LI Yong (1,2), FRIELINGHAUS Monika (1), BORK Hans-Rudolf (1), WU Shuxia (2), ZHU Yongyi (2)

Scientific registration n : Symposium n : 31 Presentation : poster Spatial patterns of soil redistribution and sediment delivery in hilly landscapes of the Loess Plateau Motifs spaciaux de zones d'érosion et d'accumulation de sédiments dans un paysage collinaire d'un plateau loessique LI Yong (1,2), FRIELINGHAUS Monika (1), BORK Hans-Rudolf (1), WU Shuxia (2), ZHU Yongyi (2) (1) Center for Agricultural landscape and Land Use Research, Institute of Soil Landscape Research, 1374 Müncheberg, Germany (2) Chinese Academy of Agricultural Sciences, Institute for Application of Atomic Energy, Beijing 194, China INTRODUCTION The spatial patterns of soil erosion and sediment delivery in the watershed have become increasingly considered as the main basis of the decision makers for future management strategies of ecological and economic natural sources. But there have been few actual measurements of the amount of erosion occurring and sediment delivery to waterways under larger mixed land uses at the landscape scale or in a watershed of any size because of the difficulty and expense in obtaining measurements. In the past decades, most of data used for soil conservation decisions in the Loess Plateau were derived from the observations on typical erosion plots with uniformed landform and homogeneous soil and sedimentation from suspended solid loads in the rivers. Is it reasonable to use these small treated-plot observations or the data of suspended sediment in the main rivers without considering the landscape locations can answer questions relevant for whole Loess Plateau landscapes? There is, evidently, a need to quantify both the magnitude and the spatial patterns in erosion and sediment delivery directly from the field measurements under entire natural agricultural landscapes. As the integrated studies of soil erosion and sedimentation assessment in the same watershed, the research reported in this paper will focus on the spatial soil redistribution caused by tillage and water erosion and sediment export in the whole hillslope landscape with different land use using the caesium-137 technique. STUDY AREA The field sampling and investigation were conducted in the Yangjuangou dam reservoir Catchment (Li et al., 1997). The Catchment has an area of 2.2 km 2 and an altitude of 12-1 m above the sea level, is oriented south-to-north and is located in the north part of the loess plateau and at the distance of 14 km from Yan an city, northern 1

Shaanxi province in China ( 36 42 N, 19 31 E ). It is the second tributary of the Yanhe River. The gully density is 2.74 km km -2. The inter-gully area comprised 44.2 per cent of the whole Catchment, and actual gully slopes comprises.8 per cent. The peculiarity of soil erosion environment in the study area is runoff scouring by exceedinginfiltration rainfall intensity, irrational activities of human reclaiming natural vegetation or cultivating on steep slopes up to 4 and the extremely weak anti-scouribility of loess lacking vegetation cover, thus no fixing and twine by plant roots (Li, 199). The study area has both typical of the loess hills and gully landscapes with long-term cultivation history in the region, and complete dam reservoir systems and therefore provides representative and integrated information about soil erosion and the source types of sediment and sedimentation rates operating in the Loess Plateau. METHODOLOGY Samples of soil for the determination of spatial patterns in 137 Cs were collected from different landscape positions over the full hillslope range in April 1997. Two downslope transects which pass all landscape positions were respectively established on a whole cultivated hillslope with a length of 27 m and a mixed land use hillslope with a length of 24 m, respectively. For the slope with mixed land use, a slope length of 14 m from the top to the middle part are supporting natural grasses and artificial forest and the remaining slope of 1 m from the lower to the foot composed of three terraced fields is used for cultivation. These two hillslopes are the same on the southwest facing slope and extremely similar in both the topography and slope length. Samples were collected using a 9.9 cm diameter hand-operated core sampler and at 1 m intervals along each transect. 2-4 cores were taken at each sampling point to a depth ensured that the total 137 Cs inventory of the soil profile was measured. The reference sites for determining the 137 Cs fall-out to the study area were respectively established at undisturbed, uneroded, level terraced fields constructed in 194 and uncultivated grassland in the same study basin with the sampled slopes above. The depth incremental samples for determining the vertical distribution of 137 Cs were collected using the method of monolithic soil in original state with a surface area of 4 cm 2 on the representative landscape locations along each downslope transect and the reference sites. The bulk density and other soil quality indicators for each sample taken were measured by the methods of soil physics and chemistry at all sampling points. The topography, vegetation coverage conditions associated with the land use type and the depths of plough layers were investigated at this time. All samples were disaggragated and then air-dried, passed through a 2 mm sieve and weighed. Measurements of 137 Cs concentration were undertaken on a sub-sample of 1-13 g of the finer fraction ( < 2 mm ) of each sample at the Institute for Application of Atomic Energy, CAAS, using a hyperpure coaxial Ge detector coupled to a multichannel analyzer. 137 Cs content of samples were detected at 662 kev and counting time, which were 8 86 4 s, provided an analytical precision of ±6 percent for 137 Cs. The amount of cesium-137 can be expressed per unit mass as the activity (mbq g - 1 ) or per unit area as the inventory (mbq cm -2 ). 2

RESULTS AND DISCUSSIONS Spatial variations in depth distribution of 137 Cs at different landscape locations The typical ranges of 137 Cs depth distributions and their landscape locations along the cultivated hillslope and hillslope with different land use are shown in Figure 1 and Figure 2. Similar profile patterns in 137 Cs depth distributions lie within the lower and foot part and but distinct profiles lie in the top to middle part for whole cultivated hillslope and slope with mixed land use. These similarity and distinction in both slope locations and land use are reflected in the shape of profile distributions in 137 Cs. Profiles from the top to the middle parts of the slope supporting forest have maximum 137 Cs activities at the depth of -1 cm and decrease sharply with depth. These demonstrate the characteristic of 137 Cs depth distribution on the uncultivated land, i.e. efficient adsorption of the fallout 137 Cs by surface soil and minimal downward movement similar to that at the undisturbed reference location with the cover of natural grasses in the study basin. But the profiles from the top to the middle parts of the cultivated slope (Figure 1) show much lower 137 Cs activities than those at the same slope locations of the uncultivated slope (Figure 2) and reference sites. There is evidence of uniform 137 Cs depth distribution throughout -3 cm plough layers in excess of the 137 Cs reference inventory at the lower and the foot of two sampled slopes. These clearly demonstrate the typical spatial patterns in 137 Cs mixing and redistribution by long-term tillage operations and water erosion processes in the hilly agricultural landscapes, Loess Plateau. Spatial distribution of 137 Cs residuals in the whole hillslope landscapes with different land uses In order to summarize the spatial patterns in 137 Cs redistribution in the study basin, 2 sampled slope data about the percentage residuals of 137 Cs from the whole cultivated hillslope and the hillslope with different land use are analyzed based on the comparisons of the point inventory at each sampled location and the reference inventory (Figure 3 and Figure 4). Negative residuals indicate the loss of soil-associated 137Cs and positive represent the gain of 137 Cs. The data from Figure 3 and Figure 4 clearly indicate the significance of landscape location and vegetation cover associated with land use in controlling spatial 137 Cs redistribution in the study area. The important features of 137 Cs redistribution can be summarized as follows: a) The spatial patterns of 137 Cs redistribution along the cultivated slope (Figure 3) have 2 evident characteristics. First, there is serious 137 Cs loss on the whole cultivated slope. Negative 137 Cs residuals greater than 4 per cent have occurred over the 92% of the sampled slope. Averages in net losses in 137 Cs inventories at the top, upper, middle, lower and foot of the slope are respectively 23, 7, 74, 7, and 3 per cent compared with the 137 Cs reference inventory ((238.96 mbq cm -2 ). This indicates that serious soil erosion and sediment delivery have occurred over the entire cultivated slopes, especially greatest net soil export at the area from the upper to the lower part of the slope. Secondly, there is very obvious evidence of 137 Cs deposition at the field boundary of the top and the field boundary between the lower and foot of the slope. Positive 137 Cs residuals at these locations are 77 and 7 per cent, respectively. This suggests a significant contribution of long-term tillage operations to 137 Cs redistribution at these landscape locations. b) The spatial patterns of 137 Cs redistribution in Figure 4 are much different from those in Figure 3 although they are still dominated by negative 3

residuals which cover the 83 per cent of the slope. Averages in negative 137 Cs residuals at the top, upper, middle, lower and foot of the slope are respectively 22, 6, 38, 68, and 22 per cent, which respectively reduce by 4.3, 92., 48.6, 2.9 and 26.7 per cent compared with same landscape locations in Figure 3. Within the uncultivated zone in Figure 4, there is a decrease of 1 per cent in the negative 137 Cs residuals in comparison with the cultivated zone of the slope. Negative 137 Cs residuals less than 3 per cent cover 71 per cent of the uncultivated steep area of the slope, and the residuals greater than 4 per cent occur over 6 per cent of the gentle cultivated slope (Figure 4). These further demonstrate that vegetation cover has play a key role in controlling 137 Cs redistribution associated soil erosion and sediment delivery in the hilly landscapes, Loess Plateau (Li, 199). There is an evident accumulation of 137 Cs (shown by positive residuals) at the boundary of top and upper part of the slope. This suggests that there is an input of 137 Cs from the top by long-term water and wind erosion. But positive the 137 Cs residual occurs at the foot of the slope demonstrates the results of long-term tillage operations. c) The data above clearly indicate the significant role of topography in affecting the transfer of mobilized (sediment-associated) 137 Cs out of the cultivated hillslope (Figure 3). In contrast, greater control of vegetation cover operates on the 137 Cs redistribution on the uncultivated steep slope (Figure 4). Spatial patterns of soil redistribution in the hilly landscapes with different land use Most of 137 Cs models previously used for estimating soil erosion rate do not consider soil redistribution by tillage operations. But recent studies have indicated that soil translocation and erosion with agricultural tillage is dominant soil degradation and geomorphologic processes operating in arable upland landscapes. To quantify the spatial patterns of soil redistribution and dominant erosion processes, the Mass Balance Model and Profile Distribution Model developed by Walling & He (1997) were respectively used for estimating the soil erosion rate by tillage and water flow on the cultivated and uncultivated slopes from the 137 Cs measurements. The estimated spatial data of soil redistribution by the past 44-year tillage and water erosion are shown in Figure and Figure 6. Evidently, there exists evidence of soil redistribution by tillage at the top of 3 m and foot of m of the slope in Figure and cultivated zone of 1 m of the slope in Figure 6. Maximum loss occurs at the upper boundaries of the top and foot part of the sampled slope in Figure and inside of terraced fields on the slope in Figure 6 and maximum gain respectively occurs at their downslope boundaries with a gradual transition between the two. The tillage translocation at the top and foot of sampled slopes accounts for 7.3 and 18. per cent of net soil loss and 17. and 49.2 per cent of the aggradation respectively. This demonstrates that soil redistribution is dominated by tillage at these slope locations. In contrast, on the slopes shown in Figure the area of maximum soil loss extends from the upper to the lower parts of the slopes and a limited aggradation by tillage only occur at the field boundaries. This demonstrates that water erosion has play very important role in controlling soil redistribution, although the extensive tillage erosion also occurs over these slope locations. Based on a comprehensive analyses of the relations of topographic and vegetation cover factors to the soil redistribution for the two sampled slopes in 4

Figure and Figure 6, the spatial patterns of soil erosion over the cultivated hillslope can be classified into 4 dominant processes, i.e. tillage erosion occurring at the top, inter-rill erosion at the upper, rill erosion at the middle, shallow gully erosion at the lower, and doubled controlling processes by tillage and shallow gully erosion at the foot of the slope. But on the uncultivated steep slope in Figure 6, soil redistribution can be seen in an increasing trend of soil loss with slope length from the upper to the middle part and evident soil loss and limited accumulation respectively at the level area and down boundary of the top. This indicates that the spatial patterns of soil redistribution are mainly dominated by double processes of wind and sheet erosion processes at the top and sheet erosion at remaining zones of the uncultivated steep slope. This quantitative classification on spatial soil erosion processes occurring in the hilly landscapes in the study basin is in keep with the qualitative results of field investigation on soil erosion forms conducted in the period of 1996-1997. Sediment delivery in the hilly landscapes with different land use In order to quantify the effects of land use changes on the sediment export, two sampled slope transects data about soil redistribution are be further calculated here. The results are summarized in Table 1 and Table 2. The results shown in Table 1 and 2 clearly demonstrate the spatial patterns of sediment export out of the whole hillslope with different land use. These data have very important implications for the decision makers for future management strategies of ecological and economic natural sources. CONCLUSIONS The spatial patterns of soil redistribution over the cultivated hillslope can be classified into 4 dominant processes including tillage erosion occurring at the top, inter-rill erosion at the upper, rill erosion at the middle, shallow gully erosion at the lower, and doubled controlling processes by tillage and shallow gully erosion at the foot of the slope. But the spatial soil redistribution on the uncultivated steep slope is mainly dominated by double processes of wind and sheet erosion processes at the top and sheet erosion at remaining zones of the slope. Changes in topography are responsible for the spatial patterns above occurring over the cultivated slope and vegetation coverage for those on the uncultivated steep slopes in the study basin. For the sediment delivery to the waterways, the most serious zone is the area from the upper to the lower parts of the cultivated steep slope, representing more than 8 per cent of sediment export of the whole cultivated hillslope. Planting artificial grasses and forests with the cover of 7 per cent at these landscape locations can reduce about 89 per cent of soil export out of these positions. ACKNOWLEGEMENT: This work is part of the project supported by Alexander von Humboldt Foundation IV CHN 139279 and International Atomic Energy Agency Research Contract No.8814 and No.942. REFERENCES [1] Quine,T.A.et al.,1994.soil erosion and redistribution on cultivated and uncultivated land near Las Bardenas in the central Ebro River Basin, Spain, Land Degradation and Rehabilitation,,41-. [1] Walling, D. E. and Quine, T. A. 1993. Use of cesium-137 as a tracer of erosion and sedimentation: Handbook for the Application of the Cesium-137 Technique. UK Overseas Development Administration Research Scheme R479.

[2] Ritchie, J. C. and McHenry, J. R.199. Application of radioactive fallout cesium-137 for measuring soil erosion and sediment accumulation rates and patterns: a review. Journal of Environmental Quality, 19,21-233. [3] Walling, D. E. and He, Q., 1997. Methods for converting 137 Cs measurements to estimates of soil redistribution rates on cultivated and uncultivated soils. 29pp. [4] Li Yong et al., 1997. Using 137 Cs and 21 Pb/ 137 Cs ratios to assess the sediment sources in a dam reservoir Catchment on the Loess Plateau, China. China Nuclear Science and Technology Report, Atomic Energy Press. CNIC-11, CSANS-113. [] Li Yong, 199, Plant Roots and Soil Anti-scouribility on the Loess Plateau, Science Press, Beijing. Key words: spatial soil redistribution, sediment delivery, caesium-137, Loess Plateau, China Mots clés : redistribution spatiale de sols, accumulation de sédiments, Césium-137, plateau loessique, Chine Table 1 Spatial patterns of sediment delivery on the cultivated hillslope in the study basin Location Top Upper Middle Lower Foot length 3 8 6 Slop angle (degree) 12 33 29 18 Mean erosion rate (t ha -1 yr -1 ) 84.21 1.97 6.33.4 44.82 Soil export (t ha -1 yr -1 ) 3.7 1.97 6.33.4 22.22 Area (%) 11.1 18. 29.6 22. 18. Sediment delivery ratio (%) 7.74 1. 1. 1. 62.8 Fraction of sediment export (%) 11. 37.9 23.36 19.68 8. Land use Cultivated Cultivated Cultivated Cultivated Cultivated Table 2 Spatial patterns of sediment delivery on the mixed land use hillslope in the study basin Location Top Upper Middle Lower Foot length 4 4 6 Slop angle (degree) 14 29 3 21 21 Mean erosion rate (t ha -1 yr -1 ) 17.8.27 17.41 63.14 123.81 Soil export (t ha -1 yr -1 ) 1.98 1.76 17.41 63.14 11.2 Area (%) 16.67 16.67 2..83.83 Sediment delivery ratio (%) 82.2 44.6 1. 1. 44.1 Fraction of sediment export (%) 1.3 1.69 16.69 6.3 1.6 Land use Grassland Forestland Forestland Cultivated Cultivated Fig 1. Spatial variations of Cs-137 depth distribution along the cultivated hillslope with a length of 27 m in the study basin 3 Cs-137 (mbq g-1) / (degree) 3 2 1 1-1cm 1-cm -3cm 3-4cm 4-cm -6cm TOP UPPER MIDDLE LOWER FOOT location 6

Fig 2. Spatial variations of Cs-137 depth distribution along the hillslope with a length of 24 m and different land use in the study basin Cs-137 (mbq g-1) / (degree) 3 3 2 1 1 Forestland Farmland TOP UPPER MIDDLE LOWER FOOT -cm -1cm 1-1cm 1-cm -2cm 2-3cm location Cs-137 residuals (%) Fig 3. Spatial distribution of Cs-137 residuals along the cultivated hillslope in the study basin 1. 8. 6. 4... -. -4. -6. -8. -1. Cs-137 residual 1 3 7 9 11 13 1 17 19 21 23 2 27 Downslope distance (m*1) 3 3 2 1 1 (degree) Fig 4. Spatial distribution of Cs-137 residuals along the hillslope with different land use in the study basin 4. 3. Farmland 3 Cs-137 residuals (%). -. -4. -6. Forestland 2 1 1 (degree) -8. Cs-137 residual -1. 1 2 3 4 6 7 8 9 1 11 12 13 14 1 16 17 18 19 21 22 23 24 Downslope distance (m*1) - 7

Fig. Rates of soil redistribution by tillage and water erosion for cultivated hillslope in the study basin Soil redistribution rate (t ha-1 yr-1) - -1-1 - 1 Tillage Water 1 2 3 4 6 7 8 9 1 11 12 13 14 1 16 17 18 19 21 22 23 24 2 26 27 Downslope distance (m*1) 3 3 2 1 1 (degree) Fig 6. Rates of soil redistribution for the hillslope with different land use in the study basin -14 3 Soil redistribution rate (t ha-1 yr-1) -1-1 -8-6 -4-4 6 Forestland Tillage Water Farmland 3 2 1 1 - (degree) 1 2 3 4 6 7 8 9 1 11 12 13 14 1 16 17 18 19 21 22 23 24 Downslope distance (m*1) 8