Published online 9 August 2006
Published in J Environ Qual 35:1756-1763 (2006)
DOI: 10.2134/jeq2005.0410
© 2006 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
TECHNICAL REPORTS
Landscape and Watershed Processes
Use of Beryllium-7 to Document Soil Redistribution following Forest Harvest Operations
Paulina Schullera,
Andrés Irouméb,
Desmond E. Wallingc,*,
Héctor B. Mancillab,
Alejandra Castilloa and
Rosa E. Trumpera
a Facultad de Ciencias, Universidad Austral de Chile, Casilla 567, Valdivia, Chile
b Facultad de Ciencias Forestales, Universidad Austral de Chile, Casilla 567, Valdivia, Chile
c Department of Geography, University of Exeter, Amory Building, Rennes Drive, Exeter EX4 4RJ, UK
* Corresponding author (D.E.Walling{at}exeter.ac.uk)
Received for publication October 27, 2005.
 |
ABSTRACT
|
|---|
Rapid and reliable methods for documenting soil erosion associated with forest harvest operations are needed to support the development of best management practices for soil and water conservation. To address this need, the potential for using 7Be measurements to estimate patterns and amounts of soil redistribution associated with individual post-harvest events was explored. The 7Be technique, which was originally developed for use on agricultural land, was employed to estimate soil redistribution associated with a period of heavy rainfall within a harvested forest area located in the Lake Region of Chile (39°44'7'' S, 73°10'39'' W; 22% slope; and mean annual rainfall 2300 mm yr–1). The results provided by the 7Be technique were validated against direct measurements of soil gain or loss during the same period obtained using erosion pins. The information produced by the two approaches was similar. The results of this study demonstrate the potential for using 7Be measurements to document event-based erosion in recently harvested forest areas.
|
INTRODUCTION
|
|---|
RAPID AND RELIABLE methods of documenting soil erosion and soil degradation within harvested forest areas are urgently needed, to provide forest managers with accurate tools for assessing the benefits of adopting best management practices for soil and water conservation, under International Organization for Standardization (ISO) and/or Forest Stewardship Council (FSC) certification requirements. In the Lake Region of Chile, final harvesting in forestry plantations commonly takes place in the drier months of the year, between October and April. However, after clearcutting and during the early stages of the establishment of a new plantation, soils remain bare during the rainy season and are, therefore, susceptible to soil redistribution processes. Several practices, including the construction of cross banks on access roads, skid trails, and landings; the rapid re-establishment of a vegetation cover on disturbed areas; and improved management of streamside zones and logging debris, have been adopted to reduce erosional impacts and to limit the delivery of sediment to watercourses, to maintain water quality standards (Park et al., 1994; Griffin, 1995; Wallbrink et al., 2002). There is, however, an important need to assess the effectiveness of such management practices in reducing soil degradation and sediment export from the harvested slopes. More particularly, there is a need for measurement techniques that are capable of documenting both gross and net erosion rates and thus the sediment delivery ratio within recently clearcut forest terrains. This paper reports an investigation aimed at exploring the potential for using 7Be measurements as a basis for quantifying event-based soil redistribution following forest harvesting and thus for assessing the effectiveness of improved management practices.
|
USING BERYLLIUM-7 MEASUREMENTS TO DOCUMENT SHORT-TERM SOIL REDISTRIBUTION RATES
|
|---|
Background
The successful use of 137Cs (t1/2 = 30.2 yr) and 210Pb (t1/2 = 22.2 yr) measurements as basis for documenting medium-term rates and patterns of soil redistribution by erosion processes has now been reported for many areas of the world (Ritchie and McHenry, 1990; Walling and Quine, 1990; Walling and He, 1999a, 1999b; Schuller et al., 2004) and the 137Cs technique is now well established in soil erosion investigations (Zapata, 2002). The 137Cs and 210Pb techniques are, however, unsuitable for documenting short-term erosion rates associated with forest harvesting, because of the relatively long half-lives of these two radionuclides. Furthermore, their use for estimating medium-term erosion rates in forest areas is also likely to be compromised by the spatial variability of fallout inputs under a forest canopy and the extensive soil disturbance frequently caused by harvesting machinery and the construction of roads, skidtrails, and landings. More recently, the potential for using the natural occurring fallout radionuclide 7Be (t1/2 = 53.3 d) to document short-term soil redistribution on agricultural land has been reported by Blake et al. (1999), Walling et al. (1999), Matisoff et al. (2002), and Wilson et al. (2003). The principles involved in using 7Be to estimate patterns of soil redistribution are essentially the same as those associated with the 137Cs and 210Pb techniques, but 7Be would appear to overcome many of the limitations facing the use of 137Cs and 210Pb in forest areas and, more particularly, for documenting erosion following harvest operations. After clearcutting, fallout inputs of 7Be will not be influenced by canopy interception and can therefore be assumed to be homogeneous within a local area receiving the same rainfall amount. Furthermore, because of its short half-life and short residence time in the soil, it can be used to document short-term soil redistribution. In addition, its restricted penetration into the soil means that a substantial proportion of the areal activity density can be removed during a single runoff event and this is likely to increase the accuracy of the resulting erosion estimates (Wilson et al., 2003).
Beryllium-7 (
-energy, E = 477.6 keV) is formed primarily in the stratosphere from cosmic-ray spallation of nitrogen and oxygen nuclei, although some is produced in the troposphere and also in situ on the surface of the earth (Lal et al., 1958; Lal and Suess, 1968; Kaste et al., 2002). The amount of cosmogenic 7Be that reaches the surface of the earth as fallout is a function of the production rate (cosmic-ray intensity), stratosphere–troposphere mixing, circulation and advection within the troposphere, and the efficiency of its removal from the troposphere by wet and dry deposition (Feely et al., 1989). The production of 7Be depends on the cosmic-ray flux, which varies with latitude, altitude, and solar activity. Solar activity maxima result in increased deflection of cosmic rays from the solar system that decreases the cosmic-ray flux to the earth, and in turn decreases the 7Be production (Gerasopoulos et al., 2003; Papastefanou and Ioannidou, 2004). Fluctuations of 7Be concentrations in surface air resulting from the 11-yr solar cycle are generally of the order of 15 to 25% (Koch and Mann, 1996). Stratosphere–troposphere exchange can increase 7Be concentrations in the troposphere and near-surface air. Maximum mixing between the stratosphere and the troposphere generally occurs in the spring or summer at mid-latitudes, and near-surface air at mid-latitudes generally has higher concentrations of 7Be at this time (Feely et al., 1989; Al-Azmi et al., 2001). Beryllium-7 attaches to airborne particles and its deposition to the earth's surface occurs continuously by wet and dry fallout (Olsen et al., 1985). Wallbrink and Murray (1994) demonstrated that 7Be fallout deposition is mainly associated with precipitation. It is delivered primarily as Be2+ in slightly acid rainfall. The Be2+ ion is extremely competitive for cation exchange sites, because of its high charge density (Kaste et al., 2002). As 7Be comes in contact with soils and vegetation, it is rapidly sequestered by exchange surfaces, and is therefore rapidly and strongly fixed by the surface soil (Hawley et al., 1986; Wallbrink and Murray, 1996; Kaste et al., 2002). Existing evidence suggests that 7Be is commonly found only in the upper approximately 10 mm of the soil profile (Wallbrink and Murray, 1996; Blake et al., 1999; Walling et al., 1999).
The successful use of 7Be to document both the spatial patterns associated with short-term (frequently event-based) soil redistribution on agricultural land and the associated erosion and deposition has been reported by Blake et al. (1999) and Walling et al. (1999). The approach employed is based on comparison of the 7Be areal activity density (Bq m–2) measured at a sampling point with a reference areal activity density determined for a nearby undisturbed and stable reference site, where neither erosion nor deposition are thought to have occurred. Depletion of the 7Be areal activity density relative to the reference value provides evidence of erosion, whereas areas of sediment deposition are characterized by increased areal activity densities. By coupling information on the degree of decrease or increase in the areal activity density with information on the characteristic depth distribution of 7Be within the surface soil, estimates of amounts of erosion and deposition can be obtained. A similar approach would seem to be applicable to forest areas.
On the basis of the known rapid and strong fixation of 7Be fallout to surface soils (Wallbrink and Murray, 1996) and of previously reported depth distributions of 7Be in soils (Blake et al., 1999; Walling et al., 1999), it can be assumed that the initial vertical distribution of the mass activity density of the 7Be within the soil will be characterized by an exponential decrease with depth. Based on this premise, Walling et al. (1999) and Blake et al. (1999) proposed a simple model for converting measurements of the 7Be areal activity density into estimates of the intensity of soil redistribution. The components of this model are described below using terms defined according to the International Commission on Radiation Units and Measurements (2001).
The Initial Depth Distribution of the Beryllium-7 in the Soil
With x, kg m–2, representing the mass depth of the soil measured from the surface (positive downward) and C(x), Bq kg–1, the mass activity density of 7Be at depth x, it is assumed that the depth distribution can be represented as:
 | [1] |
where C(0) is the initial mass activity density of the surface soil (at x = 0) and ho, kg m–2, is the relaxation mass depth.
The reference areal activity density, Aref, Bq m–2, is defined as the initial total areal activity at an uneroded stable site or reference site in the study area:
 | [2] |
The areal activity density below depth x, A(x), Bq m–2, for the initial distribution is therefore:
 | [3] |
The relaxation mass depth describes the shape of the initial depth distribution of both the mass activity density (Eq. [1]) and areal activity density (Eq. [3]) of 7Be in the soil. By setting x = ho and exp(–1) = 0.368 in Eq. [3], it follows that the areal activity density below the relaxation mass depth is:
 | [4] |
and that 63.2% of the total areal activity density of 7Be will be found within the 0 to ho soil layer, or above ho. Consequently, the greater the value of ho, the deeper will be the penetration of the radionuclide into the soil.
By measuring the mass activity density, C, in different depth increments of soil collected from the reference site and establishing the mass depth of each depth increment, the values of A(x) for corresponding mass depths x down the reference profile can be calculated. Logarithmically transforming Eq. [3], ho and the areal activity density Aref can be deduced from a linear regression between Ln[A(x)] and x.
Soil Loss at a Sampling Point
Assuming that erosion has removed a thin layer of mass depth h, kg m–2, at a sampling point within the study site, the 7Be areal activity density remaining at this eroded point, A, Bq m–2, will be lower than Aref. The soil mass eroded per unit area, R, kg m–2, is equal to the mass depth h removed. By setting x = h in Eq. [3], the remaining areal activity density at this point can be calculated as:
 | [5] |
The mass of soil per unit area eroded at the sampling point, R, kg m–2, can therefore be calculated as:
 | [6] |
Sediment Deposition at a Sampling Point
When the measured 7Be areal activity density, A', Bq m–2, for a sampling point within the study site is higher than Aref, deposition is assumed to have occurred at this point. An estimate of the sediment mass deposited per unit area, R', kg m–2, can be derived by dividing the areal activity density in excess of Aref by the mean 7Be mass activity density of the deposited sediment, Cd, Bq kg–1, that is:
 | [7] |
The 7Be mass activity density of the sediment eroded from a point, Ce, Bq kg–1, can be estimated from the areal activity density lost at that point divided by the mass of sediment eroded per unit area, that is:
 | [8] |
Consequently, the mean 7Be mass activity density of the deposited sediment Cd can be estimated as the weighted mean mass activity density, Ce, of sediment mobilized from the upslope contributing area S:
 | [9] |
Using the parameters Aref and ho established for the initial 7Be vertical distribution in the soil (i.e., for the reference site) and Eq. [6] and [7], the amounts of soil eroded or deposited at individual sampling points within a study area, and thus the spatial pattern of soil redistribution, can be established.
Key Assumptions for Applying the Beryllium-7 Technique
The procedure for using 7Be measurements to estimate erosion rates described above involves three key assumptions. The first is that the deposition of 7Be fallout associated with the erosive event is spatially uniform. The second is that any preexisting 7Be present within the surface soil of the study area is also uniformly distributed across the area. The third is that the 7Be deposited during the event is rapidly fixed by the soil at its point of receipt and can therefore only be mobilized by erosion of soil particles. The first condition is likely to be fulfilled over a relatively small area, where rainfall intensity and therefore 7Be deposition can be realistically treated as uniform. In the second case, the short half-life of 7Be means that any preexisting spatial variability in the 7Be areal activity density caused by past erosive events will quickly be lost through radioactive decay, provided that erosive events are separated by a period of sufficient length (e.g., two half-lives or 106 d). Contributions to the 7Be areal activity density associated with intervening rainfall events that do not cause significant erosion can be assumed to be spatially uniform. Further experimental work is required to provide a comprehensive validation of the third assumption, but existing evidence suggests that in most environments 7Be fallout inputs will be rapidly and firmly fixed by the surface soil on receipt (Wallbrink and Murray, 1996; Blake et al., 1999). In addition, the procedure necessarily assumes that erosion is restricted to the depth of 7Be penetration and that any variation in soil properties between the reference site and the study area or across the study area does not influence the vertical distribution of 7Be in the soil.
|
MATERIALS AND METHODS
|
|---|
Field and Laboratory Methods
The procedures described above have been applied to document soil redistribution associated with a period of heavy rainfall within an area of clearcut forest shortly after harvesting. The study site is located within the Forest Research Centre of the Universidad Austral de Chile (39°44'7'' S, 73°10'39'' W), near the city of Valdivia, in the Lake Region of Chile. Soils in the area are Typic Paleudults (Ultisol) (CIREN, 2001), developed from old eolic sediments deposited on micashists of the metamorphic coastal complex. This provides deep reddish brown clayey soils with moderate to high infiltration capacity. The site is characterized by a mean slope of about 22%, a temperate climate, and a mean annual rainfall of approximately 2300 mm yr–1 (Huber, 2004). Most rainfall falls between autumn and spring, when high rainfall intensities can occur. A digital rain gauge installed at the site provides a continuous record of precipitation with a resolution of 0.2 mm.
Clearcutting of the forest plantation ended in January 2003 (summer). Immediately after harvesting, the site was prepared for a new plantation and the soils were further disturbed and exposed as heavy machinery entered the site to move the debris remaining after harvesting. The trash was gathered into rows along contour lines, approximately 50 m apart, to reduce sediment movement from the area. A new pine (Pinus radiata D. Don) forest was planted in May 2003, but the soil remained bare for several months after planting. The low rainfall recorded during the summer and early autumn months ensured that the 7Be areal activity density was spatially uniform across the study site before the onset of the late autumn rains in early June 2003. Although substantial amounts of rainfall were recorded during June and July 2003 (see Fig. 1
) the substantial soil moisture deficit produced by the preceding dry summer, the moderate to high infiltration capacity of the soil, and the lack of intense rainfall meant that there was no visible evidence of soil redistribution during this period. Most rainfall events during July 2003 were characterized by maximum intensities of less than 5 mm h–1. There was little rainfall in early August 2003, but between mid August and mid September 2003, there was a period of very heavy rainfall, producing a total of 311 mm in 25 d (Fig. 1A) and including three 1-h periods in mid August where rainfall intensity exceeded 13.0 mm h–1 (Fig. 1B), that caused visible erosion at the study site. The absence of any visual evidence of surface runoff or surface erosion and soil redistribution before this period of very heavy rainfall meant that it was reasonable to assume that the 7Be areal activity density of the site immediately before the period of heavy rainfall was characterized by minimal spatial variability. Furthermore, the much lower intensities associated with the portion of the rainfall that occurred during early September, when the maximum intensity was only 6.8 mm h–1 and thus only about 50% of that in mid August, means that it is also reasonable to assume that most of the erosion associated with this period of heavy rainfall occurred during the storms in mid August.
View larger version (19K):
[in this window]
[in a new window]
|
Fig.1. Daily precipitation totals for the study period and the preceding month (A) and hourly precipitation totals for the period with maximum rainfall intensities (B).
|
|
Immediately after the period of heavy rainfall, the soils of the study site were sampled for 7Be measurement on 12 September, to explore the potential for documenting the soil redistribution that occurred during this period. A plot 3 m wide and 12 m long, with its longer side parallel to the predominant direction of water flow, located equidistant between two rows of woody residue was selected for detailed investigation. Forty shallow soil cores (4 cm deep) were collected from the plot at the intersections of a 0.5- x 1-m grid, using 10.6-cm-diameter cylindrical plastic core tubes. An adjacent flat area located within the harvested stand that showed no evidence of erosion or deposition was selected as a reference site and two sets of six bulk cores were collected at the intersections of a 1- x 1-m grid using the same plastic core tubes, to determine the reference areal activity density and to characterize the relaxation mass depth of 7Be for the initial depth distribution. For the latter purposes, each group of six cores was sectioned into 2-mm slices in the laboratory and the slices representing specific depth increments were bulked for measurement as a single sample. A special device was constructed for sectioning the bulk cores. This comprised a plunger (with the same diameter as the soil core) moved by a screw thread, which extrudes the soil core from the core tube by 1 mm per turn. The soil extruded by rotating the screw two turns (2 mm) was then separated from the core using a sharp pallet knife. All the soil samples were initially air dried and then dried at 105°C in an oven, before being ground and homogenized in readiness for measurement of their 7Be content by gamma spectrometry. To homogenize its 7Be content, each sample was mixed for 25 min using a shaker mixer (Turbula T2 F; Willy A. Bachofen Maschinenfabrik, Basel, Switzerland) and the homogenized samples were transferred to 81.3-mL Petri dishes for gamma assay. The gamma counting was undertaken using a Canberra high-purity Ge detector (Canberra Industries, Meriden, CT) of 28% relative efficiency. The detector was calibrated for the selected measuring geometry using standards, produced by labeling soil with known amounts of standard solution supplied by the Physikalisch-Technische Bundesanstalt (PTB, Braunschweig, Germany). The standard solution contained radionuclides with gamma emitters covering the range 81 to 1836 keV, which permitted the counting efficiency for the 477.6 keV emission of 7Be to be estimated by interpolation. Spectra were analyzed using Genie 2000 software (Canberra Industries). Due to the low 7Be concentrations in the samples analyzed, the count time was set to 20 h per sample, which provided a measurement precision of better than ±10% at the 95% level of confidence. The 7Be detection limit for this measuring time and for the measuring geometry employed was estimated to be approximately 10 Bq kg–1.
As indicated in Eq. [6] and [7], the estimates of erosion and deposition provided by the 7Be measurements are heavily dependent on the values of ho and Aref employed. Special attention was therefore directed to determining these two parameters, within the constraints imposed by the counting times required, the short half-life of 7Be, and the number of samples that needed to be analyzed to produce a representative estimate of soil redistribution within the study plot. To estimate the reference areal activity density and the relaxation mass depth of the 7Be initial depth distribution, the mean mass activity density (C, Bq kg–1) of the composite 2-mm slices obtained from the two groups of cores collected at the reference site was calculated. The mean areal activity density A(x) below each mass depth x was then calculated from these values. The mass depth of each slice or depth increment was calculated by dividing the total dry mass of the composite sample by the total internal cross-sectional area of the coring cylinders used to collect the samples. The reference areal activity density Aref and the relaxation mass depth ho were deduced from a linear regression between the values of Ln[A(x)] and x, as inferred from Eq. [3].
The vertical distribution of the 7Be mass activity density observed at the reference site was also used to determine the depth down to which the 7Be concentration exceeded the detection limit. This depth was used to define the portion of the bulk cores collected along the slope transects to be analyzed for 7Be, to maximize the counting efficiency. Soil from above this depth plus an additional 4 mm, to account for vertical extension of the 7Be depth distribution as a result of sediment deposition, was analyzed. The remainder of the core was discarded.
To assess the validity of the soil redistribution amounts documented using the 7Be measurements, independent estimates of soil redistribution associated with the period of heavy rainfall were obtained for 30 regularly distributed points located within a grid superimposed on the same plot from which the soil cores were collected, using erosion pins. The metal erosion pins had been inserted into the soil at the intersections of a 0.5- x 1-m grid on 12 August, immediately before the onset of the period of heavy rainfall. When inserted, 10.0 cm of the pin was left exposed above the soil surface and the position of the soil surface was remeasured after the period of heavy rainfall using a Vernier caliper. These measurements were made on the same day as the soil cores for 7Be analysis were collected. The mass depth, kg m–2, of soil lost or gained at each pin position (grid intersection) was calculated by multiplying the change in the position of the soil surface, relative to the original 10.0-cm reference level, by the bulk density of the surface soil layer. For this purpose, the bulk density of the surface soil layer in the vicinity of each pin was estimated from the measurements of dry mass and volume of the upper 16-mm slice obtained from the two adjacent cores collected for 7Be analysis.
|
RESULTS
|
|---|
Reference Areal Activity Density and Relaxation Mass Depth of the Initial Beryllium-7 Depth Distribution
Figure 2
shows the linear regression between the natural logarithm of the areal activity density, Ln[A(x)], and the mass depth, x, based on the samples collected from the reference site. The relaxation mass depth estimated from this regression (r = 0.9999**) is 2.14 kg m–2, and the reference areal activity density is 573 Bq m–2.
View larger version (7K):
[in this window]
[in a new window]
|
Fig. 2. The linear regression between the natural logarithm of the 7Be areal activity density and mass depth derived for the reference site.
|
|
Figure 3A
depicts the depth distribution of the mean 7Be mass activity density for the undisturbed reference area. Figure 3B shows the vertical distribution of the mean 7Be areal activity density, as deduced from the mass activity distribution shown in 3A. The functions fitted to the downward decrease of the 7Be mass activity density [C(x) = 268 exp(–x/2.14)] and of the areal activity density [A(x) = 573 exp(–x/2.14)] at the reference site are represented by the continuous curves shown in Fig. 3A and 3B, respectively.
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 3. The vertical distribution of 7Be at the reference site. (A) A plot of the mass activity density versus mass depth. (B) The reduction of the areal activity density with increasing mass depth.
|
|
The total areal activity density (Aref) calculated from the direct measurements of the 7Be mass activity density in the depth increments at the reference site is 564 ± 60 Bq m–2. This result is in very close agreement with the 7Be areal activity density (562 Bq m–2) estimated by subtracting from the total areal activity density calculated by linear regression (573 Bq m–2) the areal activity density below 8.5 kg m–2, where 7Be was found to be below the detection limit:
 | [10] |
Because the estimates of soil redistribution are based on the values of areal activity density measured along the slope transects, which exclude the 7Be content below the depth where 7Be activity falls below the detection limit, a reference areal activity of 564 Bq m–2 was used when applying the 7Be model.
Estimation of Soil Redistribution within the Study Plot
Since there was no visual evidence that the rainfall that occurred before the period of heavy rainfall in mid August and early September rains had caused significant erosion or soil redistribution, it is reasonable to assume that at the beginning of the study period the 7Be areal activity density showed little variation across the study site and that the variability of the areal activity values documented after the period of heavy rainfall reflected soil redistribution that occurred during this period, and primarily during the heavy rainfall that occurred during mid August. However, the possibility that some limited erosion and 7Be redistribution occurred before the period of heavy rainfall cannot be totally excluded and must be seen as introducing some uncertainty into the results obtained from the 7Be measurements.
The magnitude of the soil redistribution rates, kg m–2, documented along the four transects using the 7Be approach described above (Eq. [6] and [7]), is shown in Fig. 4A
. Erosion is shown as negative values and deposition as positive values. These results indicate that the amounts of erosion and deposition are randomly distributed along the slope transects, in response to the local microtopography.
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 4. The spatial distribution of the soil erosion (negative values) and sedimentation (positive values), kg m–2, along the slope transects of the study site. The decimal point of each number coincides with the position of the corresponding sampling point. The dashed lines represent contour lines. (A) Soil redistribution values determined using the 7Be approach, and (B) those obtained from the erosion pins.
|
|
A summary of the results obtained from the transects is provided in Table 1. Soil loss was documented at 27 out of the 40 sampled points, which were characterized by a mean erosion of 0.92 ± 0.18 kg m–2 and sediment deposition was documented at the remainder of the sampled points, for which the mean deposition was 0.72 ± 0.14 kg m–2. The period of heavy rainfall resulted in a gross erosion (total erosion divided by the total area of the study plot) of 0.62 ± 0.12 kg m–2 and a net erosion from the study plot of 0.39 ± 0.08 kg m–2. The overall sediment delivery ratio for the plot was estimated to be 62%.
View this table:
[in this window]
[in a new window]
|
Table 1. The soil redistribution documented for the study site using 7Be measurements and direct erosion pin measurements.
|
|
The significant amounts of erosion documented by the 7Be measurements demonstrate that, although the construction of linear trash barriers may be effective in reducing sediment delivery to watercourses, the soil redistribution occurring within the intervening areas could still represent a significant cause of soil degradation.
The magnitudes of the soil redistribution rates presented above are reasonably consistent with those reported by Iroumé et al. (1989), for a plot study of soil erosion immediately after clearcutting, undertaken in a nearby area with a similar soil type and similar bare soil conditions. That study recorded the soil loss from a series of three 20-m-long x 2-m-wide erosion plots with different slope gradients over a 120-d period extending from August to November 1988, when a total rainfall of 715 mm was recorded. The plots had slopes of 30, 50, and 60% and the net soil loss measured at their outlets was 0.16, 0.19, and 0.39 kg m–2, respectively. In this case more than the 90% of the erosion was associated with events accounting for a total rainfall of 340 mm during the 120-d study period. The somewhat higher estimate of net soil loss provided by the 7Be measurements for the period of heavy rainfall in 2003 may reflect (i) differences in the amount and duration of rainfall associated with individual storm events; (ii) differences in forest management practices, whereby heavier equipment was used to clear and prepare the soil of the clear-felled area for planting in 2003 than in 1988; and (iii) the isolation of the erosion plots on the overall slope and the prevention of upslope run-on.
Comparison of the Beryllium-7 Results with those Obtained by Direct Measurements using Erosion Pins
The values of soil redistribution provided by the 7Be measurements are also compared with those generated by the direct measurements using erosion pins in Table 1 and in Fig. 4. As with the 7Be results, the values of erosion and deposition rate derived from the pins (Fig. 4B) are randomly distributed along the slope transects and there is evidence of considerable small scale variability in the pattern of erosion and deposition. In view of this local variability, it is not possible to make a direct pairwise comparison between the results provided by the 7Be measurements and the erosion pins, because they relate to different locations within the plot (Fig. 4). However, the net erosion of 0.32 ± 0.06 kg m–2 derived from the erosion pin measurements is of a very similar magnitude to that provided by the 7Be measurements (0.39 ± 0.08 kg m–2). Equally, the values for the proportion of the measuring points characterized by erosion and deposition and the mean erosion and deposition amounts for those points are also very similar to those provided by the 7Be technique, particularly when the uncertainties associated with these values and the associated standard error statistics are taken into account. Furthermore, it is important to recognize that the results provided by the 7Be measurements and the erosion pins are not directly comparable, since the values relate to different points within the study plot and the erosion pin measurements covered only 75% of the plot area sampled by the 7Be measurements. Moreover, whereas the erosion pins provide an estimate of surface lowering or accretion in the immediate vicinity of the pin, the 7Be results represent an average amount of erosion or deposition for the 88.2-cm2 surface area of the individual cores. Bearing in mind these considerations, the close similarities between the results provided by the 7Be technique and the erosion pins appear to provide a clear validation of the results obtained from the 7Be measurements.
|
DISCUSSION AND CONCLUSIONS
|
|---|
The results presented above provide a clear demonstration of the potential for using 7Be measurements to assemble information on the magnitude of event-based or short-term erosional losses and sediment redistribution within forest areas immediately after clear-felling. Although the approach has a number of limitations, including the need to ensure that the areal activity density was uniform across the study site immediately before the period of heavy rainfall investigated and the need for specialist gamma counting facilities, it also offers a number of important advantages over the use of more conventional approaches, such as erosion plots and the use of erosion pins. In the case of erosion plots, high installation and operation costs, the need to install the plots well in advance of the study period, and uncertainties over the extent to which a small bounded plot is representative of the broader landscape will frequently limit their application. Equally, with erosion pins, there is again a need to install the pins in advance of the period of measurement and important uncertainties exist regarding the extent to which the disturbance of the soil caused by inserting the pin and the disruption of the hydraulic conditions caused by the presence of the pin might produce unrepresentative results. Key advantages of the 7Be approach include the ability to undertake retrospective investigations initiated after a period of heavy rain, the lack of a need for expensive field installations or the introduction of unrepresentative conditions, and the ability to undertake the necessary sampling during a single site visit, shortly after the period of heavy rainfall to be investigated. A clear advantage over the erosion plot is the ability to assemble spatially distributed information on amounts of erosion and deposition and thus to document soil redistribution within the study site, as well as the net soil loss from the site. Equally, an important advantage over the use of erosion pins for assembling spatially distributed data is the fact that the results represent average values for the surface area represented by the core, rather than the much smaller area represented by the pin, and are thus less likely to be influenced by small scale anomalies and/or variability associated with soil redistribution.
One important requirement of the 7Be approach is the need to ensure that the areal activity density of the radionuclide is essentially constant across the study site, before the erosional event under investigation. This is achieved either by ensuring that the event investigated is separated from previous events by a period of sufficient length to ensure that any spatial heterogeneity caused by soil redistribution associated with those events is removed by radioactive decay, or, as in this study, by ensuring that the rainfall during the preceding period did not cause significant soil redistribution and thus 7Be redistributrion. These conditions may be difficult to fulfil in study areas, such as that associated with the investigation reported here, where the annual precipitation is high and heavy rainfall is a frequent occurrence during the wet season. Further work is required to explore the potential for developing improved conversion models for use with the 7Be approach, that are capable of documenting soil redistribution produced by several successive periods of heavy rainfall.
|
ACKNOWLEDGMENTS
|
|---|
The work reported in this paper was undertaken as part of investigations on soil erosion funded by the IAEA Coordinated Research Programme D1-50-08 through the Contracts CHI-12321 and UK-12094. It was also supported by the Centro Experimental Forestal and Dirección de Investigación y Desarrollo, Univesidad Austral de Chile. The paper benefited from the comments and suggestions provided by three anonymous reviewers and Associate Editor Professor Satish Gupta.
|
REFERENCES
|
|---|
- Al-Azmi, D., A.M. Sabed, and H.A. Yatim. 2001. Variations in the 7Be concentrations in the atmosphere of Kuwait during the period 1994 to 1998. Appl. Radiat. Isot. 55:413–417.[CrossRef][ISI][Medline]
- Blake, W.H., D.E. Walling, and Q. He. 1999. Fallout beryllium-7 as a tracer in soil erosion investigations. Appl. Radiat. Isot. 51:599–605.[CrossRef][ISI]
- CIREN. 2001. Estudio agrológico, X Región. Descripción de suelos materiales y símbolos, Tomo I. Centro de Información de Recursos Naturales, Corporación de Fomento de la Producción, Santiago, Chile.
- Feely, H.W., R.J. Larsen, and C.G. Sanderson. 1989. Factors that cause seasonal variations in beryllium-7 concentrations in surface air. J. Environ. Radioact. 9:223–249.[CrossRef]
- Gerasopoulos, E., C.S. Zerefos, C. Papastefanou, P. Zanis, and K. O'Brien. 2003. Low-frequency variability of beryllium-7 surface concentrations over the Eastern Mediterranean. Atmos. Environ. 37:1745–1756.[CrossRef]
- Griffin, C.B. 1995. Uncertainty analysis of BMP effectiveness for controlling nitrogen from urban nonpoint sources. Water Resour. Bull. 31(6):1041–1050.
- Hawley, N., J.A. Robbins, and B.J. Eadie. 1986. The partitioning of 7-beryllium in fresh water. Geochim. Cosmochim. Acta 50:1127–1131.[CrossRef][ISI]
- Huber, A. 2004. Archivos meteorológicos. Univ. Austral de Chile, Valdivia, Chile.
- International Commission on Radiation Units and Measurements. 2001. Quantities, units and terms in radioecology. Nucl. Tech. Publishing. Rep. 65. IRCU, Ashford, Kent, UK.
- Iroumé, A., J. Gayoso, and L. Infante. 1989. Erosión hídrica y alteración del sitio en cosecha a tala rasa. Rev. Ecol. Biol. Sol 26(2):171–180.
- Kaste, J.M., S.A. Norton, and C.T. Hess. 2002. Environmental chemistry of beryllium-7. Rev. Mineral. Geochem. 50:271–289.
- Koch, D.M., and M.E. Mann. 1996. Spatial and temporal variability of 7Be surface concentrations. Tellus Ser. B 48B:387–396.[CrossRef]
- Lal, D., P.K. Malhotra, and B. Peters. 1958. On the production of radioisotopes in the atmosphere by cosmic radiation and their application to meteorology. J. Atmos. Terr. Phys. 12:306–328.[CrossRef]
- Lal, D., and H.E. Suess. 1968. The radioactivity of the atmosphere and hydrosphere. Annu. Rev. Nucl. Sci. 18:407–434.[CrossRef][ISI]
- Matisoff, G., E.C. Bonniwell, and P.J. Whiting. 2002. Soil erosion and sediment sources in an Ohio watershed using beryllium-7, cesium-137, and lead-210. J. Environ. Qual. 31:54–61.[Abstract/Free Full Text]
- Olsen, C.R., I.L. Larsen, P.D. Lowry, N.H. Cutshall, J.F. Todd, G.T.F. Wong, and W.H. Casey. 1985. Atmospheric fluxes and marsh-soil inventories of 7Be and 210Pb. J. Geophys. Res. 90:10487–10495.
- Papastefanou, C., and A. Ioannidou. 2004. Beryllium-7 and solar activity. Appl. Radiat. Isot. 61:1493–1495.[CrossRef][ISI][Medline]
- Park, S.W., S. Mostaghimi, R.A. Cooke, and P.W. McClellan. 1994. BMP impacts on watershed runoff, sediment, and nutrient yields. Water Resour. Bull. 30(6):1011–1023.
- 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.[ISI]
- Schuller, P., D.E. Walling, A. Sepúlveda, R.E. Trumper, J.L. Rouanet, I. Pino, and A. Castillo. 2004. Use of 137Cs measurements to estimate changes in soil erosion rates associated with changes in soil management practices on cultivated land. Appl. Radiat. Isot. 60: 759–766.[CrossRef][ISI]
- Wallbrink, P.J., and A.S. Murray. 1994. Fallout of 7Be in south eastern Australia. J. Environ. Radioact. 25:213–228.[CrossRef]
- Wallbrink, P.J., and A.S. Murray. 1996. Distribution and variability of 7Be in soils under different surface cover conditions and its potential for describing soil redistribution processes. Water Resour. Res. 32:467–476.[CrossRef]
- Wallbrink, P.J., B.P. Roddy, and J.M. Olley. 2002. A tracer budget quantifying soil redistribution on hillslopes after forest harvesting. Catena 47:179–201.
- Walling, D.E., and Q. He. 1999a. Improved models for estimating soil erosion rates from 137Cs measurements. J. Environ. Qual. 28: 611–622.[ISI]
- Walling, D.E., and Q. He. 1999b. Using fallout lead-210 measurements to estimate soil erosion on cultivated land. Soil Sci. Soc. Am. J. 63(5):1404–1412.[Abstract/Free Full Text]
- Walling, D.E., Q. He, and W. Blake. 1999. Use of 7Be and 137Cs measurements to document short- and medium-term rates of water-induced soil erosion on agricultural land. Water Resour. Res. 35(12):3865–3874.
- Walling, D.E., and T.A. Quine. 1990. Calibration of caesium-137 measurements to provide quantitative erosion rate data. Land Degrad. Rehabil. 2:161–175.
- Wilson, C.G., G. Matisoff, and P.J. Whiting. 2003. Short-term erosion rates from a 7Be inventory balance. Earth Surf. Processes Landforms 28:967–977.[CrossRef]
- Zapata, F. 2002. Handbook for the assessment of soil erosion and sedimentation using environmental radionuclides. Kluwer Academic Publ., Dordrecht, the Netherlands.
TOP
ABSTRACT
INTRODUCTION
