| GISdevelopment.net ---> AARS ---> ACRS 1989 ---> Agriculture & Soil |
Soil erosion mapping using
Remote Sensing data and Geographic Information System
Manu Omakupt
Land use planning division
Land development department
Ministry of Agriculture and Cooperatives
Manu Omakupt
Land use planning division
Land development department
Ministry of Agriculture and Cooperatives
Abstract
The study area coverage an area of 60 km x 60 km in the chiang mai basin and the surrounding hilly area in the northern Thailand.
The method used consists of interpretation of SPOT and Land sat data which enables units characteristics by combination geographical units and land use pattern .Field measurements of universal soil loss equation factor field measurements of universal soil erodibility slope steepness and length vegetation cover and conservation practice allowed the productions of an erosion map. the integration of these five erosion factors was performed on the spatial analysis system data base .Digital terrain model DTM and satellite image data were electronically transferred from the MERIDIAN image processing system . The processing and or interpretation work involved to derive each factor and the final soil erosion susceptibility map are produced.
The study showed that maximum erosion risk as expected occurs in the mountains regions where the forest has been cleared to make place for the cultivations of field crops all tree plantation mostly pine control practices as many exbhit a seeding pattern parallel to the directions of the slope that a measure that is very prone to erosion.
Introduction
Thailand is like many other developing countries situated ina tropical environment the problem of all soil erosion is of major concern from study carried area out in the 1985 by the bank it was estimated that 39% of the total area of designed forest land had been encroached deforestation in hilly areas where the risks of soil erosion are high has reached disturbing proportions even only in the last decade.
The sixth national plan emphasizes the importance of the proper utilization of renewable natural resources .Among the numbers agencies involved in the with the aspects related to the development of land resources including land use planning soil and water conversation and land improvement..
LDD is responsible for identifying areas requiring programmes of soil conservation and processing the measures to implement these being aware of this the mapping of erosion susceptibility is a suitable information for soil and water conservations work.
Back ground
Remote sensing techniques have been used as an aid soil survey Thailand since the early most of the research dealing with the use the satellite images for erosion studies has been done in the context of soil surveying a well Known parametric equation the universal soil loss equation is currently used to assess soil loss due to water erosion. It is written as:
This equation explained inn detail is Wish Meier and smith describes soil loss a as a function of rain fall R its amount and intensity soil erodibility related to the texture percentage organic matter structure and permeability of the soil topography especially the slope gradient and the length of the slope the vegetative cover and the erosion control practices (p).
Because of its simplicity the relative ease of evaluating each factor and because in most parts of the world it gives acceptable results a number of authors have used this equation to evaluate soil loss due to water erosion easily adapted to many different scales it has been used by arnoldus for morocco E1 scarify in Hawaii say ago for a region in Argentina brail et al for a small sector in the province of Quebec such as the food and agricultural organization and the land directorate of the Canadian Ministry of Environment.
Figure.1. Soil Erosion susceptibility methodology
Methodology
The study area is located in the chiang Mai region in northern Thailand it corresponds to the area covered the multi band Spot scene acquired on Jan 11 1997 and overlaps six map sheets at the scale of 1:50,000.
The five parameters of the Equation were evaluated for the study area using remote sensing ground observation and existing map data the SPOT multi band image was edged enhanced contract stretched and enlarged at the scale of 1:100.000 to be used to map land use land cover and to derive crop management factor and erosion control practice factor.
Existing maps and image interpretation over lays w were digitized and integrated in to the spatial analysis system data base and integrated digital terrain model DTM and satellite image data were electronically transferred from the MERIDIAN image processing system the processing and or interpretation work involved to derive each factor and the final soil erosion susceptibility map are described below.
Figure.2. Rainfall factor (R)
Rainfall factor
A number of rain fall erosion indicates exist but most require an enormous amount of data and considerable calculation time in many developing countries where limited amounts of precipitation data are available an index such as the modifies Fournier's index can be used to give satisfactory results this index takes in to account mean monthly or mean annual precipitation.
Data records were taken from 19 meteorological stations distribution in and around the study area were brought in to a regression analysis which provided the mean annual rainfall this relationships defined by srikhajorn et al 1984 between the amount of rainfall recorded at the station and the elevation was used to extrapolate the mean annual rain fall over generated area using a digital elevation model the DTM was generated by digitizing the contour lines on the topographic base map at the scale of 1;250,000 and interpolated to a 100 meter grid cell. The rainfall index was by using a linear regresion generally expressed as y=a+bx or in our case:
The rainfall factor r was formulated by Srikhajon M. et al. from the analysis of the amount and intensity of rainfall using the KE > 1 index (Hudson, 1971). The kinetic energy of erosive rainfall (i.e. rainfall at intensities greater than 25 m/hr ) was calculated by using Wischmeier and mith ( 1978 ) energy equation Y = 210.3 + 89log10 I where Y is the kinetic energy of rain and I is the rainfall intensity measured from 6 stations. The intensity (a) equation is:
Soil erodability
The 1:100,000 soils map for the mail and land provinces geological maps and geomorphologic maps were used to derive this factor a geo unit or gynecological units was produced combining soils geology and geomorphology units was produced with a span system and served as a basis of the system assignment .This map exemplified the ability of the system assignment .this map to integrate various data sets the detailed soil map information with attributes provided information for the chiang Mai valley up lands soils units derived from the geological and geomorphologic maps. there joined together to form the geo units a nomogram developed by Wish Meier et al was used to obtain the value of the K factor on the basis of the percentages of silt very fine sand and % organic matter spoil structure and permeability field observations were gathered in some cases when information on the soil profile was in sufficient the resulting soil erdilibility is shown in figure 3.
Figure.3. Soil erodibility factor (K)
Slope factor ( LS)
The Slope gradient was digitally from thye digitized 1:250,000 scale map. Five slope gradient classes were defined as follows:
| Slope class | Gradient |
| 1 | 0-2% |
| 2 | 3-5% |
| 3 | 6-12% |
| 4 | 13-35% |
| 5 | > 35%. |
The slope length was estimated according to the geologic and geomorphologic units in the study area. A mean slope used for each units as follows:
| Geologic Geomorphology units | Length (M) |
| Alluvium flood plan (Qa) | 50 |
| Terraces and alluval fans (QT)
-QT Middle terrace -QT3 low terrace | 40 75 100 |
| Undivided sand stone and shale ( P, C, Trh, CP ) | 100 |
| Limestone ( o ) | 100 |
| Quartzite (SD) | 200 |
| Gneiss, Shist, marble ( PE ) | 150 |
| Granite (Trgr, Cgr ) | 100 |
The slope factor which is combination of slope gradient and slope length are presented in the following matrix and computed using LS = Öl ( 0.0138 + 0.0095s + 0.00138s ) when l is slope length (m) and S = Slope stepness (%). Figure 4. shows the slope factor map.
Figure.4. Slope Factor (LS)
| Length (m) | 0-2 % | 3-5 % | 6-12% | 13-35% | >35% |
| 40 | .24 | .61 | 2.08 | 12.86 | 24.96 |
| 50 | .27 | .68 | 2.32 | 14.40 | 27.90 |
| 75 | .33 | .83 | 2.84 | 17.64 | 34.17 |
| 100 | .36 | .96 | 3.28 | 18.87 | 39.46 |
| 150 | .46 | 1.17 | 4.02 | 20.37 | 48.33 |
| 200 | .54 | 1.36 | 4.64 | 21.80 | 55.81 |
Vegetative cover (C)
Information relating to the vegetative cover was combined from a multi temporal Land sat MSS band ratio 1988 published in this forest cover maps produced from supervised classifications of MSS images plala 1988 also publishes in this symposium and existing land use map. The resulting map is shown in fig 5.
Figure.5. Vegetative cover (C)
Erosion control practice factor (P)
Erosion control practices factor like the vegetation cover factor was derived from the land use and vegetation cover maps as described in the table 1 fig 6 shows the erosion control practice factor map.
Figure.6. Erosion control practice factor (p)
Computation of soil loss using USLE
The simple modeling function defined as a product of each factor was developing and applied to obtain the final soil erosion values are given un ton bar per year.
| Land use and vegetation cover classes | C factor | P factor |
| U 1 City and town | 0 | 0 |
| U 2 Village with orchard (longan) | 0 | 0 |
| U 3 transformational land | 0 | 0 |
| U 4 Industrial 1 land | 0 | 0 |
| A 1.1 paddy one crop a year (rainfed) | 0.28 | 0.28 |
| A 1.2 1st crop -paddy / 2nd bean (irrigat) | 0.28 | 0.28 |
| A 1.3 1st crop -paddy / 2nd bean (soy bean) | 0.28 | 0.28 |
| A 1.4 1st crop -paddy / 2nd garlic & onion | 0.28 | 0.28 |
| A 1.5 1st crop -paddy / 2nd tobacco | 0.28 | 0.28 |
| A 1.6 1st crop -paddy / 2nd truck crops | 0.28 | 0.28 |
| A 1.7 1st crop -paddy / 2nd mixed field crops | 0.28 | 0.28 |
| A 1.8 other | 0.28 | 0.28 |
| A 2.1 Tobacco | 0.28 | 0.28 |
| A 2.2 Soy ben | 0.28 | 0.28 |
| A 2.3 ground nut | 0.28 | 0.28 |
| A 2.4 Corn | 0.28 | 0.28 |
| A2.5 Mied field crops | 0.28 | 0.28 |
| A 2.6 other | 0.01 | 0.28 |
| 3.1 Longan | 0.01 | 0.28 |
| 3.2 Citrus | 0.01 | 1 |
| A 3.3 Mixed | 0.01 | 1 |
| A 3.4 other | 0.01 | 1 |
| A 4.1 (banana plantation) | 0.09 | 1 |
| A 5.1 truck crops | 0.28 | 0.27 |
| A 5.2 Ornamental | 0.28 | 1 |
| A 5.3 Mixed | 0.28 | 1 |
| A 5.4 Other | 0.28 | 1 |
| A 6.1 Cultivated land | 0.45 | 1 |
| A 6.2 Bush fallow | 0.45 | 1 |
| F1 Ever green forest | 0.003 | 1 |
| F2 Deciduous forest | 0.048 | 1 |
| F3 Dipterocarp forest | 0.004 | 1 |
| F4 Degraded evergreen forest | 0.3 | 1 |
| F5 Degraded deciduous forest | 0.4 | 1 |
| F6 Degraded Dipterocarp forest | 0.50 | 1 |
| M1 range land (grass, bushes and shrubs ) | 0.72 | 1 |
| M2 wet land (marsh and swamp) | 0.28 | 0.27 |
| M3 Mine | 1 | 1 |
| M4 Other - bare land | 1 | 1 |
| w1 Natural water resource | 0 | 0 |
| W2 built up water resource | 0 | 0 |
| A1. 1 / A1.3 | 0.28 | 0.027 | A3.3/A1.3 | 0.091 | 0.71 |
| A1. 1 / A1.4 | 0.28 | 0.027 | F5/F3 | 0.091 | 1 |
| A1. 3 / A1.4 | 0.28 | 0.027 | U2/A3.1 | 0 | 0.003 |
| A2. 1 / A1.4 | 0.28 | 0.027 | A1.7/A1.5 | 0.28 | 0.027 |
| A3.1. 1 / A1.3 | 0.091 | 0.71 | A3.1/A2.5 | 0.28 | 1 |
| A3.3/M1 | 0.001 | 1 | A3.1/U2 | 0.01 | 0.71 |
| M1/A3.1 | 0.001 | 1 | A1.7/A3.3 | 0.19 | 0.71 |
| M1/M4 | 0.31 | 1 | M1/A2.5 | 0.12 | 1 |
| A2.5/A3.3 | 0.28 | 1 | A1.1/M2 | 0.28 | 0.027 |
| A2.1/A3.3 | 0.19 | 1 | M1/A3.3 | 0.01 | 1 |
| A3.1/A1.4 | 0.91 | 0.71 | A1.3/U2 | 0.91 | 0.71 |
| A1.1/M2 | 0.28 | 0.027 | |||
Five erosion classes have been defined as followes:
| Class # | Soil Loss (t/ha/yr) | Area % | Area sq.km |
| 1. very slight | 0-1 | 49 | 1656 |
| 2. slight | 1-5 | 20 | 697 |
| 3. moderate | 5-20 | 15 | 525 |
| 4. Severe | 20-100 | 10 | 325 |
| 5. very severe | >100 | 6 / 100 | 200 / 3403 |
Fig 7. shows show the soil erosion susceptibility map ass the final map produced fig illustrate a perspective view of the soil erosion susceptibility map. a grid was over land on the map to enhance the relief displacement the viewer is looking almost in north west direction 80 km from the center of scene and evaluated at an altitude of 25 km.
Figure.7. Soil Erosion susceptibility
Results and discussion
It can be seen from the map in fig 7 and 8 that maximum erosion risk as expected occurs in the mountains regions where the forest has been cleared to make place for the cultivation has been cleared corn and upland rice all tree plantations mostly pine in hilly regions are all tree plantations mostly pine for erosion control and the practics as exbhit a seeding pattern parallel to the direction of the slope a measure that is very erosion prone.
Figure.8. Perspective view of soil susceptibility
Conclusions
This multithematic study is an interesting example of the use of remote sensing data in combination with punctual data and map data . The integration of the different components was done in this case entirely manually but one can imagine that the speed and accuracy of the technique could be improved tremendously by using a geographic information system such as been recently made operational at the national research council with manual technique and also to produce average annual soil erosion maps for other parts of Thailand.
Soil erosion is a major concern in Thailand .This demonstrations project has shown that shown remote sensing combined with other data sources can provide unique and useful information to determine soil erosion risks at a scale between 1 : 50,000 to 1 : 100.000.
The SPOT imagery has been found appropriate and equivalent to other sources for erosion mapping through mapping the identification of land and vegetation cover.
The geographic information system used in conjunction with the widely accepted universal soil loss equation allows an efficient identification of the erosion prone areas.
In the slop chiang Mai area soil erosion occurs mainly on step slopes in shifting cultivation area.
References
- Aronldus H.M.J 1977 methodology used to determine the maximum potential average annaual soil loss due to sheet and rill erosion in Morocco.
- Asian development bank 1986 land resource evaluation and planning project final report, vol. 1 main report. 106 p.
- Ateshian J.H.K., 1974 Estimation of rainfall erosion index. J. Irr. drain . div., 100: 293-307.
- Baril D Perras, S., Pesant, A., and Bonn, F., 1988, Cartographie des zones de cultures sarclees a hauts risques d'erosion hydrique par integration de donnees multisources a un system d'information geogrpahique. Vie Congres de 1'A.Q.T. Sherbrooke, Canada, Mai, 1988
- El Swaify., S.A., 1977, Susceptibility of certain tropical soils to erosion by water. From soil conservation and management in the humid tropics. John Wiley and Sons.
- Environment Canada 1984 a manual for regional targeting of agriculture soil erosion and sediment loading to streams. Working paper no. 36. Lands directorate. 38 p.
- Fao 1979 A providional methodology for soil degradition assessments
- Fournier F 1960 climat et erosion. Presses Univeritaries de France, Paris.
- Lal R., 1976 Soil erosion on alfisols in western Nigeria. III: Effects of rainfall characteristics. Geoderma 16 (5) : 389-401.
- Nitaya, K. and Pala, S., 1988 The use of land sat MSS fore forest cover type mapping in Chiang Mai. 11th Asian Conference on Remote Sensing, Bangkok
- Saengwan, B., Suvanphetch, C., Rodprom, C., Bonn, F., and Vincent, P., 1988, Landuse change detection in the Chiang Mai watershed area. 11th Asian Conference on Remote Sensing, Bangkok.
- Sayago J.M., 1986, Small scale erosion hazard mapping using Remote Seinsing for resources development and enviornemnt managemant, Enschede August 1986, pp. 669-674
- Simard, R. Rochon, Gosselin, C., Sris saeng D., Vilbrusersth S., 1987. The Landsat - Thailand Project: An example in Candian Remote Sensing technology transfer 10th Asian Conference on Remote Sensing. methodology diagram.