Published in J. Environ. Qual. 34:514-523 (2005).
© ASA, CSSA, SSSA
677 S. Segoe Rd., Madison, WI 53711 USA
TECHNICAL REPORTS
Landscape and Watershed Processes
Characterizing Land Surface Erosion from Cesium-137 Profiles in Lake and Reservoir Sediments
Xinbao Zhanga and
Desmond E. Wallingb,*
a Institute of Mountain Hazards and Environment, PO Box 417, Chengdu, China, and SKLLQG, Institute of Earth Environment, Chinese Academy of Sciences, Xian, China
b Department of Geography, University of Exeter, Amory Building, Exeter, EX4 4RJ, UK, and visiting professor at Institute of Mountain Hazards and Environment, Chengdu, China
* Corresponding author (d.e.walling{at}exeter.ac.uk)
Received for publication June 4, 2004.
 |
ABSTRACT
|
|---|
Recognition of the threat to the sustainable use of the earth's resources posed by soil erosion and associated off-site sedimentation has generated an increasing need for reliable information on global rates of soil loss. Existing methods of assessing rates of soil loss across large areas possess many limitations and there is a need to explore alternative approaches to characterizing land surface erosion at the regional and global scale. The downcore profiles of 137Cs activity available for numerous lakes and reservoirs located in different areas of the world can be used to provide information on land surface erosion within the upstream catchments. The rate of decline of 137Cs activity toward the surface of the sediment deposited in a lake or reservoir can be used to estimate the rate of surface lowering associated with eroding areas within the upstream catchment, and the concentration of 137Cs in recently deposited sediment provides a basis for estimating the relative importance of surface and channel, gully, and/or subsurface erosion as a source of the deposited sediment. The approach has been tested using 137Cs data from several lakes and reservoirs in southern England and China, spanning a wide range of specific suspended sediment yield. The results obtained are consistent with other independent evidence of erosion rates and sediment sources within the lake and reservoir catchments and confirm the validity of the overall approach. The approach appears to offer valuable potential for characterizing land surface erosion, particularly in terms of its ability to provide information on the rate of surface lowering associated with the eroding areas, rather than an average rate of lowering for the entire catchment surface.
|
INTRODUCTION
|
|---|
SOIL EROSION and associated off-site sedimentation are widely recognized as a serious threat to the sustainable use of the earth's resources (Lal, 2001; Pimentel et al., 1995; Clark et al., 1985). Soil erosion depletes the global soil resource, causing a reduction in both soil productivity and the area available for agriculture. Equally, the transfer of the eroded sediment to river systems leads to a range of detrimental off-site effects, including reservoir sedimentation, siltation of irrigation canals, waterways, and harbors, and degradation of aquatic habitats. Furthermore, any increase in the transfer of sediment and associated nutrients and contaminants from the land surface of the globe to the oceans will affect the global geochemical cycle. Against this background, there is an increasing need to assemble reliable information on soil erosion and sediment mobilization at the regional and global scale, to provide an improved assessment of the nature, magnitude, and broad-scale significance of the problem.
A range of measurement techniques have been developed for assembling information on rates of soil loss and sediment mobilization in landscapes dominated by fluvial processes. These techniques can conveniently be grouped into three main approaches. The first involves the use of erosion plots, from which the on-site sediment loss (Mg ha–1) can be readily measured during both individual events and over longer periods (see Lal, 1994). The second comprises measurements of the suspended sediment output from catchments or river basins (see Walling, 1994). In this case, the measured sediment load (Mg yr–1) can be converted to a specific sediment yield (Mg km–2 yr–1), which in turn provides an estimate of the average rate of erosion within the upstream catchment. The third involves the use of environmental radionuclides as sediment tracers and thus estimating rates of soil redistribution by erosion. Cesium-137, excess 210Pb, and 7Be have each been successfully used for this purpose (see Ritchie and McHenry, 1990; Wallbrink and Murray, 1996; Walling, 2002; Walling and He, 1999a; Walling et al., 1999; Zapata, 2002).
Each of these approaches possesses important limitations (see Loughran, 1989). Although erosion plots afford a means of generating measured values of soil loss from a clearly defined area, with known topographic and land use characteristics, they possess a key limitation in terms of the representativeness of the plot involved. In most cases plot lengths will be significantly less than those in the natural landscape. Furthermore, the simple rectilinear form associated with plots will frequently be unrepresentative of the natural topography, which is generally considerably more complex. Equally, a plot will only provide information for a very small area and a substantial number of plots would be necessary to sample the diversity of topographic and land use characteristics associated with most environments. The need to operate the plots for an extended period of time and the substantial costs involved also represent a further limitation of this approach.
By providing a spatially aggregated or averaged measure of sediment generation from a landscape, measurement of the sediment yield at the outlet of a catchment affords a means of overcoming the sampling problems associated with erosion plots, by producing an integrated value for the entire upstream catchment. However, this feature is, in itself, also a limitation, because the resultant estimate of sediment yield affords no indication of the degree of spatial variability or the range of erosion rates involved. Furthermore, the suspended sediment flux at a catchment outlet cannot be used directly to estimate rates of soil loss from its surface. In the first place, this flux will commonly be significantly lower than the on-site rate of soil loss within the catchment, because a large proportion of the sediment mobilized by erosion will be deposited within the catchment, at the base of slopes, in valley bottoms, and on river floodplains, before reaching the basin outlet (see Walling, 1983; Walling et al., 2001, 2002). Such conveyance losses are frequently represented in terms of the sediment delivery ratio, which expresses the sediment export from the catchment as a fraction of the gross mass of sediment mobilized by erosion within the catchment. For most larger river basins, this fraction can be expected to be of the order of 10% or less. In addition, it cannot be assumed that the sediment flux measured at a catchment outlet is derived exclusively from erosion of the catchment soils. In many cases, there will be a substantial contribution from channel erosion. Information on the relative importance of major sediment sources is clearly of critical importance for the design of effective sediment control strategies, because control measures must be targeted toward those sediment sources within a catchment that contribute large amounts of sediment. It must also be recognized that a lengthy period of intensive monitoring may be required to derive a reliable estimate of the sediment yield from a catchment. Such monitoring may prove costly.
Recent advances associated with the use of environmental radionuclides as sediment tracers (see Walling, 2003) can be seen as overcoming at least some of the limitations associated with the first two approaches discussed above. By providing the potential to assemble spatially distributed estimates of soil redistribution rates within the natural landscape, several of the key limitations associated with the use of erosion plots can be overcome. Equally, the potential to obtain retrospective estimates of longer-term erosion rates, on the basis of a single site visit, avoids the need for expensive long-term monitoring. Furthermore, the ability to assemble information on both erosion and deposition rates and thus the overall catchment sediment budget (see Walling et al., 2002) overcomes several of the limitations associated with the use of measurements of catchment sediment yield to obtain information on rates of soil loss. However, the use of environmental radionuclides for erosion rate assessment also possesses limitations. These relate primarily to the difficulty of applying the approach to anything other than a relatively small catchment and the substantial costs involved in analyzing the large number of soil cores that are likely to be required.
Against this background, there is a need to explore other approaches to assembling information on soil degradation and associated erosion rates at the regional and global scale and to explore the possibility of using surrogates for direct measurements. Assessments of sedimentation rates in lakes and reservoirs have frequently been used as an alternative means of establishing the longer-term sediment yield from the upstream catchment, which avoids the need for long-term monitoring of sediment loads. By collecting multiple cores from the sediment deposits, dating these cores, using core correlation techniques to estimate the total volume of sediment associated with specific horizons extending across the lake or reservoir basin, and taking account of the trap efficiency of the lake or reservoir, it is possible to estimate the total sediment input to the water body, and thus the sediment yield from the upstream catchment over different periods (see Dearing et al., 1981; Foster et al., 2003). Use of lake and reservoir sediment deposits to obtain information on sediment yields and sources for the upstream catchment is, however, commonly highly demanding in terms of both the time and resources required to obtain data for a single drainage basin. The potential advantage of this approach lies primarily in the ability to document longer-term changes in catchment response, rather than contemporary sediment fluxes and erosion rates (see Walling and Webb, 1983).
The ability of lake and reservoir sediments to integrate the response of the upstream catchment and to provide a retrospective estimate of the catchment sediment yield based on a limited program of contemporary measurements, must, however, be seen as a key advantage that merits further exploitation. This paper explores the potential for extending the use of environmental radionuclides for documenting catchment erosion rates to sediment deposits in lakes and reservoirs. Cesium-137 measurements have been undertaken on cores collected from a large number of lakes and reservoirs in many different areas of the world, to establish core chronologies. Because such data are readily available for many such water bodies, attention focuses on 137Cs and the use of these data to provide information on both rates of soil loss within the upstream catchment and the relative importance of major source types in contributing to the sediment output from the catchment above the lake or reservoir. In making use of existing 137Cs records for sediment cores, the approach taken should be seen as providing a reconnaissance tool applicable at the regional or global scale to the large number of lakes and reservoirs, rather than as an approach requiring detailed field sampling and associated laboratory analysis. In this respect, existing 137Cs data should be seen as potentially providing a surrogate for more direct measurements of erosion rates and sediment sources. However, scope clearly exists to extend existing 137Cs data sets to include additional lakes and reservoirs.
|
THE BASIS OF THE APPROACH
|
|---|
Cesium-137 is a man-made fallout radionuclide with a half-life of 30.2 yr that is present in the global environment, primarily as a result of the atmospheric testing of nuclear weapons in the late 1950s and early 1960s. Radiocesium produced by weapons testing was transferred into the stratosphere and the associated fallout was globally distributed. Deposition fluxes varied globally in response to global atmospheric circulation and the location of weapons testing activity as well as the global pattern of precipitation (see Walling, 2002). Fallout was greatest in the middle latitudes of the northern hemisphere and declined toward the equator, with deposition fluxes in the southern hemisphere being as much as an order of magnitude lower than in the northern hemisphere. The temporal pattern of annual fallout was, however, broadly similar across the globe and closely related to the intensity of weapons testing (see Fig. 1)
. Significant fallout was first recorded in the mid-1950s, maximum fallout occurred in the early 1960s, and fallout declined rapidly through the mid- and late 1960s and early 1970s, as a result of the nuclear test ban treaty imposed in 1963.
In most environments, 137Cs fallout reaching the land surface was rapidly and strongly fixed by the surface soil and its subsequent redistribution in the landscape was closely associated with erosion and associated soil redistribution processes (Ritchie and McHenry, 1990; Walling and Quine, 1995). In areas experiencing soil loss, inventories would have progressively reduced relative to those for stable sites, whereas in areas of deposition, inventories would have progressively increased. Typical vertical distributions of 137Cs in surface soils are shown schematically in Fig. 2
for both uncultivated (top graphs) and cultivated (bottom graphs) soils. In each case, vertical distributions associated with stable areas (A) and with areas experiencing erosion (B) and deposition (C) are presented.
View larger version (29K):
[in this window]
[in a new window]
|
Fig. 2. Characteristic depth distributions of 137Cs in uncultivated soils (top graphs) and cultivated soils (bottom graphs) for stable (A), eroding (B), and depositional (C) sites.
|
|
In the case of an uncultivated soil (Fig. 2, top graphs), the 137Cs depth distribution found at a stable site (A) is typically exponential in form, with concentrations declining rapidly with depth and most of the radiocesium inventory occurring in the upper 10 cm. The slow downward migration of the radiocesium fallout reflects both physical and chemical processes, and particularly the influence of soil biota in causing physical mixing within the surface horizons of the soil. Eroding sites (B) are typically characterized by "truncated" 137Cs profiles, and therefore both reduced 137Cs concentrations at the surface and a reduced total 137Cs inventory, whereas at depositional sites (C), the progressive accumulation of additional soil containing 137Cs results in an upward "stretching" of the radiocesium profile and an increased inventory relative to the stable site. As erosion proceeds, the 137Cs content of mobilized sediment will progressively decline, because sediment is mobilized from further down the exponential profile. Similarly, the 137Cs in deposited sediment will progressively decline and the 137Cs depth distribution associated with a depositional site could be expected to evidence reducing concentrations toward the surface.
At cultivated sites (Fig. 2, bottom graphs), the mixing associated with plowing and other tillage practices causes the 137Cs to be mixed into the plow horizon. Radiocesium concentrations are therefore commonly relatively constant throughout the plow layer, but decline rapidly below the plow depth. At eroding sites (B), the depth distribution will again be characterized by an essentially constant 137Cs concentration within the plow layer, but this concentration will be lower than that associated with a stable site (A), because removal of soil from the surface by erosion causes soil from below the original plow depth, which contains little or no 137Cs, to be mixed into the profile by plowing and associated tillage. For a site experiencing deposition (C), the 137Cs profile will again provide evidence of upward "stretching" associated with the progressive deposition of soil containing 137Cs eroded from upslope. As with uncultivated sites, as erosion and deposition proceed, the 137Cs content of mobilized sediment will progressively decline and the 137Cs concentrations associated with the upper part of the profile from a depositional site will similarly decline. Overall, 137Cs concentrations associated with cultivated soil profiles can generally be expected to be lower than those associated with uncultivated soils, due the mixing of the 137Cs fallout input within a greater soil depth or volume by tillage activities.
Figure 3
provides a schematic representation of a typical 137Cs depth distribution in a sediment core recovered from a lake or reservoir. Such profiles have been widely used to estimate sedimentation rates or to date recent horizons, because the depth at which the peak 137Cs concentration is found can be equated with the year of peak 137Cs fallout (i.e., 1963 in the northern hemisphere and 1964 in the southern hemisphere). The precise shape of the 137Cs profile recovered from a lake will reflect both the rate of sedimentation and the relative importance of direct 137Cs fallout to the lake surface and of 137Cs inputs associated with sediment mobilized from the surrounding catchment and deposited in the lake (see Walling and He, 1992). In most lakes and reservoirs, both contributions will be significant and a general indication of their relative importance can be obtained by comparing the total 137Cs inventory associated with a lake core with that for an undisturbed site within the catchment, which will provide an estimate of the direct fallout input to the lake surface. Assuming there is a significant contribution of 137Cs associated with sediment eroded from the upstream catchment, it is possible to interpret the 137Cs profile obtained from a lake or reservoir to provide information on the nature of the erosion processes operating within the upstream catchment and the erosion rates involved.
View larger version (13K):
[in this window]
[in a new window]
|
Fig. 3. A schematic representation of a typical 137Cs depth profile in the bottom sediment of a lake or reservoir. The term Cdn represents the 137Cs concentration found at the surface of the deposited sediment and Cd0 the concentration found at the depth associated with sediment deposited in 1970.
|
|
As indicated in Fig. 1, 137Cs fallout associated with weapons testing became insignificant after the early 1970s and it is thus possible to assume that the 137Cs activity in the upper part of the profile obtained from a lake or reservoir deposited post 1970, is influenced primarily by inputs of 137Cs bound to sediment mobilized from the surrounding catchment and transported to the lake. Attention therefore focuses on this upper part of the profile, which can be identified by establishing the depth of the 1963 peak in 137Cs concentration, estimating the annual rate of sediment accretion (by assuming a constant rate of sedimentation between the 1963 peak and the surface of the sediment and a uniform bulk density for this portion of the core), and estimating the depth in the core representing 1970. In those areas of Europe and adjacent regions influenced by 137Cs fallout from the Chernobyl disaster, this additional input will complicate the simple procedure outlined above. Although this additional input could be readily taken in account in the interpretation described below, attention is here focused on bomb fallout and those areas of the world where Chernobyl fallout was effectively absent or at least insignificant.
|
INTERPRETING CESIUM-137 PROFILES IN LAKE AND RESERVOIR SEDIMENTS TO PROVIDE INFORMATION ON CATCHMENT EROSION RATES AND SEDIMENT SOURCES
|
|---|
The sediment delivered to a lake or reservoir from its catchment could be derived from two major groups of sources, first, the catchment surface where the sediment will be mobilized from the upper portion of the soil profile, and, second, channel erosion and other locations where the sediment is mobilized from lower in the soil profile. Because the 137Cs content of the latter group of sources is likely to be close to zero, any upcore trend in the 137Cs concentration of the deposited sediment will reflect changes in the 137Cs content of sediment eroded from surface sources. These changes in concentration will in turn directly reflect the erosion rates associated with the main sediment source areas on the catchment surface. If these erosion rates are low, the concentration of 137Cs in the upper portion of the sediment core could be expected to remain essentially constant. However, where erosion rates are higher, the 137Cs content of the mobilized sediment could be expected to decline through time. The degree of reduction of 137Cs concentration in the upper part of a sediment core (
C) can therefore provide information on the magnitude of the surface erosion rate associated with the main sediment source areas within a catchment. These source areas may represent only a small proportion of the catchment and the associated erosion rate is likely to be substantially greater than the average erosion rate for the entire catchment surface. Procedures for estimating the erosion rate from information on the rate of reduction of 137Cs concentration within the upper portion of a sediment core are provided later in this paper.
The magnitude of the 137Cs concentrations associated with the upper portion of a sediment core (i.e., deposited post-1970) can also provide information, albeit generalized, on the relative importance of different sediment sources within the catchment. If channel and subsurface sources are dominant, 137Cs concentrations in the deposited sediment will be low. Where surface sources are more important, concentrations will be higher. Procedures for estimating the relative importance of surface and channel and subsurface sources will again be considered in a subsequent section of this paper.
Estimating Surface Erosion Rates from C
Cultivated Soils
The change in the concentration of 137Cs within the plow layer (depth H, cm) of a cultivated soil over a period of n years after 1970, in response to a given annual rate of surface lowering (h, cm yr–1), can be expressed as follows:
 | [1] |
where Cm0 and Cmn represent the mean 137Cs concentration in the soil (Cm) at the beginning (Cm0) and end (Cmn) of the period of n years, respectively. By rearranging Eq. [1], the rate of surface lowering h can be expressed as a function of the change in the 137Cs concentration in the plow layer over the period of n years, that is:
 | [2] |
In a typical cultivated catchment, most of the land will have been cultivated for a long period and the sediment yield at the catchment outlet, and thus the sediment deposited in the lake or reservoir, is likely to comprise contributions from surface erosion of the cultivated land and from channel erosion and subsurface sources. Assuming that the relative contributions from surface and subsurface sources have remained essentially constant over the period of n years and that the grain size composition and 137Cs content of the component of the sediment entering the lake or reservoir derived from surface sources are similar to those of the eroding source material, the change in the 137Cs content of the deposited sediment can be used to provide an estimate of the average change in 137Cs concentration within the plow layer of the cultivated soils within the catchment. By modifying Eq. [2], these concentration values can then be used to estimate the annual rate of soil loss h from the eroding cultivated soils in the catchment, that is:
 | [3] |
where Cd0 and Cdn represent the 137Cs concentrations within the sediment core at given depths (d), representing the beginning of the period under consideration (e.g., 1970) (Cd0) and the time that the core was collected (Cdn) (i.e., the surface), respectively.
Assuming a typical value for H of 20 cm and a value of n of 30 yr, the relationship between the annual rate of surface lowering or soil loss and the change in the 137Cs concentration in the core over this period, expressed as the ratio of the final to the initial concentration (Cdn/Cd0), is presented in Fig. 4
. The annual rate of surface lowering or soil loss is estimated to be 0.07 cm yr–1 for Cdn/Cd0 = 0.9, 0.5 cm yr–1 for Cdn/Cd0 = 0.5, and 1.48 cm yr–1 for Cdn/Cd0 = 0.1. These values should not be seen as representing the mean rate of soil loss from the cultivated areas of the catchment, but rather the average rate of soil loss from the eroding areas. The latter can be expected to be substantially greater than the former.
View larger version (9K):
[in this window]
[in a new window]
|
Fig. 4. The relationship between the mean annual depth of surface lowering on the eroding areas within a cultivated catchment (H = 20 cm) and the rate of decline of 137Cs concentrations (Cdn/Cd0) in the upper (post-1970) horizons of a sediment core collected from a lake or reservoir at the outlet of the catchment.
|
|
Uncultivated Soils
The depth distribution of 137Cs in a soil profile at an uneroding site on uncultivated land is commonly represented by an exponential function of the form:
 | [4] |
where x is the depth measured from the soil surface, C0 is the concentration at the surface, Cx is the 137Cs concentration at depth x (Bq kg–1), and 
is a coefficient describing the profile shape. At eroding sites, the 137Cs content of the surface soil will progressively decline as soil is removed from the surface of the profile. Over a given time of n years, the change in the 137Cs content of soil mobilized from the surface of the profile will reflect the depth of soil removed by erosion x, which will in turn be directly related to the rate of surface lowering or soil loss h (cm yr–1), that is:
 | [5] |
and:
 | [6] |
As with a core retrieved from a lake or reservoir with a cultivated catchment, the downcore change in the 137Cs concentrations associated with a sediment core retrieved from a representative location in a lake or reservoir at the outlet of an uncultivated catchment can be assumed to represent the change in the 137Cs concentration eroded from the surface of the eroding areas over the same period. The erosion rate h (cm yr–1) can therefore be calculated as:
 | [7] |
where Cdn and Cd0 represent the 137Cs concentrations at the surface of the core and at depth, for the portion of the sediment core deposited during the period of n years since 1970.
Assuming a period n of 30 yr and a typical value for of 0.25, based on the authors' existing work, the relationship between the annual rate of surface lowering or soil loss and the change in the 137Cs content of the deposited sediment represented by Eq. [7] is presented in Fig. 5
. The annual rate of surface lowering or soil loss is estimated to be 0.02 cm yr–1 for a value of Cdn/Cd0 = 0.9, 0.09 cm yr–1 when Cdn/Cd0 = 0.5 and 0.31 cm yr–1 when Cdn/Cd0 = 0.1. Again it must be emphasized that these values of h represent the rate of surface lowering or soil loss from the actively eroding areas, rather than the mean rate of soil loss from the upstream catchment. The former rate is likely to be substantially greater than the latter. Comparison of Fig. 4 and 5 indicates that for a given value of Cdn/Cd0, the annual rate of surface lowering is much greater for a cultivated than for an uncultivated catchment, because in the case of a cultivated catchment the 137Cs is mixed into the plow layer and a substantially greater rate of soil loss is required to produce a given reduction in the 137Cs content of the eroded soil.
View larger version (11K):
[in this window]
[in a new window]
|
Fig. 5. A typical relationship between the mean annual depth of surface lowering on the eroding areas within an uncultivated catchment ( = 0.25 cm) and the rate of decline of 137Cs concentrations (Cdn/Cd0) in the upper (post-1970) horizons of a sediment core collected from a lake or reservoir at the outlet of the catchment.
|
|
Estimating Source Contributions
The magnitude of the 137Cs concentration (Cdn) associated with the top of the upper (i.e., post-1970) portion of a sediment core recovered from a lake or reservoir can also provide information on the relative contribution of surface and channel or subsurface sources to that sediment, providing it is known whether the surface sources are likely to be dominantly cultivated or uncultivated soils. Application of a simple mixing model provides the following relationships:
 | [8] |
 | [9] |
where Ps and Pb are the relative contributions from surface and bank and subsurface sources, respectively, and Cs and Cb are the 137Cs concentrations associated with sediment inputs to the lake or reservoir derived from surface and bank and/or subsurface sources in the upstream catchment, respectively.
Assuming that the sediment mobilized from bank and/or subsurface sources contains no 137Cs, due to the accumulation of 137Cs fallout within the surface soil and the limited fallout receipt by the vertical or near-vertical surfaces of river banks, Eq. [8] can be rewritten as:
 | [10] |
or:
 | [11] |
To use Eq. [11] to estimate Ps, it is necessary to provide a value for Cs. To establish a value for Cs, it is important to consider the following. First, it is necessary to know whether the surface sources in the catchment are dominated by cultivated or uncultivated sources, because this will exert an important influence on the value of Cs. Second, in most situations it will be difficult, if not impossible, to establish the value for Cs directly, because this represents the 137Cs content associated with recent sediment derived from surface sources that has been deposited in the reservoir. Such information is very unlikely to be available, because, as indicated in the introduction to this paper, the aim of the approach being described is to provide information for catchment areas where detailed monitoring of erosion and associated sediment fluxes has not been undertaken. A meaningful estimate of Cs can, however, be obtained by considering Eq. [1] and [5], calculating the 137Cs concentration existing at the surface of a hypothetical cultivated or uncultivated soil profile eroding at rate h, and assuming that the relationship between Cs and the 137Cs content of the surface soil takes the form:
 | [12] |
where P is a particle size correction factor and Css is the 137Cs content at the surface of the eroding soil, with Cssu representing an uncultivated soil and Cssc a cultivated soil. The particle size correction factor P is required to take account of contrasts in grain size composition between the soil and the depositing sediment, because it is well known that 137Cs is preferentially associated with the finer fractions. If the depositing sediment is finer than the original soil, its 137Cs concentration is likely to be enhanced, whereas, if it is coarser, the concentration is likely to be reduced, relative to the original soil. In most instances it will be difficult to establish the value for P without detailed field measurements, but existing work suggests that it is not unreasonable to assume, at least as a first approximation, that P = approximately 1 for most lake catchments.
If h is estimated for a particular lake catchment using Eq. [2] or Eq. [7] and the local reference inventory is known, an estimate of Cssu or Cssc can be derived by considering the behavior of the eroding soil. If, in the case of a cultivated catchment, it is assumed, as a simplification, that the total bomb-derived input was received in 1963, Cssc can be estimated using a simple mass balance model (see Walling and He, 1999b), namely:
 | [13] |
where A0 is the local 137Cs reference inventory (Bq m–2) and 
is the bulk density of the surface soil. For an uncultivated catchment, again assuming as a simplification that the total 137Cs input was received in 1963, Cssu can be estimated from the exponential depth distribution, namely:
 | [14] |
The reference inventory A0, which represents the inventory for a stable uneroding site where the value directly reflects the local fallout input, will be available for many lake and reservoir catchments where 137Cs measurements have been undertaken. However, in the absence of direct measurements of this parameter, it can be estimated from information on the location of the catchment and the mean annual precipitation using the procedure implemented by Walling and He (2001).
|
TESTING THE APPROACH
|
|---|
The River Danube near Vienna
As a first step in testing the approach outlined above, attention was directed to examining data on the 137Cs content of fine sediment transported by a river, to confirm that this declined through time in response to ongoing erosion in the upstream catchment. Such data have been reported by Maringer and Hrachowitz (personal communication, 2003) for the river Danube near Vienna, Austria, which drains a catchment area of approximately 100000 km2, for the period 1970 to 1985 (i.e., before the Chernobyl accident). These data, which represent one of the few available records of the changes in the 137Cs content of fine sediment transported by a river during this period, indicate that the 137Cs content of the suspended sediment progressively declined during this period. The rate of decline was faster than the radioactive decay, providing a general confirmation of the reasoning presented above. In the absence of direct fallout, the 137Cs content of the fine sediment reflects that of the sediment mobilized from within the catchment by erosion. The concentration reported for 1970 was approximately 25 Bq kg–1, whereas by 1985 this had reduced to approximately 12 Bq kg–1. If it is assumed that this sediment was deposited in a lake and account is taken of radioactive decay, these values would be represented by concentrations of 12.6 and 8.5 Bq kg–1, respectively, in a sediment core collected in 2000. Insertion of these values into Eq. [3] and [7] provides estimates of the average annual rate of surface erosion within the catchment of 0.51 cm yr–1 for the cultivated model (H = 20 cm) and 0.10 cm yr–1 ( = 0.25) and 0.05 cm yr–1 ( = 0.50) for the uncultivated model. Because the land use within this large catchment comprises substantial proportions of both cultivated and uncultivated (e.g., permanent pasture and forest) land, the rate of surface erosion is likely to lie between these two sets of values (e.g., approximately 0.3 cm yr–1). In view of the large catchment area involved and thus the considerable spatial variability of both land use and 137Cs fallout inputs, it is not possible to estimate the relative importance of surface and channel and/or subsurface erosion in contributing to the suspended sediment load of the river Danube, but this element of the approach will be tested in subsequent examples.
Five Small Lakes and Reservoirs in Southern England
In considering test cases representing individual lakes and their catchments, an attempt has been made to span a range of examples that are known to be characterized by different rates of surface erosion. The first example is provided by the 137Cs measurements undertaken on sediment cores collected from five small lakes and reservoirs in southern England reported by He et al. (1996) (Table 1). Specific sediment yields in this region are relatively low by world standards. The 137Cs profiles associated with sediment cores collected from these water bodies are presented in Fig. 6
and provide clear evidence of a subsurface 137Cs peak that can be related to the peak of bomb fallout in 1963. Above this depth, concentrations show an initial sharp decline, followed by a more gentle decline after the main period of bomb fallout. Attention focused on the post-1970 portions of the cores, with the 1970 level being established by assuming a constant sedimentation rate between 1963 and the time the core was collected. As indicated in Table 1, the land use of each of the catchments was dominated by permanent grassland and other undisturbed land and the model for uncultivated land provided by Eq. [7] was used to estimate the surface erosion rates associated with the individual catchments. Typical values for the profile shape coefficient of both 0.25 and 0.5 have been used, although it is suggested that a value of 0.5 is more appropriate bearing in mind that the exponential depth distribution is likely to develop progressively through time in response to a slow downward "diffusion" of the fallout input and is thus best characterized by a value for that is greater than that currently associated with a stable undisturbed site. On this basis, annual rates of soil loss from the eroding areas within the lake and reservoir catchments were estimated to range between 0.04 and 0.09 cm yr–1 (see Table 2) and thus to be relatively low. Based on existing knowledge of the suspended sediment yields of small catchments in this area of the UK (see Walling and Webb, 1987), values in the rage 25 to 60 Mg km–2 yr–1 are judged to be typical for the five study catchments. This is equivalent to an annual rate of surface lowering for the overall catchment of approximately 0.0025 to 0.0060 cm yr–1, assuming a soil bulk density of approximately 1 g cm–3. These values are more than an order of magnitude lower than the values of h listed in Table 2, but this difference is not unexpected and is likely to reflect two factors. First, it is well known that a substantial proportion of the sediment mobilized by erosion within a catchment will be deposited before reaching the catchment outlet or the lake or reservoir, and thus the values of h are likely to be substantially greater than the values of surface lowering estimated from the sediment yield. Second, the difference will reflect the fact that surface erosion within the catchments is likely to be concentrated within a relatively small area, rather than occurring over most of the catchment. This is again consistent with existing knowledge of surface runoff generation and erosion within such catchments.
View larger version (29K):
[in this window]
[in a new window]
|
Fig. 6. Cesium-137 depth distributions derived for sediment cores recovered from five lakes and reservoirs in southern England.
|
|
View this table:
[in this window]
[in a new window]
|
Table 2. Estimates of the rate of surface lowering h for eroding areas and the proportion of sediment contributed by surface sources for the five small lakes and reservoirs in southern England.
|
|
Using the values of 137Cs concentration associated with the upper portions of the cores from the five small lakes or reservoirs and estimates of the 137Cs reference inventories provided by He et al. (1996) and assuming that the land use of each of the catchments was dominated by undisturbed or uncultivated soils, Eq. [11] and [14] have been used to estimate the relative contributions of surface and bank and/or subsurface sources to the sediment output from the five catchments. The results presented in Table 2 indicate that surface erosion is the dominant sediment source (>70%) in all of the catchments, and that in three of the catchments surface sources contribute 100% of the total sediment input to the water bodies. This is again consistent with existing knowledge of sediment sources in small UK catchments (see Walling et al., 2001).
The Zhaojia Gully Catchment, China
The second example is provided by a small reservoir catchment in China, namely the Zhaojia Gully (3.86 km–2), previously investigated by the authors (see Zhang et al., 1997). The Zhaojia Gully catchment is located in the Rolling Loess Plateau of China, a region characterized by some of the highest erosion rates and specific suspended sediment yields in the world. The catchment has a mean altitude of 1050 m and a relative relief of 300 m and is underlain by thick loess deposits. The relatively gently sloping plateau surface, which occupies 53% of the catchment area, was extensively cultivated up to the mid-1990s, when the development of a local oil industry that provided employment for the local people and the introduction of soil conservation programs caused a reduction in agricultural activity. The remaining steep gullies are largely uncultivated, except for a few small cultivated areas at the toe of the gully slopes. Sheet and rill erosion were dominant on the cultivated land of the rolling plateau, whereas gully erosion and mass movements are dominant in the gully areas. The relative sediment contributions to the sediment yield at the catchment outlet from the plateau and gully areas during the period of extensive cultivation, before the reduction in agricultural activity on the plateau surface, were estimated by Zhang et al. (1997) to be approximately 25 and 75%, respectively, providing a specific sediment yield from the plateau surface of approximately 6300 Mg km–2 yr–1, before the mid-1990s. Sampling of sediment deposits in sediment detention reservoirs, and dating of these deposits using historical records of dam construction and knowledge of recent events, provided estimates of 137Cs activity (corrected to 2001) in the deposits of 1.15 Bq kg–1 in 1973 and 0.75 Bq kg–1 in 1993. Using the model for cultivated catchments (Eq. [3]), the annual rate of soil loss from the cultivated land in the catchment was estimated to be approximately 0.61 cm yr–1 (H = 20 cm, n = 20). This value is convincingly close to the value of 0.63 cm yr–1 provided by the estimate of specific sediment yield from the plateau surface, assuming a bulk density of 1 g cm–3 and a sediment delivery ratio close to 100%, as reported for the gullied loess plateau by Mou and Meng (1980). The close coincidence of the two estimates of the annual rate of soil loss from the plateau surface further suggests that erosion occurs across the entire area, rather than being concentrated in more limited areas. This is in turn consistent with existing reports of widespread severe erosion on the plateau surface before 1995.
Using the value of the 137Cs concentration associated with sediment deposited in the sediment control reservoirs in 1994 and an estimate of the 137Cs content of soil eroded from the plateau surface, Cssc, derived using Eq. [13], the proportion of the sediment deposited in the reservoir derived from surface erosion (Ps) was estimated to be 22%. This is again consistent with the estimated 25%:75% contributions from the plateau and the gully areas respectively, reported previously for this catchment (Zhang et al., 1997).
The Wujia Gully Catchment, China
The final example is provided by the Wujia Gully, a small catchment in the hilly Central Sichuan Basin of China recently investigated by the authors. This catchment extends to 0.22 km2 and has an average elevation of 450 m and a relative relief of 100 m. The catchment is underlain by horizontally bedded Mesozoic mudstones, siltstones, and sandstones, which weather to produce the characteristic "purple" soils. The topography typically comprises steep sandstone cliffs with slopes of approximately 30° separated by gentle terraces of approximately 5°, underlain by mudstones and siltstones. The gentle terrace slopes and the steep slopes account for approximately one-third and two-thirds of the catchment area, respectively. The gentle terraces have been cultivated for centuries, whereas the steep slopes were originally covered by natural grassland but have gradually been afforested with cypress trees. A small reservoir was constructed at the outlet of the catchment in 1956 and nearly all the sediment mobilized from the upstream catchment after that date has been deposited in the reservoir. A sampling program undertaken in 2002 indicated that about 5000 m3 of sediment had been deposited in the reservoir since 1956, with a maximum sediment depth of 1.3 m. Based on the volume of sediment deposited in the reservoir, the mean annual specific sediment yield of the catchment was estimated to be approximately 640 Mg km–2 yr–1.
Cesium-137 measurements undertaken on a core collected from the reservoir in May 2002 indicated that the 137Cs concentration in the sediment decreased from 3.6 Bq kg–1 in 1977 to 2.9 Bq kg–1 in 1997. Because most of the sediment mobilized from the catchment is likely to have been derived from the cultivated terrace areas, the cultivated model (Eq. [3]) has been used to estimate the rate of surface erosion, h. A value of 0.18 cm yr–1 was obtained (H = 20 cm). If the contribution from the forested steeper slopes of the catchment, which is likely to be relatively low compared with that from the cultivated terrace areas, is ignored, this is equivalent to a sediment yield from the catchment of approximately 720 Mg km–2 yr–1. This value is consistent with the annual sediment yield calculated from the reservoir deposits (approximately 640 Mg km–2 yr–1), because although the forested steeper slopes and gully areas are likely to contribute additional sediment, the overall sediment delivery ratio for the catchment, although high, will be <1.0.
Using the value of 137Cs concentration associated with sediment deposited in the sediment control reservoir in 1997 and an estimate of the 137Cs content of soil eroded from the cultivated terraces, Cssc, derived using Eq. [13], the proportion of the sediment deposited in the reservoir derived from surface erosion (Ps) was estimated using Eq. [11] to be 70%. This is substantially greater than the value for the Zhaojia Gully, but is consistent with the lack of extensive gully or channel erosion in the Wujia Gully catchment and the perception that the cultivated terraces are the most important sediment source.
|
DISCUSSION AND CONCLUSIONS
|
|---|
The results from the several case studies presented above serve to validate and confirm the viability of the proposed approach for using data on 137Cs profiles in lake and reservoir sediments to estimate both rates of soil loss in the upstream catchments and the relative importance of surface erosion and channel and gully erosion and other subsurface sources in contributing to the sediment input to the water body. The case studies include lake and reservoir catchments representing a range of conditions extending from southern England, where sediment yields and rates of soil loss are relatively low, through Sichuan Province in China, where sediment yields are an order of magnitude greater, to the loess region of China, which is characterized by some of the highest sediment yields and erosion rates recorded anywhere in the world (see Walling and Webb, 1996). The successful application of the approach to such a wide range of conditions clearly adds further weight to its validation.
While the estimates of soil loss rates are derived directly from measurements of the rate of change of 137Cs concentration in the sediment profile, the information on the relative importance of surface and channel and gully erosion and other subsurface sources involves additional calculations to estimate the 137Cs concentration in eroded sediment, which necessarily introduce some additional uncertainty into the results obtained. In both cases, however, it is necessary to assume that the catchment surface is dominated by either cultivated or uncultivated land, and this could be seen as a limitation of the approach. In many catchments this assumption will be acceptable, but in others, where both cultivated and uncultivated land account for appreciable proportions of the total catchment area, it could compromise the key assumptions underlying the approach. In many regions of the world, this problem could be expected to increase as catchment area increases, because larger catchments commonly embrace greater diversity of land use.
A key feature of the approach described is its ability to provide an estimate of the rate of surface lowering or soil loss from the eroding areas within a catchment. This is in contrast to attempts to estimate erosion rates from measurements of sediment yield from a catchment, where it is necessary to estimate the sediment delivery ratio to take account of conveyance losses of eroded sediment within the catchment, and it is generally assumed that the sediment is mobilized from the entire catchment surface. In many cases, the rate of surface lowering associated with the areas of active erosion within a catchment could be several times greater than the average rate of surface lowering from the catchment surface. This is because the main sediment source areas within a catchment are frequently restricted in terms of spatial extent, reflecting either erosion from specific fields under a particular land use or specific areas of the catchment, the location of which is controlled by the topography and the existence of saturated areas. For the lake and reservoir catchments in southern England, there is more than an order of magnitude difference between the rates of surface lowering associated with the main sediment source areas and the average rate of surface lowering for the catchment as a whole, as indexed by the specific sediment yield. As the intensity of erosion within a catchment increases, the specific sediment yield also increases. In this situation, the sediment source areas are likely to occupy a greater proportion of the catchment surface and the difference between the rate of surface lowering estimated using the approach described above and the average rate for the catchment surface will decline. In the case of the Zhaojia and Wujia Gully catchments, erosion occurs over most of the cultivated area and there is little difference between the rate of surface lowering estimated using the new approach described in this paper and the average value for the catchment surface derived from the measured specific sediment yield.
The approach described in this paper is heavily dependent on obtaining a reliable estimate of the upcore decline in 137Cs activity within the sediment deposits of a lake or reservoir. The sediment cores, to which the 137Cs data relate, must therefore be representative of the sediment deposits from which they were collected. Equally, the sediment deposits themselves should provide a meaningful record of the changing 137Cs content of sediment deposited in the water body over the last 30 to 40 yr. This in turn requires that there is no significant remobilization and redistribution of the sediment deposit within the waterbody and that the 137Cs profile is not significantly influenced by bioturbation or other mixing processes. In this context, cores collected from the deeper central areas of a water body are likely to produce the best results. However, these requirements are likely to be common to most, if not all, studies requiring data on the 137Cs profile in a sediment core and thus most existing information is likely to meet these requirements. Furthermore, although deposition rates, and thus the depth of the accumulating sediment, are likely to vary spatially within a lake or reservoir basin, this variability should have little effect on estimates of the upcore decline in 137Cs activity (i.e., C), which is effectively independent of deposition rate.
|
ACKNOWLEDGMENTS
|
|---|
The authors gratefully acknowledge the financial support provided by the University of Exeter, the Institute of Mountain Hazards and Environment in Chengdu, the China National Natural Science Foundation (NNSF 40271015), the Chinese Academy of Sciences (KZCX3-SW 422 and 330), and the International Atomic Energy Agency (IAEA 12322/RO), to facilitate their collaboration and the research reported, and the help of Helen Jones in producing the figures.
|
REFERENCES
|
|---|
- Clark, E.H., J.A. Haverkamp, and W. Chapman. 1985. Eroding soils: The off-farm impacts. The Conservation Foundation, New York.
- Dearing, J.A., J.K. Elner, and C.M. Happey-Wood. 1981. Recent sediment flux and erosional processes in a Welsh upland lake catchment based on magnetic susceptibility measurements. Quat. Res. 16:356–372.
- Foster, I.D.L., A.S. Chapman, R.A. Hodgkinson, A.R. Jones, J.A. Lees, S.E. Turner, and M. Scott. 2003. Changing suspended sediment and particulate phosphorus loads and pathways in underdrained lowland agricultural catchments; Herefordshire and Worcestershire, UK. Hydrobiologia 494:119–126.
- He, Q., D.E. Walling, and P.N. Owens. 1996. Interpreting the Cs-137 profiles observed in several small lakes and reservoirs in southern England. Chem. Geol. 129:115–131.
- Lal, R. (ed.) 1994. Soil erosion research methods. Soil and Water Conserv. Soc., Ankeny, IA.
- Lal, R. 2001. Soil degradation by erosion. Land Degrad. Dev. 12:519–539.
- Loughran, R.J. 1989. The measurement of soil erosion. Prog. Phys. Geogr. 13:216–233.
- Mou, J., and Q. Meng. 1980. Sediment delivery ratio as used in the computation of the watershed sediment yield.
- Pimentel, D., C. Harvey, P. Resosudarmo, K. Sinclair, D. Kurz, M. McNair, S. Crist, L. Shpritz, L. Fitton, R. Saffouri, and R. Blair. 1995. Environmental and economic cost of soil erosion and conservation benefit. Science (Washington, DC) 267:1117–1123.[Abstract/Free Full Text]
- Playford, K., J. Toole, and T. Sanders. 1995. Radioactive fallout in air and rain: Results to the end of 1993. DOE Rep. DOE/RAS/94.004. Dep. of the Environ., London.
- Ritchie, J.C., and J.R. McHenry. 1990. Application of radioactive fallout cesium-137 for measuring soil erosion and sediment accumulation rates and patterns: A review. J. Environ. Qual. 19:215–233.[Abstract/Free Full Text]
- Wallbrink, P.J., and A.S. Murray. 1996. Determining soil loss using the inventory ratio of excess lead-210 to cesium-137. Soil Sci. Soc. Am. J. 60:1201–1208.[Abstract/Free Full Text]
- Walling, D.E. 1983. The sediment delivery problem. J. Hydrol. (Amsterdam) 65:209–237.
- Walling, D.E. 1994. Measuring sediment yield from river basins. p. 39–80. In R. Lal (ed.) Soil erosion research methods. Soil and Water Conserv. Soc., Ankeny, IA.
- Walling, D.E. 2002. Recent advances in the use of environmental radionuclides in soil erosion investigations. p. 279–301. In Nuclear techniques in integrated plant nutrient, water and soil management. IAEA C&S Papers Ser. 11/C. Int. Atomic Energy Agency, Vienna.
- Walling, D.E. 2003. Using environmental radionuclides as tracers in sediment budget investigations. p. 57–78. In Erosion and sediment transport measurement in rivers: Technological and methodological advances. Proc. of the Oslo Workshop. June 2002. Int. Assoc. of Hydrol. Sci. Publ. 283. IAHS Press, Wallingford, UK.
- Walling, D.E., A.L. Collins, H.M. Sichingabula, and G.J.L. Leeks. 2001. Integrated assessment of catchment sediment budgets. Land Degrad. Dev. 12:387–415.
- Walling, D.E., and Q. He. 1992. Interpretation of cesium-137 profiles in lacustrine and other sediments: The role of catchment-derived inputs. Hydrobiologia 235/236:219–230.
- Walling, D.E., and Q. He. 1999a. Using fallout lead-210 measurements to estimate soil erosion on cultivated land. Soil Sci. Soc. Am. J. 63:1404–1412.[Abstract/Free Full Text]
- Walling, D.E., and Q. He. 1999b. Improved models for estimating soil erosion rates from cesium-137 measurements. J. Environ. Qual. 28:611–622.[Abstract/Free Full Text]
- Walling, D.E., and Q. He. 2001. Models for converting 137Cs measurements to estimates of soil redistribution rates on cultivated and undisturbed soils (including software for model implementation). Report to International Atomic Energy Agency. Univ. of Exeter, UK.
- Walling, D.E., Q. He, and W. Blake. 1999. Use of Be-7 and Cs-137 measurements to document short- and medium-term rates of water-induced soil erosion on agricultural land. Water Resour. Res. 35:3865–3874.
- Walling, D.E., and T.A. Quine. 1995. The use of fallout radionuclide measurements in soil erosion investigations. p. 597–619. In Nuclear techniques in soil-plant studies for sustainable agriculture and environmental preservation. IAEA Publ. STI/PUB/947. Int. Atomic Energy Agency, Vienna.
- Walling, D.E., M.A. Russell, R.A. Hodgkinson, and Y. Zhang. 2002. Establishing sediment budgets for two small lowland agricultural catchments in the UK. Catena 47:323–353.
- Walling, D.E., and B.W. Webb. 1983. Patterns of sediment yield. p. 69–100. In K.J. Gregory (ed.) Background to palaeohydrology. John Wiley & Sons, Chichester, UK.
- Walling, D.E., and B.W. Webb. 1987. Suspended load in gravel-bed rivers: UK experience. p. 691–723. In C.R. Thorne, J.C. Bathurst, and R.D. Hey (ed.) Sediment transport in gravel-bed rivers. John Wiley & Sons, Chichester, UK.
- Walling, D.E., and B.W. Webb. 1996. Erosion and sediment yield: A global overview. p. 3–9. In J. Bogen, T. Fergus, and D.E. Walling (ed.) Erosion and sediment yield: Global and regional perspectives. Proc. of the Exeter Symp. July 1996. Int. Assoc. of Hydrol. Sci. Publ. 236. IAHS Press, Wallingford, UK.
- Zapata, F. (ed.) 2002. Handbook for the assessment of soil erosion and sedimentation using environmental radionuclides. Kluwer, Dordrecht, the Netherlands.
- Zhang, X., D.E. Walling, T.A. Quine, and A. Wen. 1997. Use of reservoir deposits and cesium-137 measurements to investigate the erosion response of a small drainage basin in the rolling Loess Plateau region of China. Land Degrad. Dev. 8:1–16.
Related articles in JEQ:
- This Issue in Journal of Environmental Quality
JEQ 2005 34: 403-407.
[Full Text]
TOP
ABSTRACT
INTRODUCTION
