awasthi (2002) land-use change in two nepalese watersheds-gis and geomorphometric analysis

7/31/2019 Awasthi (2002) Land-use Change in Two Nepalese Watersheds-gis and Geomorphometric Analysis

1/19

land degradation & development

Land Degrad. Develop. 13: 495513 (2002)

Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/ldr.538

LAND-USE CHANGE IN TWO NEPALESE WATERSHEDS:

GIS AND GEOMORPHOMETRIC ANALYSIS

K. D. AWASTHI,1 B. K. SITAULA,1* B. R. SINGH1 AND R. M. BAJACHARAYA2

1Department of Soil and Water Sciences, Agriculture University of Norway, As, Norway2Kathmandu University, Dhulikhel, Nepal

Received 18 January 2002; Revised 2 October 2002; Accepted 6 October 2002

ABSTRACT

Accurate measurement of land-use/land-cover and geomorphometric parameters is important for evaluating watershedconditions, yet these are surprisingly difficult quantities to measure accurately over large areas. Watershed analysis based onthe geographic information system (GIS) was carried out in two watersheds in the western development region of Nepal. Land-use maps were prepared after interpretation of 1978 and 1996 aerial photographs. Digital data for deriving geomorphometric

parameters were prepared from topographical maps of scale 1: 25 000. The dynamics of land-use and land-cover change withinthe Mardi and Fewa watersheds were investigated by performing spatial analysis of digital land-use maps in ArcView 3.1desktop environment. There was a net increase in forest cover of 24 per cent and 11 per cent in the Mardi and Fewa watershedsrespectively, with a corresponding decrease in shrub and rainfed agriculture. Land use was found to be highly dynamic withsignificant internal trading among the land-use classes. A significant area under agriculture in 1978 was found abandoned in1996 in both watersheds most likely due to increased out migration of the labour force. Geomorphometric parameters such ashypsometric curves, hypsometric integrals (HI), drainage density and length of overland flow were analysed to explain thewatershed conditions. The results of geomorphometric analysis revealed that the watersheds have been subjected in the past tohigh erosion and are still susceptible to lateral surface erosion hence soil degradation. Some suggestion for management can bederived from this study. Copyright # 2002 John Wiley & Sons, Ltd.

key words: GIS; hypsometry; land-use change; Nepal; watershed degradation

INTRODUCTION

Estimating temporal land-use and land-cover changes is essential to assessing the rate at which these changes

advance and the problems or impacts they cause and, hence, prediction of future impacts and trends (Lambin,

1997). Land-use and land-cover modification have important environmental consequences through their impacts

on soil and water quality, biodiversity, microclimate, methane emission and reduced CO2 absorption and, hence,

contribute to watershed degradation (Lambin et al., 2000; Schneider and Pontius, 2001). Increased consciousness

of these impacts enhanced their estimating, forecasting and modelling at the global, regional or watershed scales

(Chen et al., 2001).

Numerous developing countries including Nepal face serious environmental degradation induced by large-scale

deforestation. The severe degradation of the middle mountains of the Nepal (Ives and Messerli, 1989) has recently

been quantified and mapped through the considerable efforts of the Nepalese Government and international

agencies. A total of 103 968 ha of forest in Siwaliks hills and plains were cleared under the governments

resettlement programme from the 1950s to the mid-1980s (MPFS, 1988). Comparison of the 19781979 maps withthose of 19941996, showed that the annual deforestation rate is 05 per cent nationwide, where as it is 17 per centfor southern Terai (plain areas) and 23 per cent for middle mountain regions, respectively (FRI, 1999). However,

Copyright # 2002 John Wiley & Sons, Ltd.

Correspondence to: Dr B. K. Sitaula, Agriculture University of Norway, Department of Soil and Water Sciences, PO Box 5028, N-1432, A s,Norway. E-mail: [email protected]


2/19

introduction of community forestry and leasehold forestry programmes during the 1980s and 1990s has resulted in

increase in forest cover in the middle mountain area (Gilmore and Nurse, 1991).

The destabilization of fragile mountain slopes through deforestation, agricultural expansion, excessive grazing

and expansion of the road network has increased land degradation and soil erosion rates (Ives and Messerli, 1989;

Thapa, 1990). Soil erosion rates have been estimated as high as 15 3Mgha1 for degraded forest and as high as213Mgha1 for uncontrolled grazing lands (Pahari, 1993). Agriculture was extended at the cost of forest/shrub,

marginal, and submarginal areas with very steep slopes without due consideration for the suitability of these landsfor cultivation (Tiwari, 2000).

Interpretation of aerial photographs taken at different intervals provides valuable information of physical

features such as land use, soils, vegetation, stream networks, and landforms at different time scales (Borrough and

McDonnell, 1998). It also helps in preparing topographic maps and creating a digital elevation model (DEM),

hence integrating slope, aspect and topography of a complex mountain environment (Trapp and Mool, 1996). The

capability of GIS to integrate and analyse temporal and spatial data helps in quantifying the land-use changes. In

areas of rugged topography and poor accessibility, remote sensing is a valuable tool for monitoring the spatial and

temporal changes in land use, as well as its impacts. Due to the spatial nature and distribution of watershed

parameters, remote sensing combined with GIS has proved effective for analysing, storing, retrieving and

displaying such biophysical and socio-cultural data (Sidhu et al., 2000).

In conjunction with land-use and land-cover changes, investigation of relevant geomorphometric characteristics

serves as a more holistic indicator of watershed status. Geomorphometric characteristics such as hypsometriccurves, hypsometric integrals (HI), drainage density and length of overland flow are important indicators of

watershed conditions (Ritter, 1986). The shape of the hypsometric curve explains whether alteration in slope has

taken place in comparison to the original basin. The HI expresses, as a percentage, the volume of original basin that

remains (Ritter, 1986). These parameters are important indicators for assessing the watershed health in the fragile

watersheds in the Himalayan region. Drainage density is closely associated with erosion processes, lithology, relief

and vegetation. It relates morphology to soil properties and climate (Berger and Entekhabi, 2001; Roth and La

Barbara, 1997) and plays an important role in shaping the watershed through erosion, deposition and sediment

transport processes (Tucker et al., 2001).

Despite the significance of these important environmental variables, our knowledge of land-cover dynamics and

influence of geomorphometric characteristics on watershed quality is poorly studied in Nepal. In this study, we

utilize GIS and include the geomorphometric characteristics of the watersheds for explaining its situation. So, the

objectives set for the study are to:

* Assess land-use/land-cover change in two middle mountain (Mardi and Fewa) watersheds of Nepal.

* Examine distribution of agricultural land by slope classes within the watersheds to determine land suitability.

* Compute geomorphometric parameters such as hypsometry, drainage density and length of overland flow using

GIS tools to explain watershed conditions.

* Develop a management strategy.

MATERIALS AND METHODS

Study Area

Two watersheds, Mardi and Fewa, in the Kaski District, Western Development Region of Nepal, were selected for

this study (Figure 1). The landscape is steeply dissected in both watersheds. Important information for these

watersheds is listed in Table I.

Land-use Study Approach

Land-use maps were developed from two series of aerial photos (1:50 000 nominal scale) taken during 1978 (photo

serial 7932-121 to 125, 7932-161 to 165, and 7938-12 to 15) and 1996 (photo serial L10825 to 27, L109-42 to 45,

L110-35 to 37 and L111-21 to 22). Before interpretation of the aerial photographs a land-use and land-cover

496 K. D. AWASTHI ETAL.

Copyright # 2002 John Wiley & Sons, Ltd. LAND DEGRADATION & DEVELOPMENT, 13: 495513 (2002)


3/19

reconnaissance was carried out in August 2000, by which a general understanding of the land-use and land-cover

status of the study area was obtained. Land use within both watersheds was classified into the categories as shown

in Figures 2a and 2b. Cover types were classified as follows:

* Class 1:


4/19

* Class 4: 4060 per cent tree cover (open forest)

* Class 5: >60 per cent tree cover (dense forest)

Based upon these classifications draft maps of land use and land cover for both watersheds were prepared for the

years 1978 and 1996. Field verification was carried out in October 2000 to improve the accuracy of the land-use

map of 1996 and corrections were made on the draft map before finalizing it.

Database Development and Analysis

The spatial database development consists of land-use and land-cover maps derived from interpretation of aerialphotos taken during 1978 and 1996 and topographical maps prepared by His Majestys Government, Department

of Survey (LRMP, 1986), Nepal. The land-use and land-cover maps were digitized and processed using ArcView

31 desktop environment. Elevation data for the DEM were derived from 1:25 000 scale topographic maps andslope inclination maps were derived from the DEM (ESRI, 1996).

Slope inclination was selected as one of the principal limiting parameters for agricultural land-use distribution.

Slopes were classified into the following ranges 41, 15, 530 and >30 degrees based on Land Resources

Mapping Project classifications (LRMP, 1986) for Nepal. Spatial analysis was performed for quantifying the land-

use changes. The flow chart of the process involved in spatial analysis is shown in Figure 3.

Calculating Hypsometric Curves and Hypsometric Integrals

The incision of the drainage basin has a profound effect on the hypsometry of the basin (Bishop et al., 2002; Ritter,

1986). The hypsometric curve is related to the volume of the rock in the basin, and the amount of erosion that had

occurred in a basin vs. what still remains (Hurtrez et al., 1999). The hypsometric integral thus helps in explaining

the erosion that has taken place during the geological time scale (Bishop et al., 2002). Comparison of the shape of

the hypsometric curve for a different basin in a similar climatic condition and approximately equal area also

provides relative insight into the past soil movement of basins.

The hypsometric curve represents the relative proportions of a basin area that lies below a given height. For a

selected basin, the range of basin was divided into equal intervals. For each interval the proportion of the basin area

Table I. Description of study areas

Fewa watershed Mardi watershed

I. Elevation variation: 7802480 m 9155590 mII. Co-ordinates: 28110000 to 281703000N 28190 to 28290N and 83500 to 83560E

and 874703000 to 835903000EIII. Mean annual temperature: 18C 15C

IV. Average annual rainfall: 4500 mm 4300 mmV. Geology: watershed lies between two major faults Major thrust plane passes through the northern part of the

running parallel to each other. The dominant bedrocks watershed. The dominant bedrocks are phyllite, quartzite,are phyllite, schist, and quartzite and dip angle of and dolomite and dip angle of bedrock varies from 15 to 60.bedrock varies from 15 to 60. Valley bottoms consist Valley bottoms consist of unconsolidated sediment. The mainof unconsolidated sediment. The main physiographic physiographic features are alluvial valley, hills, mountainsfeatures in the watershed are lake, alluvial valley, hills and narrow valleys.and mountains

VI. Dominant vegetation: broad-leaved mixed hardwood Broad-leaved mixed hardwood forest at lower elevations andforest coniferous forest at higher elevations

VII. Climate: monsoonal with hotdry subtropical in valleys Monsoonal with hotdry subtropical in valleys to alpineto warm moist temperate at higher elevations in higher elevations

VII. Common soils: Dystric Luvisols, Dystric Cambisols Dystric Luvisol and Dystric Cambisol in gentle slopes,in gentle slopes, Regosols in steep slopes, and Fluvisols Rigosols in steep slopes, and Fluvisols in flat and river

in flat and river valleys valleys




5/19


6/19

Figure 2b. Land use map of Mardi watershed (1996).




7/19

Figure 3. Flow chart showing processes involved in land-use and land-cover change analysis.

LAND-USE CHANGE IN TWO NEPALESE WATERSHEDS 501



8/19

was calculated. Elevations and areas were divided by the relief and total watershed area so that they range from 0

to 1. The HI represents the area under the hypsometric curve and is computed as shown in Equation (1) (Hurtrez

et al., 1999)

HIElev Elevmin

Elevmax Elevmin1

where elev is the mean elevation of the watershed, elevmin and elevmax are minimum and maximum elevations

within the watershed. Hypsometric curves were interpreted as youth (convex upward curves), mature (S-shaped

curves) and peneplain or distorted (concave upwards curves) stages of landscape evolution. Convex hypsometric

curves are more likely typical of plateaux with little erosion, which can evolve into an S shape, while

concave hypsometric curves indicate greater importance of erosion (Hurtrez et al., 1999), such as Himalayan

mountain watersheds, which are vulnerable to erosion due to high monsoon rainfall and intense pre-monsoon

storms.

Drainage Density and Length of Overland Flow

Drainage density is a fundamental property of the natural terrain, which reflects local geology, climatic condition,

topography, vegetation and soils (Berger and Entekhabi, 2001; Ritter, 1986), It represents the degree to which alandscape was dissected by stream channels (Tucker et al., 2001). The drainage density and length of overland flow

are computed as shown in Equations (2) and (3) (Ritter, 1986)

Dd L

A2

L0 1

2Dd3

where L is the total length of the channel within the watershed area A and L0 is the length of overland flow in

kilometres. Dd has dimensions of inverse length and varies with climate, vegetation and other factors. A low value

ofDd corresponds to terrain with long hill slopes and a high Dd indicates a dissected terrain (Berger and Entekhabi,2001).

The soil erosion is a function of slope length and increases with increasing slope length (Gabriels, 1999,

Wischmeier and Smith, 1978). Clarke and Rendell (2000) documented that badland slopes of initially >35 degrees

remodelled to form more gentle slopes of 1217 degrees, increased the effective area for rainfall by increasing

slope length. The increase in slope length contributed 143 m3 of soil loss in the form of slope and rill erosion.Thus, watersheds with long lengths of overland flow have the potential for significantly increasing soil erosion and

land degradation.

Land Use and Land Cover

In both watersheds, on mountains with elevations >3000 m land cover was mainly forest with (Quercus sp.) and

(Rhododendron sp.). There were some pastures and meadows. In the mountains with elevations of 20003000 m

the forest type was (Betula alnoides) and////rakchan (Daphniphyllun himalense). There were some areas cultivated

with millet and buckwheat. In the mountains with elevations of 15002000 m the forest type mainly consisted of

////rakchan, //katus (Castanopsis indica) and ////chilaune (Schima walichii). The primary crops grown within the

watersheds were maize and millet (//bari) and rainfed rice (non-irrigated rice or

///ghaiya). Mountains with low

ridges and narrow valleys (10001500 m) were dominated by//katus and////chilaune forest together with maize, millet

and rainfed rice in cultivated areas. Valley bottoms were dominated by //katus, ////chilaune in association with //sal

(Shoria robasta) forest and cultivation of early variety of maize and irrigated rice (//khet ).




9/19

RESULTS

Land-Use and Land-Cover Changes

The land uses of watersheds have undergone significant alteration and transformations from 1978 (Figure 4) to 1996

(Figure 5). In general areas under forest and rice increased, whereas those under rainfed agriculture and shrub

decreased in both watersheds. Farmers abandoned substantial areas of marginal agricultural land in both watersheds.

The magnitude of abandoned land was found to be fourfold higher in the Mardi watershed than that in Fewa.

Abandoned lands had weed and shrub cover and were used for grazing and grass collection. The grazing and pasture

areas increased marginally in Fewa watershed whereas they remained unaltered in the Mardi watershed.

Land-cover change in both watersheds is shown in Figures 6 and 7. A decrease in cover class 1 and increase in

cover class 3 was observed in both watersheds. A slight increase in cover class 2 in the Mardi and cover Class 5 in

Fewa watersheds was also noted, while cover class 4 decreased marginally in the Mardi watershed but remained

unchanged in the Fewa watershed. From 1978 to 1996, the net land cover in the Mardi watershed increased by

more than twofold (24 per cent) of that of the Fewa watershed (11 per cent).

Internal Trading of Land Uses

Internal trading among different land uses was investigated in both watersheds and is illustrated in Figur 8a and b.

About 80 per cent of the net forest cover increase in the Mardi watershed occurred at the expense of shrub and

Figure 4. The area under different land uses in 1978 and 1996 in the two watersheds.




10/19

grazing land, while only 20 per cent of the increase was at the expense of agricultural land. In the Fewa watershed,

however, shrub and grazing areas contributed roughly half, while other land uses such as plantation, pasture and

barren, contributed the remaining proportion of the net increase in forest cover. Meanwhile a significant area of

forest (91 and 118 ha in the Mardi and Fewa watersheds respectively) was converted to agricultural land.

On the whole, internal trading between other land uses and shrub/grazing land was minor in the Mardi

watershed. In the Fewa watershed, however, 30 ha of shrub land changed to other land uses (plantation and pasture)

along with 110 ha from other land uses converted to shrubs/grazing land.

Distribution of Agricultural Land by Slope Category

In both watersheds, flat lands with less than 5 degrees slopes accounted for only a small fraction (6 per cent) of the

total agricultural area (Table II) whereas, large proportions of agricultural land (79 per cent in the Mardi watershed

and 88 per cent in the Fewa watershed) was observed to be on steep to very steep slopes (530 degrees).

Approximately 15 and 6 per cent of the agricultural land in the Mardi and Fewa watersheds, respectively, was

located on inclinations >30 degrees, despite a very high risk of soil erosion and landslides. The land use for crop

production on these steep slopes typically had suffered from extensive topsoil loss and nutrient depletion,

presumably leading to the observed abandonment of the cultivated land by farmers.

Hypsometric Curve and Hypsometric Integral

The concave shape of the hypsometric curves (Figure 9) and the low values of HI (Table III) indicate that both

watersheds are in the peneplain or at the deteriorating stage. This suggests that there has been significant incision,

downslope movement of topsoil and bedrock material and washout of the soil mass since their formation.

Drainage Density and Length of Overland flow

The results from the analysis of drainage density and length of overland flow revealed that the watersheds have low

drainage density with long length of overland flow (Table III). Thus the watersheds are subjected to high rates of

surface soil erosion during the high-intensity monsoon rainfall.

Figure 5. Change in land use from 1978 to 1996 in the two watersheds.




11/19

Cover Change with Respect to Terrain

Terrain analysis with respect to cover change revealed that terrain slope was the limiting factor for the latter.

Though the cover change spreads over a wide range of terrain slope, most of it had taken place either on steep

slopes or on gentle slopes. Moderate slopes between 10 and 20 degrees were found to be less affected (Table IV).

Of the increased forest cover, the steeper slopes had gained more than the gentle slopes in both watersheds.

DISCUSSION

Land-Use and Land-Cover Changes

As the data for forest conditions in Nepal before 1954 are scarce the period after could be divided into two phases.

Between 1950 and 1978 was a phase of agricultural expansion. During the 1950s Nepals forest resources were

estimated at 48 per cent of the total land area. Later reports showed a continuous decline of the forest cover (FRS,

1967; FRS, 1973; LRMP, 1986; MPFS, 1988). The forest area shrank to 273 per cent in 1996 (FRI, 1999). In 1957,the Forest Nationalization Act was enacted and all the non-cultivated lands were placed under the jurisdiction of

the Forest Department. This Act along with //birta (government registering the large areas of land in the name of a

person as an appreciation of ones contribution to the country) Abolition Act (1959) and Land Reform Act (1964)

formally reinforced the government ownership over the forest and in practice encouraged villagers to clear the

forest to maintain ownership of the land (Mahat et al., 1986). At the same time introduction of modern healthcare,

and a malaria suppression programme in the subtropical valleys and plains resulted in an unpredictable growth of

Figure 6. The area under different cover classes in 1978 and 1996 in the two watersheds.




12/19

population (Ives and Messerli, 1989), which led to increasing demand for fuelwood, timber, fodder and agricultural

land. The expansion first accelerated in the subtropical valleys, southern plains and marginal lands, until physical

and ecological constraints (bogs, fens, shallow soils and very steep slopes) were encountered, and the rapid

expansion also affected the forest; these are the main factors for the decline of the forest cover.

The second phase, after 1978, witnessed increasing government intervention for integrated land use and natural

resources management. Driving forces were the implementation of hill community forestry, integrated soil

conservation and watershed management programmes and an operational forest management plan. In 1976 the

////Panchayat (local village authority) forest regulations were passed and communities began to act on the basis of these

regulations (Fox, 1993). Afforestation efforts in the 1980s resulted in a significant increase in forest cover (Brown

and Shrestha, 2000) as line agencies, national and international non-governmental organizations, started working at

grassroots level. Consequently there was a decline in expansion of agricultural land and a total of 622 178 ha of

degraded forest was handed over to the community organizations for regeneration, protection and management

(Mahat et al., 1987; Kathmandu Post, 10 October 2001). Thus, due to community protection and management of the

forests that had been handed over, some thin forests had been converted to dense ones and degraded forests

regenerated to immature ones, indicating an increase in forest cover. This marginal increase results from the

implementation of the community forestry programme through the Fewa Watershed Management Programme in the

Fewa watershed and the Annapuran Conservation Area Programme in the Mardi watershed. In these watersheds,

degraded lands were handed to the community for protection and management and massive plantation was also

carried out in badly degraded lands during the 1980s to establish the vegetative cover (Fox, 1993).

These optimistic results are found to be consistent with findings of the similar studies in the hills of Nepal. Fox

(1993) mentioned a significant improvement in private and government forest in the mid-hill area of the Gorkha

District. He attributed this primarily to the introduction of community management of the forests. The studies

Figure 7. Change in tree cover from 1978 to 1996 in the two watersheds.




13/19

Figure 8. Land-use dynamics from 1978 to 1996 (a) Mardi watershed (b) Fewa watershed () sign indicates increase in land use whereas ()sign indicates decrease in land use type.




14/19

carried out by Gilmore and Nurse (1991) and Brown and Shrestha (2000) in the Jhikhu Khola catchments reported

increase in the forest cover largely associated with planting of trees in degraded lands and abandonment of

marginal terraces at low elevations. However, the results differ compared with the earlier studies of land-use

change in the plains and hills in the eastern region (Sah et al., 1997) and in the subtropical riparian corridors within

the Makalu Barun National Park and Conservation Area (Zomer et al., 2001) of Nepal where it has been shown that

land cover has decreased in their respective study areas.

Internal Trading of Land Uses

Internal trading among land-use categories in the watersheds are due to abandoning of shifting cultivation at the

higher elevations, converting of low-lying shrub lands into rice fields, increase of forest cover in community

protected forests, plantation on degraded private lands, clearing of low-lying forests for agricultural expansion and

timber harvest for fuelwood and timber for construction. A similar trend of internal trading among different land-

use categories over a similar time interval had also been reported in middle mountains and tropical riparian forest

of eastern Nepal (Tamrakar et al., 1991; Zomer et al., 2001).

Table II. Distribution of the agricultural area by slope category in Mardi and Fewa watersheds. Numbers in the parenthesesindicate the percentage of total agriculture land in the watersheds

Slope gradients Type of agricultural land Mardi watershed (ha) Fewa watershed (ha)

41//Khet 6 (02) 22 (04)

>1 and45//Khet 166 (53) 316 (59)

>5 and430

//Bari 1036 (328) 1758 (330)

//Khet 1474 (467) 2906 (546)

>30//Bari 313 (99) 182 (34)

//Khet 160 (51) 142 (27)

Total 3155 (100) 5326 (100)

Note://Khet irrigated rice;

//Bari rainfed maize and millet.

Figure 9. Hypsometric curves of Fewa and Mardi watersheds.




15/19

Table IV. Summary of cover change with respect to terrain from 1978 to 1996

Mardi watershed Fewa watershed

From To Area (ha) Slope range From To Area (ha) Slope range

(degree) (degree)

Forest//Bari 244 1525 Forest

//Bari 40 1525

//Khet 666 212

//Khet 744 111

Shrub 24 2530 Shrub 112 811Grazing 86 3045 Grazing 29 512

//Bari Forest 892 2565

//Bari Forest 651 2555

//Khet 194 412

//Khet 337 712

Shrub 386 2535 Shrub 276 2535Grazing 257 3040 Grazing 22 3038Abandoned 656 2045 Abandoned 97 2243Urban 17 1017 Plantation 273 548

Pasture 277 2339

//Khet Forest 601 112

//Khet Forest 524 812

//Bari 46 1222 Grazing 53 912Shrub 29 115 Abandoned 77 37Grazing 77 111 Plantation 103 812Abandoned 42 820 Pasture 22 711

Shrub Forest 1851 1533 Shrub Forest 662 2555

//Bari 25 1020

//Bari 49 1522

//Khet 03 15

//Khet 43 59

Grazing Grazing 46 3045Plantation 215 2057

Grazing Forest 118 2550 Grazing Forest 11 2648

//Bari 43 2035

//Bari 09 2028

//Khet 266 510

//Khet 296 110Plantation 125 1544Pasture 143 1953

Plantation Forest 457 2349

//Bari 64 2031

//Khet 135 611Grazing 997 1854

Barren Forest 215 135

//Khet 35 19

//Khet irrigated rice;

//Bari rainfed maize and millet.

Table III. Drainage density, length of overland flow and hypsometric integrals

Parameters Mardi watershed Fewa watershed

Length of drainage (L) km 485 412Total drainage area of watershed (A) km2 145 120

Drainage density Dd L

Akm1 33 34

Length of over land flow 12Dd

km 0152 0145Maximum elevation (m) 5590 2480Minimum elevation (m) 915 780Average elevation (m) 3135 1687Hypsometric integrals (HI) 047 053




16/19

Distribution of Agricultural Land by Slope Category

Not surprisingly, a large portion of agricultural land is located on steep to very steep slopes and it is practised under

the most difficult circumstances. Traditional crop farming is the main occupation and only source of livelihood for

the Himalayan mountain farmers, but the productvity of the crops is very low and is not able to meet more than 30

per cent of food grain requirement of the local people (Tiwari, 2000). The land holdings are of less than 1 ha and

are scattered over long distances on the mountain slopes. The irrigated land (30 degrees and the rest lies between 5 and 30

degrees slopes (Table II). A study carried out by Thapa and Weber (1995) in the upper Pokhara valley in Nepal, had

documented similar results and had stated that nearly 60 per cent of all agricultural land is ecologically unsuitable

for cultivation and about 22 per cent of that land has a slope gradient greater than 30 degrees.

Hypsometric Curve and Hypsometric Integral

The comparison between the two curves in Figure 9 shows only a marginal difference in mass removal from the

Mardi and Fewa watersheds. For centuries, the soil erosion from large drainage basins is derived primarily from

the incision of channel beds and cutting of stream banks. Topographic evidence shows landscape concavity where

river incision is active. The hypsometric integral shows the landmass volume remaining for the whole watershed

(Bishop et al., 2002) since the formation of drainage basin. Our analysis estimated hypsometric integrals of 0 47and 053 indicating only 47 and 53 per cent of original rock mass remaining in Mardi and Fewa watershedsrespectively (Table III).

Drainage Density and Length of Overland flow

Analysis of DEM supported by DSWC (1980) showed that watershed slopes approaches to 80 to 85 degrees in the

southern part of the Fewa and northern part of the Mardi watersheds. Like elsewhere in Nepal, geological erosion

occurs in the watersheds under study, but this is accelerated by anthropogenic activities such as deforestation of

steep slopes, overgrazing, unsound agricultural practices and concentrations of the surface run-off on trails.

Erosive rainfall (rain storms higher than 125 mm of rainfall or 625 mm of rainfall in 15 min) is very common inthe area (DSWC, 1980) and rainfall data for the years 1993, 1994 and 1996 from the Fewa watershed recording

station revealed 1025, 1050 and 950 mm of erosive rainfall respectively. Annual erosive rainfall in the area was

found to be much higher than 550 mm as was reported by DSWC (1980). About 47 and 39 per cent of the

watershed is under permanent cover in Mardi and Fewa watershed respectively. Hence, torrential rainstormscombined with steep slopes, long length of overland flow and shattered lithology accelerates the process of land

degradation (Ives, 1987; Ives and Messerli, 1989). The estimated annual soil erosion rates for the middle

mountains ranges from 08 to 70 Mgha1 in agricultural lands, from 20 to 273 Mg ha1 for grazing land and from29 to 180 Mg ha1 for shrub lands (Pahari, 1993; Shrestha and Zinck, 1999; Sah et al., 1997). The average annual

sediment contribution to the Fewa Lake from the watershed draining into it has been estimated to be as high as

15Mgha1 (Stapit and Balla, 1998).

The estimates of the soil loss (or net soil erosion) at the field scale, however, do not include the other component

of the sediment flux to a channel network, namely the part delivered by the geomorphic or gravitational process.

Over the time scale of decades many stream channels draining the alluvial, agricultural fields and other disturbed

watersheds become unstable and exhibit wide fluctuations in the rate of soil erosion, due to the formation of new

drainage systems (Osterkamp and Toy, 1997). The rate of soil erosion in the Himalayan mountains which are

geodynamically unstable and ecologically sensitive is proportional to steepness, drainage density and the slope

lengths (Ives, 1987). Drainage density varies from 3 to 1300 per unit area depending on climate and geology of the

region, and the low value of the drainage density corresponds to a landscape with an average long length of

overland flow (Berger and Entekhabi, 2001; Ritter, 1986)). The lengths of overland flow are potentially long in

watersheds of Nepal (DSWC, 1980) and our analysis of length of overland flow and drainage density also showed

long lengths of overland flow 152 and 145 m and low drainage densities 33 and 34 (Table III) for Mardi and Fewawatersheds respectively. Hence, the watersheds studied are vulnerable to soil erosion and land degradation.




17/19

Error Analysis

The process by which our special database was created from a source map was complex and errors of various types

may have been introduced at each step, such as error due to land-use classification, transformation, cursor

positioning, point selection and modelling the thematic layers created from the maps of different scales. Results

from analysis of digital data point to an insignificant increase in the percentage of forest cover in the study areas

between 1978 and 1996. Although we were able to detect distinct patterns of change in the percentage of land use

and forest cover between 1978 and 1996, it is clear that our estimates of the changes of forest cover and other land

uses were not error free. Hall et al. (1991) distinguished between two sources of error that may influence the

estimates of land-use changes derived from a comparison of aerial photographs taken at different time intervals: (a)

errors in the classification of individual aerial photograph (classification error) and (b) error due to spatial

mismatching (positional error). The magnitude of each source of error can be estimated using standard techniques

such as overall accuracy or the kappa statistics and the root mean square (RMS) error, respectively. Evaluating the

combined accuracy of these errors on the empirically derived estimates of land-use change is not simple (Kadmon

and Harari-Kremer, 1999).

The spatial mismatching of the photographs used for determining change in land use may lead to overestimation

of the land-use changes and to reduce this source of error, only changes of more than 10 per cent in the percentage

land use of particular land use or vegetation will then be considered as true changes (Callaway and Davis, 1993).

However, Hall et al. (1991) noted that uncertainty in the data used to estimate the land-use changes may lead to

two contrasting types of errors: (a) when, in fact there is change, an erroneous observation of no change and (b) an

erroneous observation of change when there is no change. Evaluating the combined effect of erroneous

classification and positional error on the accuracy of land-use changes from digital processing of the maps

derived from the interpretation of the aerial photograph is an important issue which is beyond the scope of this

paper and requires considerable further innovation and effort.

Management Strategy

It can be inferred from the results that these watersheds are undergoing continuous degradation. Some points can

be singled out here that could be helpful for minimizing the degradation and improving the watershed health. First,

it is necessary to control the population growth, so as to attain both economically and environmentally sustainable

development. Household economies in the Nepalese mountains are predominately based on subsistence

agriculture and require very high demand of labour to cultivate field crops, repair terraces, take care of livestock,

collect fuelwood and fodder (Mahat et al., 1987; Thapa and Weber, 1995; Brown and Shrestha, 2000). In designing

a population programme due consideration should be given to the alternatives that are favourable to reducing

agricultural labour demand and promoting locally viable non-farming employment opportunities, e.g. promotion

of bamboo-, fruit-, herbs- and beverage-based cottage industries and of tourism. It can be predicted that the

availability of adequate non-farming employment and income generating opportunities could help in adaptation of

family planning and effectively change the increase of population pressure on the land resources.

Second, as the current cropping practice requires regular ploughing and tilling of the soil on the steep lands,

which is one of the primary causes of land degradation, a change in land use especially from the current existing

subsistence farming system to tree-crop-based system is essential (Thapa, 1990; Thapa and Weber, 1995). The

focus on tree-crop-based activities (fruits, herbs, spices) would not only be effective to the environment, but it

would also generate wider income opportunities. More than 90 per cent of the agriculture is practised on the critical

slopes of greater than 5 degrees in both watersheds; 50 per cent of the watershed area is under intense agriculture

(DSWC, 1980), have acute problem of soil erosion due to loss of soil, especially during the monsoon (Thapa and

Weber, 1995). The DSWC (1980) had recommended that the areas between 22 and 30 degrees, 30 and 40 degrees

and >40 degrees slope gradients should be kept under silvipasture, horticulture and permanent protection forest,

respectively.

Third, terracing is essential for the slope gradients between 9 and 22 degrees. If implemented carefully, it would

not only reduce the vulnerability of soil erosion, but also help restoration of the watershed environment and boost

the household income as well.




18/19

Fourth, there is evidence in the form of furrows and soil erosion in barren areas due to grazing animals close to

settlements up to 35 Mg ha1 (Thapa and Weber, 1995). Free grazing should be restricted, encouraging or

enforcing better management, e.g. stall-feeding or rotational grazing.

Finally, attention should be given to further strengthening the local community organizations to manage the

scrub and grazing land more effectively. Weak institutions and government policies should be recognized as

important contributing factors to watershed degradation. Formulation and implementation of a comprehensive

strategy to meet the challenges of 21st century is essential.

CONCLUSION

Overall results of this study indicate that interpretation of historical aerial photographs may serve as an effective

tool for the detection and quantification of long-term patterns of land-use change dynamics and change in

geomorphological characteristics. The ability of GIS to integrate the digital maps of land use derived from

interpretation of temporal sequences of aerial photographs with DEM provides a new opportunity for analysing

patterns of long-term land-use dynamics with respect the terrain. The geomorphometric analyses revealed that the

watersheds had undergone severe erosion during the past and are susceptible to surface erosion and soil

degradation. Land-use and land-cover changes alone are not sufficient for explaining the watershed situation.

Thus, geomorphometric based analysis combined with land-use changes may be useful in the high rainfall areas

within which a wide variety of analytical results can be effectively integrated using GIS for explaining the currentsituation and future vulnerability of watershed to erosion and land degradation. Some suggestions for management

can be derived from this study.

acknowledgements

We gratefully acknowledge the financial support of the Research Council of Norway (through the Nepal Project

131692/730) for conducting the research work in Nepal. We are also grateful to Professor M. K. Balla and

Professor A. K. Das of the Institute of Forestry, Pokhara (IOF), Nepal, for their overall guidance and support for

conducting the research in Nepal. We would also like to acknowledge Mr Laxman Shrestha and Mr N. R.

Chapagain for their technical help.

references

Berger KP, Entekhabi D. 2001. Basin hydrologic response in relation to distributed physiographic descriptors and climate. Journal of Hydrology247: 169182.

Bishop MP, Shroder JF, Bonk R, Olsenholler J. 2002. Geomorphic change in high mountains: a western Himalayan prespective. Global andPlanetary Change 32: 311329.

Borrough PA, McDonnell RA. 1998. Principles of Geographical Information System. Oxford University Press: Oxford.Brown S, Shrestha B. 2000. Market-driven land-use dynamics in the middle mountain of Nepal. Journal of Environmental Management59:

217225.Callaway RM, Davis FW. 1993. Vegetation dynamics, fire, and the physical environment in the coastal central California. Ecology 74: 1567

1578.Chen L, Wang J, Fu B, Qiu Y. 2001. Land-use change in small catchment of northern Loess plateau, China. Agriculture, Ecosystems and

Environment 86: 63172.Clarke ML, Rendell HM. 2000. The impact of the farming practice of remodeling hillslope topography on badland morphology and soil erosion

processes. Catena 40: 229250.DSWC. 1980. Management plan for integrated development of Phewa watershed. FED/NEP/74/020 Field Document: Fewa Tal Technical

Document No. 2. Ministry of Forest, Department of Soil Conservation and Watershed Management: Kathmandu.ESRI. 1996. Using the ArcView Spatial Analyst. Environmental Systems Research Inc.: Redlands, CA.Fox J. 1993. Forest rsources in a Nepali village in 1980 and 1990: the positive influence of population growth. Mountain Research and

Development13: 8998.FRI. 1999. Forest resources inventory. Ministry of Forest and Soil Conservation, His Majestys Government of Nepal: Kathmandu.FRS. 1967. Statistics of Terai and Adjoining Regions. His Majestys Government of Nepal, Department of Forestry, Forest Resources Survey:

Kathmandu.FRS. 1973. Statistics for the Hill Regions. His Majestys Government of Nepal, Department of Forestry, Forest Resources Survey: Kathmandu.




19/19

Gabriels D. 1999. The effect of slope length on the amount and size distribution of the eroded silt loam soils: short slope laboratory experimenton inter rill erosion. Geomorphology 28: 169172.

Gilmore DA, Nurse MC. 1991. Farmer initiatives in increasing the tree cover in central Nepal. Mountain Research and Development11: 23292337.

Hall FG, Botkin DB, Strebel DE, Woods KD, Goetz SJ. 1991. Large-scale patterns of forest succession as determined by remote sensing.Ecology 72: 628640.

Hurtrez JE, Sol C, Lucazeau F. 1999. Effect of drainage area on the hypsometry from an analysis of small-scale drainage basins in the Siwalikhills (central Nepal). Earth Surface Process and Landforms 24: 799808.

Ives JD. 1987. The theory of Himalayan environmental degradationits validity and application challenged by recent research. MountainResearch and Development7: 189199.

Ives JD, Messerli B. 1989. The Theory of Himalayan Environmental Degradation. What is the Nature of the Perceived Crisis? The HimalayanDilemma Reconciling Development and Conservation. Routledge: New York.

Kadmon R, Harari-Kremer R. 1999. Studying long-term vegetation dynamics using digital processing of historical aerial photographs. RemoteSensing Environment68: 164176.

Kathmandu Post. 10 October 2001. A daily newspaper. Kantipur Publications: Kathmandu.Lambin EF, Rounsevell MDA, Geist HJ. 2000. Are current agriculture land use models able to predict changes in land-use intensity?

Agriculture, Ecosystem and Environment82: 321331.Lambin EF. 1997. Modeling and monitoring land-cover change processes in tropical region. Progress in Physical Geography 21: 375393.LRMP. 1986. Land Resources Mapping Project. Kenting Earth Science: Kathmandu.Mahat TBS, Griffin DM, Shepard KR. 1986. Human impact on some forest of the Middle Hills of Nepal. Part1. Forests in the context of the

traditional resources of the state. Mountain Research and Development6: 223232.Mahat TBS, Griffin DM, Shepard KR. 1987. Human impact on some forest of the Middle Hills of Nepal. Part 3. Forests in the subsistence

economy in the Sidhu Palchowk and Kabhre Palanchok. Mountain Research and Development7: 5370.MPFS. 1988. Master Plan for Forestry Sector Nepal. Ministry of Forest and Soil Conservation: Kathmandu.

Osterkamp WR, Toy TJ. 1997. Geomorphic considerations for erosion predictions. Environmental Geology 29: 152157.Pahari KJ. 1993. Soil Erosion Susceptibility Using Remote Sensing and GIS: A Case Study of Andikhola Watershed Nepal . MSc dissertation NR

9316. Asian Institute of Technology: Bangkok.Ritter DF. 1986. Process Geomorphology. Brown: Dubuque, IA.Roth G, La Barbara P. 1997. Morphological characterization of channel initiation. Physics and Chemistry of the Earth 22: 329332.Sah BP, Honda K, Murai S. 1997. Sub watershed prioritization for watershed management using remote sensing and GIS. In Proceedings, 18th

Asian Conference on Remote Sensing, October 1997, Malaysia, ACRS: Kuala Lumpur; B-2-1B-2-6.Schneider LC, Pontius RG. 2001. Modeling land-use change in the Ipswich watershed, Massachusetts, USA. Agriculture, Ecosystems and

Environment 85: 8394.Sidhu DS, Das TH, Singh RS, Sharma PK, Jadaun SPS, Ravishankar T. 2000. Soil resource inventory of watershed for land use planning: a GIS

approach. Indian Journal of Soil Conservation 128: 2229.Shrestha DP, Zink JA. 1999. Land degradation assessment Using GIS. A case study in the middle mountain region of the Nepalese Himalaya.

Proceedings of the International Conference: Geoinformatics-Beyond 2000. http://pages.hotbot.com/edu/geoinformatics/f64.htmlStapit KM, Balla MK. 1998. Sedimentation monitoring of Fewa Lake. Project PD 103/90 Rev.1 (F). Institute of Forestry/International Tropical

Timber Organization: Pokhara.Tiwari PC. 2000. Land-use changes in Himalaya and their impact on the plains ecosystem: need for sustainable land use. Land Use Policy 17:

101111.Thapa GB. 1990. Integrated Watershed Management in the Upper Pokhara Valley Nepal. Doctoral thesis. Asian Institute of Technology:

Bangkok.Thapa GB, Weber KE. 1995. Natural resources degradation in small watersheds of Nepal. Natural Resources Forum 19: 285296.Tamrakar RM, Jabeju K, Shrestha B. 1991. Land use changes in the Jhikhu Khola Watershed area. Soil Fertility and Erosion Issue in the Middle

Mountains of Nepal, Shah PB, Schreier H, Brown SJ, Riley KW (eds). Workshop Proceedings, April 2225, ICIMOD: Kathmandu.Trapp H, Mool P. 1996. Lamjung District Information System for Local Planning and Assessment of Natural Resources Using GIS and Remote

Sensing Technology. MENRIS Case Study Series No. 4. ICIMOD: Kathmandu.Tucker GE, Catani F, Rinaldo A, Bras RL. 2001. Statistical analysis of drainage density from digital terrain data. Geomorphology 36: 187202.Wischmeier WH, Smith DD. 1978. Predicting the Rainfall Erosion Losses From Cropland East of the Rockey Mountains. Agriculture Handbook

No. 282. US Department of Agriculture: Washington, DC; 47.Zomer RJ, Ustin SL, Carpenter CC. 2001. Land cover change along the subtropical riparian corridors within the Makalu Barun National Park

and Conservation Area, Nepal. Mountain Research and Development21: 175183.



awasthi (2002) land-use change in two nepalese watersheds-gis and geomorphometric analysis

Documents