Soil Use and Management, December 2007, 23, 329337

doi: 10.1111/j.1473-2743.2007.00102.x

Field assessment of soil structural quality a development of the Peerlkamp test

B. C. Ball1, T. Batey2 & L. J. Munkholm3
SAC Crop and Soil Systems Research Group, West Mains Road, Edinburgh EH9 3JG, UK, 2Plant & Soil Science Department, University of Aberdeen, Aberdeen AB24 3UU, UK, and 3Department of Agroecology and Environment, Arhus University, PO Box 50, DK-8830 Tjele, Denmark
1

Abstract
Increased awareness of the role of soil structure in dening the physical fertility or quality of soil has led to the need for a simple assessment relevant to the environmental and economic sustainability of soil productivity. A test is required that is usable by farmer, consultant and researcher alike. Here an assessment of soil structure quality (Sq) is described which is based on a visual key linked to criteria chosen to be as objective as possible. The inuences of operator, tillage and crop type on Sq value were tested. The test takes 515 min per location and enough replicates were obtained for statistical comparison of data sets. The assessments of individual operators were inuenced to an extent by differences between elds, making the use of multiple operators desirable. Differences in soil management were revealed by the test and related to differences in soil physical properties (bulk density, penetration resistance and porosity) and crop growth. Indicative thresholds of soil management are suggested. The assessment should be viewed as complementary to conventional laboratory assessments of soil structure. Visual soil structure assessment can indicate to the soil scientist where to sample and what soil measurements are likely to be worthwhile.

Keywords: Visual analysis, soil structure, soil management, soil resilience, tillage, crop rotation

Introduction
Soil structure is a complex soil property, partly related to inherent characteristics of particle size and clay mineralogy and partly to anthropogenic inuences related to land use and management. According to Kay & Angers (2001), soil structure can be described in terms of: (i) structural form; the heterogeneous arrangement of pores and solids at any given time; (ii) structural stability, the ability of soil to retain its structural form after exposure to stress; and (iii) structural resilience: the ability of soil to recover its structural form through natural processes. Soil structural form may be described visually in terms of the size, shape and porosity of individual units of structure (aggregates) to which additional physical properties such as strength and stability can be added. However, from such a description, it may be difcult to derive an indication of the quality of the structure for agricultural production. Ideally, description of structural quality should include elements of form, stability and resilience.
Correspondence: B. C. Ball. E-mail: bruce.ball@sac.ac.uk Received March 2007; accepted after revision July 2007

In the 1950s and 1960s, Peerlkamp and his co-workers including Boekel, developed a scale to assess the quality of soil structure (Peerlkamp, 1959; De Boodt et al., 1967). Although their St score was related to soil organic matter, clay content and crop performance, it was not widely used outside of the Netherlands and the UK, especially in recent years. During the same period, many indices of structure were developed, based on laboratory tests for stability of individual aggregates, their dispersion in water or pore size distribution or on eld permeability assessments, for example those of Low (1954), Childs et al. (1957), Williams & Cooke (1961) and Currie (1965). Aspects of soil structure are commonly described and quantied from water retention, water and gas permeability, gas diffusivity and aggregate stability (De Boodt et al., 1967; Kay & Angers, 2001). More recent methods involve micromorphology (Ringrose-Voase, 1991), 3D visualization and advanced modelling (Young & Crawford, 2004). Laboratory studies are normally carried out using soil samples of small volume and can be time consuming and expensive. Laboratory methods are highly valuable in quantifying specic aspects of soil structure. However,

2007 The Authors. Journal compilation 2007 British Society of Soil Science

329

330 B. C. Ball et al.

they are generally unable to provide a rapid, immediate assessment of the quality of soil structure. Assessments made directly in the eld are done on the spot; little equipment is needed and the results are immediately available. Such assessments are of great value to land users and consultants who need to make soil management decisions in the eld based on spot measurements of soil structural quality. They are also of value in enabling researchers to identify which layers to sample to identify soil properties associated with differences in soil structure. In recent years, eld-based assessments of soil structure quality have been developed virtually simultaneously in several countries France (Roger-Estrade et al., 2004), Denmark (Munkholm, 2000), Germany (Beste, 2003), New Zealand (Shepherd, 2000), Australia (Lawrie et al., 2000; McKenzie, 2001), Scotland (Ball & Douglas, 2003), Switzerland (Hasinger et al., 2004) and England (NSRI, 2001). Their use has been encouraged by the activities of Working Group F Visual soil examination and evaluation of ISTRO (International Soil Tillage Research Organisation). The Working Group promoted a meeting held in 2005 in northern France when 10 methods were used to assess soil structure on the same soil directly in the eld (Boizard et al., 2005). One of the methods evaluated was the original Peerlkamp test. This is a system which is rapid enough to provide sufcient samples for assessment of standard deviations of

Table 1 Method of assessment Step Block extraction and examination 1. Extract soil block Option Procedure

Loose soil

Firm soil

2. Examine soil block

Uniform structure Two or more horizontal layers of differing structure

Remove a block of soil 15 cm thick directly to the full depth of the spade and place spade plus soil onto the sheet, tray or the ground. Dig out a hole slightly wider and deeper than the spade leaving one side of the hole undisturbed. On the undisturbed side, cut down each side of the block with the spade and remove the block as above. Remove any compacted soil or debris from around the block. Estimate the depth of each layer and prepare to assign scores to each separately. Gently manipulate the block using both hands to reveal any cohesive slabs of soil or clumps of aggregates. If possible separate the soil into natural aggregates and man-made clods. Clods are large, hard, cohesive and rounded aggregates. Break larger pieces apart and look at the internal structure of the cross-section. A crumb-like appearance with rounded aggregates easily broken apart and embedded in a ner matrix would be indicative of a well-developed natural structure and lower score. Clods that can be broken into non-porous aggregates with angular corners are indicative of poor structure and higher score. Match the soil to the pictures (Figure 1) category by category to determine which ts best. Factors increasing score: Difculty in extracting the soil block Larger, more angular, less porous, presence of large worm holes Clustering, thickening and deections Pockets or layers of grey soil, smelling of sulphur and presence of ferrous ions Multiply the score of each layer by its thickness and divide the product by the overall depth, e.g. for a 25-cm block with 10-cm depth of loose soil (Sq1) over a more compact (Sq3) layer at 10- to 25-cm depth, the block score is [(1 10) 25] + [(3 15) 25] = Sq 2.2.

Block break-up 3. Break-up block (take a photograph optional)

4. Break-up aggregates

Soil scoring 5. Assign score 6. Conrm score from Block extraction Aggregate shape and size Roots Anaerobism 7. Calculate block scores for two or more layers of differing structure

2007 The Authors. Journal compilation 2007 British Society of Soil Science, Soil Use and Management, 23, 329337

Field assessment of soil structure 331

structural scores. Several authors in the Working Group had based their method on the early work of Peerlkamp. However, during discussions after the presentation of the results, it became apparent that they had encountered several difculties during its use. They felt that the criteria for the classes were not robustly dened and that a scale of 10 was too long. Particular problems were the emphasis on size and strength of aggregates after breaking apart a spadeful of soil and the need to include further visual observations such as colour, shape and orientation of aggregates, pores and roots. A test was required that was objective, readily understood and could be taught quickly to non-soil scientists. A visual key was thus developed to meet some of these needs. The objective was to incorporate simplied structural descriptions into a scale of structural quality. Further investigations to evaluate the test have been made in Denmark and Scotland. In this paper, a new method based on Peerlkamp is described which has benets in terms of ease of use, reproducibility and sensitivity to soil management.

Methods and materials

Development of the modied Peerlkamp test
Our test incorporates parts of the original method; it is based on the appearance, strength and structure of a block of soil dug out with a spade. In the original Peerlkamp test, scores ranged from St1 (poorest structure) to St10 (best structure). After discussion with colleagues, the optimum number of scores for classication was chosen as ve. Many systems used for land classication are based on the concept that grade 1 or class 1 is the highest quality and that higher numbers represent lower quality. This concept is well established and accepted both scientically and politically. The same concept for structure quality was chosen, thereby reversing the original order. A different acronym (Sq) was used to avoid confusion with the original St scores. First, a range of soil types from loamy sands to clay loams in north-east Scotland was assessed after consulting several colleagues and a draft description was agreed. Second, the description was rened by assessing a

St r u c t u r e q u al i t y

Eas e o f b r eak u p (m o i s t s o i l )

Si ze an d ap p ear an c e o f ag g r eg at es Mo s tl y < 6 m m after crumbling

Vi s i b l e p o r o s i t y

Ro o t s

A p p ear an c e af t er b r eak -u p : v ar i o u s soils

A p p ear an c e af t er b r eak -u p : s am e s o i l d i f f er en t t i l l ag e

Di s t i n g u i s h i n g f eat u r e

Sq 1 Fr i ab l e (t en d s t o f al l off the s p ad e)

Aggregates readily crumble with fingers

Highly porous

Roots throughout the soil

Fine aggregates Sq 2 In t ac t (r et ai n ed as a block on t h e s p ad e) Aggregates easy to break with one hand A mixture of porous, rounded aggregates from 270 mm. No clods present Most aggregates are porous Roots throughout the soil

High aggregate porosity Sq 3 Fi r m Not difficult A mixture of porous aggregates from 2mm -10 cm; less than 30% are <1 cm. Some angular, non-porous aggregates (clods) may be present Mostly large > 10 cm and subangular nonporous; horizontal/platy also possible; less than 30% are <7 cm Mostly large > 10 cm, very few < 7 cm, angular and non-porous Macropores and cracks present. Some porosity within aggregates shown as pores or roots. Most roots are around aggregates

Low aggregate porosity Few macropores and cracks All roots are clustered in macropores and around aggregates Distinct macropores Very low; macropores may be present; may contain anaerobic zones Fe w, i f a ny , restricted to cracks

Sq 4 Co m p ac t

Quite difficult

Sq 5 Ver y c o m p ac t

Difficult

Grey-blue colour

Figure 1 Description and illustration of soil structure and distinguishing features in each of the ve categories of quality.

2007 The Authors. Journal compilation 2007 British Society of Soil Science, Soil Use and Management, 23, 329337

332 B. C. Ball et al.

range of tillage treatments giving contrasting structure types on the same soil type in Denmark. The resultant key contains guideline photographs of the appearance of the soil where structure differs due to soil type and where it differs due to different tillage on the same soil. The scale runs from Sq1 (best structure) to Sq5 (worst structure). The information given below is available on a laminated waterproof plastic card which includes the visual key for use in the eld.

(A) Foulum, Denmark (5630N 934E)

Main test objective at the site: suitability of the method for assessing structure within different tillage systems. Soil type: sandy loam (Mollic Luvisol according to the WRB classication). Annual rainfall: 704 mm. Tillage treatments: direct drilled soil (D), harrowed 810 cm soil (H810), and mouldboard ploughed soil (P). Existing crop: winter wheat (Triticum aestivum, L.). Previous crops: all winter sown. Number of tests: three tests per plot; four plots per tillage treatment. Timing: May 2006. Associated tests: In situ dry bulk density and penetration resistance measured in late autumn 2005. Crop performance was assessed 1 day before the Sq evaluations as spectral reectance and the relative vegetation index (RVI). The RVI is a measure of the green leaf area index and thus of the light interception of the crop canopy. These tests used the methods of Olesen & Munkholm (2007) .

Description of the test

Equipment. Flat-faced spade approximately 20 cm wide, 22 25 cm long, light-coloured plastic sheet, sack or tray approximately 50 80 cm, measuring tape, small knife and digital camera (optional). Timing of assessment. Assessments can be made at any time of the year, but preferably when the soil is moist so that a block of soil can be dug out without altering the structure. If the soil is dry, it may feel harder and be more difcult to dig out and to break apart so that an inaccurate score may be given. Roots are best seen in an established crop or for some months after harvest. Location of samples. Select an area of uniform crop or soil colour; within this, plan a grid to sample the soil at 10, preferably 20 locations. On small experimental plots, it may be necessary to restrict the number to three or ve per plot. Method of assessment. The method, summarized in Table 1, is split into three sections corresponding to block extraction and examination, block break-up and soil scoring. The method differs from the original Peerlkamp method in allowing for assessment of horizontal layers of different structure within the spadeful. The block is broken up and some of the resultant aggregates are broken apart before scoring by comparison with a visual key (Figure 1). This key includes photographs of typical samples after break-up for different soil types in each quality category (Sq) or for different tillage types on the same soil. Each quality category also has a distinguishing feature, shown in Figure 1. The Sq score is then conrmed by consideration of the ease of block extraction, shape and size of aggregates, distribution of roots and presence of any anaerobic zones.

(B) Locations 1 and 2, Craibstone Estate, near Aberdeen, Scotland (5710N 216W)
Main test objective at the site: inuence of operator in elds of contrasting soil structure. Soil type: sandy loam (Leptic Podsol according to the WRB classication). Annual rainfall: 865 mm. Fields: Location 1: relatively uniform soil structure except for a zone near one boundary fence which was compacted during establishment of the ley in 2003 (Ball et al., 2007). Location 2: variable soil structure due to storage of builders materials and soil and associated trafc during construction of an adjacent building prior to 2005. The site was subsoiled in 2005 and sown to spring barley in 2006. Existing crop: Location 1: grassclover ley grazed by sheep. Location 2: stubble of spring barley (Hordeum vulgare, L.). Previous crop: Location 1: spring barley (2003). Location 2: bare soil. Operators: Four operators assessed the soils. Two operators (1 and 3) were developers of the test and could be regarded as experts. The other two (2 and 4) were technical support staff who had been taught the test a few weeks prior to the assessments and were relatively inexperienced and could be regarded as non-experts. Number of tests: 10 tests per location by each operator. Timing: October 2006. Analysis of data: The signicance of the differences between sites and operators was assessed using a two-way analysis of variance.

Testing the method

Three sites (AC) were used to test the method under different soil, climatic and management situations and to assess the variation in score between operators.

2007 The Authors. Journal compilation 2007 British Society of Soil Science, Soil Use and Management, 23, 329337

Field assessment of soil structure 333

(C) Tulloch long-term ley-arable rotations trial, Craibstone Estate, near Aberdeen, Scotland (5710N 216W)
Main test objective at the site: inuence of crop type and crop vigour. Soil type and annual rainfall: as for site B. Crop types: third-year grass clover and third-year arable (undersown oats) (normal crop vigour). First year grass clover and fourth year grass clover (areas of plot where the crop growth was impaired due to suspected poor drainage). Previous crop: Swedes preceded third-year arable; oats undersown with grass clover preceded the rst-year grass clover. Number of tests: ve tests per plot by operators 1 and 2. Two plots per crop type where crop vigour was normal, one plot per crop type where growth was impaired. Timing: June and October 2006. Associated tests: In situ measurements of dry bulk density from 0- to 5-cm and 5- to 10-cm depth made in October 2006 using intact cores of volume 209 cm3. Macroporosity was also estimated in these cores from the volume of pores air-lled at )10 kPa. See Taylor et al. (2006) for further site details.

condence and stimulate users to consult the key to conrm their estimates. It is best to assess spadefuls together with the trainees for some time before allowing them to work alone. Each test takes between 5 and 15 min to perform depending on the compactness, moisture content and stone content of the soil and on the need to take photographs. Assessments in ploughed seedbeds can be done even more quickly. The results for comparison of operators at locations 1 and 2 are shown in Table 2. Data were checked to conrm normality of distribution. The average values for each eld varied between operators, with the range for location 2 (0.6) being double that for location 1. From the analysis of variance (Table 3), the differences in Sq values between locations were much greater than the differences between operators. All operators identied the poorer structure and the greater variability of structure at location 2 as shown from standard errors and ranges (Table 2). Although the differences between operators were not signicant (Table 3), the signicant interaction between

Table 3 Results of two-way analysis of variance for operators and locations Degrees of freedom 3 1 3 72 79 Sums of squares 0.724 7.117 3.433 27.03 38.30 Mean square 0.241 7.117 1.442 0.375 Variance ratio 0.64 18.96 3.05 F probability 0.59 <0.001 0.034

Source

Results
Ease of use and reproducibility
It takes about 1 h to teach this method to a technician or student with some scientic experience. The photographs which clearly show the gradation between Sq categories instil
Table 2 Comparison of Sq values for operators working at the same time at two elds at Craibstone in moist soil Location 1

Operator Location Interaction Error Total

Range Operator 1 2 3 4 Average over operators 1 2 3 4 Average over operators Location Location operator Mean 2.2 2 1.7 1.9 1.9 2.3 2.6 3.0 2.3 2.5 0.27 0.54 SE 0.17 0 0.20 0.08 0.07 0.27 0.26 0.25 0.15 0.12 Minimum 1 2 1 1.8 1 1.5 2 2 1.68 1.5 Maximum 2.8 2 3 2.6 3 4 4 4.5 3 4.5

Least signicant difference (P = 0.05)

Location 1 is a grass clover eld which is the site of a former ploughing date experiment, and location 2 is a eld under barley stubble after restoration from damage during builders works. SE = standard errors determined on the basis of plot averages.

2007 The Authors. Journal compilation 2007 British Society of Soil Science, Soil Use and Management, 23, 329337

334 B. C. Ball et al.

location and operator shows that the Sq value given by individual operators is inuenced by site characteristics. From the LSDs for location operator (Table 2), the differences between locations identied by operators 1 and 4 were not signicant. operator 4 gave the most favourable scores at both sites. Operator 2 gave results closest to the overall average even though his scores for location 1 were mostly uniform. In the ley-arable rotations experiment at Tulloch, operators working separately at different dates and soil moisture conditions gave signicantly different results. Operator 1 scored the soil structure less favourably than operator 2 (results not shown). Nevertheless, both operators identied the better structure under third-year grass than under oats undersown. The inuence of poor drainage in impairing structure was also detected in the rst- and fourth-year grass plots. rst- and fourth-year grass plots. The two lowest Sq values on the oats undersown, 3 and 3.5, were found in an area of a few square metres where the crop was yellower and less vigorous than elsewhere. At the Foulum site, topsoil structure quality decreased signicantly with decreasing tillage intensity from Sq1.1 for ploughed soil to Sq3.1 for direct drilled soil (Table 5). For the H810 treatment a loose, well-structured tilled layer (0 10 cm) overlaid a dense, poorly structured layer (1020 cm). Soil physical properties measured in the eld also indicated a poorer structure, viz. denser and stronger soil under D and H810 than under P. The D treatment displayed signicantly poorer growth in comparison with the P and H810 treatments at the time of Sq measurements.

Discussion
Sensitivity to differences in soil management
At Craibstone, poorer structure and greater variability of structure were found at location 2, the soil affected by builders work (Table 2). In the long-term organic ley-arable rotations experiment at Tulloch, a better structure was identied under third-year grass than under oats undersown (Table 4). There was a good correspondence between bulk density and macroporosity measurements (Table 4). It was also found that poor drainage had impaired structure in the

Ease of use and reproducibility

The test is easily used as was shown by the speed with which trainees appeared to understand it, mainly because of the photographs dening each Sq class. The test appears to be more quickly learned than the original Peerlkamp test because the descriptions of classes are more detailed and results are weighted for layering in terms of actual layer depths. This probably makes the test slightly slower to use than the Peerlkamp. However, the test is much faster than

Table 4 Comparisons of Sq values and soil characteristics for a grass and oat crop in two poorly drained grass plots of the Tulloch rotation experiment Sq score Range Crop plots Third-year grass clover Oats undersown Poorly drained grass plots Mean 1.6a 2.2b 2.3b SE 0.10 0.11 0.10 Minimum 1.3 1.7 2.0 Maximum 2.7 3.5 3.5 Bulk density (010 cm g cm)3) 1.16a 1.22b Macroporosity (010 cm m3 100 m)3) 23a 21b

SE = standard errors determined on the basis of plot averages. Within a column, values followed by the same letter do not differ signicantly.

Table 5 Comparisons of Sq values and soil and crop characteristics for three different tillage systems at the Foulum tillage experiment Sq score Range Treatment P H8-10 D Mean 1.1a 2.1b 3.1c SE 0.04 0.04 0.10 Minimum 1.0 1.6 2.0 Maximum 1.6 2.6 4.0 Bulk density (515 cm g cm)3) 1.25a 1.39b 1.36ab Penetration resistance (017 cm MPa) 0.4a 1.2b 1.4c Relative vegetation index 21.4b 17.4ab 9.2a

SE = standard errors determined on the basis of plot averages. Within a column, values followed by the same letter do not differ signicantly.

2007 The Authors. Journal compilation 2007 British Society of Soil Science, Soil Use and Management, 23, 329337

Field assessment of soil structure 335

either the visual soil assessment spade analysis of Denmark (Munkholm, 2000) or the soil quality scoring procedure of Scotland (Ball & Douglas, 2003). Nevertheless, the method is still subjective in that the soil has to be broken up and comments made on the resultant soil fragments. The visual soil assessment technique of Shepherd (2000) has a major advantage in that it involves an objective drop-shatter test with assessment of the ordered fragments from photographs. Nevertheless, this latter technique is still fairly time consuming and takes less account of the inherent soil structure in terms of rooting, layering and macropore continuity. At Craibstone operators varied in their ability to identify the poorer and more variable structure at location 2 (Tables 1 and 2). The reason for the variability between operators was due to differences in sampling regime and in the interpretation of the method, particularly the denition of difculty of breakdown of the slice of soil. Operators were free to sample wherever they chose in the eld, as would happen in practice. Expert operators did not appear to give different assessments to non-experts. When using his test, Shepherd (2000) found good agreement between expert and nonexpert groups from a wide range of disciplines. As the Sq value given by individual operators was inuenced by site characteristics at Craibstone, replication is probably better achieved by using more operators than by more measurements by a single operator. Expert operators were not necessarily better than non-experts, although they tended to observe more details of the soil structure. At Tulloch, the difference between operators (operator 1 nding conditions less favourable than operator 2) may have arisen from the difference in soil moisture conditions at sampling and from difference in the time of sampling. The soil was unusually dry when operator 2 was sampling. This dryness may have reduced cohesion so that aggregates broke down readily and producing an intact sample of soil was also difcult. In addition, operator 1 included some areas where growth of the undersown oat crop was poor, thereby increasing the overall average score.

oats undersown than in the third-year grass, contrary to the results found here. However, this may be due to various factors including the treatments being located on different plots and the low replication of the Ball & Douglass (2003) test. In the assessments in France, most spade methods gave detailed descriptions of the structure within different layers, often with accompanying photographs, but with only single or duplicate replication (Boizard et al., 2005). The original Peerlkamp test gave only a general description with no indication of layering, albeit with 10-fold replication. The test represents a compromise by providing information on layering of structure and replication. Thus, although spade tests revealed treatment differences, it is important to establish their signicance and to cover spatial variability by adequate replication.

Indicator and management thresholds

To make the method useful in practice, it is important to assess critical threshold values for Sq and to give recommendations on how to manage soil structural quality (soil physical fertility). Grading of soils by indicators across soil types, climates and cropping systems is difcult (Schjnning et al., 2004). For example, there are problems with varying spatio-temporal scales and with demonstrating causal relationships between soil quality and ecosystem functions. However, based on experiences from arable and ley-arable cropping systems on medium- to light-textured temperate soils, the threshold values given in Table 6 for Sq are proposed. The threshold values may be different for other soils, climates, cropping systems and production intensities. The test has mainly been applied to soils of clay loam texture or coarser. A modied key may be required for ner textured soils where pores and cracks and their continuity contribute more to structural function than aggregates. Sq values between 1 and 2 indicate a soil with good structural quality in relation to productivity and no changes in management are needed. Sq values between 2 and 3 indicate a soil with fair structural quality that can be improved. In many cases there is no need for sudden changes but to consider changes, for example, in crop rotation, residue management, tillage and trafc that might be expected to yield an improvement in the medium to long term. Sq values over 3 indicate poor soil structural quality and generally marked changes in management are needed to improve the quality for sustained high productivity.
Table 6 Threshold Sq values for sustained agricultural productivity Soil structural quality Good Fair Poor

Sensitivity to detect differences in soil management

At Tulloch, Craibstone and Foulum, the differences in Sq score were generally in correspondence with differences in other soil physical properties and crop growth. At Foulum, poor growth at the time of sampling was found for direct drilled soil where the mean Sq was 3.1. At Tulloch, poor growth was detected where Sq was 3 or 3.5. The inuence of poor drainage in impairing structure detected in the grass clover plots (Table 4) shows the importance of identifying the correct location to sample. Areas of poor growth or wet zones should be sampled and assessed separately from the main eld. At Tulloch, the soil quality scoring procedure of Ball & Douglas (2003) was used in the same treatments in 2001. This technique revealed that structure was better in the

Sq score 12 23 35

Management needs No changes needed Long-term improvements Short-term improvements

2007 The Authors. Journal compilation 2007 British Society of Soil Science, Soil Use and Management, 23, 329337

336 B. C. Ball et al.

A key challenge is to assess management thresholds for soil structural quality for a given soil type and climate from responses to management options. For instance, the results from the Foulum tillage experiment indicate that direct drilling is problematic under the current combination of soil type (sandy loam), climate (moist and cool) and cropping system (arable) (i.e. high Sq scores and poor crop growth). For this soil type, it appears that shallow tillage is the minimum needed to produce a desirable soil structural quality. That is, shallow tillage can be considered as a management threshold for the Foulum soil. Soil resilience, a dynamic property, may be assessed as a snapshot from visual assessment of soil structure (structural form) (Litterick, 2005). Although the absolute value of structure is an indication of resilience, it is liable to uctuate with operators. The variability of structure, expressed as standard error (Table 2), is less subject to change with operator and may be a more sensitive indicator of eld resilience than the mean Sq value. grateful to our colleagues Alex Sinclair, Ian Dickson, Colin Crawford, Derek Simpson, SAC and Stig T. Rasmussen, Bodil B. Christensen, Holger Bak, Arhus University and Amanda Liesch, University of Wisconsin River Falls for their help in developing or assessing this test.

References
Ball, B.C. & Douglas, J.T. 2003. A simple procedure for assessing soil structural, rooting and surface conditions. Soil Use and Management, 19, 5056. Ball, B.C., Watson, C.A. & Baddeley, J. 2007. Soil physical fertility, soil structure and rooting conditions after ploughing organically managed grass clover swards. Soil Use and Management, 23, 20 27. Beste, A. 2003. Erweiterte Spatendiagnose: Weiterentwicklung einer Feltmetode zur Bodenbeurteilung, PhD Thesis. Justus-Liebig-Universitat, Giessen, Germany. Boizard, H., Batey, T., McKenzie, D.C., Richard, G., RogerEstrade, J., Ball, B.C., Bradley, I., Cattle, S., Hasinger, G., Munk holm, L.J., Niewergelt, J., Peigne, J. & Shepherd, G. 2005. Field meeting Visual soil structure assessment detailed report. INRA, France. Available at: http://iworx5.webxtra.net/istroorg/download/WG%20Visual%20Soil%20Structure%20Assessment_Field% 20meeting.pdf [last accessed 10 July 2007]. Childs, E.C., Collis-George, N. & Holmes, J.W. 1957. Permeability measurements in the eld as an assessment of anisotropy and structure development. Journal of Soil Science, 8, 2741. Currie, J.A. 1965. Diffusion within soil microstructure: a structural parameter for soils. Journal of Soil Science, 16, 279289. De Boodt, M., De Leenheer, L., Frese, H., Low, A.J. & Peerlkamp, P.K. 1967. West-European methods for soil structure determination. State Faculty of Agricultural Sciences, Ghent, Belgium. Hasinger, G., Nievergelt, J., Petrasek, M. & Weisskopf, P. 2004. Observer et evaluer la structure du sol. Cahiers de la FAL 50, FAL, Switzerland. Kay, B.D. & Angers, D.A. 2001. Soil structure. In: Handbook of soil science (ed. M.E. Sumner), pp. 229276. CRC Press, Boca Raton, FL. Lawrie, J., Murphy, B., Packer, I.J. & Harte, A.J. 2000. Soils and sustainable farming systems. In: Soils their properties and management (eds P.E.V. Charman & B.W. Murphy), pp. 133149. Oxford University Press, Melbourne. Litterick, A. 2005. Indicators of soil resilience for Scotland and Northern Ireland. Final Report for SNIFFER Project LQ06, SNIFFER, Greenside Place, Edinburgh, UK. Low, A.J. 1954. A study of soil structure in the eld and in the laboratory. Journal of Soil Science, 5, 5774. McKenzie, D.C. 2001. Rapid assessment of soil compaction damage. I. The SOILpak score, a semi-quantitative measure of soil structural form. Australian Journal of Soil Research, 39, 117 125. Munkholm, L.J. 2000. The spade analysis a modication of the qualitative spade diagnosis for scientic use. DIAS Report no. 28. Plant production, Danish Institute of Agricultural Sciences, Tjele, Denmark. NSRI 2001. A guide to better soil structure. National Soil Resources Institute, Craneld University, Silsoe, UK.

Conclusions
The use of a visual key with well-dened descriptions of criteria for each category and inclusion of layering constitute the main improvements over the original Peerlkamp test. Operators scores were inuenced to an extent by differences between elds so that replication is best achieved by using more operators than by a single operator making more measurements. Alternatively, the scores of a single operator should be occasionally veried by another to eliminate bias. Differences in management either as crop or tillage type were revealed by the test or related to differences in soil physical properties and crop growth. The test may be able to provide thresholds of soil management and indicators of soil physical fertility and soil resilience. The test offers immediate structural assessment in the eld and is useful for assessing tillage need or drainage status. Sampling of dry soils should be avoided as the difculty in sample extraction and break-up may underestimate the soil score. Scores from markedly different moisture states should be kept separate. Sampling locations should be chosen according to management requirement, for example to assess areas of poor growth separately from the main eld. The holistic, qualitative eld approach should be viewed as complementary to the quantitative laboratory approach to soil structure assessment. There are advantages in using both approaches. Visual soil structure assessment can be used to indicate where to sample and what measurements are likely to be worthwhile.

Acknowledgements
This work was sponsored jointly by the Scottish Executive Environment and Rural Affairs Department and by the Danish Ministry of Food, Agriculture and Fisheries. We are

2007 The Authors. Journal compilation 2007 British Society of Soil Science, Soil Use and Management, 23, 329337

Field assessment of soil structure 337

Olesen, J.E. & Munkholm, L.J. 2007. Subsoil loosening in a crop rotation for organic farming eliminated plough pan with mixed effects on crop yield. Soil and Tillage Research, 94, 376385. Peerlkamp, P.K. 1959. A visual method of soil structure evaluation. Meded. v.d. Landbouwhogeschool en Opzoekingsstations van de Staat te Gent. XXIV No. 24, pp. 216221. Ringrose-Voase, A.J. 1991. Micromorphology of soil structure: description, quantication, application. Australian Journal of Soil Research, 29, 777813. Roger-Estrade, J., Richard, G., Caneill, J., Boizard, H., Coquet, Y., Defossez, P. & Manichon, H. 2004. Morphological characterisation of soil structure in tilled elds: from a diagnosis method to the modelling of structural changes over time. Soil and Tillage Research, 79, 3349.

Schjnning, P., Elmholt, S. & Christensen, B.T. 2004. Managing soil quality: challenges in modern agriculture. CABI Publishing, Wallingford, UK. Shepherd, T. G. 2000. Visual soil assessment. Volume 1. Field guide for cropping and pastoral grazing on at to rolling country. Horizons.mw & Landcare Research, Palmerston North, New Zealand. Taylor, B.R., Younie, D., Matheson, S., Coutts, M., Mayer, C., Watson, C.A. & Walker, R.L. 2006. Output and sustainability of organic ley arable crop rotations at two sites in northern Scotland. Journal of Agricultural Science, 144, 435447. Williams, R.J.B. & Cooke, G.W. 1961. Some effects of farmyard manure and of grass residues on soil structure. Soil Science, 92, 30. Young, I.M. & Crawford, J.W. 2004. Interactions and self-organization in the soilmicrobe complex. Science, 304, 16341637.

2007 The Authors. Journal compilation 2007 British Society of Soil Science, Soil Use and Management, 23, 329337