Printer Friendly

An investigation of spatial variation in soil erosion, soil properties, and crop production within an agricultural field in Devon, United Kingdom.


Both the existence and the importance of spatial variation in soil properties and crop production are axiomatic; nevertheless, current understanding of the causes and sources of variation is incomplete. This understanding is essential for development of cost-effective and environmentally-sustainable farming practices. It has been recognized that variation is the product of the interaction of natural properties and processes (intrinsic variability) with human practices (management-induced variability) and developments in precision farming have sought to overcome the former through spatial variation in the latter. This approach assumes that "natural" soil variation is a static variable on the time scales of concern to land managers. This assumption is called into question if a pedo-geomorphological perspective is taken in which the pattern of soil variation is considered to be the product of continuing landform evolution processes, in particular soil erosion and aggradation. The potential importance of this per spective has become more evident with the growing recognition of tillage erosion as an important erosion process per se rather than a contributor to water erosion (cf. review by Govers et al. 1999). The aim of this study was, therefore, to employ a pedo-geomorphological perspective in the investigation of interrelationships and potential causal links between processes of landscape change and the spatial distributions of soil properties and crop production.

This study builds on progress in the understanding of soil and crop variability and of soil erosion processes. Early studies of the relationships between soil properties and crop variability reflect recognition of catenary variation in soils and were often based on qualitative subdivision of the landscape (Stone et al. 1985; Ovalles and Collins 1986). More recently developments in terrain analysis have permitted quantitative description of the landscape and landscape elements through the use of grid-based terrain models and derived attributes (Miller et al. 1988; Moore et al. 1993; Franzen et al. 1997; Yang et al. 1998; Timlin et al. 1998). Although slope descriptions including gradient, aspect, and curvature are the most commonly used attributes, Moore et al. (1993) demonstrated the potential of using process-based compound terrain attributes in the prediction of spatial variation in the properties of surface soil materials. This study employs these developments in quantitative landscape analysis.

Early studies of the effect of erosion on crop production relied on qualitative or indirect measures of erosion, such as profile truncation and soil color (Stone et al. 1985; Daniels et al. 1985). More recently, advances in the measurement and assessment of erosion, in particular the development of the caesium-137 ([Cs.sup.137]) technique (Ritchie and McHenry 1975; Loughran 1989; Walling and Quine 1991), have facilitated the collection of spatially distributed erosion rate data that may be compared to soil property and crop production data (Moulin et al. 1994). Most studies of erosion-productivity relationships have focused either explicitly or implicitly on water erosion, but there is a growing recognition of the potential importance of tillage erosion especially in topographically complex terrain (Miller et al. 1988; Moulin et al. 1994). More recently, attempts have been made to separate the effects of tillage and water erosion on soil variability (Quine, Walling, and Zhang 1999). This latter development h as been made possible through the growing understanding of tillage erosion derived from experimental investigations (Lindstrom et al. 1990, 1992; Govers et al. 1994; Lobb et al. 1995; Poesen et al. 1997; Quine, Covers, et al. 1999; Van Muysen et al. 1999) and the resultant progress made in the simulation of tillage erosion (Govers et al. 1993, 1994, 1996; Quine et al. 1994, 1997; Quine, Govers, et al. 1999; Lobb et al. 1995; Lobb and Kachanoski 1999).

In a recent review, Goderya (1998) suggested that the soil properties influencing crop yield may be divided into three groups: static soil properties, water transport properties, and fertility properties. In this study, spatially-distributed data were collected for soil properties from the first and third of these groups, in addition to data concerning net soil redistribution, water erosion, and crop production. This study aims to advance understanding by employing this wide range of detailed geo-referenced spatially distributed data in the investigation and explanation of spatial variation in soil properties and crop production.

Methods and Materials

Study area. The field under study is located in the southwest of England 13 km (8.1 mi) from Exeter (3[degrees]35'46" W 50[degrees]51'28" N). The soils have developed from non-calcareous Permian Breccia and Conglomerate and, throughout the field, are classified as typical Brown Earths (Clayden 1971; Findlay et al. 1984) or Dystric Cambisols in the United Nations Food and Agriculture Organization system (Dystric Eutrochrept in the United States Department of Agriculture soil classification system). The only variation mapped by Clayden (1971) within the field is a division according to the texture of the parent material into Shaldon series developed from "Clayey" breccia over the upper part of the field (above the 114 m [347 ft] contour - Fignre 1a) and the Crediton series developed from "Loamy" breccia on the lower part of the field. Despite this mapped variation and the complex relief, the soils of the field show very little physical variation. In particular, no systematic variation in texture of the Ap or t he B horizons was observed, and there was no evidence for higher clay contents over the upper part of the slope (the weak, statistically insignificant trend was the reverse).

The climate is maritime, but the study area has a long-term mean annual rainfall of 845 mm (33.3 in) (Applegarth and Cornish 1982) that is amongst the lowest in the region because it lies in the rain-shadow of the Dartmoor upland to the south. The area is well suited to agricultural cropping with its mild temperatures and high sunshine rates, although frost incidence and occasional summer drought affect crop growth on certain topographic positions. Water erosion of soil on the arable land is usually associated with low-intensity, long-duration rainfall during winter months.

Coombe Barton Farm is part of a typical family-run, heavily mechanized farm. Cereal production has been practiced for several decades with a rotation system including barley, wheat, oat, and green cover grass. The fields are delineated by well-maintained living hedges. The annual tillage cycle includes one pass of a moldboard plow and two or more passes of a disk or cultivator harrow Composite chemical fertilizer and organic manure are uniformly applied throughout the field. In the two years preceding the study the sampled field had been planted to winter wheat.

Field survey and sampling strategies. In the autumn of 1997, the field was surveyed and soil samples obtained. Crop samples were collected during the autumn of 1998. The field terrain was surveyed on an approximate 10-m grid using a TOTAL station (Wild, Leica Geosystems, Heerbrugg Switzerland) composed of an electronic theodolite (Wild T1000), infrared distomat, and a data logger (GRE3); the same equipment was used to record all sample locations. Soil core samples were collected using steel core tubes with an internal diameter of 6.9 cm (2.7 in) and length of 65 cm (25.6 in). These tubes were driven into the ground using a motorized percussion corer. In total, 245 soil samples were collected from 19 transects parallel with the tramlines (to facilitate later collection of crop samples from the same locations). Each sample was separated in situ into current plow layer ([A.sub.p] samples) and subplow layer (B samples) based on knowledge of the plow depth and on contrasts in color, structure, texture, and compac tness. The same sampling locations were revisited in the autumn of 1998 to collect crop samples. In practice, relocation of some sample points was problematic because of the standing crop; however, all data used are from crop samples located within 3 m of the associated soil sample. Figure 1 shows the field topography and locations of soil samples. For each crop sample, the wheat within a 1-x-1-[m.sup.2] area was cut with garden shears immediately above the ground and returned to the laboratory for measurement. During November 1998, after seeding of the study field, a prolonged period of rainfall resulted in extensive water erosion. The extent of the rill network and location of the major channels were recorded using the TOTAL station.

Crop sample treatment. For each crop sample, the grain was separated from the stalk, and both were weighed to provide field-dry weights per 1-[m.sup.2] sampled area of grain ([F.sub.g]) and stalk ([F.sub.s]). A subsample of each grain and stalk sample was then oven-dried to establish field moisture content of grain ([W.sub.g]) and stalk ([W.sub.s]). These data were used to derive dry aboveground biomass and dry yield:

AGB = ([F.sub.g](1 - [W.sub.g])) + ([F.sub.s](1 - [W.sub.s])) (1)

Y = ([F.sub.g](1-[W.sub.g])) (2)


AGB is aboveground biomass (kg [m.sup.-2]).

F is field dry weight per unit area (kg [m.sup.-2]).

W is moisture content of field dry samples (kg [kg.sup.-1]).

Y is yield (kg [m.sup.-2]).

Soil sample treatment and analysis. All soil samples were oven-dried for 24 hours at 105[degrees]C and then disaggregated by hand and machine and sieved through a 2-mm mesh. The >2-mm fraction was weighed and stored, and the <2-mm fraction was weighed (a kg for plow - [A.sub.p] - samples; b kg for sub-plow - B - samples) and used in all subsequent analyses.

A subsample of the total core was then prepared for [Cs.sup.137] analysis by careful mixing of (a/(a+b)) kg of the plow layer <2-mm sample with (b/(a+b)) kg of the sub-plow layer <2-mm sample This total core subsample was then packed into a marinelli beaker, and [Cs.sup.137] was measured by gamma spectrometry, using a coaxial germanium detector and multi-channel analyzer system (EG&G ORTEC, PerkinElmer, Inc., Shelton, CT). Counting time of ca. 30,000s gave an analytical precision of 10%. These measurements of [Cs.sup.137] activity were converted into [Cs.sup.137] inventories as follows:

[C.sub.p] = [C.sub.a] (a+b) / [X.sub.c] (3)


[C.sub.p] is the [Cs.sup.137] inventory at the sampling point (Bq [m.sup.-2]).

[C.sub.a] is the measured [Cs.sup.137] activity of the analyzed subsample (Bq [kg.sup.-1]).

[X.sub.c] is the cross-sectional area of the core tube ([m.sup.2]).

Plow and sub-plow samples (<2-mm fraction) were analyzed for total carbon and total nitrogen with an NA2500 C/N Analyzer (CE Instruments, Milan, Italy) and for total phosphate and inorganic phosphate using a SP6 UV/VIS Spectrophotometer (UNICAM, Thermo ONIX Corp., Angleton, TX) following extraction with 1.0 N sodium hydroxide and 1.0 N hydrochloric acid. Samples analyzed for particle size composition were pretreated with hydrogen peroxide and 0.4% sodium hexametaphosphate and then measured using a Malvern Mastersizer (Malvern Instruments Ltd., Worcestershire, UK). The percentages of sand (<2-mm and >0.06mm), silt (.06 mm and >0.002-mm), and clay (<0.002-mm) fractions were calculated according to the British Standard.

Pattern simulation, data integration, and data analysis. All the statistical analyses were carried out using MINITAB software (version 13.1). The TBGIS interface (Zhang 1996) was used to access ARC/INFO sub-routines in the analysis of the topographic data and in the management and integration of soil property and crop production data. A digital terrain model (DTM) was created from the topographic survey data using the spline interpolation method, and the DTM was used to derive the following topographic attributes: slope, profile curvature, planform curvature, and upslope contributing area. TBGIS was also used to simulate tillage erosion using the simple diffusive model that was proposed by Govers et al. (1993, 1994, 1996) and has since been used extensively (Quine et al., 1994, 1996, 1997; Quine, Walling, and Zhang 1999; Quine, Walling, et al. 1999):

Qt = [k.sub.1]S (4)


Qt is the soil flux due to tillage (kg [m.sup.-1]).

[k.sub.1] is the tillage flux coefficient (kg [m.sup.-1]).

S is the local slope gradient (tangent).

Where only the pattern of tillage erosion is of interest, the value of [k.sub.1] may be set to 1 to provide relative values of Qt.

TBGIS was also used to simulate the pattern of potential overland flow erosion based on the stream power index used by Moore et al. (1993):

[OMEGA]=S [A.sub.s] (5)


[OMEGA] is the stream power index.

[A.sub.s] is the specific catchment area ([m.sup.2] [m.sup.-1]).

The [Cs.sup.137] data may be used directly to examine qualitative patterns of soil redistribution by calculating [Cs.sup.137] percentage residuals:

[C.sub.r] = 100 ([C.sub.p] - [C.sub.f])/[C.sub.f] (6)


[C.sub.r] is the [Cs.sup.137] percentage residual

[C.sub.f] is a measure of the fallout inventory of [Cs.sup.137] (in this case 250[+ or -]10 Bq [m.sup.-2]), derived by sampling an undisturbed, uneroded site adjacent to the sampled field.

Negative percentage residuals indicate erosion, and positive residuals indicate aggradation. The magnitude of the residual is approximately proportional to the severity of erosion or aggradation. In this study, the [Cs.sup.137] residuals have been used directly as measures of net soil redistribution. This avoids the need to make assumptions regarding process and erosion rate calculation.

A range of approaches has been proposed for the conversion of percentage residuals into quantitative estimates of erosion and these have been reviewed elsewhere (e.g., Walling and Quine 1990; Quine 1995; Quine et al. 1996). In this study, the only use of [Cs.sup.137] for erosion rate estimation was in the calibration of the tillage erosion model to derive the appropriate value of [k.sub.1]., which followed the approach explained in detail in Quine et al. (1997). The measured [Cs.sup.137] data were compared with [Cs.sup.137] data derived as a result of simulated tillage erosion using TBGIS for a range of values of [k.sub.1] (Equation 4) until optimum agreement was obtained between simulated and measured [Cs.sup.137] residuals. Optimum agreement was obtained with a [k.sub.1] value of 520 kg [m.sup.-1] [year.sup.-1] (350 lb [ft.sup.-1] [year.sup.-1]), and this was used to produce a spatial distribution of tillage erosion rates over the last four decades. However, it is important to emphasize that in the analysi s that follows, the use of [Cs.sup.137] residuals as indicators of net soil redistribution makes no assumption regarding erosion process, and the simulated tillage erosion pattern is independent of the magnitude of the [k.sub.1] value.

Results and Discussion

The integrated site-specific data were divided into four broad groups: terrain attributes, soil properties, crop production parameters, and measures of soil redistribution. Only the crop production parameters satisfied the Anderson-Darling test for normality, and therefore, non-parametric statistical tests were used in the exploration of interrelationships. A list of summary statistics for terrain attributes and measured data can be seen in Table i.

Topography and soil variability. The terrain attributes derived from the DEM provide a satisfactory basis for the subsequent non-parametric analysis, even though the absolute values are scale-dependent. The range of slope angles cultivated in the region is well represented by the large variation encountered at the field site, and the topographic complexity of the landscape is reflected in the very high variability in plan and profile curvature.

The soil physical properties, such as bulk density and particle size composition, demonstrate lower variability than the chemical properties. The calculated coefficients of variance range from 7-18% and 17-25%, respectively, in the plow soil. Variability is higher and differences between physical and chemical properties are more pronounced in the sub-plow layer, where variance ranges are 20-30% and 45-80%, respectively. Crop production parameters exhibit similar variability to the chemical properties of the plow soil. When the data for this site are compared to the studies reviewed by Goderya (1998), variability in this field is seen to be higher for hulk density (6-14%) and crop yield (8-15 and 29%), lower for sand (17-83%) and clay (9-45%), and within the range of previously observed values for the other properties.

The soil and terrain variables have been identified as non-normal in distribution, and it has been suggested by Ovalles and Coffins (1986) that this is an indication of systematic rather than random variation. Further investigation of potential systematic variation was undertaken by examination of bivariate relations between soil properties and topographic attributes. The Spearman's correlation coefficients for these relations are shown in Table 2. Profile curvature is seen to be strongly related (95%) to all of the soil properties and very strongly related (99%) to total N, total C, and total P. In profile curvature, a negative sign for the correlation coefficient indicates depletion on convexities and enrichment in concavities, while for plan curvature this association is reflected in a positive sign for the coefficient. The relationships between the soil properties and plan curvature are weak, but the topographic association indicated by the two measures of curvature is consistent. Convexities are associa ted with depletion of total N, total C, total P, inorganic P, and silt, and with enrichment of sand. It is interesting to note that slope is only statistically significantly related to P (total and inorganic)--the sign of the relationship indicating enrichment in areas of low slope angle and depletion on steep slopes.

While the terrain and soil data and the associated relations provided evidence for the existence of systematic variation in soil properties, it remained necessary to identify the cause of this variation. This was investigated by examining relations between the soil properties and measures of erosion.

Soil redistribution and variability of soil properties. The most direct evidence for soil redistribution within the field site is provided by the [Cs.sup.137] percentage residuals. Comparison of these with the soil properties makes no assumption regarding erosion process, but allows evaluation of a possible link between soil variability and spatial variation in erosion and deposition intensity. Since non-parametric correlation analysis is used, the [Cs.sup.137] percentage residual is a particularly appropriate and robust indicator of soil redistribution intensity, as no assumptions need to be made concerning conversion of [Cs.sup.137] data to erosion rate estimates.

The correlation of [Cs.sup.137] with the measured soil properties can be seen in Table 2. These data show that all the measured soil properties are significantly ([alpha] = 0.05) correlated to the [Cs.sup.137] residual, and all except inorganic P are strongly ([alpha] = 0.01) correlated. This indicates that the variation of soil properties within the field is significantly related to the pattern of net soil redistribution that has occurred over the last four decades. Furthermore, for all soil properties except inorganic P, the correlation coefficients suggest that interrelations with [Cs.sup.137], and therefore net soil redistribution, are stronger than with individual terrain attributes. The nature of these interrelations is illustrated in Figure 2 for total P, N, and C.As would be expected from the positive correlation coefficients, the scatter plots show a gradual increase in the P, N and C with increasing [Cs.sup.137] In the case of total P, there is relatively little scatter but a marked change in gradi ent of the relationship for [Cs.sup.137] residuals above 100%. However, caution is needed in interpreting this because of the small number of sites with such high aggradation rates. More scatter is seen in the plots of total N and total C, but both evidence consistently low levels in the most eroded locations ([Cs.sup.137] <-50%).

In summary, eroded areas tend to be nutrient-poor and aggraded areas relatively nutrient-rich. This pattern is clearly seen when the spatial distributions of total N (Figure 3) and [Cs.sup.137] residuals (Figure 4a) are compared. Although the [Cs.sup.137] residual data provide evidence for a link between soil redistribution and variability in soil properties, without further analysis they do not identify the erosion process or processes responsible. This can be investigated from two directions: first, by identifying the erosion process that best explains the [Cs.sup.137] residual pattern, and second, by examining relations between soil properties and process-specific measures of soil redistribution.

The spatial distributions of [Cs.sup.137] residuals, simulated tillage, and simulated overland flow erosion potential are shown in Figure 4. The location of the rill network surveyed in 1998 is shown in Figure 5. The rill network and overland flow potential distribution show good qualitative agreement, indicating that the latter may be used as a reliable indicator of spatial variation in water erosion intensity in the non-parametric analysis. Nevertheless, comparison of the single-process distributions with the pattern of [Cs.sup.137] residuals suggests that tillage erosion is the dominant soil redistribution process within the field. This is further supported by quantitative analysis, which indicates that 40% of the observed variance in [Cs.sup.137] residuals can be explained by simulated tillage erosion, but only 10% by simulated water erosion. This close agreement between [Cs.sup.137] and simulated tillage suggests that tillage erosion may be identified as the dominant soil redistribution process causing variation in soil properties. This suggestion is supported by the strong statistically significant relationships between measured soil properties and simulated tillage erosion (Table 2). In contrast, the water erosion index exhibits no statistically significant (95%) relationships with soil properties.

Qualitative interpretation of crop variability. The spatial distributions of aboveground biomass and crop yield are shown in Figure 6. Over the whole field except the lowest margin, the crop production distributions are similar to those of [Cs.sup.137] residuals and soil redistribution by tillage (Figures 4a and 4b, respectively). The crop production distributions exhibit low values where highest erosion rates were noted on the upper field perimeter and on the mid-field convexities. Furthermore, high yield and biomass are noted where soil aggradation was identified in the mid-field concavities. Therefore, over most of the field, variability in biomass and yield appears to be closely correlated with and, by implication at least, partially controlled by soil redistribution. Nevertheless, the relationship between soil redistribution and crop production is not simple in this field. At the base of the field, especially in the lowest corner, sample sites with high rates of soil aggradation and associated nutrient e nrichment are not characterized by elevated rates of crop production. Field observations suggest that the area is subject to water-logging, weed infestation, and occasional ground frost. Low crop yields in similar areas have been reported elsewhere (Franzen et al. 1997; Schumacher et al. 1999). The presence of low crop production in this area of high aggradation weakens the relationship between soil redistribution and crop production (Figure 7). In light of this complexity, it is not surprising that non-parametric correlation analysis tends to show weaker correlation between crop variability and soil properties and erosion measures than has been seen between the latter groups. Nevertheless, statistically significant ([alpha] = 0.05) correlation coefficients are found between above ground biomass and both planform curvature and simulated soil redistribution by tillage. Both aboveground biomass and crop yield also exhibit statistically significant ([alpha] = 0.05) correlation coefficients with total phosphorus.

Despite the close visual similarity between patterns of crop production parameters and measures of tillage erosion and total soil redistribution ([Cs.sup.137]), there remains much unexplained variability in crop production. Even when the data are log-transformed, variation in [Cs.sup.137] explains only 16% and 11%, respectively, of the variation in aboveground biomass and yield. This reflects the fact that the growth and development status of a crop population depends on the coordination of soil-plant-air continuity, which is temporally and spatially variable (Franzen et al. 1997; Timlin et al. 1998). Under conditions of relatively uniform management and with the adoption of the same crop cultivar, crop variability at the field scale is most likely to be determined by the interaction of soil properties and microclimates. The growing season for the winter wheat under study was the wettest for 20 years, with a recorded rainfall of 959 mm (37.8 in), compared with the average of 845 mm (33.3 in). Therefore, this single-year dataset only represents the crop variability in a wet year, under conditions in which nonlinear relationships between crop production and soil properties (and erosion) are likely.

High rates of soil redistribution within agricultural fields have been identified in earlier studies using (137) Cs (Martz and de Jong 1987; Walling and Quine 1990; Quine and Walling 1991; Quine et al. 1994). This study is consistent with these earlier investigations, evidencing patterns and rates of soil redistribution than cannot be explained by water erosion. The pattern of soil redistribution, characterized by loss from convexities and deposition in concavities, is typical of tillage erosion and agrees with earlier studies that have indicated the dominance of tillage in soil redistribution on rolling topography subject to mechanical cultivation (Covers et al. 1993, 1996; Quine et al. 1994, 1996, 1997). Tillage erosion operates like a conveyor belt, transferring soil and associated constituents from convexities to concavities. On this field, rates of erosion and accumulation of as much as 4 kg [m.sup.2] [year.sup.-1] or 40 t [ha.sup.-1] [year.sup.-1] (16 tons [acre.sup.-1] [year.sup.-1]) (Figure 4b) were noted, equivalent to 1.5% of the plow layer per year. During cultivation, plow layer depth is maintained on the convex slope elements by incorporation of nutrient-poor and, on this field, usually more coarse subsoil into the plow soil. Therefore, the plow soil on these eroded Convexities becomes depleted in surface-applied or surface-immobilized nutrients, the products of weathering, and fine soil fractions (Table 2). (It is interesting to note that, on this field, soil modification through tillage erosion results in whole-field correlation between nutrients and texture that would probably be attributed to size-selective transport of fines and associated nutrients, if water erosion was assumed to be dominant). The depleted plow soil is also translocated away from the convexities, and therefore, areas of no (or limited) net soil loss on linear slope elements below convexities may also be characterized by nutrient depletion of the plow soil. Conversely, plow soil accumulates in concavities through downslope tra nslocation from the upslope landscape elements. The significance of these processes for profile anisotropy may be seen in the ratio of the total carbon concentration in the plow soil to the concentration in the sub-plow soil (Figure 8). In areas of aggradation, most of the ratios lie between 1.5 and 3, with a decline towards 1 as aggradation rates increase. This reflects the presence in the sub-plow soil of increasing quantities of former plow soil (chemically similar to the current plow soil) isolated below the plow layer by gradual surface elevation due to aggradation. In contrast, in eroding areas ratios are much higher, reflecting the surface origin and, therefore, surface concentration of carbon and the pure sub-soil origin of the sub-plow soil. It is suggested that translocation of soil by tillage erosion is, therefore, a major contributor to within-field variability in soil properties. The strong correlation between soil properties (C, N and P) and both net soil redistribution ([Cs.sup.137]) and simula ted tillage erosion supports this hypothesis.

Clearly, continuing tillage erosion will result in more extreme spatial variability in soil properties. In particular, greater depletion of the soils on the convexities and an expansion of the area of depleted soils is anticipated. In order to evaluate the severity of this phenomenon, TBGIS was used to simulate [Cs.sup.137] and soil redistribution due to tillage from 1997-2040. The frequency distributions of [Cs.sup.137] residuals over the sampled area for 1997 and 2040 are shown in Figure 9a. These distributions show an increase in skewness and in the proportion of the field that is severely depleted ([Cs.sup.137] residuals <-50%), from 13-38%. The area subject to very high rates of aggradation ([Cs.sup.137] residuals >100%) also increases from less than 1% to over 7%. This represents a significant increase in within-field variability. Evolution of variability in other properties will be more complex, due to artificial and "natural" additions over the time period, but the pattern is likely to be similar.

The economic implications of continually-developing soil variability are more difficult to gauge because the data obtained in this study represent a single exceptionally wet year, and the relationship between erosion and crop production has been observed to be more complex than that between erosion and the soil properties investigated. Nevertheless, a qualitative evaluation may be made on the basis of the operation of tillage erosion discussed above. On the basis of the increase in the proportion of the field in the most denuded areas (Figure 9a) and the low yields seen in such areas at present (Figures 9b and 7), it is anticipated that significant yield decline may occur on both the currently eroded zones and on the adjacent downslope landscape elements. Areas of current aggradation will continue to be receptor sites for soil and nutrients. As has been seen, these aggradation areas may not be optimal for crop production, and it is anticipated that, at best, these areas will not produce sufficient additional yield to compensate for eroded zones and, at worst, that continued soil accumulation may contribute to yield decline in these areas as well (cf. Figures 7 and 9).

Summary and Conclusion

This study employed a pedo-geomorphological approach to the investigation of soil and crop variability. A close relationship between spatial patterns of soil properties and the spatial distribution of net erosion and tillage erosion was demonstrated. It was suggested that soil redistribution by tillage is the most important erosion process contributing to within-field soil variability and that, due to its continuous operatino and high magnitude, higher within-field soil variability can be expected to develop. This has the important implication for precision farming approaches that "natural" soil variation can no longer be considered a static variation on the time scales of concern to land managers--indeed the management of the land is contributing to the pattern of soil variability that managers attempt to overcome. It is clear that attention must be given to tillage erosion in the development of strategies to manage soil variability.

Patterns of soil redistribution by tillage also showed close visual similarity to patterns of crop production. However, the strength of the statistical relationship between these parameters was influenced by lower than expected yields in low-lying aggradational zones near the field boundary. More data for years with different weather conditions are required to explore these trends more fully. Nevertheless, it is apparent that yields in eroded zones are unlikely to match those in uneroded areas and are unlikely to be fully compensated for by elevated yields in aggradational areas. It is suggested that tillage erosion represents a threat to cost-effective and environmentally-sustainable agriculture in areas of rolling terrain because it results in the transfer of resources from areas where they are needed to areas where they are surplus to requirements and may have a deleterious effect.


This study was undertaken as part of the TERON (Tillage Erosion: current state, future trends and prevention) project (FAIR CT96 1478), and the financial support provided by the European Commission is gratefully acknowledged. We are also very grateful to Mr. John Lee for providing ready access to the field site on several occasions and for his assistance with field logistics.

[Figure 2 omitted]

[Figure 7 omitted]

[Figure 8 omitted]

[Figure 9 omitted]
Table 1

Descriptive statistics of measured field and crop properties.

Properties                                        min.   max.  median

Terrain        slope (degree)                      1.4   19.8     8.3
               plan curvature (m/100 m) (+)     -0.905   1.11    0.02
               profile curvature (m/100 m) (+)  -1.639  1.023   0.034
Plow soil      bulk density (g/[m.sup.3])        1.046  2.352   1.478
               sand fraction (%)                  28.7   53.2    41.5
               silt fraction (%)                  35.1   58.9    43.8
               clay fraction (%)                  11.2   24.0    14.6
               total C (%)                        0.64   3.37    1.36
               total N (%)                        0.09   0.31    0.15
               total P (ppm)                       385   1392     590
               inorganic P (ppm)                   280   1200     439
Sub-plow soil  sand fraction (%)                  25.3   60.1    43.4
               silt fraction (%)                  29.9   55.3    42.1
               clay fraction (%)                   7.7   20.1    14.3
               total N (%)                        0.02   0.23    0.10
               total C (%)                        0.07   2.56    0.69
Crop           AGB (g/[m.sup.2])                   658   2444    1583
               yield (g/[m.sup.2])                 270   1051     587
               ratio of grain to stalk            0.37   1.32    0.60
Erosion        Caesium-137 residual (%)          -79.2  148.0    -3.4

Properties     Q1 (a)  Q3 (a)  C.V. (b)

Terrain           6.5     9.8      38.9
               -0.112   0.225     > 100
               -0.135   0.211     > 100
Plow soil       1.364   1.624      17.6
                 39.4    43.7      10.5
                 42.1    45.5       7.7
                 13.8    15.5      12.0
                 1.24     1.5      18.8
                 0.13    0.17      24.2
                  523     667      24.5
                  396     472      17.1
Sub-plow soil    39.0    52.0      29.8
                 36.2    44.9      20.7
                 11.8    16.1      30.1
                 0.07    0.12      46.8
                 0.39    0.94      78.6
Crop             1376    1795      26.5
                  484     671      31.9
                 0.54    0.66      20.6
Erosion         -24.0    15.8     > 100

(a)Q1 and Q3 are the 25th and 75th of quartile, respectively.

(b)Coefficient of variance (C.V. %) is calculated as: ((Q3. Q1)/ median)
(*) 100

(+)Concavity is negative and convexity is positive for plan curvature.
The opposite sign is adopted for profile curvature
Table 2

Tabular list of Spearman's correlation coefficients

                                     Total N       Total C

Plow soll properties

Total N                                              0.876 (**)
Total C
Total P
Inorganic P

Topographic attributes

Slope                                 -0.111        -0.174
Plan curvature                        -0.376 (*)    -0.367 (*)
Profile curvature                      0.464 (**)    0.585 (**)

Erosion measures

[Cs.sup.137] residual (- = erosion)    0.613 (**)    0.671 (**)
Simulated Tillage redistribution
 (- = erosion)                         0.512 (**)    0.586 (**)
Simulated water erosion
 (+ = erosion)                         0.180         0.168

                                     Total P               P

Plow soll properties

Total N                                0.576 (**)      0.264
Total C                                0.622 (**)      0.356 (*)
Total P                                                0.652 (**)
Inorganic P

Topographic attributes

Slope                                 -0.435 (**)     -0.493 (**)
Plan curvature                        -0.226          -0.018
Profile curvature                      0.461 (**)      0.314 (*)

Erosion measures

[Cs.sup.137] residual (- = erosion)    0.605 (**)      0.407 (*)
Simulated Tillage redistribution
 (- = erosion)                         0.481 (**)      0.238
Simulated water erosion
 (+ = erosion)                         0.146           0.006

                                       Sand         Silt

Plow soll properties

Total N                              -0.516 (**)   0.466 (**)
Total C                               0.524 (**)   0.516 (**)
Total P                               0.570 (**)   0.566 (**)
Inorganic P                          -0.214        0.231
Sand                                               0.948 (**)

Topographic attributes

Slope                                 0.199       -0.203
Plan curvature                        0.344 (*)   -0.326 (*)
Profile curvature                    -0.362 (*)    0.38 (*)

Erosion measures

[Cs.sup.137] residual (- = erosion)  -0.518 (**)   0.503 (**)
Simulated Tillage redistribution
 (- = erosion)                       -0.470 (**)   0.458 (**)
Simulated water erosion
 (+ = erosion)                       -0.170        0.159

                                     Ground        Crop
                                     Biomas       Yield

Plow soll properties

Total N                               0.120       0.157
Total C                               0.165       0.205
Total P                               0.303 (*)   0.354 (*)
Inorganic P                           0.036       0.164
Sand                                 -0.288      -0.286
Silt                                  0.237       0.243

Topographic attributes

Slope                                -0.057       0.002
Plan curvature                        0.345 (*)  -0.262
Profile curvature                     0.272       0.212

Erosion measures

[Cs.sup.137] residual (- = erosion)   0.283       0.254
Simulated Tillage redistribution
 (- = erosion)                        0.370 (*)   0.310
Simulated water erosion
 (+ = erosion)                       -0.096      -0.133

(*)statistically significant at [alpha] = 0.05

(**)statistically significant at [alpha] = 0.01


Applegarth, S.M. and L. Cornish. 1982. Agricultural land classification of England and Wales - Report to accompany sheet 176, Exeter, 1975. United Kingdom Ministry of Agricultural Fisheries and Food. 19p.

Clayden, B. 1971. Soils of the Exeter District, Memoirs of the Soil Survey of England and Wales. Soil Survey of England and Wales, Harpenden, United Kingdom. 254p.

Daniels, R.B., J.W. Gillam, D.K. Cassel, and L.A. Nelson. 1985. Soil erosion class and landscape position in the North Carolina piedmont. Soil Science Society of America Journal 49:991-995.

Findlay, D.C., G.J.N. Colborne, D.W. Cope, T.R. Harrod, D.V. Hogan, and S.J. Staines 1984. Soils and their use in South West England. Soil Survey of England and Wales, Harpenden, United Kingdom. 4l9p.

Franzen, D.W., A. Halvorsen, and V. Hofman. 1997. Variability of soil nitrate, phosphate, chloride and sulfate-S under different landscapes-a report of work in progress. North Dakota State University Extension Service, Fargo, ND.

Goderya, F.S. 1998. Field scale variation in soil properties for spatially variable control: a review. Journal of Soil Contamination 7:243-264.

Govers, G., T.A. Quine, and D.E. Walling. 1993. The effect of water erosion and tillage movement on hillslope profile development: a comparison of field observations and model results. Pp 285-300. In: S. Wicherek (ed). Farm Land Erosion in Temperate Plains Environment and Hills. Amsterdam: Elsevier.

Govers, G., K. Vandaele, P.J.J. Desmet, J. Poesen, and K. Bunte. 1994. The role of tillage in soil redistribution on hill-slopes. European Journal of Soil Science 45:469-478.

Govers, G., T.A. Quine, P.J.J. Desmet, and D.E. Walling. 1996. The relative contribution of soil tillage and overland flow erosion to soil redistribution on agricultural land. Earth Surface Processes and Landforms 21:929-946.

Govers, G., D.A. Lobb, and T.A. Quine. 1999. Tillage erosion and translocation: emergence of a new paradigm in soil erosion research. Soil and Tillage Research 51:167-174.

Lindstrom, M.J., W.W. Nelson, T.E. Schumacher, and G.D. Lemme. 1990. Soil movement by tillage as affected by slope. Soil and Tillage Research 17:255-264.

Lindstrom, M.J., W.W. Nelson, and T.E. Schumacher. 1992. Quantifying tillage erosion rates due to moldboard plowing. Soil and Tillage Research 24:243-255.

Lobb, D.A., R.G. Kachanoski, and M.H. Miller. 1995. Tillage translocation and tillage erosion on shoulder slope landscape positions measured using 137Cs as a tracer. Canadian Journal of Soil Science 75:211-218.

Lobb, D.A. and R.G. Kachanoski. 1999. Modelling tillage translocation using step, linear-plateau and exponential functions. Soil and Tillage Research 51:317-330.

Loughran, R.J. 1989. The measurement of soil erosion. Progress in Physical Geography 13:216-233.

Martz, L.W. and E. de Jong. 1987. Using caesium-137 to assess the variability of net soil erosion and its association with topography in a Canadian Prairie landscape. Catena 14:439-451.

Miller, M.P., M.J. Singer, and D.R. Nielsen. 1988. Spatial variability of wheat yield and soil properties on complex hills. Soil Science Society of America Journal 52:1133-1141.

Moore, I.D., P.E. Gessler, G.A. Nielsen, and G.A. Peterson. 1993. Soil attributes prediction using terrain analysis. Soil Science Society of America Journal 57:443-452.

Moulin, A.P., D.W. Anderson, and M. Mellinger. 1994. Spatial variability of wheat yield, soil properties and erosion in hummocky terrain. Canadian Journal of Soil Science 74:219-228.

Ovalles, F.A. and M.E. Collins, 1986. Soil-landscape relationships and soil variability in north central Florida. Soil Science Society of America Journal 50:401-408.

Poesen, J., B. van Wesemael, G. Govers, J. Martinez- Fernandez, P.J.J. Desmet, K. Vandaele, T.A. Quine, and G. Degraer, 1997. Patterns of rock fragment cover generated by tillage erosion. Geomorphology 18:183-197.

Quine, T.A. 1995. Estimation of erosion rates from caesium-137 data: the calibration question. Pp 307-329. In: I.D.L. Foster, A.M. Gurnell, and B. W. Webb (eds). Sediment and Water Quality in River Catchments. Chichester, United Kingdom: John Wiley.

Quine, T.A. and D.E. Walling. 1991. Rates of soil erosion on arable fields its Britain: quantitative data frosts caesium-137 measurements. Soil Use and Management 7:169-176.

Quine, T.A., P.J.J. Desmet, K. Vandaele, G. Govers, and D.E. Walling. 1994. A comparison of the roles of tillage and water erosion its landform development and sediment export on agricultural land near Leuven, Belgium. Pp 77-86. In: L.J. Olive, R.J. Loughran, and J.A. Kesby (eds). Variability its Stream Erosion and Sediment Transport (Proceedings of the Canberra Symposium, December 1994). International Association of Hydrological Sciences Publication 224..

Quine, T.A., D.E. Walling and G. Govers. 1996. Simulation of radiocaesium redistribution on cultivated hillslopes using a mass-balance model: an aid to process interpretation and erosion rate estimation. Pp 561-588. In: M.G. Anderson and S.M. Brooks (eds). Advances its Hillslope Processes. Chichester, United Kingdom: John Wiley.

Quine, T.A. G. Govers, D.E. Walling, X. Zhang, P.J.J. Desmet, and Y Zhang. 1997. Erosion processes and landform evolution on agricultural land-new perspectives frosts caesium-137 data and topographic-based erosion modelling. Earth Surface Processes and Landforms 22:799-816.

Quine, T.A., D.E. Walling. and X. Zhang. 1999. Tillage erosion, water erosion and soil quality on cultivated terraces at Xifeng in the Loess Plateau, China. Land Degradation and Development 10:251-274.

Quine, T.A., G. Govers, J. Poesen, OF. Walling, B. vast Wesemael, and J. Martinez-Fernandez. 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, T.A., D.E. Walling. Q.K. Chakela, O.T. Mandiringana, and X. Zhang. 1999. Rates and patterns of tillage and water erosion on terraces and contour strips: evidence frosts caesium-137. Catena 36:115-142.

Ritchie, J.C. and J.R. McHenry. 1975. Fallout Cs-137: a tool in conservation research. Journal of Soil and Water Conservation 30:283-286.

Schumacher, T.E., M.J. Lindstrom, J.A. Schumacher. and G.D. Lemme. 1999. Modeling spatial variation in productivity due to tillage and water erosion. Journal of Soil and Tillage Research 51:331-339.

Stone, J.R., J.W. Gilliam, D.K. Cassel, R.B. Daniels, L.A. Nelson, and H.J. Kleiss. 1985. Effect of erosion and landscape position on the productivity of Piedmont soils. Soil Science Society of America Journal 49:987-991.

Timlin, D.J., Y.A. Pachepsky. V.A. Synder, and R.B. Bryany. 1998. Spatial and temporal variability of corn grain yield on a hillslope. Soil Science Society of America Journal 62:764-773.

Van Muysen, W., G. Govers, J. Poesen. G. Bergkamp and M. Roxo. 1999. Measurement and modelling of the effects of initial soil conditions and slope gradient on soil translocation by tillage. Journal of Soil arid Tillage Research 51;303-316.

Walling, D.E. and T.A. Quine. 1990. Calibration of caesium-137 measurements to provide quantitative erosion rate data. Land Degradation and Rehabilitation 2:161-175.

Walling, D.E. and T.A. Quine. 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.

Yang, C., C.L. Peterson, G.J. Shropihire, and T. Otawa. 1998. Spatial variability of field topography and wheat yield in the Palouse region of the pacific northeast. Transactions of the American Society of Agricultural Engineers 41:17-27.

Zhang, Y. 1996. GIS-based modelling of field-scale erosion and deposition patterns. Unpublished Ph.D. thesis, University of Exeter, United kingsdom.

Dr Timothy A. Quine is a reader in Earth Surface Processes and Dr. Zhang Yusheng is a research fellow with the Department of Geography, University of Exeter, Exeter, Devon, United Kingdom.
COPYRIGHT 2002 Soil & Water Conservation Society
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2002 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Quine, T.A.; Zhang, Y.
Publication:Journal of Soil and Water Conservation
Article Type:Abstract
Date:Jan 1, 2002
Previous Article:Modeling wind and water erosion in Northern China under climate and land use changes.
Next Article:Conservation: overhead or value added? A viewpoint from the Soil and Water Conservation Society's executive director. (Home Front).

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters