GISdevelopment.net ---> AARS ---> ACRS 1996 ---> Application of RS and GIS to Land Degradation

Assessment of watershed degradation and its socioeconomic impacts using Remote Sensing and GIS: A case study of Trijuga watershed, Nepal

Bhuwneshwar P. Sah1, Shunji Murai2, Kiyoshi Honda1, Karl E. Weber1, Haja Andrianasolo1
12SAR/SERD, Asian Institute of Technology
PO Box 4, Klongluang, Pathumthani 12120, Thailand
2Institute of Industrial Science, University of Tokyo
7-22-1 Roppongi, Minato-ku, Tokyo 106, Japan


Abstract:
This paper is based on a research to develop and test a methodology for assessing the watershed resources degradation over time and seeking its socioeconomic impacts. Universal Soil Loss Equation in conjunction with Remote Sensing and GIS had been utilized for resources monitoring, while household survey had been conducted for socioeconomic status assessment. The land use change had been exceeded the permissible limit along with 44 percent increment in soil erosion Rae between 1978 to 1991 and. The analysis of sensitivity and socioeconomic status had been found strongly correlated with resource degradation. A multiple linear regression model has been developed from these parameters which can be used to simulate the resource degradation speed under the various socioeconomic conditions. The study concluded that, development activities should be concentrated in valley while conservation activities should be focused to the hills, by considering and formulating land use plan.

1. Introduction:
The utilization of a watershed area beyond its carrying capacity to provide food, fiber, and shelter for the exploding population has resulted in its deterioration in most part of the world (FAO, 1985). However, such deterioration is more severe in developing countries including Nepal (Thapa and Weber, 1990). Being an integral part, the natural resources and socioeconomic status of a watershed should be paid equal attention (Erickon, 1995). Unfortunately until now most of the people are confined only to the resources degradation, keeping the social factors aside.

Watershed degradation is a phenomena by which the potentiality of the watershed is getting reduced over time, which can be confined to the forest loss and the rate of soil erosion increment, if other factors are negligible (Kelly, 1983). These resources can be monitored by using Remote Sensing (RS) and Geographic Information Systems (GIS) in conjunction with Universal Soil Loss Equation (USLE). For the assessment of the socioeconomic conditions, household survey along with other ancillary data can be used. These two aspects can be correlated for the better understanding of the degradation phenomena of the watershed. Under this context, the present study was carried out with the objectives (i) to analyze the locational sensitivity based on resources monitoring. (ii) to establish the relationship between resources monitoring and socioeconomic status and evaluate the applicability of RS and GIS for this purpose (iii) to allocate the suitable zones in the watershed for their specific use and management purpose.

2. The Study Area:
The study area, encompassing 732 sq. km., lies between 26o 42’ and 26o 59’ N latitude and 86o 33’ 46” and 86o 59’ 48” E longitude in the Eastern Region of Nepal (Map. 1). The altitude varies from 75 m to 1942 m. The tropical climate of the low-lying valley gradually passes into the sub-tropical towards higher elevation; north. The average annual temperature is 20oC (WEC, 1982) with 1942 mm rain fall. More than 70 percent rainfall is concentrated from may to October. The forest cover is nearly 58 percent area which is dominated by tropical Sal (Shorea robusta) forest, followed by 24 percent area of agriculture. Agriculture along with livestock are the important source of income and livelihood the of the people. Population per ha. Arable land comes to be 4.53 which is some what lowever as compared to the Tarai region of the Nepal.


Map 1. Location of the Study Area

3. Methodology:
An integrated approach of digital image processing of satellite data and visual interpretation of airphoto combined with GIS and USLE was carried out for resources assessment. Household questionnaire survey was conducted for the socioeconomic status assessment. The ancillary data were used where ever it was relevant.

3.1 Collection of Data and Assessing Resources:
  1. Socioeconomic survey: out of 18,000 household, 113 were interviewed with well structured questionnaire. The sample size was calculated as described by Cochran (1977).
  2. Physiographic survey: 35soil samples were taken for soil fertility and texture analysis.
  3. Satellite dataL Landsat TM of path/row 140/041, acquired on 21 Dec: 1991 and MSS of April 1984.
  4. Airphotos: Scale 1:50,000, Nov., 1978
  5. Reconnaissance survey maps: Topographic map: 1:25,000, 1995 and 1:63,360, 1958, Land use map: 1:50,000, 1982, Land system map: 1:50,000, 1984, and Political map: 1:250,000, 1987
By using these data, the general methodology was followed as presented in Fig. 1. The socioeconomic survey was done in three strata, which were described as Hills (High altitude), Midlands (Medium altitude) and Valley (Low latitude). Land use map of 1978 was obtained by using visual interpretation of airphoto, while surpervised digital image processing was adopted for 1984 and 1991 satellite data. After comparison of land use of 198, 1984 and 1991, the change trends had been obtained.



Fig 1 General Methodology

Widely adopted USLE model was taken to estimate the soil loss (Schawab et.al., 1993). The equation is written is written as

E=RKLSCP

Where,
E= Mean annual soil loss (tons/ha/yr.), R= Rainfall erosivity index, K= Soil erodibility, L = Slope length, S = Slope steepness, C= Crop management and vegetation cover, P = Erosion control practice factor,

From the average annual rainfall (1942 mm) and the maximum 30 minute intensity (100 mm) of the year 1991, the R value was calculated (Forster, 1981 and Morgan, 1986). The K factor for 20 soil units and the P factor for different land use were determined by using Schawab et al. (1993). The slope length for different land use were adopted from the DSCWM/HMG, Nepal 1992, while the slope inclination factor was determined by using the Digital Terrain Model (DTM) which was interpolated after digitizing ht 20 meter contour interval lines. The C value was adopted from suggested by Morgan, (1986). The layers of USLE model were created and integrated with the help of GIS (Fig. 2). The soil samples were analyzed for soil nutrients, texture and permeability at RARS, Tarahara, Sunsari, Nepal. As the socioeconomic study involves both qualities and quantitative information, both descriptive as well as analytical statistics measures had been used. Furthermore, weight age index were also formulated where ever essential.



Fig 2 USLE Model in conjuction with RS & GS to Estimate Soil Erosion

4. Result and Discussion:

4.1 Resource Monitoring:

Overall land cover change for the duration of 13 years’ between 1978 and 1991, is given in Table 1. For the management of natural resources, the land cover change should be less than 0.1 per cent per year (Murai 1993, citied) in Pahari, 1993) on sustainability basis. The rate of forest degradation of study area was 0.57 per cent per year, and is too high for sustainable use of resources. Although the rate of soil erosion was the highest for shrubs land the contribution was maximum (54% during year 1991) from agriculture land. The temporal variation of soil erosion rate from different land use, may be attributed to the change in spatial location of land (Table 3). (Map 2 and 3)


Table 1: Land Use Change
Land cover 1978 % 1984 % 1991 % Time interval % change
1978-84 1984-91 1978-91
Shrubs 2.27 4.08 3.45 1.81 -0.63 1.18
Degraded forest 4.19 12.49 8.90 8.30 -3.59 4.71
Forest 65.21 57.31 57.86 -7.90 0.55 -7.35
Agriculture 22.49 21.11 23.66 -1.38 2.55 1.17
River 5.84 5.00 6.14 -0.83 1.13 0.30
Total 100.0 100.0 100.0 0.00 0.00 0.00


Map 2. Soil Erosion Map Year 1991


Map 3. Land Use Map Year (Derived from Landsat TM, 1991)

4.2 Socioeconomic Status:
The several socioeconomic factors were analyzed and few of them were taken for the location comparisons of the socioeconomic status. Table 3 shows the midlands is in better condition as compared to Hills and Valley.


Table 2: Soil Erosion form Different Land use
Land use Ave. soil loss of year (t/ha/yr) %contribution to soil loss
1978 1984 1991 1978 1987 1991
Shrubs 123.36 166.16 180.24 22.47 42.41 34.61
Degraded forest 10.96 6.01 10.53 3.68 4.70 5.21
Forest 1.72 2.01 1.85 8.99 7.20 5.96
Agriculture 35.99 34.63 41.20 64.86 45.69 54.21
River 0.00 0.00 0.00 0.00 0.00 0.00
Total 12.48 15.99 17.98 100.00 100.00 100.00


where, x = Locational and y = Total value and c = constant which is calculated as the ε y/y after assigning the – or + (by considering the negative or positive impacts upon resources) value to the indication.

4.3 Impacts:
The average lad holding (LH) and herd size (LSU) of Udaypur district was 1.6 ha/HH and 4.46 LSU/HH respectively during 1981 (CBS 1993). The decreasing size of land holding with increasing size of herd is not friendly for forest. (fig. 3). Removal of top soil resulted in depletion of soil nutrients with decrease in crop productivity (fig.4). The average loss of top soil from agricultural field is estimate to 3.05 mm/ha/yr. It is quite higher rate than tolerance level; 1.5 mm/ha/yr.



Fig 3 Impacts of Social Factors on Watershed



Fig 4 Soil Loss and crop Productivity


Table 3 Locational Socioeconomic Status
Indicator Units Locations Index
Valley Mdand Hills Total Valley Madand Hills
Mgation (Net in) % 60 55 5 42 -1.43 -1.31 -0.12
Landholding Ha 1.06 1.81 0.86 1.27 0.83 1.43 0.63
Cooping intensity % 123 114.49 178 136.6 0.89 0.83 1.28
LSU Patio 3.92 4.96 4.86 4.57 -0.86 -1.09 -1.06
HH Size No 5.58 6.33 6.76 5.92 -0.94 -1.07 -1.14
Education
Male
Female
 
Index 40 40 30 40 1.00 1.00 0.75
Index 20 20 10 20 1.00 1.00 0.50
Firewood consumption % 4.1 3.73 6 4.61 -0.89 -0.81 -1.30
Soil fertility Amount 384 538.49 540 487.50 0.79 1.10 1.11
Occupation
Agricultural activities
Other activities
 
%popu 53 48 56 52 -1.02 -0.92 -10.08
%popu 47 52 44 48 0.98 1.08 0.92
crop productivity t/ha 4.81 4.72 6 5.18 0.93 0.91 1.16
Fodder supply
>60% fram forest
>60% fram farmfand
 
%HH 80 56 93 75 -1.07 -0.75 -1.24
%HH 20 44 7 25 0.80 1.75 0.28
Gross cash income Rs. “000” 13 22.1 11.5 15.5 0.84 1.43 0.74
Gross cash expenses Rs “000” 10.5 20.4 6.7 12.5 0.84 1.63 0.54
Total Sodoeconamic Status index 2.69 6.23 2.01

4.4 Integration of Resource Monitoring and Socioeconomic Status:
the integration of resources monitoring and socioeconomic status has been done by formulating the indexes of socioeconomic status (see Tab 3), resources monitoring and the sensitivity analysis. Degradation and sensitivity indexes were calculated for the locations to compare with its socioeconomic status (Table 4). Finally, these indexes were used to formulate the Degradation Speed Model (DSM).


Table 4: Locational Resources Status
Locations Forest cover (%) Soil loss (t/ha/yr) %Contribution to soil loss Dl Sl
year 1978 1991 Change 1978 change 1991 1978 1991 Change
Hedya(Vally) 70.17 62.22 7.95 5.53 7.55 2.02 3.96 4.11 0.15 3.33 0.25
Beltar(Mdland) 20.20 15.58 4.62 13.11 15.05 1.94 2.86 3.65 0.80 2.46 0.42
Jalpedhi(Hill) 46.44 43.34 3.10 62.07 70.21 8.14 15.65 17.08 1.43 4.95 2.63

The sensitivity analysis (Fig.5) shows that the hills is more sensitive as slight loss of forest produced tremendous amount of soil.



Fig 5 Locational Sensitivity of Soil Loss

From Table 3 and 4, Fig. 6 shows the negative impact of resources degradation on socioeconomic conditions while Fig.7 sensitive area degraded more and vice versa.



Fig 6 Resources Degradation and Social Status



Fig 7 Resources Degradation Vs Sensitivity

4.5 Prediction of Degradation speed:
From above discussion, the degradation speed is the function of sensitivity and socioeconomic conditions of the watershed. In other words, Degradation speed = f(Sensitivity & Socioeconomic Status)

Thus,

z = ax + by + c

Where,
z = Degradation Speed, x = Sensitivity, y = Socioeconomic Status, a and b are x and y coefficient respectively and c = Constant

By using index value of z,x and y, multiple regression has been done to predict degradation speed (Fig. 8).



Fig 8 Linear Regression Model for Degradation Speed (DSM)

The value of x and y coefficient are 0.60 and –0.27 respectively while the constant value is obtained as 3.92. Thus the degradation speed model can be written as

z = 0.6*x – 0.27*y + 3.92

The negative coefficient of socioeconomic status and the positive coefficient of sensitivity reveals their specific relationship with degradation speed; degradation speed is inversely proportional to the socioeconomic status. and directly proportional to the sensitivity of locations. This model can be utilized to simulate the degradation speed of the watershed with different level of socioeconomic conditions

4.6 Locational Recommendations for land use planning:
Crop and livestock breed improvement, energy use efficiency, female education, forest preservation awareness with community forestry and family planning/mother care the general recommendations for all locations. The specific locational recommendations are given in Table 5.


Table 5: Locational Recommendations for proper Land Use
Field Activities
Valley Midlands Hills
Agriculture Crop Introduction of wheat.
Cash crop; jute., Legumes;
Improve cropping intensity
Introduction of wheat.
Cash crop pulse/mustard
Improve cropping intensity
Discouraging cereal crops
Cash crop: cardamom, pulses/mustard
Vegetable Fruits & Trees Kitchen gardening
Mango orchard
Commercial/Kitchen gardening
Mango orchard
Kitchen gardening
Tejpat
Mango/jackfruit/citrus/litchi
Commercial orchards
Irrigation Community canal
Land filling control
Community canal
Land filling control
Erosion control
Agro forestry Fire wood and fodder tree with Cereals Fire wood and fodder tree with cereals Forest tree and cardamom
Combination of fruit tree &frit plant.
Fodder grass, fodder tree and cereals
Soil management Composting
Green manuring
Chemical fertilizer.
Composting
Chemical fertilizer
Composting
Livestock: Cow& Buffalo
Establishment of milk collection center, Poultry, Fishery.
Commercial chicken and pig farming Gradual reduction of cattle and buffalo Discouraging cow/buffalo
Commercial pig/goat farming
Encouraging poultry rearing
Feeding system Feed Stall and open grazing
Concentrate.
Urea treated wheat straw.
Agricultural by products.
Stall feeding
Concentrate
Urea treated wheat straw
Agricultural by products
Stall and open grazing
Fodder tree & grass
Agricultural by products
Urea treated wheat straw
Land cover management Skukumbasi settlement.
Thinning productive forest.
Plantation of encroached forest.
Allocation of forest land, for settlement of hills area people.
Fodder tree and grass
Checking illegal logging
Shifting the Shukumbasi to valley
Increment of forest cover by 1 Percent /year
Checking of forest loss
Forest fire control program
Replantation degraded forest
Range management for shrubs and Grazing land
Place crop<20% slope
Place fruit>20% slope
Shift some HH of valley
Conservation: Flood control
River bank erosion checking
Maintaining good crop converge
Flood control
River bank erosion control
Gully establishment
Terracing
Good crop cover
Gully control
Terracing agricultural land
Safe disposal
Rehabiliation
Plantation in sensitive areas
Landslide control
Others: Establishment of forest based Industries. Prohibition immigration
Cottage industries
Handicraft industries
Employment in Quarry

5. Conclusion:
Assessment of resources degradation, and finding out the socioeconomic conditions in a watershed, are essential factors for proper understanding and planning for watershed conservation and development purposes. Remote Sensing and GIS in combination with USLE model, are found to be appropriate tools for resources monitoring, while household questionnaire surveys yielded better results of socioeconomic status. The analysis of sensitivity and socioeconomic status have been found strongly correlated with resource degradation in this study. Highly sensitive areas suffered from speedy resource degradation, which resulted from poor socioeconomic status. Again, poor socioeconomic status. Again, poor socioeconomic conditions resulted into speedy resource degradation which may be irrespective of sensitivity. A multiple linear regression model has been developed from these parameters which can be used to simulate the resource degradation speed under the various socioeconomic conditions.

6. References:
  • Central Bureau of Statistics, CBS (1993), Statistical Year Book of Nepal, National Planning Commission, Kathmandu.
  • Cochran, W.G. (1977) Sampling Technique, 3rd edition, Jhon Wiley & Sons, NY.
  • DSCWM, (1992), Soil conservation and Watershed Management Operation Plan for Subwatersheds of Palpa District, DSCWM/HMG, Nepal.
  • Erickson, D.L. (1995), Rural Land Use and Land Cover Change in Land Use Policy, Vol. 12, No. 3, Elsevier Science Ltd. UK
  • FAO (1995), Tropical Forestry Action Plan, Committee on Forest Development in the Tropics, FAO, UN, Rome.
  • Foster, G.R. (1988), Modeling Soil Ersosion and Sediment Yield, In oil Erosion Research Methods, Edited by R. Lal, ISSS, The Netherlands.
  • Kelly H.W. (1983), Keeping the Land Alive. Soil Erosion-its Causes and Cures, FAO Soil Bulletin No. 55, FAO, Rome.
  • Morgan, R.P.C. (1986), soil erosion and Conservation, Longman group UK Ltd., London. Pahari, K.J. (1993), Soil Erosion Susceptibility by Using Remote Sensing and GIS: A Case Study of Andhikhola Watershed. Nepal, AIT Thesis No. NR-93-16, AIT, Bangkok.
  • Schawab, G.O., Fangmeier, D.D., Elliot, W.J. and R.K. Frevert (1993), Soil and Water Conservation Engineering, 4th edition, John wiley & Sons, Inc., NY.
  • Thapa, G.B. & K.E. Weber (1991), Soil Erosion in Developing Countries: Causes, Policies and Programs, HSD working paper 35, AIT, Bangkok.
  • Water and Energy Commission (WEC) (1982), Hydrological Studies of Nepal, Vol. I & II, WEC, Ministry of Water Resource, HMG, Nepal.
  • Wishmeier, W.H. and D.D. Smith (1978), Predicting Rainfall Erosion Losses A Guide to conservation Planning, USDA Handbook 537, Washington.