Content uploaded by Dhruba Pikha Shrestha
Author content
All content in this area was uploaded by Dhruba Pikha Shrestha on Nov 26, 2014
Content may be subject to copyright.
1
ASSESSMENT OF SOIL EROSION IN THE NEPALESE HIMALAYA,
A CASE STUDY IN LIKHU KHOLA VALLEY, MIDDLE MOUNTAIN
REGION♣
♣♣
♣
Dhruba Pikha Shrestha
International Institute for Aerospace Survey and Earth Sciences (ITC)
7500 AA Enschede, The Netherlands
ABSTRACT
Soil erosion is a crucial problem in Nepal where more than 80% of the land area is mountainous and
still tectonically active. Although deforestation, overgrazing and intensive agriculture, due to
population pressure, have caused accelerated erosion, natural phenomena inducing erosion, such
as exceptional rains, earthquakes and glacial-lake-outburst flooding in the high Himalayas are also
common. It is important to understand the erosion process under normal conditions and to assess
the magnitude of problem so that effective measures can be implemented.
Results provided by running a soil erosion assessment model (Morgan et al., 1984) in a GIS
environment show that annual soil loss rates are the highest (up to 56 tonnes/ha/yr) in the areas with
rainfed cultivation, which is directly related to the sloping nature of the terraces. The lowest soil
losses (less than 1 tonne/ha/yr) are recorded under dense forest. In the degraded forest, the soil loss
varies from 1 to 9 tonnes/ha/yr and in the grazing lands it is estimated at 8 tonnes/ha/yr. The rice
fields seem to trap the sediments brought from up-slope. Erosion rates are higher on the south
facing subwatershed than on the north facing one. The index of structural instability of the topsoil,
calculated by the amount of dispersible clay content, seems not to vary so much among the soils,
whether developed on gneiss or micaschist, the main rock types of the study area.
Under normal climatic conditions, soil losses can be considered low although in heavy monsoon,
with exceptional rain, the situation might be different. The study shows that soil erosion can be
modelled in the mountainous region and that the results confirm the soil loss data obtained by means
of experimental erosion field plots in the area, and the study of suspended sediment delivery from
small catchments.
Key Words:
Soil erosion, erosion model, landuse, sloping and level terraces, digital elevation model,
GIS
♣ Land Husbandry, volume 2, no. 1, 1997. Oxford & IBH Publishing Co. Pvt. Ltd, pp.59-80
2
INTRODUCTION
Land resource degradation in the Himalayan region is mainly caused by landslides, mudslides,
collapse of man-made terraces, soil loss from steep slopes and decline of forest/pasture areas
(ICIMOD, 1994). In the world map on the status of human-induced soil degradation (UNEP/ISRIC,
1990), deforestation, removal of natural vegetation and overgrazing are reported to be the main
reasons for loss of topsoil and terrain deformation due to soil erosion in the mountainous regions
of Nepal. Deforestation in the middle mountains is, however, not a recent phenomenon. Clearing
of forests was not only for timber or firewood collection but also to maximise agricultural surplusses
and land taxes according to government land use policy (Mahat et al., 1986). Deforestation
continues mainly for subsistence agriculture. Soil degradation resulting from conversion of forest
land into agriculture in the Chitwan district of Nepal is reported by Burton et al. (1989). In contrast,
in the middle mountains, no significant reduction in forest area has taken place during the recent
decades (Gilmour, 1991). This might be because farmers are well aware of the impact of
deforestation. In some villages, the farmers have begun to develop their own method for resolving
the problem through community management (Fox, 1993). Despite population growth, the
condition of the forest seems to improve and, in some areas, clearing of forest is compensated by
the increase of trees on agricultural land (Carter, 1992; Gilmour, 1991; Gill, 1991; Severinghans
and Adhikary, 1991). Removal of topsoil occurs generally through sheet erosion. Slope length and
steepness, vegetation cover, surface soil condition, amount of rainfall are important factors
determining the rates of soil erosion. Apart from these, particle size distribution, effect of slope
exposition and terrace farming seem to have substantial influence on soil erodibility and
development of erosion features in the study area.
The study aims to evaluate the magnitude of soil erosion in the Middle Mountain region of Nepal.
Although various researchers have undertaken studies related to erosion issue (Kunwar, 1995;
Tamrakar, 1993 ;Likhu Khola Project, 1995; Shah and Schreier (eds), 1991 ), some attention with
regards to erosion modelling is essential considering the inaccessibility of the mountainous areas.
The present study attempts to evaluate the applicability of an erosion model in mountainous terrain.
In addition, it aims to analyse the effect of land use, slope exposition and terrace farming on soil
erosion.
DESCRIPTION OF THE STUDY AREA
The investigated area is within the Middle Mountain Region of Nepal, in the watershed belonging
to the river Likhu Khola (figures 1). The area is chosen because of bio-climatic diversity due to
elevation differences from valley floor to mountain summits, and related land use changes having
influence on soil erosion which is considered typical for the Middle Mountains of Nepal. The
watershed occupies about 160 sq.km. and lies between 27
o
53'55" and 27
o
48'15" North latitude and
between 85
o
13'01" and 85
o
2751" East longitude. The climate varies from subtropical in the main
valley and footslopes through warm temperate at mid-elevations to cold temperate in the higher
mountains. In the lowlands, average summer temperature is 26
o
C with hotter months from April
to September and average winter temperature is 15
o
C (Trisuli station). At higher elevations, average
summer temperature is 19
o
C and average winter temperature is 11
o
C with extreme values of -4
o
C in December (Kakani station). Annual precipitation also varies according to elevation changes,
from 1000 mm in the lowlands (Chhahare, 780 m asl) to 2800 mm at higher elevations (Kakani,
2064 m asl). Most of the rain falls during the months of May to September.
The area is characterised by mountain ridges, having very sharp crests on Precambrian augen and
banded gneiss of various grade of metamorphism and mixtures of micaschist, phyllite and gneiss,
mainly east-west oriented. Likhu Khola is the main drainage system, fed by many tributaries
3
entering the Likhu Khola from both sides. The valley is narrow and elongated, but widens
downstream where rice cultivation prevails. The river joins Tadi Khola before draining into Trisuli
Ganga, which finally joins the river Narayani.
At higher elevations, land cover is mainly forest which consists of chir pine (Pinus roxburghii) and
broad leaf trees (Schima wallichii, local name Chilaune). In the cultivated areas, rainfed maize and
millet are grown. At lower altitudes, sal forest (Shorea robusta) dominates; crops include irrigated
rice and rainfed maize and millet.
Figure 1: Study area (area 2)
Because of the effect of elevation on bio-climatic variations and the presence of old erosion surfaces
due to vertical displacements caused by crustal movements (Iwata et. al., 1983), the study area can
be divided as follows: mountains with very high ridges and narrow valleys (1000-2600 m asl);
mountains with high ridges and narrow valleys (850-2200 m asl); mountains with medium ridges
and narrow valleys (550-1200 m asl); mountains with low ridges, hills and alluvial fans (550-900
m asl); main valley (530-950 m asl); and tributary valleys (630-1000 m asl) (figures 2).
Two sample subwatersheds, namely the south facing subwatershed of Mahadev Khola and the north
facing subwatershed belonging to Jogi and Bhadare Khola, were chosen. Both subwatersheds
belong to the Likhu Khola system. The two subwatersheds were chosen to analyse the relation
between slope exposition and soil erosion. The subwatershed of Jogi and Bhandare Khola covers
a surface area of 256 ha and the elevation varies from 600 to 1225 m. The subwatershed includes
mountains, with medium ridges and narrow valleys, and low ridges, hills and colluvial/alluvial fans.
The subwatershed of Mahadev Khola covers a surface area of 346 ha and the elevation varies from
655 to 1510 m. It includes mountains with high and middle ridges, escarpments, hills and
colluvial/alluvial fans.
4
Figure 2: Topographic cross-sections (AB, CD) through the Likhu Khola valley,
Middle Mountains, Nepal. The arrows indicate the assumed presence of old
erosion surfaces. Notice the narrow valley bottom (cross-section AB) indicating
river incision. The wide valley floor if used for rice cultivation (cross-section CD)
5
EROSION MODELLING
Several empirical and physical models are available to assess soil erosion. Some models, applicable
to a particular area, may not be directly applicable to other areas as they were designed for specific
applications. The Universal Soil Loss Equation, (USLE) (Wischmeier and Smith, 1965), allows to
assess soil loss from agricultural fields in specific conditions. It has been adapted to other conditions
through modified versions such as MUSLE (Williams and Berndt, 1977) and RUSLE (SWCS,
1993) for sediment yield estimation. SLEMSA, the Soil Loss Estimation Equation for Southern
Africa (Stocking, 1981) was developed in Zimbabwe on the basis of the USLE model. WEPP, the
Water Erosion Prediction Project (Nearing et al., 1989) is a process based erosion model, designed
to replace the Unversal Soil Loss Equation.
To compute soil erosion within a watershed, models such as ANSWERS, Areal Non-point Source
Watershed Environment Response Simulation (Beasley et.al., 1980), and AGNPS, Agricultural
Non-Point Source Pollution Model (Young et al., 1987) are available. These models are based on
grid cells and were developed to estimate runoff quality, with primary emphasis on sediment and
nutrient transport. Since they can be linked to a geographic information system (GIS), their
application in a watershed environment may be more interesting for data integration. De Roo (1993)
gives an example of the application of ANSWERS (Beasley et.al., 1980) linked to a GIS to
simulate surface runoff and soil erosion in South Limburg, The Netherlands, and in Devon, United
Kingdom. Gully erosion is also modelled using GIS (Bocco et al., 1990).
Selected erosion model
Although USLE has been widely used through various modified versions, its application in
mountainous terrain with steep slopes is still questionable. Some models, such as AGNPS or
ANSWER, may not be suitable in the Nepalese context because of very high data demand and
AGNPS in particular is not adapted well enough to the Nepalese Middle Mountain conditions
(Kettner, 1996). On the other hand, modelling results may be often impressive but difficult to
interpret (Meyer and Flanagan, 1992) and to validate because of model complexity. Considering all
these, the model developed by Morgan, Morgan and Finney (Morgan et al., 1984) is used in the
present study to assess soil loss from hillslopes in the middle mountain region of Nepal. It was
selected because of its simplicity, flexibility and strong physical base. It separates the soil erosion
process into a water phase and a sediment phase.
- Water phase
In the water phase, the annual precipitation is used to determine the rainfall energy available for
splash detachment and the volume of runoff. The rainfall energy is computed from the total annual
rainfall and the hourly rainfall intensity for erosive rain, based on the relationship established by
Wischmeier and Smith (1978). The annual volume of overland flow is predicted using the model
by Kirkby (1976). In this model, the runoff is assumed to occur whenever the daily rainfall exceeds
a critical value corresponding to the storage capacity of the surface soil layer. Equations used are
given below:
6
For calculating the rainfall energy:
E = R ( 11.87 + 8.73 log
10
I) (1)
where
E = kinetic energy of rainfall (J m
-2
)
R = annual rainfall (mm)
I = rainfall intensity (mm/hr)
For computing the overland flow:
Q = R * exp
(-Rc/Ro)
(2)
where,
Q = volume of overland flow (mm)
R = annual rain (mm)
Rc = Soil moisture storage capacity under actual vegetation (mm)
Ro = mean rain per day (mm)
Soil moisture storage capacity is computed considering soil moisture content at field capacity (MS),
bulk density (BD), rooting depth (RD) and the ratio of actual to potential evapotranspiration (Et/Eo),
as follows:
Rc = 1000 MS.BD.RD (Et/Eo)
0.5
(3)
Mean rain per rainy day (Ro) is calculated by dividing the average annual rain by the number of
rainy days in a year.
- Sediment phase
In the sediment phase, splash detachment is modelled as a function of rainfall energy, soil
detachability and rainfall interception effect by crops. The transport capacity of the overland flow
is determined using the volume of overland flow, slope steepness and the effect of vegetation or
crop cover management (Kirkby, 1976). The equations used are as follows:
For computing splash detachment:
F = K (E exp
-aP
)
b
.10
-3
(4)
where,
F = rate of splash detachment (kg/m
2
)
K = soil detachability index (g J
-1
), defined as the weight of soil detached from the
soil mass per unit of rainfall energy
P = percentage rainfall intercepted by crops
values of exponents: a = 0.05, b = 1.0
For computing the transport capacity of overland flow:
G = C* Q
2
* sin S* 10
-3
(5)
where,
G = transport capacity of overland flow (kg/m
2
)
C = crop cover management factor
Q = overland flow volume (mm)
7
sin S = sine of the slope gradient
For estimation of the soil loss:
soil loss = minimum value of the two: transport capacity of overland flow (G) and
the estimated rate of soil detachment (F).
Materials and softwares used
Use was made of aerial photographs at the scale of 1:40,000, November 1992, covering the whole
watershed of Likhu Khola and of aerial photographs at the scale of 1:20,000, February 1991, for the
subwatersheds of Jogi, Bhandare and Mahadev Khola. Topographic base maps, at scale 1:5000 and
contour intervals of 5 metres, prepared by the Topographical Survey Branch, Dept. of Survey,
Nepal, were used. The erosion model was run in the Integrated Land and Water Information System
(ILWIS), a raster based GIS software package, capable of combining conventional GIS procedures
with image processing and using a relational database (Valenzuela, 1988). For processing the
rainfall data, the spreadsheet software package called Quatro-Pro was used.
Data collection, structure and input for running the model
For running the erosion model, soil data (detachability, moisture content at field capacity of the
surface soil layer, bulk density, rooting depth), rainfall data (annual rainfall, rain intensity and
number of rainy days), land cover data (types of crop, forest, pasture land, and the management
practices) and topographic data (slope gradient) are required.
- Soil data
Soil data were collected from mini-pit and auger hole observations (total 85) along transects in the
Likhu Khola valley. The mountains with very high ridges and narrow valleys have steep to very
steep slopes (>60% slope). Soils are mainly shallow (Lithic and Typic subgroups of Ustorthents,
Ustochrepts, Dystrochrepts, Haplumbrepts, etc.). The mountains with high ridges and narrow
valleys have steep to very steep slopes (30-80% slope). Soils are shallow to moderately deep (Lithic
Ustorthents, Typic Ustochrepts, Typic Ustipsamments, etc.). The mountains with middle ridges and
narrow valleys have moderately steep to steep slopes (15-70% slope). Soils are moderately deep to
deep (Typic Ustochrepts, Typic Kanhaplustalfs, etc.). In the mountains with low ridges, hills and
fans, the slope varies from slightly steep to steep (5-60% slope). The soils are moderately deep to
deep (Typic Ustochrepts, Typic Kanhaplustalfs, (Alfic) Ustarents, etc.). In the main valley of Likhu
Khola and in the tributary valleys, the soils are generally deep (Anthraquic subgroup of Ustifluvents
and Ustochrepts, Fluventic Ustochrepts, etc.)
In addition to the transect studies, soil observations were concentrated in the two sample
subwatersheds of Jogi, Bhandare and Mahadev Khola. Soil data generated by a detailed soil survey
carried out by the Soil Science Division, Nepal Agriculture Research Council (Soil Science
Division, 1992), were also used. Aerial photo interpretation of the subwatersheds, based on
geopedologic analysis (Zinck, 1988), soil survey data and field studies resulted in a soil map legend
as shown in tables 1.
8
- Land use and cover information
In the Nepalese mountains, agriculture uses terracing. Terraces conserve moisture and protect the
land from erosion. Depending on water availability, terraces are either level or sloping. The sloping
terraces, with slopes up to 20%, are for growing rainfed crops such as maize, millet and wheat. The
level terraces are used for rice cultivation. In general, 2 crops of rice are grown, but at lower
elevations where temperature is favourable, up to 3 crops of rice are harvested.
In the mountains with very high ridges and narrow valleys, land cover is mainly forest with chir pine
(Pinus roxburghii) and broad leaf trees (Schima wallichii, local name Chilaune). There are some
cultivated areas with maize and millet. In the mountains with high ridges and narrow valleys, the
forest type is pine (Pinus roxburghii) in association with broad leaf trees (Schima wallichii) at
higher elevations and sal forest (Shorea robusta) at lower elevations. Crops grown are rainfed
maize, millet and rice. In the mountains with medium ridges and narrow valleys, the sal forest
dominates, together with rainfed maize and millet. Some areas are occupied by rice cultivation. In
the mountains with low ridges, hills and fans, the forest type consists mainly of sal trees and
cultivation of rice dominates. In the main and the tributary valleys, irrigated rice is the main crop.
In the Morgan, Morgan and Finney model, the soil parameters used are soil moisture content at field
capacity (MS), bulk density of the top soil in g/cm3 (BD), rooting depth (RD), and the soil
detachability index (K), defined as the amount of soil detached from the soil mass per unit of rainfall
energy per unit area. For the two study areas, the soil parameters used are based on the average
values of the laboratory data from the soil samples collected in the Likhu Khola valley. The soil
detachability index is based on the value suggested in Morgan et al. (1982). The selected soil
parameters are given in table 2.
Landuse maps were generated for the subwatersheds of Jogi and Bhandare Khola and Mahadev
Khola, using interpretation of aerial photographs at 1:20,000 scale and ground control. The
following land cover/landuse classes were established: dense forest, degraded forest, grazing,
rainfed agriculture and irrigated rice cultivation. The natural forests are degraded by the gathering
of fodder and firewood, resulting in low canopy and litter covers. Rice fields were easily separated
because level contour terraces give a special pattern on aerial photographs. The landuse
interpretations were checked in the field and updated. The final map was digitized, rasterized and
georeferenced to fit the base map, with a grid size of 4m.
The percentage of rainfall contributing to permanent interception (P), the ratio of actual to potential
evapotranspiration (Et/Eo) and the crop cover management factor (C) were used as plant parameters
(table 3).
- Digital elevation model and slope gradient map
Since slope gradient is an important parameter in the model (Morgan et al., 1984), especially in
calculating the transport capacity of overland flow, a digital elevation model was generated by
digitizing contour lines at 5m intervals from a topographic base at scale 1:5000. Using the height
values of the contours as attributes, an interpolation procedure was followed to generate a digital
elevation map with spatial resolution of 4 m. Grid size of 4 m was selected as a compromise
between the maximum spatial detail which can be obtained and the possibility to fit the resulting
image on a screen with resolution of 1024 columns and 768 lines. Differential filters (in X and Y
directions) were applied to the elevation map, to generate height differences in X and Y directions.
Finally, a slope gradient map was computed using the height difference maps.
- Rainfall data
9
Rainfall data, recorded during a three year period (1992 to 1994) by automated rain loggers, were
available through the courtesy of the Division of Soil Science, Nepal Agriculture Research Council.
Because of incomplete yearly data at the rain stations, available data from the 3-year period were
pooled to create a set of rain data for a simulated year. In addition, average rainfall data from a
period of ten years (Dept. of Meteorology, 1984) in Kakani (2064 m asl) and Nuwakot (1003m asl),
both located in the vicinity of the study area, were used. The annual rainfall data, available from the
7 stations (table 4), were correlated with elevation. A positive correlation (r = 0.84) was obtained,
indicating that an elevation increase of 100 m increases the amount of annual rainfall by 104 mm.
The rainfall maps were then generated for Jogi and Bhandare Khola and Mahadev Khola
subwatersheds, using the digital elevation model.
For assessing soil erosion, rainfall intensity is very important since splash detachment is a function
of rainfall energy, soil detachability and rainfall interception by crops. The rainfall energy is directly
related to rain intensity (Wischmeier and Smith, 1978). However, not all rainfall events are erosive.
Rain showers of less than 12.5mm are assumed too small to have practical significance and are not
considered erosive (Wischmeier and Smith, 1978). Thus, for estimating the intensity of erosive rains
in the study area, rainfall was first recorded at 5, 10, 15, 30, 45, 60, 120, 180, 360, 720 and 1440
minutes after the first rain on a given rainy day. If the total rain was less than 12.5 mm in a given
day, it was not considered in the calculation of rain intensity. The rain intensities (mm/hour) at
various time intervals were then calculated for rain showers with more than 12.5 mm. For estimating
the rainfall energy, the intensity of 30 minutes was used (table 5).
10
Table 1 Geopedologic legend of the subwatersheds of Mahadev Khola and Jogi and Bhandare Khola.
Landscape Relief type Lithology
/facies
Landform Map
unit
symbol
General
slope
(%)
Dominant soil types Area
Mahadev
Khola
(ha)
Area
Jogi &
Bhandar
e Khola
(ha)
Mountains High
elevation
ridges
Gneiss Summit-
shoulder
complex
Mo111 5-30 Typic Ustochrepts
Lithic Ustorthents
Typic Kanhaplustalfs
10 n/a
Slope facet
complex
Mo112 30-60 Typic Ustochrepts
Lithic Ustochrepts
Dystric Ustochrepts
Typic Ustipsamments
84 n/a
Strongly
dissected
slopes
Mo113 >60 Typic Ustorthents
Lithic Ustorthents
Typic Ustochrepts
54 n/a
Medium
elevation
ridges
Gneiss Summit-
shoulder
complex
Mo211 5-15 Typic Ustipsamments
Typic Ustochrepts
12 n/a
Slope facet
complex
Mo212 20-80 Typic Ustipsamments
Dystric Ustochrepts
23 n/a
Gneiss/
micaschist
Summit-
shoulder
complex
M0221 5-20 Typic Kanhaplustalfs
Typic Ustochrepts
Typic Ustipsamments
9 8
Backslopes Mo222 20-70 Typic Ustochrepts
Typic Kanhaplustalfs
Oxyaquic Ustochrepts
78 51
Footslopes Mo223 10-30 Typic Ustochrepts
Typic Ustorthents
Anthraquic
Ustochrepts
25 11
Low
elevation
ridges
Gneiss/
micaschist
Summit-
shoulder
complex
M0321 10-15 Typic Kanhaplustalfs
Typic Ustochrepts
Typic Ustipsamments
10 n/a
Slope facet
complex
Mo322 30-70 Typic Ustochrepts
Typic Kanhaplustalfs
14 n/a
Micaschist Summit-
shoulder
complex
Mo331 8-15 Typic Ustochrepts 1 48
Backslopes Mo332 15-60 Typic Kanhaplustalfs
Typic Kanhaplustults
Anthraquic
Ustochrepts
Oxyaquic Ustorthents
13 99
Footslopes Mo333 10-40 Dystric Ustochrepts
Typic Kanhaplustalfs
Anthraquic
Ustochrepts
Oxyaquic Ustorthents
6 28
Escarp-ment Quartzite/
micaschist
Scarp, talus
complex
Mo441 >60 Typic Ustochrepts n/a 11
Vales Colluvial/
alluvial
River bed/
alluvial land
Mo541 5-10 --------- 8 n/a
11
Table: 2 Soil parameters used in the model (Morgan et. al., 1984)
Surface texture Soil moisture
content at field
capacity (%)
Bulk density
(g/cm3)
Soil detachability
index
Coarse texture (less than 15%
clay: sandy loam, loam)
0.30 1.1 0.3
Medium texture (less than 35%
clay: loam, sandy clay loam,
silty clay loam)
0.34 1.27 0.4
Fine texture (more than 35%
clay: silty clay, sandy clay)
0.37 1.3 0.4
Note: The soil moisture content and bulk density values are based on the laboratory analysis data. Soil
detachability values are taken from the typical values adapted by Morgan et al (1982).
Table: 3 Plant parameters used in the model (Morgan et al., 1984)
Landuse P (%) Et/Eo C
Grazing land 35 0.80 0.01
Dense forest 35 1.00 0.001
Degraded forest 35 0.90 0.01
Rainfed crops 25 0.67 0.07
Rice cultivation 43 1.35 0.01
Note: The C values for rainfed crops and rice are adjusted by multiplying by 0.15
because of conservation measure through terracing (Morgan, 1982)
Table 4 Annual rainfall at various locations of the Likhu Khola valley
Rainfall station Elevation
(m asl)
Annual rain
(mm)
RL7 780 998
RL1 810 1671
RL4 840 2000
RL2 890 1895
Nuwakot 1003 1872
RL10 1200 1894
Kakani 2064 2839
12
Since detailed rainfall data are only available from a 3-year period from five stations, it is not
possible to compute the rain intensities as a function of elevation. Thus, the 30-minute rain intensity,
averaged from all five stations, was taken as input value (9.86mm/hour rain) to calculate the rainfall
energy. Similarly, the number of rainy days, a necessary parameter for the erosion model, was
averaged from the five stations. The average number of rainy days resulted in 137, which was used
to compute the mean daily rainfall amount.
T
able 5 Rainfall intensity at various stations
Station Elevation Total rain Erosive rain
(m asl) Rainy days Amount
rain
(mm)
Total rain
(mm)
Average rain
intensity
(mm/hr)
RL1 810 104 1671 1460 11.7
RL2 890 159 1895 1426 7.8
RL4 840 159 2000 1305 11.7
RL7 782 107 998 831 12
RL10 1201 157 1894 1496 6.1
Average for
the watershed
137 1692 1304 9.86
Running the model
Once all the attribute maps indicating rain (annual rain, rainfall energy and mean daily rain),
topography (slope gradient), soil (soil moisture content at field capacity, bulk density and soil
detachment index) and plant parameters (percentage rainfall contributing to permanent interception,
ratio of actual to potential evapotranspiration, and crop management factor) were generated, the
model was applied in a GIS environment using map calculation procedures. Two results were
obtained: the predictions of detachment by rainsplash and the transport capacity of the runoff (table
6). The prediction of detachment is compared with the transport capacity of the runoff and the lower
of the two values is assigned as the annual rate of soil loss, denoting whether the detachment or the
transport is the limiting factor. The resulting annual soil loss rates for the subwatersheds of Mahadev
Khola and Jogi and Bhandare Khola are shown in table 7 and the maps of soil losses, calculated by
the model, are given in figure 3 and 4.
13
Table 6 Soil detachment and transport capacity
Subwatershed Mahadev Khola Jogi and Bhandare Khola
Landuse Detachment
(tonnes/ha)
Transport
capacity (tonnes/ha)
Detachment
(tonnes/ha
Transport
capacity (tonnes/ha)
Average St.dev Average St.dev. Average St.dev. Average St.dev.
Rainfed crops
(maize, millet)
38.1 6.9 57.8 36.8 35.2 5.6 19.0 11.7
Grazing land 22.8 3.3 8.1 4.4 20.4 3.5 0.8 0.7
Dense forest 21.3 0.5 0.3 0.1 n/a n/a n/a n/a
Degraded
forest
23.3 3.3 2.5 2.1 20.1 3.3 0.5 0.6
Rice 14.7 2.7 0.3 0.2 13.7 2.2 0.2
0.1
Table 7 Soil loss prediction and comparison between the two subwatersheds
Mahadev Khola Subwatershed
(south-facing)
Jogi & Bhadare Khola Subwatershed
(North-facing)
Soil loss (tonnes/ha) Soil loss (tonnes/ha)
Landuse
Area
(ha)
Range Average St.dev. Area
(ha)
Range Average St.dev.
Rainfed crops
(maize, millets)
56 6.1-56.2 32.0 11.0 60 2.9-34.6 17.7 8.7
Grazing land 96 1.6-19.8 8.1 4.3 9 0.1-4.4 0.8 0.7
Dense forest 13 0.1-0.4 0.3 0.1 - - -
Degraded forest 91 0.1-8.6 2.5 2.1 46 0.1-3.4 0.5 0.6
Rice 84 0.1-0.8 0.3 0.2 141 0.1-0.5 0.2 0.1
14
Figure 3: Soil loss estimation in the subwatershed of Mahadev Khola
Figure 4: Soil loss estimation in the subwatershed of Jogi and Bhandare Khola
15
RESULTS AND DISCUSSIONS
Soil losses are comparatively lower (less than 10 tonnes/ha/yr) under landuse types, such as forest,
grazing land and rice cultivation. Annual soil loss rates are maximum (up to 56 tonnes/ha/yr) in
areas under rainfed cultivation. The lowest soil losses (less than 1 tonne/ha/yr) are recorded in rice
fields and under the dense forest. In the degraded forest areas, soil losses vary from 1 to 9
tonnes/ha/yr and in the grazing lands, it is about 8 tonnes/ha/yr.
The modelled soil losses also confirm the data obtained with other methods using field plots in the
Likhu Khola valley which was carried out by the Soil Science Division, Nepal Agriculture
Research Council (Likhu Khola Project, 1995). Soil erosion was monitored in field plots, of size
varying from 25 m
2
. to 535 m
2
, under different land uses and management types, including rainfed
agriculture, dense forest and degraded forest on different slope aspects and gradients. Altogether
24 plots were monitored during the pre-monsoon and monsoon rains in 1992 and 1993. Results
highlight that runoff can be generated under all land uses by rainfall of low magnitude and
intensity. Forest canopy has a positive effect on controlling excess runoff. It is reported that soil
loss of less than 5 g/m2 is recorded under grassland and relatively slightly degraded secondary
forest. Soil loss on non-cultivated land is estimated at 11 tonnes/ha/year. Under rainfed cultivation,
soil losses range from 2.7 to 8.2 tonnes/ha for the period of May to September 1993. The highly
degraded forest shows intermediate levels of soil loss. Under dense forest and grassland cover, soil
is lost at a long-term sustainable rate.
If a soil loss of up to 25 tonnes/ha/yr is considered tolerable in mountainous areas where the natural
rate of soil loss is high (Morgan, 1986), both study watersheds have moderate soil losses. This is
also confirmed by the result of a study on the suspended sediment delivery from a small catchment
area having different landuses (Ries, 1995), where soil erosion rates were observed to be low. In
heavy monsoon, the situation might be different since a single rainstorm can generate a soil loss
as high as 300g/m2, as shown by the result obtained on the erosion plot under rainfed agriculture
in the Likhu Khola valley (Likhu Khola Project, 1995).
-Effect of the sloping nature of terraces
The high soil loss rates under rainfed agriculture are directly related to the sloping nature of the
terraces. Making sloping terraces is cheaper than making level terraces. The cost involved in level
terraces is not justified by growing of rainfed crops. Farmers are willing to invest more for growing
a cash crop like rice, if water supply and temperature conditions are favourable. Rainfed crops are
usually grown in a relatively drier environment, in soils with lower organic matter content and
reduced structural stability.
It is also interesting to note that only a few rills are observed on sloping terraces. Rills disappear
through cultivation practices, as labour input in the Nepalese agriculture is quite considerable. But
it is also worth considering the way sloping terraces are made. The slope of the terraces varies from
10 to 15%. The width of the terrace is determined by the slope of the land. The steeper the slope,
the higher the slope gradient of the terrace and the narrower the width of the terrace. There is no
bund on the outer edge of the sloping terraces. In the Likhu Khola valley, the width of the sloping
terraces varies from 2 to 3 m. A ditch at the foot of the terrace riser diverts the runoff. In this way,
surface runoff cannot concentrate, as the effective slope length is too short (2 to 3 m). However,
a gully may develop towards the lower reaches of the side stream because of the high volume of
runoff collected from the ditches. This shows that sheet erosion is dominant.
The high erosion rates under rainfed cultivation on sloping terraces is corroborated by the analysis
of micro-topographic features and the application of simple field tests in the Likhu Khola valley
(Kunwar, 1995). The field tests carried out were the crumb test, pin-hole test and rainfall
acceptance test (Bergsma, 1990).
16
-Role of level terraces
Rice cultivation dominates the lower ridges, hills, coalescing alluvial fans and the valley floors
because of water availability. Rice is cultivated with rain or irrigated water and excess water is
removed from the field by means of a small opening on the terrace bund. The water is then allowed
to flow to terraces at lower elevations. In this way, the water passes a number of terraces before
entering a stream. In a hill slope, an average of 15 to 20 terraces may exist, but in the main valley
a sequence may include no more than 10 terrraces because of the larger width of the rice fields.
Because of this way of water management, most of the sediments brought from upslope are
trapped. At the bottom of the Likhu Khola valley, rice fields are harvested before the rainy season
and used for trapping sediments.
-Influence of particle size distribution
According to the soil erodibility factor (K) of USLE, the combination of sand and silt with low clay
content and low organic matter content indicates moderate to high erodible conditions. Study of
particle size distribution of topsoils taken from 48 locations in the two subwatersheds shows high
sand content, followed by silt and clay contents (table 8), creating a textural class likely to promote
soil erodibility. On the other hand, the presence of water-dispersible clay indicates not only
structural instability but also the availability of material for erosion. It has been found on the basis
of the laboratory analysis results obtained from soil samples collected from the area that the relative
amount of water-dispersible clay is higher in soils derived from gneiss than in soils developed on
micaschist. This is probably due to differences in the type of clay minerals. In contrast, the index
of structural instability calculated by the ratio of dispersible clay to total clay of the topsoil seems
not to vary so much among the soils, whether developed on gneiss or on micaschist. This shows
that the external factors, namely climate and human activities, play an important role.
Table 8 Particle sizes distribution of the topsoils in Mahadev, Jogi and Bhandare Khola.
Mahadev Khola Jogi and Bhandare khola
Particle size
class
Average
content (%)
Standard
deviation
No. of
observatio
ns
Average
content
(%)
Standard
deviation
No. of
observatio
ns
Sand 56 8.2 23 51 11.9 25
Silt 32 5.9 32 8.6
Clay 12 6.6 16 8.1
- Effect of slope aspect
Erosion rates are relatively higher on the south-facing subwatershed than on the north-facing one
(table 7). This is also confirmed by the shallowness of soils on the south-facing slopes than on the
north-facing ones. Analysis of the effect of slope aspect on the depth to the B and C horizons were
carried out based on some 71 soil profile descriptions of the area (36 profiles from the
subwatershed of Jogi and Bhandare and 35 profiles from the subwatershed of Mahadev Khola).
Many of the profile descriptions of the area were obtained from semi-detailed soil survey of the
area (Soil Science Division, 1992). The average depth to the B and C horizons are 19 cm and 95
cm respectively in the soils developed from gneiss on the north facing Jogi and Bhandare
subwatershed, while they are 20 cm and 122 cm in soils developed on micaschist. Comparatively,
on the south facing slope of the Mahadev Khola subwatershed, the average depths to the B and C
horizons are 16 cm and 49 cm respectively in soils developed from gneiss, while they are 17 and
78 cm in soils developed on micaschist (table 9). The south facing slopes are drier because of more
17
radiation and higher evapotranspiration, which decrease weathering and retard soil development.
But, a drier environment also has a scarcer vegetative cover which promotes erosion. Although the
rock type is the main factor controlling the weathering rate, slope aspect effects not only
weathering but also soil erosion.
Table 9 Depth to the B and C horizons in soils developed on various rock types in the
subwatersheds of Jogi and Bhandare Khola and Mahadev Khola
Subwatershed Lithology Average depth to
B horizon
(depth range)
Average depth to C
horizon
(depth range)
Gneiss 19 cm
(10-34 cm)
95 cm
(50-117 cm)
Jogi and Bhandare Khola
(North aspect)
Micaschist 20 cm
(11-36 cm)
122 cm
(64-187 cm)
Gneiss 16 cm
(9-31 cm)
49 cm
(8-147 cm)
Mahadev Khola
(South aspect)
Micaschist 17 cm
(11-35 cm)
78 cm
(24-217 cm)
CONCLUSIONS
It is shown that erosion is more pronounced in sloping terraces under rainfed agriculture in the
Middle Mountain Region of Nepal. Soil losses are minimal in dense forest and level irrigated rice
fields. In the rice fields the problem is not only minimal or absent, but the rice fields seem to trap
the sediments brought from upper slopes. The study also demonstrates that soil erosion can be
modelled in the mountainous areas.
Under the prevailing climatic conditions, soil losses can be considered low. However, during heavy
monsoon, the situation might be different since a single rainstorm can generate heavy soil losses.
If an exceptionally high rainfall event with rain amount higher than 400 mm in a day, like the one
of 20 July 1993 ( Dhital et al., 1993) takes place, enormous soil losses can be expected, in addition
to heavy damages to infrastructure, human lives and property. The recurrence of such a rainfall
event (of more than 400 mm per day) is estimated at 60 years, and that of 100 mm rain per day is
about 1.5 years (Kakani station).
In conclusion, the erosion issue in Nepal seems to be more related to nature than to human
influence. Bruijzeel and Bremmer (1989) reached a similar conclusion when stating that the impact
of land rehabilitation programmes will be mainly felt "on-site" and the effects will be negligible
or minor even at the scale of relatively small catchment areas.
ACKNOWLEDGEMENTS
The author would like to express his sincere gratitude to Prof. Alfred Zinck from the International
Institute for Aerospace Survey and Earth Sciences (ITC) and Prof. Salle Kroonenberg from the
Technical University of Delft, The Netherlands for their valuable suggestions and critical remarks.
Anonymous reviewers of this paper are gratefully acknowledged. The helpful company of the late
Mr. Lok Bahadur Kunwar in the field is very much appreciated. Ms. Anneke Fermont is
acknowledged for her help in analysing the rainfall data.
18
REFERENCES
Beasley, D. B., L. F. Huggins and E. J. Monke, 1980. ANSWERS: a model for watershed
planning. Transactions of the ASAE, 23(1980)4, pp. 938-944..
Bergsma, E., 1990. Approximation of soil erodibility using simple tests. ITC lecture notes
Bocco, G. et. al., 1990. Gully erosion modelling using GIS and geomorphic knowledge, ITC
Journal 1990-3, pp.253-261
Bruijnzeel, L. A. and C. N. Bremmer, 1989. Highland-lowland interactions in the Ganges
Brahmaputra river basin: A review of published literature. ICIMOD occasional paper no.
11, Kathmandu, Nepal.
Burton, S., P. B. Shah and H. Schreir, 1989. Soil degradation from converting forest land into
agriculture in the Chitwan district of Nepal. Mountain Research and Development,
9(4):393-404.
Carter, E. J., 1992. Tree cultivation on private land in the middle hills of Nepal; lessons from some
villagers of Dolakha district. Mountain Research and Development, 12(3):241-255.
Dept. of Meteorology, Nepal, 1984. Climatological records of Nepal, 1971-1982, vol. 1, HMG,
Kathmandu, Nepal.
De Roo, A. P. J., 1993. Modelling surface runoff and soil erosion in catchments using Geographic
Information Systems. Netherlands Geographical Studies 157, University of Utrecht, Utrecht,
The Netherlands.
Dhital, M. R, N Khanal and K. B. Thapa, 1993. The role of extreme weather events, mass
movements and land use changes in increasing natural hazards. A report of the preliminary
field assessment and workshop on cause of the recent damage incurred in south-central
Nepal (July 19-20, 1993). ICIMOD, ISBN 92-9115-175-0.
Fox, J, 1993. Forest resources in a Nepali village in 1980 and 1990: the positive influence of
population growth. Mountain Research and Development, Vol. 13(1):89-98.
Gill, G. J., 1991. Indigenous erosion control systems in the Jhikhu Khola Watershed, workshop
proceedings: Soil fertility and erosion issues in the middle mounts of Nepal, Topographical
Survey Branch, HMG, Nepal.
Gilmour, D. A., 1991. Trends in forest resources and management in the middle mountains of
Nepal; workshop proceedings: Soil fertility and erosion issues in the middle mounts of
Nepal, Topographical Survey Branch, HMG, Nepal.
ICIMOD, 1994. Constraints and opportunities. Proceedings of the International Symposium on
Mountain Environment and Development, International Centre for Integrated Mountain
Development, Kathmandu, Nepal.
Iwata, S, T. Sharma and H. Yamanaka, 1983. A preliminary report on geomorphology of Central
Nepal and Himalayan Uplift. Journal of Nepal Geological Society, vol. 4, Special Issue,
1984, p. 141-149.
Kirkby, M. J. , 1976. Hydrological slope models; the influence of climate. In Derbyshire, E. (ed.),
Geomorphology and climate. Wiley, p. 247-267.
19
Kettner, A., 1996. Simulated man-induced erosion in the Middle Mountains of Nepal, a case study
on the relation between land use, land tenure and erosion with the use of the AGNPS-model
in the Mahadev Khola watershed. Msc. thesis (unpublished), Dept. Of Soil and Water
Conservation and Irrigation, Agricultural University, Wageningen.
Kunwar, L. B., 1995. Study of the relationship between land use and soil erosion hazard using
simple erosion field test, surface microtopographic features and remote sensing data. MSc.
thesis (unpublished), Soils Division, ITC.
Likhu Khola Project, 1995. Landuse, soil conservation and water resource management in the
Nepalese Middle Hills. Institute of Hydrology, Nepal Agriculture Research Council and
Royal Geographic Society United Kingdom.
Mahat, T. B. S., et. al., 1986. Human impact on some forests of the middle hills of Nepal; Forestry
in the context of the traditional resources of the state. Mountain Research and Development,
Vol.6, No.3, 1986, p.223-232.
Meyer, C. R. and D. C. Flanagan, 1992. Application of case-based reasoning concepts to the
WEPP soil erosion model. AI Applications, 6(3)
Morgan, R. P. C., 1986. Soil erosion and conservation. ed. by D. A. Davidson. Longman Scientific
and Technical, Longman Group UK Limited. 298 p. ISBN 0-582-30158-0
Morgan, R.P.C., D.D.V. Morgan and H.J. Finney, 1982. Stability of agricultural ecosystems:
documentation of a simple model for soil erosion assessment. International Institute for
Applied Systems Analysis Collaborative Paper CP-82-50
Morgan, R.P.C., D.D.V. Morgan and H.J. Finney, 1984. A predictive model for the assessment of
soil erosion risk. J. Agric. Engng. Res., 30, 245-253.
Nearing M. A., G. R. Foster, L. J. Lane, S. C. Finkner, 1989. A process-based soil erosion model
for USDA-Water Erosion Prediction Project technology. Transactions of the American
Society of Agricultural Engineers 32: 1587-93.
Ries, J. B., 1995. Soil erosion in the high mountain region of the eastern Nepalese Himalayas. Z.
Geomorph.N.F. Suppl.-Bd.99, p.41-52
Severinghans, J. and B. R. Adhikary, 1991. A study on Leucaena leucocephala and cereal crop
biomass production in Majhigaun village of Sindhupalchowk district, Nepal; workshop
proceedings: Soil fertility and erosion issues in the middle mounts of Nepal, Topographical
Survey Branch, HMG, Nepal.
Shah, P. B. and H. Schreier (eds), 1991. Soil fertility and erosion issues in the Middle Mountains
of Nepal. Workshop proceedings Jhiku Khola watershed, 22-25 April, 1991. Integrated
Survey Section, Topographical Survey Branch, Kathmandu, Nepal.
Soil Science Division, 1992. Soils of Bore, Mahadev, Dee, Khahare and Jogi & Bhandare Khola
subwatersheds. Landuse, soil conservation and water resource management project, Likhu
Khola watershed area, Nuwakot District, Nepal.
Stocking, M., 1981. A working model for the estimation of soil loss suitable for underdeveloped
areas. Development studies, occasional paper No. 15, University of East Anglia, Norwich
(UK).
20
SWCS, 1993. User's guide. Revised Universal Soil Loss Equation, version 1.03. Soil and Water
Conservation Society, Iowa, USA.
Tamrakar, R., 1993. A comparative study of land use change in the Shivapuri integrated watershed
development area between 1981-1993. Dept. Of Soil Conservation and Watershed
Management, Min. of Forest and Soil conservation, Nepal.
UNEP/ISRIC, 1990. World map on status of human induced soil degradation; scale 1:10,000,000.
UNEP, Nairobi, Kenya.
Valenzuela, C.R., 1988. ILWIS overview. ITC Journal 1:4-14
Williams, J. R. and H. D. Brendt, 1977. Sediment Yield prediction based on watershed hydrology.
Transactions of the ASAE vol. 20, p. 1100-1104.
Wischmeier W. H. and D. D. Smith, 1965. Predicting rainfall erosion losses from cropland. A
guide for selection of practices for soil and water conservation. USDA Agri. Handbook no
282, Washington DC., USA.
Wischmeier W. H. and D. D. Smith, 1978. Predicting rainfall erosion losses. A guide to
conservation planning. USDA Agricultural Handbook, No. 537
Young R. A. et. al., 1987. AGNPS, Agricultural Non-Point-Source Pollution Model, Agricultural
Research Service, Conservation Research Report 35, USDA, USA.
Zinck, J. A., 1988. Physiography and soils. ITC lecture notes on soil survey course.
