Soil & Tillage Research 140 (2014) 29–39

Changes in soil cone resistance due to cotton picker traffic duringharvest on Australian cotton soils

M.V. Braunack *, D.B. Johnston

CSIRO, Plant Industry and the Cotton Catchment Communities CRC, Locked Bag 59, Narrabri, NSW 2390, Australia

A R T I C L E I N F O

Article history:

Received 14 August 2013

Received in revised form 30 January 2014

Accepted 4 February 2014

Keywords:

Soil compaction

Irrigation

Rainfed

Simulation

OZCOT

Cone penetrometer

A B S T R A C T

Australian cotton growers have rapidly adopted new picking technology of round module balers on dual

tyres. These machines weigh twice that of previous basket pickers, usually on single tyres, being

replaced. This raises some concern about implications for subsoil compaction (>0.4 m depth) from

harvest traffic. The objective of this study was to quantify changes in soil strength due to picker traffic

during harvest. Measurements of soil strength were undertaken before and after traffic by new round

module baler (32 t) and current basket (16 t) pickers during one cotton picking season. Soil cone

resistance, water content and plastic limit (PL) were measured in the upper 0.6 m depth at eight sites

during normal picking operations. Results showed that soil strength increased after traffic of either

picker compared with before traffic and increases were detected to a depth of 0.6 m. Despite differences

in soils and profile water content, the change in strength was similar under the round module baler and

the basket pickers. A zone of greater soil strength (3 MPa) occurred closer to the soil surface under the

round module baler (0.3 m) compared with the basket picker (0.4 m). Zones of increased soil strength

were also detected at 0.6 m depth under both pickers indicating possible subsoil compaction. The OZCOT

cotton simulation model was used to determine the frequency at which the soil profile was wetter than

the PL for both irrigated and dryland systems. Simulations showed that the soil profile could be expected

to be wetter than the PL 75% and 14% of the time under irrigated and dryland systems, respectively, at

harvest over the period from 1960 to 2012. This indicates that cotton picking in irrigated systems has a

high probability of occurring when the soil is susceptible to compaction, with the risk of subsoil

compaction greater with the round module baler.

� 2014 Elsevier B.V. All rights reserved.

Contents lists available at ScienceDirect

Soil & Tillage Research

jou r nal h o mep age: w ww.els evier . co m/lo c ate /s t i l l

1. Introduction

It is well known that crop yield is reduced by soil compaction(Soane and van Ouwerkerk, 1994) and subsoil compaction (>0.4 mdepth) is of concern due to the increase in size and weight ofagricultural equipment (Horn et al., 2000). Cone penetrometers areoften used to assess soil strength in relation to soil compaction androot growth (Bengough et al., 2001). Resistance to penetration isaffected by soil type; soil texture, organic matter content and claymineralogy (Stitt et al., 1982), while within a soil type it is affectedby soil water content, bulk density and structure. Soil resistancegreater than 2 MPa is considered to limit root growth (Hamza andAnderson, 2005). However, soils with a resistance less than 2 MPahave been shown to reduce cotton yield (Carter and Tavernetti,

* Corresponding author. Tel.: +61 2 6799 1500; fax: +61 2 6793 1186.

E-mail addresses: michael.braunack@csiro.au (M.V. Braunack),

david.b.johnston@csiro.au (D.B. Johnston).

http://dx.doi.org/10.1016/j.still.2014.02.007

0167-1987/� 2014 Elsevier B.V. All rights reserved.

1968) while root growth in repacked soil columns ceased at2.5 MPa (Rosolem et al., 2008). Recently Kulkarni et al. (2010)indicated that although cotton growth was affected by soilresistance as low as 1.6 MPa (measured range 1.6–2.9) on a loamsoil in Arkansas, there was no yield penalty.

The plastic limit is an arbitrary measure of the soil watercontent where the soil changes from brittle and fracturing tobecoming plastic and ductile. The soil plastic limit (PL) has beenused to define the point where the soil is susceptible todegradation from tillage operations or harvesting traffic and thatdamage will be limited when the soil is drier than the PL (Kirby,1990, 1991). Soil degradation due to picker traffic will beinfluenced by picker and soil parameters; whether the pickerhas tyres or tracks, the total load of the picker, the contact area ofthe tyres or tracks and the speed of travel while soil factors includethe soil strength which is a function of water content, texture andstructure (Kirby, 1988). As the clay fraction and soil organic matterincrease, so will the PL as these parameters affect soil watercontent. Soil beneath a tyre or track is subject to both compression

M.V. Braunack, D.B. Johnston / Soil & Tillage Research 140 (2014) 29–3930

and shear force and soil degradation is due to the soil response tothese stresses.

A difficulty in assessing the risk of soil degradation isidentifying when the soil is susceptible to compaction. By usingcrop simulation models and simulating the cropping system over along period it is possible to gain an idea of the frequency that a soilprofile may be wetter or drier than the PL at harvest. Augment thiswith the amount of rainfall in the five days prior to harvest and thedata provides information on the risk for compaction on anoperational basis (Littleboy et al., 1998).

To reduce the effect of machinery traffic on soil degradation,manufacturers have used dual tyres and tracks to reduce groundpressure. The cotton industry uses both rubber tyres and tracks ontractors for planting and in-crop operations and rubber tyres onpickers. Soil stress beneath two and four wheel drive cotton pickerswas similar and much greater than beneath a rubber trackedtractor (Kirby et al., 1991). With respect to dual versus singlewheels Kirby and Blunden (1993) have demonstrated thatcompaction near the soil surface is dependent on ground contactpressure while that at depth is dependent on total axle load. Themeasurement of stress transmission/distribution through field soilprofiles also reflects this (Lamande and Schjønning, 2011a,b). Inreality as the weight of equipment increases, the tyre size shouldincrease to maintain ground contact pressure and minimisesurface compaction and the number of axles should increase toshare the increased weight to minimise subsoil compaction(Hakansson and Reeder, 1994).

Research was undertaken during the period from 1981 to 2002in response to planting and harvesting on wet soils and theamelioration of soil degradation (Daniells, 1989; McGarry andChan, 1984; Stewart et al., 2002). Sullivan and Montgomery (1998)concluded that subsoil compaction in cotton fields was due to in-field traffic and not clay translocation. Although it is claimed thatVertosol soils (Isbell, 1996) are self repairing due to the shrink–swell behaviour, it may take in the order of 11 wet/dry cycles torepair structural degradation as blocks of compressed soil remainbetween the large cracks in the profile, and especially in subsoildue to overburden (Sarmah et al., 1996). This raises an issue as towhether cotton growers still suffer from subsoil degradation frompast years when operations were undertaken on wet soils.

Table 1Details of sites, soil type and equipment measured.

Site Soil Equipment

Auscott (1) (Narrabri) Vertosola (grey cracking clay)

Light clay (35% clay, 0–0.1 m)

to heavy clay (>50% clay, 0.1–1.2 m)

Round module (du

Auscott (2) (Narrabri) Vertosol (grey cracking clay)

As above

Round module (du

Hillston (1) Chromosol (red brown clay)

Silty-clay (35% clay, >25% silt,

0–0.1 m)

to Clay (45% clay, 0.1–0.9 m)

Round module (du

Hillston (2) Chromosol (red brown clay)

(As above)

Basket (single tyre

Boggabilla Vertosol (black cracking clay)

Medium clay (45% clay, 0–0.9 m)

Basket (single tyre

Bourke Kandosol (red earths)

Loamy-clay (30% clay, 0–1.0 m)

Basket (single tyre

Myall Vale Vertosol (grey cracking clay)

Medium clay (40% clay, 0–0.1 m)

to heavy clay (>50% clay, 0.1–1.0 m)

Basket (dual tyres)

St George Sodosol (solodic soils)

Silty-Loam (25%, >25% silt, 0–0.1)

to Clay-loam (30%, clay, 0.1–1.2 m)

Basket (single tyre

PL, plastic limit measured (%); WP, wilting point (%, 15 bar); DUL, drained upper limit

a Australian Soil classification (Isbell, 1996). Field texture & approximate clay or silt

Growers are adopting new harvesting technology of roundmodule balers. These pickers build a round module on the go in anintegrated baling mechanism and drop the wrapped module whilebuilding another, compared to a basket picker which collectscotton seed in a basket on the go and then transfers this to amodule builder on the headland. The new pickers offer severaladvantages over the basket picker: a reduction of in-field labour,greater picking efficiency (Willcutt et al., 2009) and less equipmentto clean down and move between locations – all reducing the costof production. These pickers are considerably larger and weighmore than current basket pickers and pose a risk in generatingsubsoil compaction, especially if wet soil conditions occur atharvest or the soil has not dried sufficiently at depth after the lastirrigation or significant rainfall. Growers need to be proactive indeveloping strategies to minimise the risk of subsoil compactionwhich is difficult to ameliorate and can limit crop performance.

The objective of this work is to quantify changes in soil strengthdue to cotton picker traffic on Australian cotton soils and identifythe potential risk of subsoil compaction using long-term cropsimulation modelling.

2. Materials and methods

2.1. Field measurements

Eight typical cotton fields were selected during the 2011 cottonharvest covering a range of soil types and soil moisture conditionsat harvest (Table 1). Soil cone resistance was measured, to depth ofup to 0.6 m at intervals of 0.02 m, with a recording penetrometerinserted at a constant rate (ASAE, 1986) (12.3 mm dia. cone, 308included angle) across twelve or eight furrows (round modulebaler and basket picker, respectively) and crop rows before andafter the passage of a fully laden cotton picker operating in the fieldat the time. It was not possible to insert the cone penetrometer to0.6 m at all sites due to dry soil or traffic pre-history of the field. Thepenetrometer was mounted in a metal frame and inserted at aconstant rate by a battery driven ram; this eliminated operatorfatigue and ensured consistent strength data. Measurement alwaysstarted and finished in a non-traffic furrow across twelve or eightrows with strength being recorded at 20 mm depth intervals; the

Weight (t) Profile (0–0.6 m) soil water (%)

Empty Full Front

Axle

Pre- &

post-traffic

PL WP DUL

al tyres) + trailer 38 47 21 32 25 22 40

al tyres) + trailer 38 47 21 24 22 22 40

al tyres) 32 34 21 19 17 13 36

s) 17 20 14 23 22 13 36

s) 16 18 12 22 19 18 44

s) 15 17 13 20 21 19 35

19 20 16 23 21 19 40

s) 15 18 13 17 19 15 26

(%) (WP & DUL from APSoil database).

content and profile depth are given in parenthesis.

M.V. Braunack, D.B. Johnston / Soil & Tillage Research 140 (2014) 29–39 31

distance between readings being 0.25 m. The mean for eachposition of measurement (before and after traffic) was used forgenerating profile contour maps, with the mean of all positionsbeing used for before and after traffic analyses. Soil samples werecollected from a furrow, a crop row and equidistant between thetwo, at the same time from 0–0.1, 0.1–0.2, 0.2–0.3, 0.3–0.4, 0.4–0.5to 0.5–0.6 m depths for gravimetric water content and assessmentof plastic limit (PL, Australian Standards Association, 1995). Twotransects five metres apart were measured 20 m in from the taildrain end of the field perpendicular to the direction of picker travel.All penetrometer readings, for all sites, were corrected for surfacegeometry using the height difference between the top of the croprow and furrow base and to a common soil water content (usingthe average profile soil water content at each site after Busscheret al., 1997). Soil resistance data were contoured using SigmaPlot12.0 (Systat Software, 2011). Picker parameters collected includedempty and loaded weights, tyre size and inflation pressure, andtrack width.

In order to determine significant changes in soil cone index dueto picker traffic the soil strength data was analysed as a multiplesplit plot (site/field position/depth) with soil water content as acovariate using Genstat v14 (VSN International, 2011). Due to thenature of the field measurements there was a wide range in pre-history and profile moistures at each site, so a multiple regressionwas fitted to soil strength after picking to account for pre-historysoil strength, soil moisture, picker weights and profile depth. Theequation was of the form:

SqrtðAHÞ ¼ 0:662 þ 0:01462 � D þ 0:456 � BH þ 0:000453

� Pw � 0:00787 � W � 0:00673 � ðBH � PwÞ

� 0:00409 � ðD � BHÞ (1)

where AH is the after harvest soil strength (MPa), D is depth (m),BH is the before harvest soil strength (MPa), W is the weight(tonnes) of the picker and Pw is the profile moisture content. Thesite by depth interaction was not included in the model as it wasnot a significant factor in regression. The regression accounted for64% of the variability in the data. Eq. (1) integrates many of thefactors associated with measured changes in soil strength afterpicking in these measurements; it indicates that the soil strengthbefore picking (reflecting pre-history of the field) had the greatesteffect on soil strength after picking.

It is not possible to make direct comparisons between roundmodule baler and basket pickers as both were not operating in thefield at the same time. The sites, soils, pickers, soil water content atthe time of traffic and the corresponding soil plastic limit (PL),wilting point (WP) and drained upper limit (DUL) are given inTable 1. The round module balers were twice the weight of thebasket pickers depending on configuration (Table 1).

2.2. Simulation

The soil compaction model SoilFlex (Keller et al., 2007) wasused to simulate vertical stress distribution in the soil under thewheels of the round module builder and basket picker assuming anelliptical contact area; this model has the advantage of using tyresize (520/85R42 & 20.8-38 for round module & basket picker), andinflation pressure, (270 kPa for both) as measured inputs.

The cotton crop simulation model OZCOT (Hearn, 1994) wasused to determine the frequency that the soil profile water content(0–0.6 m depth) at harvest, for the seven sites at which soilstrength was measured, was higher or lower than the measured PLto determine the risk of soil compaction at harvest (which canoccur up to three weeks after defoliation at 60% open bolls).Rainfall may occur in this intervening period which will re-wet the

surface soil. OZCOT uses the Ritchie water balance model (Ritchie,1972) and has been widely calibrated and validated across cottonsoils and growing regions in Australia (Cull et al., 1981; Hearn andConstable, 1981, 1984; Hearn, 1994).

Simulations were conducted for both irrigated and drylandsystems on the same soil using long-term weather data (1960–2012) for the seven sites that the PL was measured. For thesimulations the parameters used for all sites were cultivar Sicot71B planted at 0.05 m depth in 1 m rows at 12 plants per metrewith 200 kg N/ha applied 60 days prior to planting on 16 October,and full profile plant available water of 233, 176, 139, 122 and130 mm for the Auscott/Myall Vale, Boggabilla, Bourke, Hillstonand St George sites, respectively. The crop was first irrigated 40days after planting with no irrigation applied after 60% open bolls;otherwise the model applied irrigation as required. The model wasbenchmarked for the Myall Vale site where the actual plant andirrigation dates were known; the regression between the observedand simulated profile water (0–0.6 m depth) resulted in RMSE of0.04 an indication of model performance. At the same site thesimulated long-term profile water at harvest was 26.4% withthe benchmarked profile soil water being 22.5% compared with themeasured profile water of 22.7%. Differences between themodelled and measured soil water reflect different actual plantingdate compared with the single plant date used in the long-termsimulations. This provides confidence that the long-term simula-tion of profile water is a good estimation of profile soil water at thetime of cotton picking.

An additional simulation was also undertaken for the MyallVale site to assess whether the timing of the last irrigation could beused as a strategy to minimise the effect of picking traffic. Theinputs were as described above with the exception that the lastirrigation was applied at 10%, 20%, 40% open bolls compared with60% open bolls with harvest occurring between 10 and 20 daysafter 60% open bolls.

3. Results

Although only one season is considered, the conditions weretypical of those normally experienced at picking; the exceptionwas Auscott (1) which was wetter than usual due to rainfall prior topicking. The soil cone resistance was measured at one point intime; immediately before and after picker traffic (to ensureminimal change in profile soil moisture) to quantify the effect ofpicker traffic on changes in profile soil strength. After harvestingthe cotton the field is slashed and a root cutting operation(0–0.1 m) is undertaken to prevent cotton from re-growing. This isfollowed by cultivation in preparation for a rotation crop; theseoperations will remove surface (0–0.1 m) compaction, whilesubsoil (0.2–0.4 m) compaction will persist. The concern is thatan increase in soil strength in the subsoil will not be remediatedand persist over time affecting future productivity.

3.1. Soil strength

Busscher et al. (1997) correction of soil strength to a commonsoil water value does not result in a significant reduction incoefficient of variation (CV) (CV before correction 29.1% vs 29.0%after correction) so the original values of soil strength and profilewater contents were used. Contour maps of soil strength provide avisual image of the effect of picking traffic in relation to the croprow and depth down the soil profile. The before picking strengthcontours reflect the pre-history from earlier traffic across the field.Changes in soil strength occurred after traffic by both cottonpickers (Figs 1a and 2a). The number of vertical zones of soilstrength less than 2 MPa was reduced after round module balertraffic (Auscott (1) site, Fig. 1b). Zones of greater soil strength have

Fig. 1. Soil strength (MPa) profiles (a) before and (b) after traffic and (c) change in profile strength by a round module baler cotton picker at the Auscott (1) site. (WT is the

wheel tracks of the picker).

M.V. Braunack, D.B. Johnston / Soil & Tillage Research 140 (2014) 29–3932

resulted beneath and between the dual wheels of the roundmodule baler (Fig. 1b) with soil strength of 3 MPa occurring closerto the soil surface after traffic (0.1 m), compared with before traffic(0.2 m) (Fig. 1a). Columns of high soil strength developed under thewheels, with the greatest change occurring beneath and betweenthe dual tyres (Fig. 1c). The pre-existing zones of high strength atdepth (4 MPa, at 0.6–0.7 m) appear to have coalesced after trafficand become more contiguous than pre-traffic (Fig. 1b and c).

The response in soil strength under a basket picker (Boggabillasite) was different due to soil texture (lower PL) and soil watercontent at the time of traffic. The zone of soil strength less than2 MPa did not greatly change after basket picker traffic (Fig. 2a andb).There was development of circular zones of greater strengthdirectly beneath wheel tracks (single tyres) (Fig. 2c) and columnsof higher strength at depth (3 MPa lines more vertical) (Fig. 2b andc). Also, a zone of high soil strength (4 MPa) has appeared beneathand to the side of the wheel tracks (Fig. 2c).

The SoilFlex simulation indicated that vertical stress (80–100 kPa) under the round module builder occurred at 0.3 m depthin the profile compared with 0.2 m under the basket picker (Fig. 3).Vertical stress extended to 0.8 m depth under the tyres of theround module builder compared with 0.6 m under the basketpicker (Fig. 3). There was no indication of an interaction in stressbetween the dual tyres of the round module builder (Fig. 3).

The ANOVA in Table 2 for soil cone resistance shows significanteffects of site (P < 0.001), depth (P < 0.001), site � depth(P < 0.001), position � depth (P < 0.05), treatment (P < 0.001),site � treatment (P < 0.01) and site � position � treatment(P < 0.001). The Auscott (2) site had the greatest soil strength(Fig. 4) with all other sites being similar and the interaction withtreatment was due to greater increase in soil strength after pickingat the Auscott (2) and Myall Vale sites compared with no increaseat Auscott (1) and Bourke. Soil strength increased with depth at allsites with low strength in the top 0.1 m of the hill (crop row)compared with the non-wheel track and the wheel track (Table 3).The Auscott (2) site had a greater increase in soil strength afterpicking in the wheel track compared with all other sites, especiallyBourke (data not shown).

Change in profile soil strength was generally greater in thesurface 0.1 m and varied with depth at all sites measured (Fig. 4).This reflected the profile water being higher than the soil plasticlimit at the time of picking (Fig. 5), with the exception of the 0.1depth at Auscott (1), Hillston (2) and Bourke and the whole soilprofile at St George. Where the profile water and plastic limit weresimilar, the change in soil strength was less (Figs. 4 and 5c–e).Where the profile water was greater than the plastic limit to depth,the change in soil strength occurs to depth (Figs. 4 and 5b and f).The change in profile soil strength also varied between and under

Fig. 2. Soil strength (MPa) profiles (a) before and (b) after traffic and (c) change in profile strength by a basket picker at the Boggabilla site. (WT is the wheel tracks of the

picker).

M.V. Braunack, D.B. Johnston / Soil & Tillage Research 140 (2014) 29–39 33

pickers at each site (Fig. 4), which reflects the variation in pre-history and difference between the plastic limit and thecorresponding profile water at the time of picking (Fig. 5).

3.2. Simulation

Simulation of profile soil water content at picking indicates thatthe profile was above the PL in irrigated systems in 75% of years(Fig. 6). There were two exceptions, Boggabilla and Bourke, whereon average the profile was drier than the PL at picking (Fig. 6b andc). In contrast, under dryland systems the soil profile was lowerthan the PL at picking in 14% of years (Fig. 6). There were however,occasions when the soil profile was wetter than the PL underdryland systems (Fig. 6f) due to autumn rain.

For the Myall Vale site, simulating the last irrigation at 10% openbolls indicated that the profile may dry to less than the soil PL on62% more occasions (Fig. 7d), depending on climatic conditions,compared with 20%, 40% or 60% open bolls (Fig. 7a–c). This shouldbe put into perspective; when the last irrigation occurs at 10%, 20%and 40% open bolls the model continues to grow the crop until itreaches 60% open bolls when defoliation is applied with harvestoccurring 10–20 days later. The measured profile water content atharvest was 22.7% (g/g) compared with the simulated profile watercontent of 22.7% (g/g). The time between the last irrigation and

picking will be longer at 10% open bolls compared with 20% and40% open bolls which means there is opportunity for rainfall to re-wet the profile before picking. The corresponding rainfall shows aperiod of wet harvests during the late 1980’s and periods whenlittle or no rain occurred in the period between defoliation andpicking the profile rarely dried to or below the soil PL (Fig. 7). Therewere more occurrences of the profile drying to the PL when the lastirrigation was at 10% open bolls compared with 20%, 40% or 60%open bolls (Fig. 7).

Yield was decreased when the last irrigation was applied at 10%,20% and 40% open bolls compared with 60% open bolls (Table 4).On average over the period of simulation this corresponded to 3.5%,1.3% and 0.3% reduction in yield, respectively with a range of 0–30%reduction (Table 4). Irrigation occurred, on average, 30, 23 and 15days earlier (Table 4).

4. Discussion

This is the first study on the effect of picker traffic on soilconditions during a cotton harvest under Australian conditions.Although growers and picking contractors believe that pickingtraffic does not result in soil degradation if the soil at harvest is dry,this may be the case with respect to the soil surface; howeversubsoil wetter than the PL may be susceptible to degradation, as

Fig. 3. Simulated vertical stress beneath tyres of (a) round module builder and (b) basket picker (using the SoilFlex model).

M.V. Braunack, D.B. Johnston / Soil & Tillage Research 140 (2014) 29–3934

occurred at four sites in this study. The cotton production system ischanging with the adoption of round module balers by growerswhich is of some concern due to the potential for subsoilcompaction from increasingly heavier picking equipment.

Table 2ANOVA for changes in soil strength due to picker traffic. Variate, cone index (MPa);

Covariate: soil water content (%).

Source of variation d.f. s.s. m.s. F pr.

Rep stratum

Covariate 1 0.6241 0.6241

Rep.Site stratum

Site 7 66.0579 9.4368 0.013

Covariate 1 2.4572 2.4572 0.213

Residual 6 7.5873 1.2646

Rep.Site.Posn stratum

Posn 2 1.8567 0.9283 0.373

Site.Posn 14 9.8875 0.7063 0.657

Covariate 1 0.2325 0.2325 0.615

Residual 15 13.1985 0.8799

Rep.Site.Posn.Depth stratum

Depth 4 97.1911 24.2978 <0.001

Site.Depth 28 23.5324 0.8404 <0.001

Posn.Depth 8 3.8032 0.4754 0.024

Site.Posn.Depth 56 12.1616 0.2172 0.376

Covariate 1 0.0466 0.0466

Residual 95 19.2254 0.2024

Rep.Site.Posn.Depth.Trt stratum

Trt 1 2.9192 2.9192 <0.001

Site.Trt 7 2.9973 0.4282 0.004

Posn.Trt 2 0.4288 0.2144 0.203

Depth.Trt 4 0.1228 0.0307 0.920

Site.Posn.Trt 14 6.4971 0.4641 <0.001

Site.Depth.Trt 28 3.0649 0.1095 0.715

Posn.Depth.Trt 8 0.7340 0.0918 0.698

Site.Posn.Depth.Trt 56 7.4656 0.1333 0.479

Residual 120 15.9109 0.1326

Total 479 305.1979

Posn, furrow, crop row, wheel track; Trt, before and after traffic.

4.1. Change in soil strength

In considering the soil strength contours, there is an amount ofprior history recorded in the before traffic maps; as a result oftraffic from planting (Grainger et al., 1997) and in-crop weed andinsect control and from in-field traffic in previous crops. Thestrength contours were similar to the vertical stress measuredunder different tyres at two inflation pressures with varied loads(Lamande and Schjønning, 2011a,b) and those simulated usingequipment parameters in this study. The results indicate that soilstrength has changed at all sites after traffic compared to thatprior to traffic which agrees with the results of Braunack et al.(2006) for a sugarcane harvesting system. Soil strength increasedto greater than 3 MPa in the 0.3–0.6 m depth which will restrictroot growth if maintained at soil moisture greater than those atthe time of traffic (Hamza and Anderson, 2005; Raper, 2005;Taylor and Gardner, 1963). If root growth is restricted, water andnutrient uptake will be reduced with consequences for cropproductivity. Also, in irrigated systems, irrigation may need to beapplied more frequently as roots may be restricted to surfacelayers of the profile. Notwithstanding the difference in soils andsoil moisture at the time of traffic, the degree of change wasgreater in surface soil (0.1–0.3 m at five of the eight sites) underboth round module baler and the basket pickers. Change in soilstrength was generally greater in the surface 0.1 m and variedwith depth at all sites measured. This reflected the general profilewater being higher than the soil PL at the time of picking at all sitesexcept St George. The change in profile soil strength also variedbetween and under pickers at each site, which reflectsthe variation in pre-history and difference between the PL andthe profile water at the time of picking, and as indicated by theregression Eq. (1) where before harvest strength had the greatestinfluence on the outcome in strength change. The two greatestchanges were at Auscott (2) and Myall Vale; both sites had soilmoisture profiles above the PL. However Auscott (2) had highbackground soil strength and traffic from a heavy round baler andMyall Vale had a lighter basket picker. This seems to agree with

(b)(a)

(c) (d)

(e) (f)

(g) (h)

Fig. 4. Before and after picker traffic soil strength (MPa) profiles for sites (a) Auscott

(1), (b) Auscott (2), (c) Hillston (1), (d) Hillston (2), (e) Bourke, (f) Myall Vale, (g)

Boggabilla and (h) St George.

(a) (b)

(c) (d)

(e) (f)

(g) (h)

Fig. 5. Profile soil water content (%, g/g) at the time of traffic and the soil Plastic Limit

(%) for sites (a) Auscott (1), (b) Auscott (2), (c) Hillston (1), (d) Hillston (2), (e)

Bourke, (f) Myall Vale, (g) Boggabilla and (h) St George. (Scale differences indicate

difference between profile moisture at picking and plastic limit between sites).

M.V. Braunack, D.B. Johnston / Soil & Tillage Research 140 (2014) 29–39 35

the results of Lamande and Schjønning (2011a) with respect tovertical stress distribution, where stress at depth was less on wetsoils and higher on dry soils; similar vertical stress occurred underwet soil conditions while greater stress occurred on dry soil at thesame depth. When the profile water was above the PL the changein soil strength occurred through the depth of the profile, whichindicates the importance of timing of harvest so the soil is as closeas possible or below the PL to limit subsoil compaction. It has beensuggested that the PL may be a useful indicator to minimisecompaction on some soils (Kirby, 1990; Kirby and Blunden, 1993).

Table 3Soil strength (MPa) by depth for all row positions (averages across all sites and

treatments).

Depth (m) Hill Non-wheel track Wheel track Mean

0.1 1.63 1.88 2.12 1.88

0.2 2.21 2.47 2.43 2.37

0.3 2.59 2.79 2.83 2.74

0.4 2.98 2.95 2.93 2.96

0.5 3.26 3.13 3.14 3.18

lsd (5%) 0.31

For a range of Vertosols in Australia that the PL was near or slightlyabove the WP (Kirby, 1991) suggesting that soil compactionshould be minimal. The simulation result suggests that this maynot be a viable option due to potential yield loss. Where thesurface soil is dry it suggests that the soil strength is sufficient tosupport pickers, limiting compaction. However, when the profileis above the PL, compaction can occur to depth under both pickers.An option to limit subsoil compaction may be monitoring soilstrength in the 0–0.3 m layer to provide some indication as towhether the soil can support the picker. The fact that changes insoil strength can be detected to 0.5 m under both pickers indicatedthat consideration should be given to dry the profile to minimisecompaction at this depth. The overall outcome is the result ofinteractions between profile moisture in relation to the PL at thetime of traffic, the type of picker and the traffic pre-history of thefield; a difficult combination to manage. Degradation in thesurface soil is readily ameliorated by tillage at the appropriate soilmoisture (Daniells, 1989). Depending on the pre-history of thefield and the soil profile water at the time of picking relative to theplastic limit, this data shows that potential exists for compactionto occur at depth.

(a) (b)

(c) (d)

(e) (f)

(g)

Fig. 6. Simulated profile soil water content (%, g/g) at harvest over time for irrigated and dryland systems at (a) Auscott (Narrabri), (b) Boggabilla, (c) Bourke, (d) Myall Vale, (e)

Hillston (round module), (f) Hillston (basket) and (g) St George. The site plastic limit is indicated by the horizontal line.

M.V. Braunack, D.B. Johnston / Soil & Tillage Research 140 (2014) 29–3936

4.2. Position of wheel traffic

An increase in soil strength was observed under the row,between and alongside the dual wheels of the round module baleron soil above the plastic limit, indicating some lateral soilmovement, which would require amelioration before the nextcotton crop. The simulation of vertical soil stress did not indicateany interaction between the dual wheels, which reflects differ-ences between structured soils compared with elastic homoge-neous semi-infinite media assumed in models. The traffickedfurrow increased in strength following the passage of the picker,similar to that observed during sugarcane harvest (Braunack et al.,2006). As round module balers are heavier, that effect would begreater than a lighter basket picker. The development of increasedsoil strength as columns under wheel tracks is seen as beneficial ifpreserved between crops. This would restrict subsequent compac-tion and enable zones between the inner wheels to be managed forroot growth. Since an increase in soil strength was detected at0.5 m in the profile under both pickers, this indicates that stress

due to traffic is transmitted to depth with the potential for subsoilcompaction, which is in agreement with the finding of Hakanssonand Reeder (1994). This is also borne out by the distribution ofvertical stress under the tyres of the two pickers, where simulatedstress is transmitted to 0.6 and 0.8 m under the basket and roundmodule picker respectively. Lamande and Schjønning (2011a,b)measured the maximum soil stress at 0.9 m, similar to thesimulated vertical stress under the round module baler. Soildegradation at this depth is difficult to ameliorate and may becomea permanent constraint to productivity (Hakansson and Reeder,1994). These studies were conducted on arable soils with annualcultivation to 0.2 m with traffic at soil field capacity; no indicationwas provided whether this was above or below the soils PL.

4.3. Profile water content at traffic

Changes in soil strength were minimised when profile soilwater was less than 20%, however this needs to be verified across arange of soils. The greater the soil strength is at any given soil water

(A) (B)

(C) (D)

Fig. 7. Simulated profile soil water content (%, g/g) at (a) 60, (b) 40, (c) 20, and (d) 10% open bolls at Myall Vale in relation to the soil plastic limit and rainfall between

defoliation and harvest.

M.V. Braunack, D.B. Johnston / Soil & Tillage Research 140 (2014) 29–39 37

content implies that there would be a smaller change in strengthdue to traffic and this is reflected in the pre-history of the fields andprofile water content at the time of traffic Since a change in soilstrength was detected at depth under both pickers, it is suggestedthat growers should aim to dry the profile as much as possible priorto harvest.

Soil water content at the time of traffic varied between sites andwith soil; while the PL ranged from 17 to 22 percent for theVertosol soils (Table 1). The difference in soil water contentbetween the DUL and the PL indicates the evapotranspirationrequired after significant rainfall or the last irrigation to reduce theprofile to the PL; this corresponds to on average 16 mm acrosssites. These values seem relatively low, however at this point in theseason the crop has been defoliated and is not transpiringsignificant water; the only mechanism for drying is evaporationfrom the soil surface. The greater the difference between the soil PLand WP the lower the risk for soil degradation as there is a widerwindow of soil water content suitable for trafficking before

Table 4Simulation (1960–2012) of the effect of the last irrigation on lint yield and the time o

Last irrigation at percent open bolls

60 40

Average lint yield (kg/ha) 2710 (1690 to 3215) 2701

% Change in yield cf 60% �0.3

Change in day of last irrigation cf 60% �15

The numbers in parenthesis are the range of values (negative indicates a reduction in

degradation occurs; the closer the PL to the WP the greater thepropensity for soil compaction to occur (Littleboy et al., 1998).

4.4. Simulation

The simulation of soil profile water content at picking showedthat on average across all sites the profile was wetter than the soilPL 75% and 14% of the time under irrigated and dryland systems,respectively. The soil water content at the time of traffic in thisstudy was greater than the soil PL at six sites and drier at two. Thisindicates that potentially the average soil conditions at whichgrowers consider being suitable for picking, the soil is susceptibleto compaction in irrigated systems compared with drylandsystems. Growers tend to accept the risk of picking on fields thatare wetter than they may like to do so, as harvest operates to a tightschedule. In a simulation study by Whisler et al. (1993) the effect ofsoil compaction (increased soil bulk density) on cotton yield variedwith location and soil type. The effect of compaction was

f the last irrigation relative to 60% open bolls for the Myall Vale site.

20 10

(1690 to 3215) 2677 (1453 to 3531) 2620 (1453 to 3531)

(7.9 to �13.8) �1.3 (11.9 to �15.2) �3.5 (12.5 to �33.3)

(0 to �76) �23 (0 to �79) �30 (0 to �82)

yield or an earlier irrigation compared with 60% open bolls), cf, compared with.

M.V. Braunack, D.B. Johnston / Soil & Tillage Research 140 (2014) 29–3938

detrimental to crop yield in some instances and were negated inothers by subsoiling or rainfall resulting in sufficient water forgrowth, even with restricted root systems.

An imperative is to dry the soil profile below the soil PL toreduce the effect of picker traffic on subsoil compaction. This isdetermined by the timing of the last irrigation before picking andwhether significant rainfall occurs prior to picking. Once the crophas been defoliated little drying of the profile occurs as the crop isnot extracting water. Simulating the timing of the last irrigationprovided some indication of the benefit of this strategy; this needsto be balanced against any effect on reduced yield potential. Thesimulation indicated that over a long period (52 years) of time thata 3% reduction (range 0–33%, Table 3) in yield may occur when thelast irrigation occurred at 10% open bolls. Hearn and Constable(1984) simulated (over a 22 year period) a yield reduction of 20%when the final irrigation was 20 days prior to the standardcommercial last irrigation, which is in the range simulated in thecurrent study. Yeates et al. (2010) measured a yield loss of 36%when cotton was water stressed at the last effective flower whichis greater than that indicated from the simulation. If irrigation iswithheld late in the season, there is still potential for rainfall to wetthe profile again; even on the day of picking. The occurrence ofrainfall after the last irrigation will result in the profile not dryingbelow the PL so picking at this time may result in subsoil (>0.4 m)compaction. Simulation results tend to indicate that this may notbe a practical option due to operational pressure of contractpicking and risk of yield loss.

4.5. Management options and future research

The current strategy to manage soil compaction is to adoptcontrolled traffic (Hulme et al., 1996) and using rotation crops(Hulme et al., 1991) to dry the profile and induce cracking.Monitoring the profile soil strength may indicate whether the soilis sufficiently strong to support a picker, especially in the wheeltrack furrows. Should soil compaction be used for benefit bymaintaining compacted wheel tracks and managing the soilresource for plant growth and continued productivity in thelong-term for the benefit of the Australian cotton industry? This isthe fundamental concept behind controlled traffic farming (CTF)systems. This is difficult to achieve in cotton systems due to thevariation in wheel configurations of equipment (singles or duals ondifferent track widths) and the number of rows planted orharvested in one pass; commonly 12 rows planted and 6 rowspicked making the adoption of true CTF difficult. Further workneeds to be undertaken to assess the effect of picker traffic on soilconditions and strategies such as maintenance of permanentcompacted traffic lanes, fully matched controlled traffic systems toinclude all crops in the system, longer breaks between cottoncrops, to manage or ameliorate subsoil structural degradation.

Field experiments in the same field are required to quantify theeffect of both pickers on changes in soil physical conditions andwhether tillage or a rotation crop will ameliorate subsoilcompaction prior to the next cotton crop.

Growers will often compromise and continue picking ascontractors need to move on or if rainfall is imminent, fibrequality may be down-graded (Bange et al., 2009). This maybe short-term expediency at the cost of soil degradation in thelong-term.

5. Conclusions

Profile soil strength increased after picker traffic compared withbefore traffic and the change was detected down to at least 0.5 m inthe soil profile. Greater strength occurred closer to the surface aftertraffic compared with that before traffic, with the effect being

greater with a heavy picker. Increases in strength also occurredunder the wheel tracks and encroached under adjacent crop rows.Simulation of profile soil water at harvest indicates that cottonpicking potentially occurs when irrigated soils are wetter andsusceptible to compaction and dryland soils are dryer and lesssusceptible to compaction. The practicality of withholding the lastirrigation (to dry the profile) is low due to potential yield loss.Further studies are required to quantify the effect of both pickerson changes in soil physical conditions and whether deep tillage or arotation crop will effectively ameliorate any soil degradation priorto the next cotton crop. Strategies need to be developed to manageand minimise soil degradation under the new pickers to maximisethe benefits of the technology for growers.

Acknowledgments

The Cotton Catchment Communities CRC provided funding. JoPrice and Darin Hodgson undertook the collection of field data.Max Barnes fabricated the frame for the penetrometer. Thanks tothe growers and contractors for their cooperation in allowingmeasurements to be undertaken at a busy time of the season.

References

ASAE, 1986. Soil Cone Penetrometer, ASAE Standard S313.3.2. American Society ofAgricultural Engineers Standards, St Joseph, MI.

Australian Standards Association, 1995. AS1289.3.2.1 Methods of Testing Soils forEngineering Purposes. Standards Association of Australia, Sydney.

Bange, M.P., Constable, G.A., Gordon, S.G., Long, R.L., Naylor, G.R.S., van der Sluijs,M.H.J., 2009. FIBREpak: A Guide to Improving Australian Cotton Fibre Quality.Narrabri, NSW, Cotton Catchment Communities CRC, pp. 105.

Bengough, A.G., Campbell, D.J., O’Sullivan, M.F., 2001. Penetrometer techniques inrelation to soil compaction and root growth. In: Smith, K.A., Mullins, C.E. (Eds.),Soil and Environmental Analysis: Physical Methods. 2nd ed. revised and ex-panded. Marcel Dekker, New York, pp. 377–403.

Braunack, M.V., Arvidsson, J., Hakansson, I., 2006. Effect of harvest traffic position onsoil conditions and sugarcane (Saccharum officinarum) response to environ-mental conditions in Queensland, Australia. Soil Till. Res. 89, 103–121.

Busscher, W.J., Bauer, P.J., Camp, C.R., Sojka, R.E., 1997. Correction of cone index forsoil water content differences in a coastal plain soil. Soil Till. Res. 43, 205–217.

Carter, L.M., Tavernetti, J.R., 1968. Influence of precision tillage and soil compactionon cotton yields. Trans. Am. Soc. Agric. Eng. 11 (1) 65–67.

Cull, P.O., Smith, R.C.G., McKaffery, K., 1981. Irrigation scheduling of cotton in aclimate of uncertain rainfall II. Development and application of a model forirrigation scheduling. Irrig. Sci. 52, 141–154.

Daniells, I.G., 1989. Degradation and restoration of soil structure in a cracking greyclay used for cotton production. Aust. J. Soil Res. 27, 455–469.

Grainger, C.M., Kosmicki, P.A., Zoz, F.M., 1997. Irrigated Cotton Yield with BeltedTrack and Wheel Tractor Traffic. Society of Automotive Engineers, TechnicalPaper Series No. 972735, pp. 8 p.

Hakansson, I., Reeder, R.C., 1994. Subsoil compaction by vehicles with high axle load– extent, persistence and crop response. Soil Till. Res. 29, 277–304.

Hamza, M.A., Anderson, W.K., 2005. Soil compaction in cropping systems: a reviewof the nature, causes, and possible solutions. Soil Till. Res. 82 (2) 121–145.

Hearn, A.B., 1994. OZCOT: a simulation model for cotton crop management. Agric.Syst. 44, 257–299.

Hearn, A.B., Constable, G.A., 1981. Irrigation for crops in a sub-humid environment.V. Stress day analysis for soybeans and an economic evaluation of strategies.Irrig. Sci. 3, 1–15.

Hearn, A.B., Constable, G.A., 1984. Irrigation for crops in a sub-humid environment.VII. Evaluation of irrigation strategies for cotton. Irrig. Sci. 5, 75–94.

Horn, R., van den Akker, J.J.H., Arvidsson, J. (Eds.), 2000. Subsoil Compaction:Distribution, Processes and Consequences, Adv. Geoecology, 32. IUSS, CatenaVerlag, Germany, p. 462.

Hulme, P.J., McKenzie, D.C., Abbott, T.S., MacLeod, D.A., 1991. Changes in thephysical properties of a Vertisol following an irrigation of cotton as influencedby the previous crop. Aust. J. Soil Res. 29, 425–442.

Hulme, P.J., McKenzie, D.C., MacLeod, D.A., Anthony, D.T.W., 1996. An evaluation ofcontrolled traffic with reduced tillage for irrigated cotton on a Vertisol. Soil Till.Res. 38, 217–237.

Isbell, R.F., 1996. The Australian Soil Classification. CSIRO, Collingwood, Australia,pp. 143.

Keller, T., Defossez, P., Weisskopf, P., Arvidsson, J., Richard, G., 2007. SoilFlex: amodel for prediction of soil stresses and soil compaction due to agricultural fieldtraffic including a synthesis of analytical approaches. Soil Till. Res. 93, 391–411.

Kirby, J.M., 1988. Degradation of cotton soils beneath vehicles – shear damage themajor problem. Aust. Cottongrower 9 (1) 33–38.

Kirby, J.M., 1990. Strength and compression properties of cotton soils. Aust. Cot-tongrower 11 (4) 77, 75.

M.V. Braunack, D.B. Johnston / Soil & Tillage Research 140 (2014) 29–39 39

Kirby, J.M., 1991. Critical-state soil mechanics parameters and their variation forVertisols in eastern Australia. J. Soil Sci. 42, 487–499.

Kirby, J.M., Blunden, B.G., 1993. Single versus dual wheels and the soil compactiondebate. Aust. Cottongrower 14 (5) 46–47, 44.

Kirby, J.M., Blunden, B.G., McLachlan, C.B., 1991. Soil stresses beneath tracked andtyred vehicles. Aust. Cottongrower 12 (4) 48–49, 46.

Kulkarni, S.S., Bajwa, S.G., Huitink, G., 2010. Investigation of the effects of soilcompaction in cotton. Trans. Am. Soc. Agric. Biol. Eng. 53 (3) 667–674.

Lamande, M., Schjønning, P., 2011a. Transmission of vertical stress in a real soilprofile. Part II: effect of tyre size, inflation pressure and wheel load. Soil Till. Res.114, 71–77.

Lamande, M., Schjønning, P., 2011b. Transmission of vertical stress in a real soilprofile. Part III: effect of soil water content. Soil Till. Res. 114, 78–85.

Littleboy, M., McGarry, D., Bray, S.G., 1998. A combination of modelling and soil datato determine risk of soil structure degradation. In: Proceedings National SoilConference, Environmental Benefits of Soil Management, 27–29 April, 1998,Brisbane Exhibition and Conference centre, Australian Soil Science SocietyIncorporated, pp. 462–465.

McGarry, D., Chan, K.Y., 1984. Preliminary investigation of clay soils behaviourunder furrow irrigated cotton. Aust. J. Soil Res. 22, 899–908.

Raper, R.L., 2005. Agricultural traffic impacts on soil. J. Terramech. 42, 259–280.Ritchie, J.T., 1972. Model for predicting evaporation from a row crop with incom-

plete cover. Water Resour. Res. 8, 1204–1213.Rosolem, C.A., Schiochet, M.A., Souza, L.A., Whitacker, J.P.T., 2008. Root growth and

cotton nutrition as affected by liming and soil compaction. Comm. Soil Sci. PlantAnal. 29 (1) 169–177.

Sarmah, A.K., Pillai-McGarry, U., McGarry, D., 1996. Repair of the structure of acompacted Vertisol via wet/dry cycles. Soil Till. Res. 38, 17–33.

Soane, B.D., van Ouwerkerk, C. (Eds.), 1994. Soil Compaction In Crop Production.Developments in Agricultural Engineering, vol. 11. Elsevier, Amsterdam, pp.662.

Stewart, C., Boydell, B., McBratney, A., 2002. Precision agriculture: where are wenow and where to next???. In: Proceedings 11th Australian Cotton Conference,13–15th August, Brisbane Convention and Exhibition Centre, Queensland, pp.829–832.

Stitt, R.E., Cassel, D.K., Weed, S.B., Nelson, L.A., 1982. Mechanical impedence oftillage pans in Atlantic coastal soils and relationships with soil physical,chemical and mineralogical properties. Soil Sci. Soc. Am. J. 46, 100–106.

Sullivan, L.A., Montgomery, L.L., 1998. Deep subsoil compaction of two crackingclays used for irrigated cotton production in Australia. Soil Use Manag. 14, 56–57.

Systat Software, 2011. SigmaPlot 12.0.Taylor, H.M., Gardner, H.R., 1963. Penetration of cotton seedling taproots as

influenced by bulk density, moisture content and strength of soil. Soil Sci.96, 153–156.

VSN International, 2011. Genstat v 14.Whisler, F.D., Reddy, V.R., Baker, D.N., McKinion, J.M., 1993. Analysis of soil

compaction on cotton yield trends. Agric. Syst. 42, 199–207.Willcutt, M.H., Buschermohle, M.J., Barnes, E., To, P., Field, J., Allen, P., 2009. In field

time in motion comparisons of conventional. In: John Deere 7760 and Case 625module express cotton pickers. Beltwide Cotton Conferences, San Antonio,Texas, January 5–8, 2009. P 462-476..

Yeates, S., Roberts, J., Richards, D., 2010. High insect protection of GM Bt cottonchanges crop morphology and responses to water compared to non Bt cotton.In: Dove, H., Culvenor, R.A. (Eds.), Food Security from Sustainable Agriculture,Proc. 15th Agron. Conf., 15–18 Nov. 2010, Lincoln, New Zealand.