GISdevelopment.net ---> AARS ---> ACRS 1990 ---> Agriculture and Forestry

Remote Sensing application in soil erosion studies on the loess plateau

Lin Pei , Liu Liming
National Center of Agricultural Remote Sensing Application and Training Beijing Agricultural University , Beijing ,China


Abstract
Remote sensing application in soil erosion studies has been made on the Loess Plateau, northwest of China, which is one of the national projects taken on in the 7th -five-year plan period. This paper tries to introduce a methodology of quantitative analysis of soil erosion taking small watershed for the study cell of the erosion system. A match of the results of remote sensing interpretation and experimental measurements was made, and a series of models based on the most important erosion factors were developed. Finally, the technological process of quantifying soil erosion with remote sensing information supported by Soil Erosion Information System was established. A trial small watershed was selected to test the method, and the results showed that the technological process is logically available.

Introduction
The Loess Plateau that is famous for its deep deposit and wide distribution of Loess is located between 103-112E and 34-38N in the Northwest of China, with an area of 357.000 km2 and average elevation of 1.000-2,500 meters. Severe erosion with a long history not only has produced a great impact to agriculture in this area, but also silted up the reservoirs and riverbeds in the lower reaches. So it has been seriously damaging the ecological systems in the Loess Plateau and the near areas. Since 1950's, this problem has been payed great attention, to study and now the remote sensing application on the loess Plateau is one of the 7th -five -year national projects.

General situation of soil erosion on the loess plateau and study programme of Remote Sensing application
The Loess Plateau is located in semi-arid region with annual precipitation of 350-500mm. About 70% of the precipitation falls in storms in July and August. Because of the deep deposited Loess and steep descent of hillslope, soil erosion is very severe in this region. However, intensive cultivation has not been limited and vegetation cover is very low (<20%). As a result, the whole plateau is incised into small scraps. According to some references , about 16-18 billion tons of soil is transported into Bohai Sea by Yellow River, over 80% is from the Loess eau , especially from the gorge between shanXi and shann Xi provinces, Reasonablely, the main trials of remote Sensing application in soil erosion were taken on in Mizhi county, Ansai county, and county, and Shenmu county in this region.

From remote sensing images, we can distinctly three types of landforms: the uplands deeply incised by gullies (Yuan), the narrow hills (Liang) and the round hills (Mao). Especially, soil erosion in the Liang and Mao regions are more serious than that in the Yuan regions, because of much steeper slope and more strap land. We usually call the Liang and Mao regions "hilly -gully regions".

According to the erosion types and characteristics in different geomorphologic locations, small watershed can be divided into three sections: rill-interrill erosion section, gully-gravity erosion section and deposit section of channel system. The land use and erosion types corresponding to the different sections can refer to the figure 1 and the Table 1.

  1. Rill-Interill Erosion Section

  2. Gully-Gravity Erosion Selection

  3. Deposit Section of Channel

This division is very important to remote sensing interpretation and thematic mapping such as soil erosion mapping such as soil erosion mapping, land use mapping, etc., corresponding to the different sections. Particularly, the boundary between rill-interrill erosion section and gully-gravity erosion section which can be portrayed along gully heads can only be precisely identified from color infrared aerial photos (on the scale of 1:10,000-1:50,000). But, on SPOT and TM images, the gullies which are narrower than 5m-10m can not be identified distinctly because of their resolution. So the rill-interrill

Table 1 Types of Land Use and Soil Erosion in Different Sections
Location Slope Land Use Soil Erosion
Hill slope 150 Hill slope field , terrace field Rill-interrill erosion
Gully slope 150 -450 Uncultivated gully slope land Gully -gravity erosion
Channel 50 Silt land Silt depositing or delivering


Erosion area portrayed on SPOT or TM images is larger than that on color infrared aerial photos. This would lead to incorrect area measurements of rill-interrill erosion section as well as of gully-gravity erosion section simultaneously. Figure 2 shows the above fact on four kinds of images.

Figure 2 Contrust of the Four kind of Remote Sensing images



According to the particular characteristics of soil erosion on the loess Plateau we devoted our research to four aspects as follows:

  1. Systematical plan and Overall Study of the Project

    Beside of independently analysing of soil erosion factors, an overall study froth one field to another must be made as the DTM and geomorphic features, the Quaternary geological structure motion, gully's stability and development, the torrential rain runoff models, vegetation cover, and the characteristics of soil spectral reflectance, etc.. Thus we tried to extract the available information from remote sensing images as far as possible. Then, a synthetic analysis was mode based on the above studies. Figure 3 show the relationship of these specialized studies.

  2. Quantitative Analysis of soil Erosion with Remote Sensing Information

    The essence of quantifying soil erosion is to determine the quantitative relationships among detachment, transport ant, and deposit, However, soil erosion intensity is usually indicated simplificately by erosion modulus in per unit area year. But there are two kinds of data of the erosion modulus. One is from runoff-erosion plots; the other is form hydrologic gauging station. Although the measurements of soil loss from experimental plots are primarily important to determine the amount of soil lost from field , and they do not express sediment yield without due consideration to the process of entrainment and transport . So the former is just the gross erosion. Sediment yield measured at hydrologic gauging station is dependent on the gross erosion in the watershed and on the transport of eroded material out of the watershed. Thus, the sediment delivery ratio is much important to determine total sediment yield based on computed gross erosion. Therefore the quantitative relationships amongst the experimental plots data, the hydrologic gauging data, the computed erosion and the results of quantifying soil erosion with remote sensing are the theme of this study.
Naturally, the total amount of sediment in the small watershed includes rill-interrill erosion and gully -gravity erosion

  1. Rill-interrill erosion can be computed by the Universal Soil Loss Equation (USLE). But the USLE must be modified in order to be used on the Loess Plateau. Several experimental plots were set up in Mizhi County, Shann Xi Province , referring to the standards of the SCS , U.S. Based on several year's measurements, a modification USALE was developed :

    A = 1.244+0.24021*RKSLCP

    Where A is the computed soil loss in per unit area, R is the rainfall factor, K is the soil credibility factor, L is the slope length factor , S is the slope gradient factor , C is the cropping management factor , P is the erosion control practice factor , and 1.244 and 0.2401 are modification coefficients of the equation is 0.97.

  2. The USLE Predicts the erosion of soil at the point of detachment of soil particles from the surface. However, only a small portion of detached particles reaches the watershed outlet. Sediment yield is the amount of soil that is transported out of a drainage basin and equals the gross erosion minus the soil deposited within transporting proceeding. The ratio of the sediment delivered to the outlet to that eroded from the upslope surface is called "Sediment Delivery Ratio (SDR)" . Especially, in the hilly-gully regions of the Loess Plateau, because of steep gully slope and the quick descent of channel as well as curing measures, depositing almost does not occur when eroded materials are delivered in gullies. So, only was the SDR in rill-interrill erosion section considered in this study. Depending on the transport capacity equation of overland flow by kirkby (1976) and the detachment rate equation by Meyer (1981), a S D R formula was developed. Factors concerning in the formula include the factors in cropping management , the erosion control practice and the soil erodiablity of the USLE , the soil moisture content at field capacity , the buld density of the top soil layer and other rainfall factors , ect..

  3. Sediment in the small water shed can be acquired from hydrologic gauging station in one hand, and in the other hand, it can be determined by measuring the stereoscopic aerial photo pairs of silt arresters with steroplotter. Then a quantitative analysis of soil erosion factors based on the above measurements was made referring to the fundamental principles of factors based on the above measurements was made referring to the fundamental principles of Mathematical Geography, while the erosion factors were interpreted quantitatively from the

    Color infrared aerial photos as well as other supplementary information sources. Finally, a geosciences model predicting sediment yield in the small watershed was developed with most important erosion factors.

    Y= R(0.306P-0.859 X-0.294 (0.062)D (1.025)S (0.985)L (0.921)F)


    Where Y is the sediment yield in per unit area per year, R is the runoff modulus, P is the % of vegetation cover and cured land area, D is the % of cultivated hillslope field area, S is the average slope gradient of the watershed , L is the average slope-length of the watershed X is the circularity ratio of the small watershed , and F is the content of 0.1-2.0mm fine sands(%)

  4. The amount of erosion in the gully-gravity erosion section is the sediment yield in the watershed minus the eroded materials delivered from rill-interrill erosion section. The formula is expressed as:

    G = Y-A* SDR


    Where G is the gully -gravity erosion in the watershed
    Y is the sediment yield in the watershed
    A is the rill-interrill erosion in the watershed
    SDR is the sediment -delivery ratio in rill-interrill erosion section
    In order to test the above technological process of quantifying soil erosion with remote sensing information, a trial watershed of QuanjiaGou in Mizhi Country was selected. The amounts of soil erosion including the sediment yield of the watershed, gully-gravity erosion , and rill-interrill erosion were computed . The S D R in rill-interrill erosion section was also calculated. The results showed that the computed values were very close to the corresponding experimental data.

  5. Mapping Soil Erosion with Remote Sensing Technique

    Quantifying soil erosion is the basis of mapping soil erosion with remote sensing technique. Therefore an overall analysis of erosion factors with remote sensing information depending on experimental data and field survey results is primarily important to set up mapping system . That is to say: (a) experimental data and field survey results is the basis of remote sensing interpretation; (b) numerous erosion factors must be considered synthetically when the legend of the soil erosion map is established. In this study, we put forward a method called "Erosion Factors Matching Proceeding" by which the grades of erosion intensity were determined.

  6. Soil Erosion information, Entropy Model and soil Erosion information System (SEIS) The ultimate objective of research on soil erosion with remote sensing information is to set up the soil erosion information system. Supported by the SEIS, a soil erosion information entropy model was developed. The model was tested in Gao XiGou watershed, MiZhi County , and was proved to be a very flexible and effective tool in soil erosion research.
Conclusions
  1. Soil erosion is caused by numerous factors in geography, geology and the mankind activities on the surface of the earth. A quantitative analysis of these factors with remote sensing information is the basis of quantifying and mapping soil erosion.

  2. Systematical experimental data and field survey results provide basic inputs to the soil Erosion Information System, on which remote sening application in soil erosion depends.

  3. Erosion models are the junction points between soil erosion study and remote sensing application. The Soil Erosion Information System (SEIS) is a practical and effective tool in soil conservation.

  4. The technologic process of quantifying soil erosion with remote sensing information in this study is proved to be logically and practically available.