ASSESSMENT OF SOIL EROSION RISK IN NORTHERN THAILAND

LV Tingting a，b, *，SUN Xiaoyua，b，ZHANG Dandana，b，XUE Zhenshana，b，GONG Jianminga

a
Institute of Geographic Science and Natural Resources Research, CAS,100101 Beijing China - (lvtt.04b, sunxy.05b,
zhangdd.06b, xuezs.07b)@igsnrr.ac.cn
b
Graduate School of the Chinese Academy of Sciences, 100039,Beijing, China

Commission VIII, WG VIII/6

KEY WORDS: Northern Thailand, Soil Erosion , Risk Evaluation, Land Cover, IMAGE\LDM

ABSTRACT:

Soil erosion is very serious in Thailand especially in northern Thailand. Important on-side effects of soil erosion may be the decline
in qualities of soil related to agricultural productivity. So it is very important to assess the soil erosion risk for the sustainable
development of agriculture. This study was conducted with objective of modelling and assessing soil erosion risk in the northern
Thailand with the application of IMAGE\LDM. Rainfall erosivity index, relief index, soil erosivity index and land cover index are
four basic factors used in IMAGE\LDM. Soil erosion risk can be grouped into six classes. Furthermore, the spatial distribution
characteristics were also analyzed with the application of GIS in the view of elevation, land use types. From the result we can find
soil erosion risk is high in the altitude between 100 and 400. Soil erosion risk is lower in the forest area than in the agriculture and
plantation area.

1. INTRODUCTION on soil erosion mainly focus their attentions on a slop or

watershed scale. Currently, research on soil erosion is changing
Soil erosion is the deterioration of soil by the physical from slope scale or small watershed to the regional scale
movement of soil particles from a given site. Weather, although soil erosion model at regional scale is not very perfect.
vegetation, soil, topography and geology and human beings are Now the mostly used soil erosion model at regional scale is the
usually the main causes of soil erosion. With the rapid growth qualitative assessment method based on expert’s knowledge.
of the population and the acceleration of the process of the For example, the Global Assessment of Human Induced Soil
industrialization, the degree of destruction to natural systems is Degradation (GLASOD) is a worldwide map of human-induced
far great than maintenance by human beings. For example, soil degradation conducted by the International Soil Reference
vegetation degradation and soil erosion have become serious and Information Centre (ISRIC). Using uniform Guidelines and
problems to hindering the sustainable development of human international correlation, the status of soil degradation (type,
beings. The northern region of Thailand is very vulnerable to extent, degree, rate and main causes) was mapped
soil erosion due to its undulating topography, steep slopes and within loosely defined physiographic units (polygons), based on
high rainfall. For the past few decades, encroachment of expert judgment throughout the world.
agricultural activities on forest areas and misuse of lands have
become serious problems in northern Thailand especially on Another model used for regional soil erosion evaluation is the
steeply sloping land with no conservation measures. These semi-quantitative methods. For example, the Integrated Model
practices not only accelerated the soil erosion but also have to Assess the Global Environment (IMAGE) is a dynamic
resulted in soils of low fertility. Chao Phraya River basin is the integrated assessment modelling framework for global change.
largest and most important river basin of Thailand which Land degraded model is one of the basic models of IMAGE.
originates in the mountain ranges in the Northern Thailand. The (Revised) Universal Soil Loss Equation (USLE/RUSLE) is also
major tributaries are the Ping, Wang, Yom and Nan rivers with used by many researchers to evaluate the regional soil erosion.
mountainous high steep slope from upper northern part of But some researchers think that USLE/RUSLE is an empirical
Thailand to lower northern part, which also accelerated the soil soil erosion model designed to predict soil erosion at slop or
erosion in these river basins. At the same time, deposition of small watershed scale and its application at a regional scale
sediment transported by rivers from the upper part of Chao exists certain problems. For example it is very difficult to
Phraya River basin will pose a great threat to the safety of the guarantee the accuracy of the evaluation or difficult to verify.
lower river basin which is the center of Thailand’s agriculture, Recent advances in geographic information systems (GIS)
economics and cultural. So assessment of soil erosion risk in technology and remote sensing (RS) technology provide us
northern Thailand has very important significance not only to strong toolsets and new data sources to evaluate the soil erosion.
the ecological environment of Thailand but also to sustainable Spatial analysis tools and model building tools help us to design
development of Agriculture. the erosion model, evaluate the erosion degree and analyze the
spatial distribution characteristics of soil erosion. By RS
The soil erosion research can be attributed to three levels technology real-time and wide range monitoring of soil erosion
according to different spatial scales, namely slope scale, small becomes true.
watershed scale and regional scale. For a long time researchers

* Corresponding author. This is useful to know for communication with the appropriate person in cases with more than one author.

703
 The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B8. Beijing 2008

2. RESEARCH AREA 3.2 Method

Our study area is located in the north part of Thailand which The Integrated Model to assess the Global Environment
covers an area of 170,000 km2 (shown in figure 1). The region (IMAGE)-land degrade model (LDM) was used to evaluation
is characterized by mountains and percent of area with the soil erosion risk in north Thailand. The LDM is based on the
elevation above 500m is about 60%. Most of the mountains are concept of soil’s susceptibility and sensitivity to water erosion.
composed of carbonic limestone. Intermountain basins alone Susceptibility to water erosion is based on the current terrain
faults are mainly composed of quaternary sediments which have erodibility and rainfall erosivity. Sensitivity to water erosion
fertile soil and are the major agricultural area in the northern describes the chance that water erosion will occur accounting
region. for the actual land use and land cover. According to LDM, soil
erosion susceptibility and sensitivity index was calculated. On
The northern Thailand has a tropical climate, influenced chiefly the basis of water erosion-sensitivity index, soil erosion risk
by monsoon. The cool and dry season is from November to grade can be determined.
February of next year. The hot and dry season is from March to
June. From July to October is the rainy season when about 89% 3.2.1 Indexes: Establishing an appropriate index system is
annual rainfall is accumulated. the first step for us to evaluate the status of soil erosion at the
regional scale. Because soil erosion is a very complex process
Northern Thailand has, by far, the largest total forest area of all involving many factors, the establishment of the index system
four regions. But with the rapid development of economic and must meet the following requirements: on the one hand, index
growth of population, the percentage of forest-covered land in system must include, as far as possible, key factors influencing
northern region decreased greatly. According to the statistics by the erosion process, on the other hand the index system must
Thailand Royal Forestry Department, in 1961, about 68.54 reflect the regional characteristics of the research area. Based
percent of the total land area of northern Thailand is covered by on these requirements mentioned above four major factors
forest, but in 2004 the percent dropped to 54.27%. The northern (rainfall pattern, topography, soil, and land use type) are finally
Thailand also has the least area of arable land, accounting for chosen for computing the soil erosion susceptibility and
about 26.39 percent of the area of the region, of which the total sensitivity index in northern Thailand in the paper.
area of rice accounted for 53.19 percent, field crop accounted
for 27.53 %, perennial crops accounted for 12.88 % according (1) Rainfall erosivity index (R-factor)
to the statistic of 2004.

Figure 2. Rainfall erosion index

Among the four major factors affecting the soil erosion, rain is
the agent for erosion, which reflects the potential rate of soil
erosion. Not all rainfall can induce soil erosion except those
Figure 1. Study area showers of high intensity. So the erosivity of rainfall is largely
determined by the intensity of rainfall events. As mentioned in
section 2, rainfall in northern Thailand is very unevenly
3. DATA AND METHOD distributed, which mainly concentrates in rainy season from
June to September, so the rainfall data from June to September
3.1 Data and Resource was used to calculate R-factor. According to IMAGE-LDM, the
monthly mean rainfall intensity (mm/day) was selected as the
Data used in the paper includes land use data at 1:250000 scale indication of rainfall intensity. If the maximum mean monthly
interpreted from ETM by Thailand Royal Forestry rainfall intension of three month exceeds 2mm/day, the R-factor
Department.(provided by Marc Souris, IRD ), precipitation data is assigned 1. If the maximum mean monthly rainfall intension
(1981-2001) from Thailand meteorological station, soil dataset of three month is belongs to 0 to 2mm per day, the R-factor is
with the resolution of 5-min provided by FAO/UNESCO, DEM assigned 0. If the value between these two extremes a linear
with the resolution of 90 m acquired from SRTM. relation shown in formulation 1 is assumed.

704
 The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B8. Beijing 2008

1 1 I t = -0.005 * CL + 0.005 * SI + 0.5 (3)

y = − + *x (1)
9 18

After experimentation with several interpolation methods, the With: CL is clay content of soil types. SI is the silt content of
inverse distance weighted interpolation method was applied to soil types.
establish the spatial layer of the R-factor (shown in figure 2) in
this research. Ib is the bulk density index. According to the research of
Baties(1996), if the bulk density exceeds 1.55g/cm3, the Ib is
(2) Relief index（Ia） assigned 1. If the bulk density is belongs to 0 to 1.55g/cm3, the
Among the four major factors, topography is relatively stable R-factor is assigned 0. If the value between these two extremes
which can remain fairly constant over time. In the soil erosion a linear relation shown in formulation 4 is assumed.
study at field or at small watershed scale, slope-length and
slope steepness are two important indexes which are used to
reflect the effect of topography on erosion. There are a number y = −2.875 + 2.5 * x (4)
of empirical formulas capable of calculating the L and S factors.
However, in study at the regional scale, with the resolution or
scale of topographical information carrier decreases, slope- Table 1 provides the soil depth index Id according to soil depth.
length and slope steepness will lost their significance on soil
erosion. The relief degree of land surface is an important factor
in describing the landform macroscopically which has been Soil depth(cm) Id
widely used in soil erosion assessment at the regional scale. 0-25 1.0
25-50 0.9
Based on the macro-scale digital elevation model data, by using 50-100 0.6
ARCGIS software, the relief degree of 2km×2km grid size is 100-150 0.25
extracted and mapped. >150 0

If the relief degree exceeds 300m, the Ia -factor is assigned 1. If Table 1. soil depth classes and corresponding values of the soil
the relief degree is less than or equal to 0m, the Ia -factor is depth index
assigned 0. If the value between these two extremes the Ia -
factor is calculated as relief degree divided by 300. Figure 3 is Figure 4 is the final map of Soil erodibility index.
the spatial distribution of Ia-factor.

Figure 3. Relief index

（3） Soil erodibility index（SE）

Figure 4. SE factor map
The soil erodibility factor-SE represents the average long-term
soil and soil-profile response to the erosive power associated
（4）Land cover index（LC）
with rainfall and runoff. Soil erodibility factor is estimated from
Land cover that represents resistance of the ground surface to
soil attributes such as particle size distribution, organic carbon
the transport of water-soil mixture is another very important
content, and density of eroded soil. In the IMAGE-LDM the SE
factor in the soil erosion model. Given conditions of climate,
is derived from indices for soil texture, bulk density and soil
soil and relief, the protection provided by land use and land
depth by taking the average of the two highest values.
cover determines the actual sensitivity to water erosion. In this
Derivation of the factors required by IMAGE-LDM is as
paper land-cover indexes for different land-cover types was
follows:
provided by Thailand Land Development Department (LDD) as
shown in table 2. Higher LC-factor values indicate higher risk
SE = average (max (I t , I b , I d ) 1, max (I t , I b , I d ) 2 ) (2) of soil erosion. From table 1 we can find forest has lower value
than agriculture mainly because forest can provide a high
Where It is soil texture index whose formulation is given by degree of protection against soil erosion than agriculture. The
formulation 3. geographic distribution of different land-cover types forms the
basis for the land-cover index. In the paper spatial vegetative

705
 The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B8. Beijing 2008

cover type was extracted from Landsat TM imagery acquired in From E1 to E6, the potential susceptibility and sensitivity to
2000 by Thailand Royal Forest Department (RFD). Based on water erosion gradually increased. The soil erosion risk map
the C-factor developed by LDD (LDD, 1999) values for the resulting from the method mentioned above in the northern
various land cover types are assigned accordingly. Figure 5 is Thailand is presented in figure 6.
the spatial distribution of LC factor.

Land cover types C

Agriculture 0.225
Dry evergreen forest 0.019
Hill evergreen forest 0.003
Tropical pine 0.019
Bamboo 0.015
Mixed Deciduous 0.001
Dipterocarps 0.020
Plantation 0.088
Building 0.000
Grass land 0.015
Water 0.000

Table 2. LC factors for each land cover types

Figure 6. Soil erosion risk map

4. RESULT AND DISCUSSION

4.1 Soil erosion distribution in different altitude belts

100- 200- 400- 600- 800-

Value 0-100 >1000
200 400 600 800 1000
E1 3.11 23.55 30.72 58.04 69.05 58.74 38.77
E2 8.96 1.77 9.58 2.09 5.56 13.37 33.10
Figure 5. LC factor map
E3 53.53 11.33 9.94 3.24 2.82 5.75 6.55
3.2.2 Model: The soil erosion model used in the paper is E4 28.92 34.13 31.83 10.76 7.55 8.49 8.92
given by formulation 2, 3 and 4.
E5 2.97 14.60 11.05 7.50 4.63 4.72 3.60
E6 2.51 14.62 6.88 18.37 10.39 8.92 9.07
Ep = (R + T)/2 (2)
Table 3. the ratio of each erosion grade under different
elevation condition (%)
T = (SE + Ia)/2 (3)
From table 3 we can find that erosion risk grade has some
Ea = Ep × LC (4) relationship with the altitude. Generally, erosion risk is lower in
higher altitude belts than that in lower altitude belts. Soil
erosion grade E3 has the highest proportion in the region with
altitude from 0 to 100m which accounts for 53.53%. The ratio
Where R is the rainfall erosivity index, SE is the Soil erodibility of erosion grade E4 in this region is also very high. Soil erosion
index, Ia is the relief index and LC is the land cover index. Ep grade E1 and E4 have the higher proportion in the region with
stands for susceptibility to erosion which is the potential altitude from 100 to 400m. In the other regions with altitude
susceptibility of the soil erosion based on indexes of rainfall greater than 400m E1 has the highest ratio which is 58.04%,
and terrain (including soil and relief indexes). Ea stands for 69.05%, 58.74% and 38.77%.
sensitivity to water erosion which is final result of the model.
By using the natural break classification method the Ea was Columns in figure 7 represent the proportion of area of forest
divided into six groups. Ei stands for risk grade and i was used and agriculture in each altitude belts. Line in figure 7 represents
to present the grade number. the highest proportion of soil erosion grade in each altitude

706
 The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B8. Beijing 2008

belts. From figure 7 we can find area of agricultural land E1 E2 E3 E4 E5 E6

decreases gradually with the altitude increases while area of Value
(%) (%) (%) (%) (%) (%)
forest land increases. The highest ratio occurs in the region with Agriculture 0.01 3.92 28.26 28.02 12.11 27.68
altitude from 100m to 400m, which is the transitional zone of
forest and agriculture and is very sensitive to the soil erosion. Building 100.0 0.00 0.00 0.00 0.00 0.00
Mixed deciduous forest 99.89 0.11 0.00 0.00 0.00 0.00
Hill evergreen forest 18.57 70.35 11.09 0.00 0.00 0.00
Grass 0.08 0.43 2.23 13.30 36.89 47.06
Dipterocarps 11.21 13.33 23.95 42.15 9.25 0.11
Second forest 0.04 1.98 10.94 20.41 23.14 43.48
Plantation 0.21 1.01 5.31 28.56 33.91 30.99
Dry evergreen forest 0.27 3.75 13.92 56.06 25.91 0.08

Table 4. The ratio of each erosion grade in different land use

type

Figure 7. Ratio of forest and farmland area in different

5. RESULT
elevation belt
From the soil erosion risk map resulting from the IMAGE-LDM
4.2 The relationship between soil erosion risk and land use mentioned above we can find that in the northern Thailand
lower risk grades area are mostly forest area while high risk
Land use data at 1:250000 scale interpreted from ETM by
grade area are mainly concentrated on the edge of Chao Phraya
Thailand Royal Forestry Department was used. According to
River alluvial plain and block basins among mountain. By
this land use data, the largest area of land use type is forest in
analyzing the relationships between soil risk grade, altitude and
the northern Thailand which accounts for 55.73% and the
land cover types, the conclusions were done as follows:
second largest area of land use type is agricultural land.

Land use type is one of key factors which could reinforce or （1）Soil erosion risk in high altitude area is lower than area in
weaken the soil erosion. By overlaying the soil erosion risk map the lower altitude region. The main reason is that main land
and land use map, the relationship between them can be cover type in high altitude area is forest land whose water
acquired. The proportions of soil erosion grade in each land use conservation capacity is higher than that of agricultural land.
type were given in table 4. The proportion of soil erosion grade Soil erosion grade is the highest in the transitional zone of
in each land use type is greatly different. For example, the forest and agriculture whose altitude is from 100m to 400m.
erosion risk of building and mixed deciduous forest is very low. One of important reasons is the encroachment of agricultural
About 70.35% of hill evergreen forest’s risk grade is E2. Risk activities on forest areas
grade of agriculture land is mainly concentrated in the range
from E3 to E6. Risk grade of dipterocarps and dry evergreen （2）Amongst the different land use types, mixed deciduous,
forest is mainly concentrated in the E4 while the biggest hill evergreen forest and building have lower erosion risk grade
proportion in plantation and grass occurs in grade E5. About while dry evergreen forest, plantation and second forest’ risk
43.48% of second forest’s risk grade is E6. levels are higher. According to the statistics, about 90 % of
slope farmland has very high soil erosion grade. Risk level of
Slop farmland is one of the land use types which are very natural vegetation is lower than that of artificial vegetation. The
vulnerable to soil erosion. Slop farmland is defined as the risk level will increase in the region where large area of natural
farmland whose slop degree is greater than 5o in this paper. vegetation was replaced by artificial vegetation.
Slop farmland map can be acquired by using slop degree map
extracting from DEM and land use type map. Overlaying slop
farmland and soil erosion grade map we can acquire each risk REFERENCES
grade’s proportions on slop farmland region（Figure 8）.
Li Hui,Chen Xiaoling. Review of soil erosion model at different
spatial scales[J]. Journal of Central China Normal University
(Natural Sciences). 2006,40（4）:621-624.

Zhou Weifeng,Wu Bingfang. Analysis of Soil Erosion Research

on Regional Scale[J]. Research of Soil and Water Conservation.
2006,13（1）:255-268.

Zhou Zhengchao, Shangguan Zhouping. Overview on Soil

Erosion Model Research[J]. Science of Soil and Water
Conservation. 2004, 2 (1):52-55.

Li Guanglu,Zhang Shengli. A Review of the Researches of Soil

Figure 8. The proportion of each erosion rate in slope Erosion Models [J]. Journal of Northwest Forestry University.
farmland 2000, 15(2):76-83.

707
 The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B8. Beijing 2008

Zhu Hongchun, Chen Nan. Extracting the Undulating Degree of Lu Yuan, Hua Cui elt.. GIS—based Sensitivity Evaluation for
Terrain by 1:10000 Scale DEM data-A Case Study on Loess Soil Erosion in Guangxi. Research of Soil and W ater
Plateau in northern Shanxi[J]. Science of Surveying and Conservation. 2007, 14(1):98-100.
Mapping. 2005, 30(4):86-88.
Wang Wanzhong, Jiao Juying. Qutantitative Evaluation on
Feng Zhiming, Tang Yan elt. The Relief Degree of Land Factors Influencing Soil Erosion in China[J]. Bulletin of Soil
Surface in China and Its Correlation with Population and Water Conservation. 1996, 16 (5):1-20.
Distribution[J]. ACTA Geographica Sinca. 2007,62(10):1073-
1082. Suthipong Sthiannopkao,Satoshi Takizawa. Soil erosion and its
impacts on water treatment in the northeastern provinces of
Pan Jianping, Gong Jianya. The Research Situation of Soil Thailand [J]. 2007, Environment International,33:706-711.
Erosion Models and its Application in GIS and RS[J]. Journal
of Geological Hazards and Environment Preservation. 2005, R.M.Hootsmans,A.F.Bouwman. Modelling land degradation in
16(1):89-93. IMAGE2, 2001.

Hu Liangjun, Shao Mingan. Review on Regional Soil Loss Merritt,W.S.,Croke,B.F.W. The Biophysical Toolbox: a
Study[J]. Journal of Mountain Research. 2001,19（1）:69-74. Biophysical Modelling Tool Developed Within the IWRAM-
DSS, 2002, iCAM Working Paper.
Zhang Youquan, Gong Lihui. A method for fast estimating soil
erosion intensity based on RS and GIS [J]. Ecology and Hua Lu,Ian P.Prosser et al. Predicting sheet wash and rill
Environment. 2007, 16(1):102-107. erosion over the Australian continent[J].Australian Journal of
Soil Research,2003,41:1037-1062.
Chen Yanhong, Pan Wenbin elt.. Assessment of Soil Erosion
Sensitivity in Watershed Based on RUSLE-A Case Study of IAN P. PROSSER, HUA LU, CHRIS J. MORAN. Assessment
Jixi Watershed. Journal of Mountain Research. 2007, 25(4)490- soil erosion and its off-site effects at regional to continental
496. scales[OL].2003, OECD web site.

Hu Liangjun, Li Rui elt.. Region-Scaled Water Erossion

Assessment Based on GIS[J]. Journal of Mountain Research.
2001,38（2）:167-175.

708