dashed lines indicate watershed divides, a black dotted
line indicates the Indo-Bhutan border. The outline of the research area presented on Fig. 2 is marked by an unbroken black line.
P. Prokop, D. Płoskonka / Journal of Environmental Management 138 (2014) 15e23
sediments faster than the assimilation of previous material into the
genetic horizons (NBSS and LUP, 1991).
The terrace area is currently covered with natural tropical moist
deciduous Sal (Shorea robusta) forest and tea plantations
(Champion and Seth, 1968). The floodplains are occupied by wide
braided river channels and rice cultivation. About 56% of the total
population (56,000 inhabitants) of the area lives in nine tea gardens, whilst the remainder of the population is settled along the
main road and railway in the Himalayan foothills (Government of
India (2011)).
3. Material and methods
3.1. Database development
Historical reports and maps from the colonial British administration were used as the sources of information concerning the
LULC and population trends of the Himalayan piedmont in the 19th
and first three decades of the 20th century. The changes in LULC
were derived from the Survey of India topographic maps at a scale
of 1:63,360 for the year 1930, panchromatic photo images acquired
by the American Corona satellite for the year 1970, and satellite
images downloaded from Google Earth Pro for the year 2010. Maps
and satellite images were supported by tea garden plans/cadastral
maps at scales of 1:3,960 for various years between 1930 and 2000.
These cadastral maps are renowned for being reliable, detailed and
useful for land use and soil studies (Ghosh et al., 2004; Sarkar et al.,
2006). Geometric corrections were performed to rectify all the
maps and images using the UTM-45N projection system. A visual
interpretation technique, combined with several field surveys
conducted between 2007e2011, was used for LULC mapping. The
applied method of interpretation is superior to automatic classification when the research output needs to be compared with map
data in a study concerned with detecting change (Petit and Lambin,
2002; Pelorosso et al., 2009). In addition, visual interpretation allows the delimiting of realistic objects, such as patches with
irregular shapes, which can represent important contrasts in
ecological structure or process in the landscape. Thus, visual
interpretation integrates environmental knowledge into image
analysis making the results more environmentally meaningful
(Jensen, 2007; Shao and Wu, 2008; Hersperger and Bürgi, 2009). In
order to standardize the values of maps from three different time
periods, seven consistent LULC categories were defined: forest,
grassland, tea plantation, rice cultivation, other crop, built-up area
and river channel. Logical rules in GIS were used for determination
where areas had the same land use class in each time period (stable), locations where there had been one point of change between
two land use classes on either sides of that area of change (stepped), areas where frequent change has occurred between two
categories (cycle); and where there was a high turnover between
many different classes (dynamic).
Soil data, based on the USDA soil classification system, were
digitized from a 1:250,000 scale soil map developed by the National Bureau of Soil Survey and Land Use Planning (NBSS and LUP,
1991).
3.2. Soil sampling and analysis
The soil sampling scheme was based on the assumption that soil
differentiation between two main landforms, terrace and floodplain, depends on long-term natural soil forming factors (NBSS and
LUP, 1991). Subsequently, within these two main landforms, the
sources of soil variability were the LULC changes. The sampling
sites were therefore located in areas of “stable” LULC (delimited on
a land use stability map prepared earlier), ensuring the same land
17
use (forest, tea and rice) during the period 1930e2010. Within
stable LULC areas of terrace, samples were taken close to the
mountain margin, in the middle and in the lower part of the alluvial
fan. Within stable LULC areas of floodplain, samples were taken
using the same scheme but also at different distances from the
main rivers, in order to take into account sediment heterogeneity.
Soils under built up areas, grassland and other crops were not
examined in detail because of their low contribution in the study
area and the high dynamic of these land use changes during the
investigated period.
Twenty soil profiles were analysed from the major landforms:
terrace with natural forest and tea plantations as well as floodplain
with natural forest and rice cultivation. The natural forests served as
a control against which changes in soil properties were compared.
Significant differences between the means of soil properties were
identified using the t-statistics, with a level of probability of 5%.
Physical (grain size composition, bulk density, compression
strength) and chemical (organic matter (OM), carbon (C), nitrogen
(N), sulphur (S), pH) soil properties were analysed. The rationale for
the selection of soil quality indicators was the detection of longterm cultivation impact on soil properties (Larson and Pierce,
1991; Panwar et al., 2011). Also important was the possibility of
comparison of the soil properties changed due to cultivation with
the soil properties under tea- and rice-based agricultural systems in
other regions of India and Southeast Asia. The grain size composition of each sample was determined using the combined sieving
method and a Fritsch laser particle sizer Analysette 22, after pretreatment with H2O and ultrasonic disaggregation. Soil bulk density was calculated by drying the soil ring samples (100 cm3) at
105 C before weighing them. Soil compression strength (attributable to soil compaction) was measured by using a base surface
cone penetrometer, and the readings of ten random repeated
penetrations were averaged. The concentration of soil organic
matter (OM) was determined using the wet digestion method
(Walkley and Black, 1934). Total C, N and S were determined for the
A soil horizon (topsoil) by combustion in a CHNS vario EL III
Element Analyzer. The pH values in H2O and KCl were measured
electrometrically in a 1:2.5 soil/water and a soil/KCl suspension,
respectively. The colour of the soil was determined according to the
Munsell system.
4. Results
4.1. Land use changes
4.1.1. General tendencies in LULC changes up to 1930
The LULC of the Sikkim Himalayas piedmont, before being ceded
by Bhutan to the British East India Company in 1864, was characterized by large tracts of forest intersected by numerous rivers with
floodplains overgrown by scattered trees and reed jungle (O’Donel
et al., 1864e68; Hunter, 1876, Fig. 2). Owing to the difficulty of procuring water in the dry winters, the area was only seasonally settled
and population density did not exceed 5e10 inhabitants km 2
(Government of Bengal, 1872). Environmental conditions in this
sparsely populated area led to the piedmont initially being categorized as wasteland by the Government of Bengal.
The perception of the region as being void of natural disasters,
and the successful introduction of tea cultivation in the adjacent
Sikkim Himalayas by 1860 resulted in the piedmont area being
leased for the foundation of numerous tea estates. Large-scale
deforestation was initiated in the outlet of the Tista River,
through the foundation of tea plantations between 1874e1885 and
the heavy demand for timber for railway construction (Allen et al.,
1906; Ghosh, 1970). From this date onwards, tea cultivation rapidly
expanded, and by 1930 the number of tea gardens had increased to
18
P. Prokop, D. Płoskonka / Journal of Environmental Management 138 (2014) 15e23
Fig. 2. Spatial distribution of LULC for 1930, 1970 and 2010 with map from 1868 (O’Donel et al., 1864e1868). Black lines indicate terrace-floodplain boundary.
142, covering an area of the Sikkim and adjacent Bhutan Himalayan
piedmont of approximately 483 km2 (The Indian Tea Association,
1930). The development of plantations induced a growing demand for a workforce and the migration of labourers towards the
extensive tracts of cultivable land east of the Tista (Allen et al.,
1906). As a result, the population density in the piedmont zone
had reached 60 inhabitants km 2 by 1931 (Government of Bengal,
1931). The migration of people also initiated the replacement of
floodplain forest by rice cultivation to support their growing food
needs. The formation of the LULC framework was finished when the
1865 Forest Act, provided the legal instrument for registering forest
as “reserved” by the Forest Department. The ‘reservation’ process
started from several tracts of natural forest with a total area of
460 km2, and continued during the period of independence
(Banarjee et al., 2010).
4.1.2. LULC changes between 1930 and 2010
The majority of the area (74.4% in the period 1930e1970 and
77.0% in the period 1970e2010) was not subject to any change
Table 1
Area proportion (%) of LULC classes for the entire study area and within geomorphic
units.
LULC
Entire area
Terrace
Floodplain
Year
1930
1970
2010
1930
1970
2010
1930
1970
2010
Forest
Grassland
Tea
Rice
Other crop
Built up
River channel
41.2
8.5
13.7
8.9
4.0
2.9
20.8
32.3
4.9
14.4
6.0
3.2
3.8
35.4
31.1
3.2
16.5
13.9
1.0
4.8
29.5
55.9
6.4
25.5
e
7.5
4.1
0.6
54.7
4.9
26.9
e
6.1
6.9
0.5
54.8
3.3
31.2
e
1.9
8.6
0.2
24.4
10.9
e
19.2
e
1.4
44.1
6.7
4.8
e
12.9
e
0.3
75.3
4.1
3.1
e
29.8
e
0.3
62.7
(Fig. 2, Tables 1 and 2). The main processes of LULC change between
1930 and 2010 were deforestation and the expansion of both
agriculture and river channels. During this period, the forest cover
decreased by 24.5%, and the rate of deforestation was significantly
higher in the period 1930e1970. About 85.2% of the forest loss
occurred at the expense of river channel widening or migration.
The rest of the deforestation was the effect of conversion to
grassland, followed by the extension of built up area and rice
cultivation. Changes in the area of rice cultivation occupying the
floodplains were also dependent on the widening or narrowing of
river channels in response to flooding events.
Extensive flooding by the Tista and its tributaries in 1954 and
1968 was an important factor in the widening of river channels by
about 70.2% during the period 1930e1970. The subsequent reduction in channel areas between 1970 and 2010 is the result of a recovery process in response to their earlier destabilization.
Nevertheless, river channels occupied an area some 41.8% larger in
2010 than in 1930. These calculations may to some extent have
Table 2
Predominant types of LULC changes (claiming over 1% of the total area).
Type of LULC change
1930e1970
1970e2010
No change
Forest to grassland
Forest to river channel
Grassland to rice
Grassland to river channel
Other crop to built up
Rice to river channel
River channel to forest
River channel to grassland
River channel to rice
74.4
1.1
7.3
1.1
4.0
77.0
3.2
1.6
1.3
5.0
1.3
2.0
1.4
8.3
P. Prokop, D. Płoskonka / Journal of Environmental Management 138 (2014) 15e23
been affected by the construction of embankments and retaining
walls which have restricted the extension of river channels in
recent decades. The first embankments were constructed only to
protect bridges crossing the Lish and Gish rivers and some parts of
the Tista. In 1978 the left bank of the Gish was protected between
the margin of the mountain and its outlet to the Tista. The length of
all embankments has recently been increased to 20 km, so that they
protect 30% of the river banks.
Besides changes in overall land use, a different pattern of LULC
changes may be observed between terrace and floodplain as
geomorphic units represent a proxy for soils and hydrology. Tea
plantations are located only on high terraces with well drained and
acidic soil (Table 2). Rice cultivation, on the other hand, requires
submersion conditions and is concentrated only on floodplains
with shallow, slightly acidic soils. The reserved forest within the
terrace shows stable cover of about 55% of the total terrace area,
between 1930 and 2010. In contrast, over the same time period,
about 19% of the total floodplain area was deforested. Grasslands
occupy both the terraces near built up areas and the floodplains
along rivers, and does not show a strong correlation with geomorphology and soils. Built up areas, on the other hand, are located
mainly on higher elevated terraces, together with other crops
planted in domestic gardens. In the 1930s, built up areas also
expanded into the distal part of the extensive Tista floodplain. The
1968 flood changed this situation, and now only temporary huts are
constructed within areas prone to inundation.
4.1.3. LULC stability
The study of LULC changes in the period 1930e2010 revealed
that stable land use dominates in the investigated region and
covers 60.3% of the total area (Fig. 3). The stable land use area
mainly consists of large blocks of reserved forest, tea gardens and
some settlement cores. About 1/3 of the study area has undergone
stepped changes, in the form of the conversion of forest or grassland to tea bushes at the margin of the tea gardens. Extensive
changes in this category are concentrated along river banks and are
related to the widening of braided river channels. Cyclical changes
Fig. 3. LULC stability between 1930 and 2010.
19
occur only in places within distal areas of the large river floodplains, such as Tista and Gish, where the margins of these wide
floodplains are occupied by rice fields. During extreme flooding, the
cultivated land is affected by debris deposition and thus excluded
from cultivation. Usually, after some time, the sediment is removed
or mixed with the soil, and the land is again taken under cultivation. The area affected by cycle changes covers 4.9% of the region.
Dynamic changes in land use occupy only 4.2% of the total area.
Land undergoing high turnover between many different land use
classes (dynamic changes) is scattered across the Tista river channels, where sand bars are frequently degraded and built up by
flowing water.
4.2. Impact of LULC change on soil properties
4.2.1. Effect of land-use systems on the physical and chemical
properties of soil at different depths
Soils within the terraces are deep and well drained (Table 3).
Their grain-size composition is characterized by the dominance of a
sandy fraction and low clay content (
Contents lists available at ScienceDirect
Journal of Environmental Management
journal homepage: www.elsevier.com/locate/jenvman
Natural and human impact on the land use and soil properties
of the Sikkim Himalayas piedmont in India
P. Prokop*, D. P1oskonka
Department of Geoenvironmental Research, Institute of Geography and Spatial Organization, Polish Academy of Sciences, Jana 22, 31-018 Krakow, Poland
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 4 December 2012
Received in revised form
6 December 2013
Accepted 22 January 2014
Available online 20 February 2014
Natural and human causes of change in land use and soil properties were studied in the Sikkim Himalayas piedmont over the last 150 years, with a special emphasis on the period 1930e2010. Analysis of
historical reports, combined with the visual interpretation of topographic maps and satellite images,
indicates that the land reforms related to the location of tea gardens caused rapid deforestation of the
higher elevated terraces in the late 19th century. Continuous population growth between 1930 and 2010
caused a shift in the major land use changes from the terraces to the floodplains. As a consequence, a
gradual extension of tea plantation and forestry development helped in stabilizing the land use of the
terraces, while the parallel deforestation of mountain catchments and floodplains for rice cultivation
intensified fluvial activity. The enlargement of river-channel area by about 42% between 1930 and 2010
excluded a large part of the floodplains from cultivation and increased risk of soil degradation. The
replacement of natural forest by monocultural tea and rice cultivation influenced the physical and
chemical properties of the soil. Statistically significant changes were observed only in some chemical
properties of the topsoil. Tea cultivation reduced the total carbon content by 26% and total nitrogen
content by 33% in the surface soil horizon. The influence of rice tillage on the soil properties is masked by
the fluvial activity. The combined effect of flooding and rice cultivation is reflected in the lower content of
total carbon and nitrogen in the surface of the soil, namely, 76% and 77% respectively. Taking into account
the long-term nature of the plantation, the soil still has the capability to support tea production. The
productivity of rice depends partly on fertilization levels and partly on the natural deposition of fresh
sediment eroded from mountains.
Ó 2014 Elsevier Ltd. All rights reserved.
Keywords:
Deforestation
Land degradation
Fluvial activity
Tea plantation
1. Introduction
The humaneenvironment interactions are characterized by
great diversity, with much feeding back, and many nonlinear processes, thresholds and time lags (Rudel et al., 2005; DeFries, 2008;
Lambin and Meyfroidt, 2010; An, 2012). Human impacts are most
frequently related to changing patterns of land use and are of
special importance in environmental studies (Dale et al., 1998;
Geist and Lambin, 2002; Carr et al., 2006). While landforms and
soils are subject to formation, change and even destruction by
natural forces over a geological time scale, land use and land cover
(LULC) changes resulting from human activity usually occur more
rapidly and have a strong impact on vegetation, water resources
and soil (Lambin et al., 2003; Vanacker et al., 2003; Geist and
Lambin, 2006; Davidar et al., 2010). On a smaller scale, there is a
wide variety of humaneenvironment relationships, with different
* Corresponding author. Tel./fax: þ48 124224085.
E-mail address: [email protected] (P. Prokop).
0301-4797/$ e see front matter Ó 2014 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jenvman.2014.01.034
patterns across regions and nations. The densely settled piedmont
zones of young mountains constitute an example of areas where
human activity is superimposed on changes induced by natural
forces (Tiwari, 2000; Liebault and Piegay, 2002; Starkel et al., 2008).
The tectonically active Sikkim Himalayas, with lithology which
is prone to mass movement, receive the highest annual rainfall
along the whole Himalayan front (Starkel, 1972; Dhar and Nandargi,
2000; Soja and Starkel, 2007). Their piedmont, as a transitional
zone down to the lowland plains, is under the strong influence of
the adjacent mountains. The nature and extent of the Himalayas
impact on their piedmont is largely a product of adjustments in
fluxes of water and sediment (Starkel and Basu, 2000; Grujic et al.,
2006; Ghosh and Carranza, 2010). Both are frequently accelerated
by various forms of human activity such as agriculture, logging,
mineral extraction or road building at the mountain margin
(Froehlich and Starkel, 1993; Tiwari, 2000). Increased fluvial sediment transport increases the flood risk, river bank erosion and
shifting of their braided courses, leading to direct changes in the
LULC system of the piedmont (Froehlich and Walling, 2007; Sarkar,
16
P. Prokop, D. Płoskonka / Journal of Environmental Management 138 (2014) 15e23
2008; Starkel et al., 2008). Therefore piedmont of the Sikkim
Himalayas creates specific environmental conditions either
favourable or restrictive for certain forms of human activity.
The present-day economy of the piedmont depends mainly on
tea estates established in the late 19th century (Ray, 2002). Cultivation of other crops (e.g. rice, millet, areca nut) is largely for local
consumption. A significant part of the piedmont area constitutes
reserved forest, national parks and wildlife sanctuaries. Forestry
based on commercially valuable trees and tourism is an important
contributor to the economy, as well as being an employer of large
numbers of people (Government of West Bengal, 2008e2009;
Madhusudan, 2011).
The presented paper attempts to determine the relative
importance of human activity as compared to natural processes in
the transformation of the environment of the Sikkim Himalayas’
piedmont over the last 150 years, with special emphasis on the
period 1930e2010. The natural and human impact is analyzed by
looking at land use change, and quantifying these changes by
comparing the soil properties under the mononocultural cultivation of tea and rice against the properties of the natural forest soil.
2. The study area
The study area is located in the piedmont of the Sikkim
Himalayas in the Jalpaiguri District of the West Bengal State in
India (Fig. 1). A hydrologic and geomorphic approach was
employed to delineate 176.6 km2 at the outlet of the Tista River
from the Himalayas. The area represents a system of Quaternary
fans decreasing in elevation from 200e300 m a.s.l. at the base of
the mountain to about 100 m a.s.l. over a distance of 15 km
(Nakata, 1972; Chakraborty and Ghosh, 2010). The fan surface,
fragmented by the wide braided channels of the Tista and its
tributaries, creates two major landforms, terrace and floodplain,
occupying an area of 93.3 km2 (53%) and 83.3 km2 (47%) respectively. The relative height of the terrace surface increases above
the present river bed, from 6e10 m at the front of the mountain to
25e30 m downstream.
The climate is subtropical monsoonal, with the warm rainy
season spanning from June to October and the dry cooler season
from November to May. The mean annual air temperature reaches
23 C, and fluctuates between 16 C in winter (January) and 28 C in
summer (August). The frontal zone of the mountains and piedmont
receives about 4000e6000 mm rainfall annually, more than 80% of
which falls during the summer monsoon. During this period, 10e15
days of rainfall, of up to 100 mm per day, is observed every year
(Soja and Starkel, 2007; Bookhagen, 2010).
The Tista river drainage area of 8640 km2 covers the whole of
the Sikkim Himalayas. In the piedmont, it receives water from the
Lish and Gish, draining 51 km2 and 157 km2 respectively of the
Himalayan margin. Every year, at the mountain’s edge, the Tista
discharge may reach 3000e5000 m3 s 1. The peak discharge in the
devastating flood of 1968 was calculated at 18,745 m3 s 1, about
45 km downstream of the mountain (Murray and Bochin, 1973).
The 1952 records show a maximum discharge of 255 m3 s 1 for the
Lish and 630 m3 s 1 for the Gish (Dutt, 1966). Several days of
continuous rainfall cause regional floods every 20e30 years, while
the high frequency of severe hourly rainfall causes flash floods
almost every year (Starkel and Basu, 2000; Sarkar, 2008). During
such events, the water level rises in the river channels up to 5e6 m
and a heavy sediment load is deposited on the floodplains (Dhar
and Nandargi, 2000). In the past three decades, the river beds
along the margins of the Himalayas have experienced aggradations
of about 6 cm year 1 in the case of the larger trans-Himalayan Tista,
and between 9e14 cm year 1 in the case of the smaller Lish and
Gish (Sarkar, 2008). Thick alluvial cover creates good conditions for
water infiltration and during the winter season the braided channels are totally dry.
Soils from relatively stable terraces and active floodplains
revealed two distinct stages of pedogenic development, A-(Bw)-C
and A-C, with the dominant soil classified as Inceptisols (Dystrudepts) and Entisols (Fluvaquents and Endoaquents) respectively
(Soil Survey Staff, 2010). Good drainage conditions of Inceptisols
result in rapid leaching and enhancement of the weathering process, giving rise to a humus horizon, thereby lowering the pH (4e5)
and development structural Bw (Cambic) horizon. The Entisols of
active floodplains, imperfectly drained, without any diagnostic
horizon except ochric epipedon, show little alteration of the original deposits and higher pH (5e7), because they receive new
Fig. 1. Location of the study area in part of the Sikkim Himalaya and piedmont, with a schematic longitudinal profile. White dashed lines indicate watershed divides, a black dotted
line indicates the Indo-Bhutan border. The outline of the research area presented on Fig. 2 is marked by an unbroken black line.
P. Prokop, D. Płoskonka / Journal of Environmental Management 138 (2014) 15e23
sediments faster than the assimilation of previous material into the
genetic horizons (NBSS and LUP, 1991).
The terrace area is currently covered with natural tropical moist
deciduous Sal (Shorea robusta) forest and tea plantations
(Champion and Seth, 1968). The floodplains are occupied by wide
braided river channels and rice cultivation. About 56% of the total
population (56,000 inhabitants) of the area lives in nine tea gardens, whilst the remainder of the population is settled along the
main road and railway in the Himalayan foothills (Government of
India (2011)).
3. Material and methods
3.1. Database development
Historical reports and maps from the colonial British administration were used as the sources of information concerning the
LULC and population trends of the Himalayan piedmont in the 19th
and first three decades of the 20th century. The changes in LULC
were derived from the Survey of India topographic maps at a scale
of 1:63,360 for the year 1930, panchromatic photo images acquired
by the American Corona satellite for the year 1970, and satellite
images downloaded from Google Earth Pro for the year 2010. Maps
and satellite images were supported by tea garden plans/cadastral
maps at scales of 1:3,960 for various years between 1930 and 2000.
These cadastral maps are renowned for being reliable, detailed and
useful for land use and soil studies (Ghosh et al., 2004; Sarkar et al.,
2006). Geometric corrections were performed to rectify all the
maps and images using the UTM-45N projection system. A visual
interpretation technique, combined with several field surveys
conducted between 2007e2011, was used for LULC mapping. The
applied method of interpretation is superior to automatic classification when the research output needs to be compared with map
data in a study concerned with detecting change (Petit and Lambin,
2002; Pelorosso et al., 2009). In addition, visual interpretation allows the delimiting of realistic objects, such as patches with
irregular shapes, which can represent important contrasts in
ecological structure or process in the landscape. Thus, visual
interpretation integrates environmental knowledge into image
analysis making the results more environmentally meaningful
(Jensen, 2007; Shao and Wu, 2008; Hersperger and Bürgi, 2009). In
order to standardize the values of maps from three different time
periods, seven consistent LULC categories were defined: forest,
grassland, tea plantation, rice cultivation, other crop, built-up area
and river channel. Logical rules in GIS were used for determination
where areas had the same land use class in each time period (stable), locations where there had been one point of change between
two land use classes on either sides of that area of change (stepped), areas where frequent change has occurred between two
categories (cycle); and where there was a high turnover between
many different classes (dynamic).
Soil data, based on the USDA soil classification system, were
digitized from a 1:250,000 scale soil map developed by the National Bureau of Soil Survey and Land Use Planning (NBSS and LUP,
1991).
3.2. Soil sampling and analysis
The soil sampling scheme was based on the assumption that soil
differentiation between two main landforms, terrace and floodplain, depends on long-term natural soil forming factors (NBSS and
LUP, 1991). Subsequently, within these two main landforms, the
sources of soil variability were the LULC changes. The sampling
sites were therefore located in areas of “stable” LULC (delimited on
a land use stability map prepared earlier), ensuring the same land
17
use (forest, tea and rice) during the period 1930e2010. Within
stable LULC areas of terrace, samples were taken close to the
mountain margin, in the middle and in the lower part of the alluvial
fan. Within stable LULC areas of floodplain, samples were taken
using the same scheme but also at different distances from the
main rivers, in order to take into account sediment heterogeneity.
Soils under built up areas, grassland and other crops were not
examined in detail because of their low contribution in the study
area and the high dynamic of these land use changes during the
investigated period.
Twenty soil profiles were analysed from the major landforms:
terrace with natural forest and tea plantations as well as floodplain
with natural forest and rice cultivation. The natural forests served as
a control against which changes in soil properties were compared.
Significant differences between the means of soil properties were
identified using the t-statistics, with a level of probability of 5%.
Physical (grain size composition, bulk density, compression
strength) and chemical (organic matter (OM), carbon (C), nitrogen
(N), sulphur (S), pH) soil properties were analysed. The rationale for
the selection of soil quality indicators was the detection of longterm cultivation impact on soil properties (Larson and Pierce,
1991; Panwar et al., 2011). Also important was the possibility of
comparison of the soil properties changed due to cultivation with
the soil properties under tea- and rice-based agricultural systems in
other regions of India and Southeast Asia. The grain size composition of each sample was determined using the combined sieving
method and a Fritsch laser particle sizer Analysette 22, after pretreatment with H2O and ultrasonic disaggregation. Soil bulk density was calculated by drying the soil ring samples (100 cm3) at
105 C before weighing them. Soil compression strength (attributable to soil compaction) was measured by using a base surface
cone penetrometer, and the readings of ten random repeated
penetrations were averaged. The concentration of soil organic
matter (OM) was determined using the wet digestion method
(Walkley and Black, 1934). Total C, N and S were determined for the
A soil horizon (topsoil) by combustion in a CHNS vario EL III
Element Analyzer. The pH values in H2O and KCl were measured
electrometrically in a 1:2.5 soil/water and a soil/KCl suspension,
respectively. The colour of the soil was determined according to the
Munsell system.
4. Results
4.1. Land use changes
4.1.1. General tendencies in LULC changes up to 1930
The LULC of the Sikkim Himalayas piedmont, before being ceded
by Bhutan to the British East India Company in 1864, was characterized by large tracts of forest intersected by numerous rivers with
floodplains overgrown by scattered trees and reed jungle (O’Donel
et al., 1864e68; Hunter, 1876, Fig. 2). Owing to the difficulty of procuring water in the dry winters, the area was only seasonally settled
and population density did not exceed 5e10 inhabitants km 2
(Government of Bengal, 1872). Environmental conditions in this
sparsely populated area led to the piedmont initially being categorized as wasteland by the Government of Bengal.
The perception of the region as being void of natural disasters,
and the successful introduction of tea cultivation in the adjacent
Sikkim Himalayas by 1860 resulted in the piedmont area being
leased for the foundation of numerous tea estates. Large-scale
deforestation was initiated in the outlet of the Tista River,
through the foundation of tea plantations between 1874e1885 and
the heavy demand for timber for railway construction (Allen et al.,
1906; Ghosh, 1970). From this date onwards, tea cultivation rapidly
expanded, and by 1930 the number of tea gardens had increased to
18
P. Prokop, D. Płoskonka / Journal of Environmental Management 138 (2014) 15e23
Fig. 2. Spatial distribution of LULC for 1930, 1970 and 2010 with map from 1868 (O’Donel et al., 1864e1868). Black lines indicate terrace-floodplain boundary.
142, covering an area of the Sikkim and adjacent Bhutan Himalayan
piedmont of approximately 483 km2 (The Indian Tea Association,
1930). The development of plantations induced a growing demand for a workforce and the migration of labourers towards the
extensive tracts of cultivable land east of the Tista (Allen et al.,
1906). As a result, the population density in the piedmont zone
had reached 60 inhabitants km 2 by 1931 (Government of Bengal,
1931). The migration of people also initiated the replacement of
floodplain forest by rice cultivation to support their growing food
needs. The formation of the LULC framework was finished when the
1865 Forest Act, provided the legal instrument for registering forest
as “reserved” by the Forest Department. The ‘reservation’ process
started from several tracts of natural forest with a total area of
460 km2, and continued during the period of independence
(Banarjee et al., 2010).
4.1.2. LULC changes between 1930 and 2010
The majority of the area (74.4% in the period 1930e1970 and
77.0% in the period 1970e2010) was not subject to any change
Table 1
Area proportion (%) of LULC classes for the entire study area and within geomorphic
units.
LULC
Entire area
Terrace
Floodplain
Year
1930
1970
2010
1930
1970
2010
1930
1970
2010
Forest
Grassland
Tea
Rice
Other crop
Built up
River channel
41.2
8.5
13.7
8.9
4.0
2.9
20.8
32.3
4.9
14.4
6.0
3.2
3.8
35.4
31.1
3.2
16.5
13.9
1.0
4.8
29.5
55.9
6.4
25.5
e
7.5
4.1
0.6
54.7
4.9
26.9
e
6.1
6.9
0.5
54.8
3.3
31.2
e
1.9
8.6
0.2
24.4
10.9
e
19.2
e
1.4
44.1
6.7
4.8
e
12.9
e
0.3
75.3
4.1
3.1
e
29.8
e
0.3
62.7
(Fig. 2, Tables 1 and 2). The main processes of LULC change between
1930 and 2010 were deforestation and the expansion of both
agriculture and river channels. During this period, the forest cover
decreased by 24.5%, and the rate of deforestation was significantly
higher in the period 1930e1970. About 85.2% of the forest loss
occurred at the expense of river channel widening or migration.
The rest of the deforestation was the effect of conversion to
grassland, followed by the extension of built up area and rice
cultivation. Changes in the area of rice cultivation occupying the
floodplains were also dependent on the widening or narrowing of
river channels in response to flooding events.
Extensive flooding by the Tista and its tributaries in 1954 and
1968 was an important factor in the widening of river channels by
about 70.2% during the period 1930e1970. The subsequent reduction in channel areas between 1970 and 2010 is the result of a recovery process in response to their earlier destabilization.
Nevertheless, river channels occupied an area some 41.8% larger in
2010 than in 1930. These calculations may to some extent have
Table 2
Predominant types of LULC changes (claiming over 1% of the total area).
Type of LULC change
1930e1970
1970e2010
No change
Forest to grassland
Forest to river channel
Grassland to rice
Grassland to river channel
Other crop to built up
Rice to river channel
River channel to forest
River channel to grassland
River channel to rice
74.4
1.1
7.3
1.1
4.0
77.0
3.2
1.6
1.3
5.0
1.3
2.0
1.4
8.3
P. Prokop, D. Płoskonka / Journal of Environmental Management 138 (2014) 15e23
been affected by the construction of embankments and retaining
walls which have restricted the extension of river channels in
recent decades. The first embankments were constructed only to
protect bridges crossing the Lish and Gish rivers and some parts of
the Tista. In 1978 the left bank of the Gish was protected between
the margin of the mountain and its outlet to the Tista. The length of
all embankments has recently been increased to 20 km, so that they
protect 30% of the river banks.
Besides changes in overall land use, a different pattern of LULC
changes may be observed between terrace and floodplain as
geomorphic units represent a proxy for soils and hydrology. Tea
plantations are located only on high terraces with well drained and
acidic soil (Table 2). Rice cultivation, on the other hand, requires
submersion conditions and is concentrated only on floodplains
with shallow, slightly acidic soils. The reserved forest within the
terrace shows stable cover of about 55% of the total terrace area,
between 1930 and 2010. In contrast, over the same time period,
about 19% of the total floodplain area was deforested. Grasslands
occupy both the terraces near built up areas and the floodplains
along rivers, and does not show a strong correlation with geomorphology and soils. Built up areas, on the other hand, are located
mainly on higher elevated terraces, together with other crops
planted in domestic gardens. In the 1930s, built up areas also
expanded into the distal part of the extensive Tista floodplain. The
1968 flood changed this situation, and now only temporary huts are
constructed within areas prone to inundation.
4.1.3. LULC stability
The study of LULC changes in the period 1930e2010 revealed
that stable land use dominates in the investigated region and
covers 60.3% of the total area (Fig. 3). The stable land use area
mainly consists of large blocks of reserved forest, tea gardens and
some settlement cores. About 1/3 of the study area has undergone
stepped changes, in the form of the conversion of forest or grassland to tea bushes at the margin of the tea gardens. Extensive
changes in this category are concentrated along river banks and are
related to the widening of braided river channels. Cyclical changes
Fig. 3. LULC stability between 1930 and 2010.
19
occur only in places within distal areas of the large river floodplains, such as Tista and Gish, where the margins of these wide
floodplains are occupied by rice fields. During extreme flooding, the
cultivated land is affected by debris deposition and thus excluded
from cultivation. Usually, after some time, the sediment is removed
or mixed with the soil, and the land is again taken under cultivation. The area affected by cycle changes covers 4.9% of the region.
Dynamic changes in land use occupy only 4.2% of the total area.
Land undergoing high turnover between many different land use
classes (dynamic changes) is scattered across the Tista river channels, where sand bars are frequently degraded and built up by
flowing water.
4.2. Impact of LULC change on soil properties
4.2.1. Effect of land-use systems on the physical and chemical
properties of soil at different depths
Soils within the terraces are deep and well drained (Table 3).
Their grain-size composition is characterized by the dominance of a
sandy fraction and low clay content (