Tillage erosion intensity in the South Canterbury Downlands, New Zealand.
There is growing awareness of the pressure on land resources at global, regional, and national scales (e.g. Greenland et al. 1997) and, especially in developed countries, there is an increasing interest in land management strategies that have low environmental cost and are sustainable in both the short and the long term. In this context, it is clearly essential to develop an understanding of the range of processes that operate in human-impacted systems and the rates of change associated with them. In cultivated agricultural environments, soil erosion is widely recognised as one of the most important threats to long-term, low-cost sustainability. Nevertheless, there is still considerable uncertainty concerning the most important processes of erosion and their associated rates. Until recently, there has been a frequently implicit and occasionally explicit (Boardman and Bell 1992) assumption that processes of erosion by water dominate soil redistribution on agricultural land. The validity of this assumption has been challenged by research demonstrating the importance of wind erosion (Basher et al. 1995; Basher and Painter 1997) and, more significantly, over the last decade by a series of studies that have demonstrated both the erosion potential of tillage (Lindstrom et al. 1990, 1992; Revel et al. 1993; Govers et al. 1994; Lobb et al. 1995; Poesen et al. 1997; Quine et al. 1999; Van Muysen et al. 1999) and the dominance, in many areas, of tillage erosion over water erosion in total soil redistribution (Quine et al. 1993, 1994, 1996, 1997; Govers et al. 1996).
The development of understanding of tillage erosion has been reviewed elsewhere (Govers et al. 1999) and elaboration lies beyond the scope of this paper. However, it is valuable to outline the meaning of tillage erosion and of tillage translocation. Whenever the soil is tilled, the cultivation layer is mobilised and there is a net movement of soil mass, usually principally in the direction of tillage with a second component perpendicular to the direction of tillage. This movement is termed tillage translocation. This can be characterised either by a tillage translocation distance, which is the mean distance over which the tillage layer is moved in the direction of tillage, or by a soil flux, which is the mass of soil translocated through an area of depth equal to the implement depth and of unit width (Fig. 1). If tillage translocation is spatially uniform, then pairs of opposing tillage operations will cancel each other and the long-term average effect will be zero net flux of soil mass, merely lateral mixing. However, where there is spatial variation in the magnitude of tillage translocation, cancelling will not take place and a net soil flux will result. Where the magnitude of the soil flux varies, net soil redistribution, i.e. erosion and aggradation will occur.
[FIGURE 1 OMITTED]
Farmers have long recognised that there is a tendency for long-term movement of the plough layer downslope leading to the formation of banks at cross-slope field boundaries. Archaeologists have also recognised the role of downslope movement of soil due to tillage in the formation of lynchets, although there has been a tendency to use terminology, such as plough-wash, that reflects uncertainty regarding process. Nevertheless, the existence of this knowledge and the significance of tillage in soil redistribution have not been widely recognised by soil scientists and geomorphologists until relatively recent experimental work addressing the need for understanding the controls on tillage translocation and the magnitude of net translocation distances and fluxes (Lindstrom et al. 1990, 1992; Revel et al. 1993; Govers et al. 1994; Lobb et al. 1995; Poesen et al. 1997; Quine et al. 1999; Van Muysen et al. 1999). In addition to confirming the significance of tillage direction, experimental work has demonstrated the importance of slope gradient in determining translocation distances. This is important when considering net soil redistribution on agricultural land, as can be seen by examination of the diffusive model for soil redistribution by tillage proposed by Govers et al. (1993, 1994):
(1) [T.sub.D] = a. - [k.sub.t].S
(2) Qs = P.[rho].[T.sub.D]
where [T.sub.D] is the tillage translocation distance in the direction of tillage (distance per unit time, for a single operation: m/pass); S is slope tangent (m/m), positive for upslope tillage and negative for downslope tillage; [k.sub.t] is the tillage translocation coefficient (m/pass), the gradient of the regression relationship between translocation distance and slope tangent; a is the tillage translocation constant (m/pass), the intercept of the regression relationship between translocation distance and slope tangent; Qs is the soil flux due to tillage (kg/m.pass); P is the plough depth (m); and [rho] is the soil bulk density (kg/[m.sup.3]).
Therefore, the mean soil flux may be established from Eqns 1 and 2, assuming successive operations are in 2 opposing operations, as follows:
(3) Qsm = 0.5. * P.[rho] [[T.sub.Dup] - [T.sub.Ddown]] = -P.[rho].[k.sub.t].S = - [k.sub.f].S
(4) [k.sub.f] = P.[rho].[k.sub.t]
where Qsm is the mean soil flux due to tillage (kg/m.pass), assuming equal numbers of passes up and down slope--note that the negative sign indicates that there is a net downslope movement of soil; [T.sub.Ddown] is the translocation distance in the direction of tillage for a downslope tillage operation (m/pass); [T.sub.Dup] is the translocation distance in the direction of tillage for an upslope tillage operation (m/pass); and [k.sub.f] is the soil flux coefficient (kg/m.pass).
On the basis of Eqns 3 and 4, net soil redistribution by tillage, controlled by spatial variation in soil flux, can be characterised as follows:
(5) Rt = dQsm/dx = -[k.sub.f] dS/dx = -[k.sub.f] [d.sup.2]z/[dx.sup.2]
where Rt is the soil redistribution rate due to tillage (kg/[m.sup.2].pass)--negative for erosion, positive for deposition; x is distance from the divide (m); and z is elevation (m).
Equation 5 implies that soil redistribution by tillage is controlled by soil curvature, with erosion occurring on convex slope elements and aggradations in concave elements.
Experimental work has also demonstrated that values of the tillage flux coefficient ([k.sub.f]) for a single mouldboard tillage operation are of the order of 200-300 kg/m, which is an order of magnitude larger than for other diffusive processes such as soil creep and splash erosion (Govers et al. 1996). More recent work has suggested that the diffusive model may be a slight over-simplification because of the absence of a significant relationship between slope and translocation for upslope tillage (Van Muysen et al. 1999), and because of non-topographic controls (including sub-plough layer resistance to erosion and plough layer cohesion) on both tillage displacement and tillage detachment (Quine et al. 1999). Nevertheless, the diffusive model has proved very valuable in identifying the processes contributing to long-term soil redistribution evidenced by caesium-137 (Quine et al. 1993, 1994, 1996, 1997; Govers et al. 1996). These studies indicate that in landscapes of rolling terrain soil redistribution by tillage often dominates the pattern of landscape evolution and is usually the most important process when considering on-site impacts of erosion (Kosmas et al. 2001; Quine and Zhang 2002).
In New Zealand, sustainable use of natural resources is a legal requirement of the Resource Management Act (1991: http://www.mfe.govt.nz/laws). If this goal is to be achieved for agriculture, documentation of rates of erosion and identification of the processes responsible is a prerequisite to establishing appropriate soil management techniques. This has been recognised for many years and national, regional, and local surveys have yielded qualitative data concerning the areal extent of visible rill, interrill, and wind erosion (e.g. Raeside and Baumgart 1947; Eyles 1983). High magnitude, low frequency wind and water erosion events have also been studied (Painter 1978; Hunter and Lynn 1988, 1990; Basher 1990). Nevertheless, Basher et al. (1995) identified the problems associated with using such data in the assessment of sustainability and the need to document long-term rates of surface erosion that incorporated all erosion processes. In their study, Basher et al. (1995) evaluated the potential of the caesium-137 ([sup.137]Cs) technique (e.g. Ritchie and McHenry 1975; Loughran 1989; Walling and Quine 1990, 1991) for erosion assessment in the South Canterbury Downlands. Their study both demonstrated the value of [sup.137]Cs in this environment and also provided new insights into long-term erosion in the area. In particular, the pattern of soil redistribution did not coincide with observed patterns of water erosion, instead it was characterised by maximum soil loss from shoulders and aggradation in footslopes, toeslopes, and swales. This is a pattern that is more consistent with tillage erosion. In the light of these findings, the growing literature concerning tillage erosion, and the particular context of New Zealand legislation regarding sustainability, this study sought to investigate the potential contribution of tillage erosion to long-term soil redistribution at the same farm investigated in the Basher et al. (1995) study. The specific aims were:
* to assess experimentally tillage erosion intensity associated with the most important tillage implement, mouldboard plough, by deriving tillage translocation and soil flux coefficients, and comparing them with published values;
* to evaluate the significance of tillage erosion by deriving tillage erosion rates for the slope units encountered and estimating the contribution of tillage erosion to the long-term soil redistribution rates obtained by Basher et al. (1995), and by examining field evidence for long-term patterns of soil redistribution.
Methods and study site
The study area is typical of the loess-mantled South Canterbury Downlands in which broad, flat to gently sloping interfluves are separated from gently sloping, normally dry, swales by convexo-concave slopes that are usually <25 m in length and have a mean slope of 5-10[degrees]. On the basis of New Zealand Meteorological Service (1985) data and farmer records, Basher et al. (1995) estimated the mean annual rainfall to be 625 mm. Although rainfall intensities as high as 80 mm/h have been recorded (Coulter and Hessell 1980), overland flow is usually generated by saturation following long duration (1-3 days) low intensity rainfall (<20 mm/h). The area is predisposed to wind erosion by seasonal and longer duration droughts and by strong north-west winds (on average 31 days per year will be characterised by gusts in excess of 60 km/h), especially in spring and summer (New Zealand Meteorological Service 1973).
The soils are formed in Quaternary loess (Tonkin et al. 1974) and are mapped (Kear et al. 1967) as Timaru series (Aquic or Typic Fragiochrepts). They have silt loam texture throughout and typically display [A.sub.p]/AB/[B.sub.w(g)]/B[C.sub.xg] horizonation. The presence in the subsoil of fragipans with high bulk density and low permeability contributes to a high water erosion hazard, as does poor structural stability under cultivation (Packard and Raeside 1952; Watt 1972).
The study site is on the same farm and within 1 km of the 'cropping site' studied by Basher et al. (1995) and is located approximately 10 km north-west of Pleasant Point (site location: NZMS 260 J38/543642). It is cultivated frequently and cropped for grass seed, wheat, barley, and peas. The usual plough depth is approximately 0.17 m.
The topography of the experimental site (Fig. 2a), recorded using a dumpy level, shows that the site is typical of the area with a slope length of approximately 25 m and a maximum slope angle of 15[degrees]. Tillage translocation was investigated using a similar methodology to Lindstrom et al. (1990, 1992), Govers et al. (1994) and others. Seven locations along the slope profile, representative of the range of angles encountered, were selected for measurement of tillage translocation. At each location, 3 red metal pins were inserted into the ground at 5 m intervals along the contour and the location of each pin was recorded. Between each pair of pins, 50 numbered aluminium cubes (edge 0.015 m) were inserted into the ground at 0.05 m intervals using a steel rod. The distance between the middle pin and the nearest cube on each side was 0.4 m. Successive cubes were inserted to depths 0, 0.075, 0.15, 0, 0.075, 0.15, 0, ... m. It was, therefore, possible to derive the initial position of each cube. Each strip of cubes was allocated a letter. Mid-way along the slope profile between each of pairs of successive outer pins, an additional white pin was inserted into the ground and the location recorded. The arrangement of pins and cubes is illustrated in Fig. 2b.
[FIGURE 2 OMITTED]
Immediately before tillage, the mid-point red pins were removed. The farmer then undertook a single pass of the mouldboard plough used for cultivation, downslope through H to N and upslope through G to A. Tractor speed was recorded along each pass. In each case the plough was lowered into the ground several metres before the first strip to ensure that working depth and speed had been achieved before the tracers were met. However, at the base of the slope the concavity was so sharp that, due to the rigidity of the frame, the plough lifted out of the soil. For this reason only one cube at site N and no cubes at site G were displaced.
The aluminium tracers were then recovered by hand excavation. The position of each recovered tracer was then recorded by measuring the distance to the tracer from the adjacent red and white pins (Fig. 2c). This information was then used to calculate the final location for each cube on the same co-ordinate system as the initial locations. Although a record was taken of the approximate depth of each recovered cube, this information was not used in the calculation of the final locations, because of the difficulty associated with depth estimation during excavation. However, because burial depths were small compared with horizontal travel distances, the maximum error that this introduces on a single cube measurement is smaller than the linear dimension of the cube. Recovery rates exceeded 95% for all strips.
For each cube, the translocation distance in the direction of tillage, perpendicular to tillage, and total were derived from the initial and final co-ordinates:
(6) [T.sub.Dj] = [x.sub.jf] - [x.sub.ji]
(7) [T.sub.Pj] = [y.sub.jf] - [y.sub.ji]
(8) [T.sub.Ti] = [square root of (([T.sub.Dj.sup.2] + [T.sub.Pj.sup.2]))]
where [T.sub.Dj] is the translocation distance of cube j in the direction of tillage (m/pass); [T.sub.Pj] is the translocation distance of cube j perpendicular to the direction of tillage (m/pass); [T.sub.Tj] is the total translocation distance of cube j (m/pass); [x.sub.ji] is the initial x co-ordinate of cube j; [x.sub.jf] is the final x co-ordinate of cube j; [y.sub.ji] is the initial y co-ordinate of cube j; and [y.sub.jf] is the final y co-ordinate of cube j.
Mean tillage translocation distances, in the direction of tillage ([T.sub.D]), perpendicular to tillage ([T.sub.P]) and total ([T.sub.T]), for each strip were then calculated. For example, the tillage translocation distance in the direction of tillage for each strip was derived as follows:
(9) [T.sub.D] = ([sup.j = 1,n][SIGMA] [T.sub.Dj])/n
where n is the number of recovered cubes.
Results and analysis
The initial and final positions of the tracers are shown in Fig. 3 with the slope gradient over the 3 m slope section immediately above and below each strip of tracers. Comparison within each pairs of strips provides a clear visual impression of the greater translocation distances travelled by the cubes as a result of tillage in the downslope direction compared with tillage in the upslope direction. Greater translocation distances on the steeper slope section (Fig. 3b) than the shallower slope section (Fig. 3a) are also clearly visible along the downslope pass. The mean translocation distances for each strip are recorded in Table 1 with the relevant topographic data.
[FIGURE 3 OMITTED]
Quantitative investigation of relationships between translocation distance and slope initially followed the pattern used in previous studies. In Fig. 4, translocation distances (in the direction of tillage, [T.sub.D]; perpendicular to the direction of tillage, [T.sub.P]; and total, [T.sub.T]) are plotted against slope tangent. This shows a pattern consistent with previous work in which there is no systematic relationship between slope tangent and translocation distance perpendicular to the direction of tillage (when the tillage direction is perpendicular to the contour), whereas there is a strong relationship ([r.sup.2] = 0.80) between translocation distance in the direction of tillage and slope tangent. This latter relationship also leads to a strong relationship ([r.sup.2] = 0.75) between tangent and total translocation distance.
[FIGURE 4 OMITTED]
Interest in relationships between translocation distance and slope tangent are largely driven by the desire to quantify and compare tillage erosion intensity through tillage translocation and soil flux coefficients (Eqns 1, 3, and 4). In the earliest studies (Lindstrom et al. 1990, 1992; Govers et al. 1994), these coefficients were derived using a single relationship between translocation distance and slope tangent and the values derived using this method are listed in Table 2. However, Van Muysen et al. (1999) re-analysed the data and suggested that the relationship between translocation and tangent was better described by treating upslope and downslope tillage separately. This analysis is illustrated in Fig. 5a and is consistent with that of Van Muysen et al. (1999). The relationship between downslope translocation and tangent is statistically significant (P = 0.04) but the slope of the relationship between upslope translocation and tangent is not statistically significant (P = 0.34). Under these circumstances Van Muysen et al. (1999) used the following relationships:
(10) [T.sub.Dm] = 0.5 * [[T.sub.Ddown] - [T.sub.Dup]]
(11) = [a.sub.m] * - [k.sub.tm] * S
(12) [k.sub.tm] = 0.5 [k.sub.tD]
(13) [a.sub.m] = 0.5*([a.sub.D] * - [a.sub.U])
where [T.sub.Dm] is the mean translocation distance downslope for a pair of tillage operations, one uplsope and one downslope (m/pass); [k.sub.tm] is the mean tillage translocation coefficient (m/pass); [k.sub.tD] is the tillage translocation coefficient for downslope tillage(m/pass); [a.sub.m] is the mean tillage translocation constant (m/pass); [a.sub.D]) the intercept of the regression relationship between downslope translocation distance and slope tangent (m/pass); and [a.sub.U] the intercept of the regression relationship between upslope translocation distance and slope tangent (m/pass).
Van Muysen et al. (1999) assumed that the intercepts are equal and that [a.sub.m] is zero and that, therefore, the value of [k.sub.tm] can then be used in Eqn 4 to derive the tillage flux coefficient. In this study, the values of the intercepts are not equal; however, the difference is only 0.08 m, resulting in a value of [a.sub.m] of only 0.04 m. It is considered that this is of insufficient magnitude to justify further elaboration of the relationships; therefore, Eqn 12 has been used to derive a mean tillage translocation coefficient (Table 2) and Eqn 4 the soil flux coefficient. When Van Muysen et al. (1999) undertook their analysis on the data of Lindstrom et al. (1992) and Govers et al. (1994), they found differences of +2% and -17%, respectively, between [k.sub.f] values calculated using two relationships compared with the values obtained based on one. In this study the difference was 37.5% (Table 2). This difference emphasises the need to identify the most appropriate method for calculation of translocation and flux coefficients. However, some of the deviation may be due to the deviations from linearity seen in the scatter of the measured translocation distances (Fig. 4). In Fig. 4, it is clear that relatively high translocation distances are seen in all directions on the downslope pass at both the first strip encountered and the last traversed before the plough lifted out of the ground. It is possible that larger than typical distances occurred in the former because of higher speed carried into the strip (the speed data are insufficiently detailed to test this) and in the latter because of low resistance to transport downslope as the plough emerged from the ground.
In the development of topographic relationships it is important to attempt to eliminate the influence of non-topographic controls on translocation distances. Therefore, further data analysis was undertaken to attempt to compensate for the 'abnormal' behaviour noted in the first and last strips encountered by the plough. Because 'abnormal' behaviour is seen in translocation both in, and perpendicular to, the tillage direction, the feasibility of correcting for this was investigated by using the ratio of [T.sub.D]:[T.sub.P] Figure 5b shows the results of this analysis. The scatter has been considerably reduced and for downslope tillage, the [T.sub.D]:[T.sub.P] ratio is very strongly related to slope tangent ([R.sup.2] = 0.93; P = 0.002). For upslope tillage, the gradient of the regression line between [T.sub.D]:[T.sub.P] ratio and slope tangent is not statistically significantly different from 0. The following relationships are, therefore, derived:
(14) [T.sub.Ddown]/[T.sub.Pdown] = b - [k.sub.r] * S
(15) [T.sub.Dup]/[T.sub.Pup] = c
where [T.sub.Pdown] is the translocation distance perpendicular to the direction of tillage for a downslope tillage operation (m/pass); [T.sub.Pup] is the translocation distance perpendicular to the direction of tillage for an upslope tillage operation (m/pass); b is the intercept of the regression relationship between [T.sub.D]:[T.sub.P] ratio for downslope tillage and slope tangent (dimensionless); [k.sub.r] is the tillage translocation ratio coefficient--the gradient of the regression relationship between [T.sub.D]:[T.sub.P] ratio for downslope tillage and slope tangent (dimensionless); and c is the intercept of the regression relationship between [T.sub.D]:[T.sub.P] ratio for upslope tillage and slope tangent (dimensionless).
The reduction in scatter and increase in the degree of explanation obtained for downslope tillage suggests that use of the ratio has removed some of the non-topographic controls on translocation distance, possibly associated with operator control and the discontinuous tillage associated with the conduct of the experiment. If it is assumed that, under normal continuous tillage (as opposed to the short discontinuous passes in the experiment) at this site with this implement, there would be relatively little variation in [T.sub.P] then these relationships can be reworked and combined with Eqns 11 and 12 to derive an improved estimate of the tillage translocation coefficient. The assumption of minimal variation in [T.sub.P] is supported by the absence of systematic differences between up and downslope tillage in the observed data (cf. Table 1; mean values for [T.sub.Pdown], [T.sub.Pup], and [T.sub.Pall] equal 0.40 m):
(16) [T.sub.Dm] = 0.5 * [T.sub.Pall] [b - [k.sub.r] * S - c]
(17) [k.sub.tm] = 0.5 * [T.sub.Pall] * [k.sub.r]
(18) [a.sub.m] = 0.5*[T.sub.Pall](b - c)
As was the case with the Van Muysen method (Van Muysen et al. 1999), the value of [a.sub.m] is not zero, but it is close enough (0.04 m) to be approximately considered as zero, given the precision of the experimental technique. The new values of [k.sub.tm] and [k.sub.f] (Table 2) lie between those derived using the single regression relationship between [T.sub.D] and tangent and those derived using the Van Muysen method. Although it is considered that the ratio-based values are the most reliable here, this method will only be applicable where there is evidence to suggest that [T.sub.P] has limited variance and does not vary systematically. Because lateral movement of soil during mouldboard ploughing is often constrained by the configuration of the last furrows, limited variance in [T.sub.P] may be quite common and the ratio method may be widely applicable to the removal of non-topographic influences. Nevertheless, this cannot be assumed without evidence, not least because in the Van Muysen et al. (1999) study [T.sub.P] was found to vary systematically with slope.
Comparison of coefficients with other studies
The tillage coefficients derived in this study are summarised in Table 3 with those derived in other published studies that have used a similar methodology in analysis. Compared with most previous studies, tillage detachment rates are low because of the relatively shallow plough depth; however, the value of [k.sub.f], the soil flux coefficient, of 265 kg/m * pass is similar to values observed on other regularly cultivated sites (150-335 kg/m * pass). This is due to the unusually high value of the tillage translocation coefficient, [k.sub.tm] (Table 3:1.16 compared with a median value from previous mouldboard plough studies of 0.69). Quine et al. (1999) suggested that the magnitude of [k.sub.tm] is determined by soil resistance to translocation, tillage implement type and speed, and that the influence of detachment magnitude on displacement was unknown. In this comparison with other mouldboard plough studies, a single implement type is under consideration (although the individual ploughs may have differed in their specific configurations, e.g. angle of the mouldboard, etc.) and, in the absence of detailed information concerning implement configuration, discussion of differences in [k.sub.tm] focuses on speed, detachment magnitude, and soil resistance to translocation.
Progress has been made in understanding the relationship between detachment magnitude (largely controlled by depth) and translocation distance, but data published to date concerning tillage speed are equivocal. Studies by both C. Kosmas and colleagues in Greece and by G. Govers and colleagues in the Belgian loess belt show an increase in [k.sub.tm] values when plough depth is increased, if other factors are held constant (TERON 1999). The following relationship applies to the data collected by Kosmas and colleagues:
(19) %[k.sub.t] = 1.26% P ([r.sup.2] = 0.94)
where %[k.sub.t] is percentage change in the value of [k.sub.tm], and %P is percentage change in plough depth.
Therefore, it seems likely that the high [k.sub.tm] value observed here occurs despite, not because of, low plough depth. In relation to speed of mouldboard ploughing operations, Govers and colleagues (TERON 1999; Van Muysen 2001) conducted a 3-way comparison of mouldboard tillage erosion with 2 depths and 3 speeds. This identified a 45% increase in [k.sub.tm] for a 16% increase in speed, however, a 38% increase in [k.sub.tm] was observed for a 7% decrease in speed when accompanied by a 19% increase in depth. Disentangling the two effects is problematic because of the small number of cases. In this context, it is difficult to interpret the data obtained in this study, nevertheless, it is noted that the tillage speeds used in this study are among the highest in Table 3 and it seems likely that speed may have contributed to the high [k.sub.tm] value observed.
Soil condition or resistance to translocation has been used as an explanation for significant differences in [k.sub.tm] values between treatments by both Van Muysen et al. (1999) and Quine et al. (1999). The [k.sub.tm] and [k.sub.f] values obtained in the latter study are larger than all the others presented in Table 3, despite the relatively small soil-contact of the tool examined--a duckfoot chisel plough. The magnitude of the coefficients was attributed to the very loose, non-cohesive nature of the stony soil examined. However, the study by Van Muysen et al. (1999) is of greater relevance here because it focussed on mouldboard plough operation. Two adjacent plots were examined that differed only in their recent use; one had been tilled in the same year and the other had been fallow for several years. Experimental control was exercised over several parameters by using the same mouldboard plough and operator, but, due to the compaction of the fallow field, it was not possible to till it to the same depth as the previously tilled field. Van Muysen et al. (1999) present an elegant analysis of the data in which they use the [T.sub.P] ratio and the [T.sub.D] ratio of the treatments. The [T.sub.P] ratio is assumed to indicate the influence of tillage depth on translocation distances and the difference between the [T.sub.P] and [T.sub.D] ratios to indicate the role of soil condition. This analysis suggests that increase in soil depth and differences in soil condition make approximately equal contributions to the 100% increase in [k.sub.tm]. Despite the intuitive strength of this analysis, it is possible that the influence of soil condition may have been overestimated. If the 120% increase in plough depth observed by Van Muysen et al. (1999) is used in Eqn 19, an increase in [k.sub.tm] of 170% is predicted, which is significantly more than the measured 100% increase. It is clearly hazardous to pursue this argument at length because it involves combination of disparate data sources, nevertheless, the data of Kosmas and colleagues certainly indicate that significant increases in mouldboard plough depth can result in large increases in [k.sub.tm], irrespective of soil conditions. It is possible that the ratio variation observed by Van Muysen et al. (1999) was caused by the different trajectories, associated with the different tillage depths that would result from the helical shape of the mouldboard.
Although a growing body of experimental data exists, it remains difficult to identify specific controls on the magnitude of the tillage translocation and soil flux coefficients; nevertheless, two observations may be made. First, the silt loam soils in the study area offer relatively little resistance to translocation and may, therefore, be considered to exhibit high sensitivity to tillage translocation. Second, the high speed employed in tillage is likely to be a significant contributor to the high [k.sub.tm] values. In order to attempt to estimate the importance of speed in this context, multiple regression analysis was undertaken using all the mouldboard plough data in Table 3. The following relationship was obtained:
(20) [k.sub.tm] = eV + gP + h ([R.sup.2] = 0.33, adjusted [R.sup.2] = 0.17)
where V is tractor speed (km/h), e and g are coefficients (0.10 and 0.03, respectively), and h is a constant (0.40).
Figure 6 shows the comparison between measured [k.sub.tm] values and those obtained using this relationship. Scatter in the data is evident in both Fig. 6 and the [R.sup.2] values. Nevertheless, there is greater confidence (P = 0.09) in the coefficient for velocity (e) than in the relationship as a whole. This was supported by low variation in the coefficient for velocity (0.12, 0.12, 0.13, respectively) obtained in multiple regression of the data after extraction of 1,2, and 4 apparent outliers (resulting in an increase in adjusted [R.sup.2] to 0.30, 0.46 and 0.47, respectively). If no reduction in plough depth is considered, then percentage change in [k.sub.tm] as a result of change in tillage speed is defined as follows:
(21) %[k.sub.t] = 100*(e*dV/[k.sub.tm])
[FIGURE 6 OMITTED]
This suggests that for this study site, a 10% reduction in [k.sub.tm] and, therefore, tillage erosion intensity could be obtained for each 1 km/h reduction in tillage speed ([k.sub.tm] = 1.16; e = 0.10 - 0.13), within the bounds of the source data.
Tillage erosion rates
Although the tillage translocation coefficient ([k.sub.tm]) and soil flux coefficient ([k.sub.f] provide an indication of the tillage erosion potential of the tillage practice, it is necessary to combine this with landscape sensitivity to tillage erosion to obtain tillage erosion rates. This is undertaken by using the derived value for [k.sub.f] (265 kg/m * pass) in Eqn 5 (for a 1-dimensional profile) and applying it to the field topography. The pattern of average soil redistribution rates (per pass for paired opposing passes) due to tillage erosion are shown in Fig. 7. This shows rates of soil loss as high as 51 t/ha * pass and rates of soil accumulation as high as 117 t/ha * pass. Figure 7 also shows the slope form, the pattern of slope angles, and qualitative slope descriptors using the same terminology as Basher et al. (1995). The pattern of tillage erosion may be summarised as follows: maximum soil loss from shoulders; significant soil erosion from the upper backslopes and significant soil accumulation on the lower backslopes; high soil accumulation in the foot, toe, and swale locations (which are of limited extent at this site). This is a very similar pattern to that documented by Basher et al. (1995) for a nearby field on the same farm using [sup.137]Cs. The latter authors noted the following general pattern: little or no net soil redistribution on the interfluves, highest mean erosion rates (13 t/ha * year) on the shoulders; both high erosion (up to 27 t/ha * year) and high deposition (up to 28 t/ha * year) on the backslopes; and deposition on toeslopes, footslopes, and in the swales. The higher rates of tillage erosion on the shoulders and higher rates of tillage deposition in the concave slope elements may reflect: increased intensity of tillage erosion (the [sup.137]Cs data provide a 40-year average); the particular configuration of the landscape (i.e. higher profile curvature on this slope segment than on the field examined by Basher); and, in the case of the concave slope elements, the removal of some of the tillage deposited soil by water erosion during high magnitude events. However, the overall agreement in pattern suggests that a large proportion of the soil redistribution documented by Basher et al. (1995) using [sup.137]Cs was due to tillage erosion. This is consistent with other studies of mechanised agricultural land that have attempted to separate the contributions of tillage and water erosion to total soil redistribution (Quine et al. 1994, 1996, 1997; Govers et al. 1996).
[FIGURE 7 OMITTED]
Further evidence for the importance of tillage erosion is seen in landscape features that are consistent with the patterns of soil redistribution described above. These include significant steps along cross-slope field boundaries (Fig. 8a) and exposure of subsoil on shoulder slope elements (Fig. 8b). The latter are particularly significant in relation to the overall 'health' of the soil because they indicate that tillage erosion and water erosion may be expected to operate synergistically. It has already been noted that the subsoil tends to have a high bulk density and poor structural stability. Continuing tillage erosion will lead to increasing concentration of subsoil material in the plough soil in shoulder slope segments (51 t/ha * pass is equivalent to replacement of 2% of the plough layer by subsoil per mouldboard plough pass). This can be expected to lead to exposure of subsoil of low permeability and an increased tendency for formation of surface seals on the shoulder slope segments. Both of these contribute to the generation of runoff onto the steeper slopes and swales. Promoted water loss has implications beyond water erosion because of the seasonal scarcity of water and the reduced moisture storage potential of the soil in the shoulder slope elements due to decline in organic matter and porosity associated with elevated subsoil content. It is, therefore, anticipated that continuing tillage erosion will have deleterious effects on soil quality and crop production over substantial areas (erosion rates approach 20 t/ha * pass over 50% of the slope section studied) and will contribute to the promotion of water erosion.
[FIGURE 8 OMITTED]
This study was justified by the need to understand and document the most important soil redistribution processes operating on agricultural land in order to fulfil the need for a sustainable land management strategy. The experimental data collected suggest that tillage erosion is a very significant soil redistribution process in the rolling landscapes of the Canterbury Downlands and that rates of soil erosion due to tillage may exceed 20 t/ha * pass over a large proportion of the landscape. This is not sustainable in an absolute sense if soil formation rates are used as a benchmark. Comparison of the experimental data derived in this study with earlier studies elsewhere indicates that high tillage speed may make a significant contribution to elevated tillage erosion rates. It is tentatively suggested that reduction of tillage speed from 7 km/h to 4 km/h could produce a 30% decrease in tillage erosion rates. Further experimentation is required to test this hypothesis.
Table 1. Treatment characteristics and translocation results s.d., standard deviations Parallel to Slope tillage, Direction Treatment tangent Speed [T.sub.D] Out In Mean (km/h) Mean s.d. Upslope A 0.02 0.02 0.02 6.50 0.31 0.23 B 0.03 0.07 0.05 0.16 0.15 C 0.09 0.12 0.10 0.18 0.16 D 0.19 0.21 0.20 0.22 0.19 E 0.22 0.26 0.24 0.18 0.17 F 0.22 0.17 0.20 0.09 0.13 G 0.10 0.00 0.05 Downslope H -0.02 -0.02 -0.02 7.40 0.46 0.46 I -0.05 -0.04 -0.05 0.33 0.27 J -0.13 -0.09 -0.11 0.46 0.30 K -0.20 -0.18 -0.19 0.66 0.67 L -0.23 -0.24 -0.24 0.73 0.47 M -0.16 -0.22 -0.20 0.91 0.53 N 0.00 -0.12 -0.05 Translocation distances (m) Direction Treatment Perpendicular to tillage, [T.sub.p] [T.sub.D]: [T.sub.p] Mean s.d. Mean s.d. ratio Upslope A 0.49 0.25 0.61 0.28 0.63 B 0.43 0.25 0.47 0.27 0.37 C 0.39 0.24 0.45 0.26 0.46 D 0.47 0.27 0.53 0.31 0.46 E 0.38 0.24 0.44 0.25 0.49 F 0.25 0.30 0.28 0.31 0.37 G Downslope H 0.49 0.43 0.75 0.53 0.95 I 0.35 0.30 0.53 0.33 0.93 J 0.37 0.37 0.67 0.35 1.23 K 0.41 0.33 0.83 0.68 1.59 L 0.35 0.30 0.86 0.50 2.12 M 0.44 0.34 1.03 0.60 2.08 N Table 2. Estimates of the tillage translocation and soil flux coefficients derived in this study Plough depth = 0.17 m; bulk density = 1350 kg/[m.sup.3] Mean tillage Tillage translocation translocation Soil flux constant, coefficient, coefficient, [a.sub.m] [k.sub.tm] [k.sub.f] (m/pass) (m/pass) (kg/m.pass) Single regression 0 (0.39-0.39) 1.41 324 relationship for [T.sub.D] v. tangent 'Van Muysen method'-- 0.04 (0.5* 1.03 235 separate regression [0.32-0.24]) (0.5*2.05) relationships [T.sub.D] v. tangent for up and downslope tillage Ratio method--regression 0.04 (0.5*0.4* 1.16 265 of TD: TP v. slope; [0.71-0.51]) (0.5*0.4*5.78) separate relationships for up and downslope tillage Table 3. Experimental data from tillage experiments Where 2 values are quoted for the coefficients, that in parentheses is based on a single regression relationship while that outside the parentheses is based on separate treatment of downslope and upslope tillage; bulk density refers to soil condition before tillage Implement/author Tillage Tillage depth speed (m) (km/h) Mouldboard plough--up and down This study 6.5 (A), 7.4 (B) 0.17 Lindstrom et al. (1992) 7.6 0.24 Govers et al. (1994) 4.5 0.28 Van Muysen et al. (1999) Cultivated land 1.8 0.33 Fallow field 2.7 0.15 TERON (1999); Van Muysen (2001) Depth and speed analysis 5.0 0.25 5.4 0.21 6.3 0.20 Lobb et al. (1995) 4.0 0.15 Lobb et al. (1999) 6.2 0.23 Revel et al. (1993) 6.5 0.27 Others Duckfoot plough, Quine et al. (1999) 0.19 Chisel plough, Govers et al. (1994) 0.15 Tillage layer Tillage bulk density detachment (kg/[m.sub.3]) (kg/[m.sub.2]) Mouldboard plough--up and down This study 1350 230 Lindstrom et al. (1992) 1350 324 Govers et al. (1994) 1350 378 Van Muysen et al. (1999) Cultivated land 1070 353 Fallow field 1650 247 TERON (1999); Van Muysen (2001) Depth and speed analysis 1500 375 1570 330 1530 306 Lobb et al. (1995) 1350 202.5 Lobb et al. (1999) 1350 310.5 Revel et al. (1993) 1350 364.5 Others Duckfoot plough, Quine et al. (1999) 1382 263 Chisel plough, Govers et al. (1994) 1350 203 Tillage translocation Soil flux coefficient, coefficient, [k.sub.tm] [k.sub.f] (m/pass) (kg/m.pass) Mouldboard plough--up and down This study 1.16 (1.41) 265 (324) Lindstrom et al. (1992) 1.035 (1.02) 335 (330) Govers et al. (1994) 0.51 (0.62) 193 (234) Van Muysen et al. (1999) Cultivated land 0.795 245 Fallow field 0.345 85 TERON (1999); Van Muysen (2001) Depth and speed analysis 0.63 236 0.46 150 0.66 202 Lobb et al. (1995) 0.91 184 Lobb et al. (1999) 1.11 346 Revel et al. (1993) 0.72 263 Others Duckfoot plough, Quine et al. (1999) 2.30 605 Chisel plough, Govers et al. (1994) 0.55 111 (A) Uphill; (B) Downhill.
We thank Colin Lyons for access to his property and for assistance with the tillage experiment and we thank Exeter University geography undergraduates on the 1999 field class for assisting with tracer excavation.
Basher LR (1990) Wind erosion and soil re-formation on a terrace in the upper Hurunui River Gorge. DSIR Division of Land and Soil Sciences Technical Record CH6.
Basher LR, Matthews KM, Zhi, L (1995) Surface erosion assessment in the South Canterbury Downlands, New Zealand using [sup.137]Cs distribution. Australian Journal of Soil Research 33, 787-803.
Basher LR, Painter DJ (1997) Wind erosion in New Zealand. In 'Proceedings of the International Symposium on Wind Erosion'. (USDA-ARS: Manhattan, Kansas) http://www.wern.ksu. edu/symposium
Boardman J, Bell M (1992) Past and present soil erosion: linking archaeology and geomorphology. In 'Past and present soil erosion'. (Eds M Bell, J Boardman) pp. 1-8. (Oxbow Books: Oxford)
Coulter JD, Hessell JWD (1980) The frequency of high intensity rainfalls in New Zealand. Part II Point estimates. New Zealand Meteorological Service Miscellaneous Publication 162.
Eyles GO (1983) The distribution and severity of present soil erosion in New Zealand. New Zealand Geographer 39, 12-18.
Govers G, Lobb DA, Quine TA (1999) Tillage erosion and translocation: emergence of a new paradigm in soil erosion research. Soil and Tillage Research 51, 167-174.
Govers G, Quine TA, Desmet PJJ, Walling DE (1996) The relative contribution of soil tillage and overland flow erosion to soil redistribution on agricultural land. Earth Surface Processes Landforms 21, 929-946.
Govers G, Quine TA, Walling DE (1993) The effect of water erosion and tillage movement on hillslope profile development: a comparison of field observations and model results. In 'Farmland erosion in temperate plains environment and hills'. (Ed. S Wicherek) pp. 285-300. (Elsevier: Amsterdam)
Govers G, Vandaele K, Desmet PJJ, Poesen J, Bunte K (1994) The role of tillage in soil redistribution on hillslopes. European Journal of Soil Science 45,469-478.
Greenland DJ, Gregory PJ, Nye PH (Eds) (1997) Land resources: on the edge of the Malthusian precipice? Philosophical Transactions of the Royal Society of London, Series B, 352 (No. 1356) 859-1033.
Hunter GG, Lynn IH (1988) Wind erosion of a soil in North Canterbury. New Zealand Journal of Experimental Agriculture 16, 173-177.
Hunter GG, Lynn IH (1990) Storm induced soil losses from South Canterbury and North Otago Downlands. DSIR Land Resources Technical Record 22.
Kear BS, Gibbs HS, Miller RB (1967) Soil of the downs and plains Canterbury and North Otago, New Zealand. New Zealand Soils Bureau Bulletin 14.
Kosmas C, Gerontidis St, Marathianou M, Detsis B, Zafiriou Th, Van Muysen W, Govers G, Quine T, Vanoost K (2001) The effects of tillage displacedsoil on soil propertie and wheat biomass. Soil and Tillage Research 58, 31-44.
Lindstrom MJ, Nelson WW, Schumacher TE (1992) Quantifying tillage erosion rates due to moldboard plowing. Soil and Tillage Research 24, 243-255.
Lindstrom MJ, Nelson WW, Schumacher TE, Lemme GD (1990) Soil movement by tillage as affected by slope. Soil and Tillage Research 17, 255-264.
Lobb DA, Kachanoski RG, Miller MH (1995) Tillage translocation and tillage erosion on shoulder slope landscape positions measured using [sup.137]Cs as a tracer. Canadian Journal of Soil Science 75, 211-218.
Lobb DA, Kachanoski RG, Miller MH (1999) Modelling translocation and tillage erosion in the complex upland landscapes of south-western Ontario, Canada. Soil and Tillage Research 51, 189-209.
Loughran RJ (1989) The measurement of soil erosion. Progress in Physical Geography 13, 216-233.
New Zealand Meteorological Service (1973). Summaries of climatological observations to 1970. New Zealand Meteorological Service Miscellaneous Publication 143.
New Zealand Meteorological Service (1985). Rainfall normals for New Zealand 1951-80. New Zealand Meteorological Service Miscellaneous Publication 185.
Packard RQ, Raeside JD (1952) Deterioration of soil structure in a South Canterbury soil. New Zealand Journal of Science and Technology 33A, 40-53.
Painter DJ (1978) Soil erosion rates on New Zealand farm land. In 'Proceedings of the Conference on Erosion Assessment and Control in New Zealand'. pp. 25-42. (New Zealand Association of Soil Conservators: Christchurch)
Poesen J, van Wesemael B, Govers G, Martinez-Fernandez J, Desmet PJJ, Vandaele K, Quine TA, Degraer G (1997) Patterns of rock fragment cover generated by tillage erosion. Geomorphology 18, 183-197.
Quine TA, Desmet PJJ, Govers G, Vandaele K, Walling DE (1994) A comparison of the roles of tillage and water erosion in landform development and sediment export on agricultural land near Leuven, Belgium. In 'Variability in stream erosion and sediment transport'. (Eds LJ Olive, RJ Loughran, JA Kesby) pp. 77-86. IAHS Publication No. 224.
Quine TA, Govers G, Poesen J, Walling DE, van Wesemael B, Martinez-Fernandez J (1999) Fine earth translocation by tillage in stony soils in the Guadalentin, south-east Spain: an investigation using Caesium-134. Soil and Tillage Research 51,279-302.
Quine TA, Govers G, Walling DE, Zhang X, Desmet PJJ, Zhang Y (1997) Erosion processes and landform evolution on agricultural land--new perspectives from Caesium-137 data and topographic-based erosion modelling. Earth Surface Processes and Landforms 22, 799-816.
Quine TA, Walling DE, Govers G (1996) Simulation of radiocaesium redistribution on cultivated hillslopes using a mass-balance model: an aid to process interpretation and erosion rate estimation. In 'Advances in hillslope processes'. (Eds MG Anderson, SM Brooks) pp. 561-588. (John Wiley: Chichester, UK)
Quine TA, Walling DE, Zhang X (1993) The role of tillage in soil redistribution within terraced fields on the Loess Plateau, China: an investigation using Caesium-137. In 'Runoff and sediment yield modelling. Proceedings of the International Symposium'. (Eds K Banasik, A Zbikowski) pp. 149-155. (Warsaw Agricultural University: Warsaw)
Quine TA, Zhang Y (2002) An investigation of spatial variation in soil erosion, soil properties and crop production within an agricultural field in Devon, UK. Journal of Soil and Water Conservation 57, 55-65.
Raeside JD, Baumgart IL (1947) Erosion on the Downlands of Geraldine County, South Canterbury. New Zealand Journal of Science and Technology 29A, 49-57.
Revel JC, Guiresse M, Coste N, Cavalie J, Costes JL (1993) Erosion hydrique et entrainement mechanique des terres par les outils dans les coteaux dus sud-ouest de la France. La necessite d'etablir un bilan avant toute mesure anti-erosive. In 'Farmland erosion in temperate plains environment and hills'. (Ed. S Wicherek) pp. 551-562. (Elsevier: Amsterdam)
Ritchie JC, McHenry JR (1975) Fallout Cs-137: a tool in conservation research. Journal of Soil and Water Conservation 30, 283-286.
TERON (1999) Tillage erosion: Current state, future trends and prevention. Unpublished Progress Report for the project FAIR3-CT96-1478.
Tonkin PJ, Runge EC, Ives DW (1974) A study of late Pleistocene loess deposits, South Canterbury. Part II Paleosols and their stratigraphic implications. Quaternary Research 4, 217-231.
Van Muysen W, Govers G, Poesen J, Bergkamp G, Roxo M (1999) Measurement and modelling of the effects of initial soil conditions and slope gradient on soil translocation by tillage. Soil and Tillage Research 51,303-316.
Van Muysen W (2001) Tillage Translocation and tillage erosion: an experimental approach. Unpublished PhD thesis, Katholieke Universiteit, Leuven, Belgium.
Walling DE, Quine TA (1990) Calibration of Caesium-137 measurements to provide quantitative erosion rate data. Land Degradation and Rehabilitation 2, 161-175.
Walling DE, Quine TA (1991) The use of Caesium-137 measurements to investigate soil erosion on arable fields in the UK: potential applications and limitations. Journal of Soil Science 42, 147-165.
Watt JPC (Ed.) (1972) 'Loess soils and problems of land use on the Downlands of the South Island, New Zealand'. Otago Catchment Board Publication No. 4.
Manuscript received 27 May 2002, accepted 19 December 2002
Timothy A. Quine (A,C), Les R. Basher (B), and Andrew P. Nicholas (A)
(A) Department of Geography, University of Exeter, Amory Building, Rennes Drive, Exeter, Devon, EX4 4RJ, Great Britain.
(B) Manaaki Whenua-Landcare Research, PO Box 69, Lincoln, New Zealand.
(C) Corresponding author; email: firstname.lastname@example.org
|Printer friendly Cite/link Email Feedback|
|Author:||Quine, Timothy A.; Basher, Les R.; Nicholas, Andrew P.|
|Publication:||Australian Journal of Soil Research|
|Article Type:||Author Abstract|
|Date:||Jul 1, 2003|
|Previous Article:||Parameterisation of physically based solute transport models in sandy soils.|
|Next Article:||Field study of pesticide leaching in an allophanic soil in New Zealand. 1: experimental results.|