Soil Compaction and its Effect on Alfalfa in Zone Production Systems1

B. D. MEEK,* E. A. RECHEL, L. M. CARTER, AND W. R. DETAR

ABSTRACTAlfalfa (Medicago sativa L.) was grown under two zone production

systems (no harvest traffic) (i) preplan! compaction and (ii) no pre-plant compaction and compared to two systems where traffic wasapplied (i) to 100% of the area during each harvest and (ii) to lanessimulating the traffic a grower would apply during harvest. The ex-periment was done in a semiarid area under irrigation on a Wascosandy loam (coarse-loamy, mixed, nonacid, thermic Xeric Torrior-thent), which is easily compacted. Without traffic the soil settled toa bulk density of 1.6 to 1.70 Mg m~\ Traffic increased the bulkdensity to the 1.8 to 1.9 Mg m~3 range. Increases in bulk density atthe 0.05- and 0.15-m depths occurred only during the first five passesof traffic, but 10 passes were required to reach a stable bulk densityat depths of 0.25 to 0.5 m. Traffic similar to what a grower wouldapply reduced yields by 10% compared with no traffic. Compacting100% of the soil surface after each harvest reduced yields 17% com-pared with no traffic.

Additional Index Words: bulk density, wheel traffic, alfalfa crowndamage, Medicago sativa L.

ZONE PRODUCTION SYSTEMS were defined by Tay-lor (1983) as those systems where the crop zone

and traffic lanes are distinctly and permanently sep-arated. Studies are being conducted at the U.S. CottonRes. Stn. on zone production systems (Rechel et al.,1987) using alfalfa (Medicago sativa L.) and (Gossy-pium hirsutum L.) cotton. Other zone production sys-tem experiments have been conducted in England, theNetherlands, Israel (Hadas et al., 1983) and South Af-rica. In England, Soane (1975) measured large yieldincreases (24%) for potato (Solanum tuberosum L.)using controlled traffic. Zone production systems forthe production of potato, wheat (Triticum aestivumL.) and sugarbeet (Beta vulgaris L.) in the Netherlandshas resulted in small increases in yield but large sav-ings in energy (Lamers et al., 1986).

Modern crop production is a continuing battle be-tween soil compaction caused by traffic and allevia-tion of this compaction by tillage. The degree of soilcompaction resulting from traffic depends on force ap-plied and soil properties, especially soil water. Harris(1971) reviewed the soil compaction process and pre-sented stress-compaction relationships but stated thatthe possibility of developing rigorous relationships byanalytical means is remote. Lambe (1962) measuredthe relationships between bulk density and soil-watercontent at the time of loading and found that the de-gree of compaction is strongly dependent on the watercontent. Texture and organic matter are two other soilproperties that have a large influence on the com-pactability of a soil. Erickson et al. (1974) calculatedthe "degree of compactness" of a soil and comparedsoils with a wide range of textures and found optimumyields to be in the range of 85 to 90%.U.S. Cotton Res. Stn., USDA-ARS, 17053 Shafter Avenue, Shafter,CA 93263. Contribution from the U.S. Cotton Res. Stn., USDA-ARS. Received 30 Apr. 1987. "Corresponding author.

Published in Soil Sci. Soc. Am. J. 52:232-236 (1988).

Soil compaction can be characterized by bulk den-sity (Canarache et al., 1984;Cassel, 1982), soil strength(Bauder et al., 1981; Swan et al., 1987), water infiltra-tion rate (Douglas and McKyes, 1982; Patel and Singh,1981), and root growth (Barley and Greacen, 1967;Taylor and Ratliff, 1969).

Numerous soil compaction studies have been re-ported. Most field experiments have involved me-dium to heavy compaction levels because it was notpossible to remove wheel traffic from areas that wereof sufficient size to obtain accurate yields. The exper-iment described below was done using a wide tractiveframe that allowed crop production under zero traffic.

In alfalfa, production yields are affected both by soilcompaction and damage to crown and shoot regrowthfrom harvest traffic (Sheesley et al., 1974; Sheesley andGrimes, 1977). This traffic may cover up to 70% offield surface (Grimes et al., 1978). Gifford and Jensen(1967) found that increasing bulk density from 1.44to 1.53 Mg m-3 in a loam soil reduced yield, especiallyunder a dry treatment.

The objectives of this study were to define and char-acterize soil density and alfalfa yield changes resultingfrom traffic (under both conventional and 100% traffic)and compare these values to those obtained under zoneproduction systems.

METHODS AND MATERIALSThe research was conducted at the U.S. Cotton Res. Stn.,

Shafter, Ca, at 35°32'00" N, 119° 16'40"W, and 111-m abovesea level. The location receives an average of 159 mm ofrainfall per year with little rainfall from May to September.The soil is a Wasco sandy loam (coarse-loamy, mixed, non-acid, thermic Xeric Torriorthent (Entisols). Unilateralcompression at 200 kPa resulted in a bulk density of 1.72Mg m-1 at a soil moisture of 13% by weight (matric potentialof -10 kPa). The bulk density was 1.96 Mg m-3 at 11.8%soil moisture when compacted by the low-compaction pro-cedure for soil material passing a 4.75 mm (no. 4) mesh sieve(ASTM D 698-58 T) (Felt, 1965).

Alfalfa, nonwinter dormant cultivar 'WL514', was sownin October 1982. All plots were initially rototilled to 0.15m; then chiseled with 0.33-m spaced shanks to a depth of0.54 m in 0.18-m increments in August 1982. This tillageoperation established a uniform soil condition for plantingby eliminating any variation due to previous soil manage-ment. In February 1983 triple superphosphate was broadcastat 162 kg P ha '. Plant tissue analyses were conducted dur-ing the study to ascertain that P levels were adequate. Thealfalfa was sprinkler irrigated in October 1982 (for germi-nation), November 1982, and early April 1983. Plots wereflood irrigated for the first time on 25 Apr. 1983, and allsubsequent irrigations were by flooding using each plot as abasin. The alfalfa was irrigated when 50% of the availablewater was depleted in at least 25% of the plots at the 0.45-m depth.

Six replications of four traffic treatments were arranged ina randomized complete block design. Each plot was 8 mwide and 30 m long. The treatments were:

No Traffic (NWT) Alfalfa was seeded into the loosenedsoil and all traffic was excluded.

Preplan! Traffic (PR) The soil surface was allowed to dry,then compacted with an International TD9 crawler tractor3

233

234 SOIL SCI. SOC. AM. J., VOL. 52, 1988

(Case-IH, Racine, WI) followed by a John Deere 4020 trac-tor (Deere and Co., Moline, IL), both trafficking 100% ofthe plot area. There was no harvest traffic applied.

Repeat Traffic (RE) The plots were treated as the PR plotsbefore planting with traffic applied after each harvest. Eachplot was trafficked (100%) 3 to 5 d after each harvest, bysingle passes from a John Deere 4020 tractor with a 18.4-34, 6 ply, 2020-kg rear tire inflated to 150 kPa and a 10.0-16, 6 ply, 823-kg front tire inflated to 138 kPa.

Grower Traffic (GR) This treatment simulated conven-tional traffic patterns in an alfalfa field and was based on alocal survey of farmer practices. The initial preplant con-dition consisted of two tracks through each plot representingthe alfalfa seed planter. Subsequent traffic during each har-vest, also simulated by a John Deere 4020, was aligned ina pattern to represent a swather, rake, baler, and bale wagon.This wheel pattern created many distinct traffic zones thelength of the plot and resulted in 48% of the soil surfacereceiving wheel traffic. Four of these traffic zones were se-lected for detailed study and were defined as (i) GR-O, azone or track having never received traffic, (ii) GR-L, a zonereceiving traffic only from two to four passes by narrow light-weight tires, (ii) GR-M, a zone representing two passes side-by-side from the rear tire and varying degrees of traffic fromthe front tire of the John Deere 4020, (iv) GR-H, a zonerepresenting two passes in the same exact same track by therear tire of the John Deere 4020, plus varying degrees oftraffic from the front tire.

The NWT and PR treatments represented two methodsof soil preparation prior to initiating the zone productionsystem. The RE and GR treatments represent two degreesof conventional traffic applied to an alfalfa field.

All plots were swathed at the same time when 50 to 70%of the regrowth buds were 10 to 20 mm in length. All trafficpatterns were applied 3 to 5 d after swathing. This timesequence represents the traffic from the baler and bale wagon,which can cause the greatest damage to alfalfa regrowth(Sheesley et al., 1974). Traffic damage from the swather andrake are minimal because of the short time period after cut-ting; for convenience this pattern was applied at the sametime as the baler operations. The first harvest was on 1 Apr.1983; but because the soil was wet from a rain, the first trafficwas not applied until 17 May after the second harvest.

All cultural operations and measurements were made witha wide tractive research vehicle (WTRV) spanning each plot.The WTRV travels on permanent wheel paths, and opera-tions are conducted without applying traffic to the plots.Some measurements were taken from catwalks that spannedeach plot. No foot traffic was allowed in any plot.

Bulk density was measured using a two-probe density gauge(Model 2376, Troxler Lab., Triangle Park, NC). Rawitz etal. (1982) described the equipment and the calibration pro-cedure. Equation [3] presented by Rawitz et al. (1982) wasused to calculate bulk density using unattenuated count rate(/„ =316 000) and mass attenuation coefficients suggestedby them for soil and water. The /„ was derived by the regres-sion equation between bulk density (soil core method) andcounts in the field. Parallel aluminum access tubes (o.d. 51mm) were driven into the soil (tubes were driven about 0.20m at a time, then pulled out, and the soil inside removed)0.30 m apart so that the tops were even with the soil surfaceand closed with rubber stoppers when measurements werenot being made. Three sets of tubes were placed in each plotof the NWT, PR, and RE treatments and eight sets in eachplot of the GR treatment (two sets in each subtreatment).Tube alignments were measured and the distance apart en-tered in the equation when bulk density was calculated.

1 Mention of a trademark product does not constitute a guaranteeor warranty of the product by the USDA, and does not imply itsapproval to the exclusion of other products that may also be suit-able.

Table 1. Bulk density at selected times and depths for thetreatments not receiving harvest traffic.

Depth1982fFall

1983

March May 1984 1985 1986

————— Mgrn-'No traffic (NWT)

0.05 1.390.15 1.440.25 1.530.35 1.450.45 1.320.55 1.350.65 1.64

1.42 1.491.45 1.551.56 1.651.52 1.631.44 1.661.42 1.661.65 1.68

1.521.611.681.661.661.661.69

1.531.611.671.661.651.651.67

1.591.631.701.701.681.651.67

Preplant traffic (PR)0.05 .650.15 .630.25 .740.35 .570.45 .510.55 .580.65 .59

(36)tLSD (P = 0.05) =

.631.65 .641.74 .731.66 .681.53 .581.56 .511.68 .63(18) (18)

0.049

1.681.671.741.701.601.561.68(36)

1.651.671.701.691.601.541.65(18)

1.661.651.731.701.591.551.65(18)

t Values are an average of all measurements taken that year except for 1983when measurements were taken before the first flood irrigation (March)and after the plots had been flooded twice (May). Measurements in 1982were made after preplant traffic had been applied.

t Number of measurements averaged for each value in column.

Bulk density measurements were made in the NWT, PRand RE treatments in 1982 (October and November), 1983(March 13-16 May [NWT and RE treatments only], 24-27May, and September [RE treatment only]), 1984 (July andSeptember), 1985 (June), and 1986 (June). Measurementswere made in the GR treatment in 1983 (June and Septem-ber), 1984 (July and September), 1985 (June), and 1986(June). Depth of measurements were made from a depthrange of 0.05 to 0.65 m in 0.10-m increments.

RESULTSThe zone production systems (no harvest traffic) re-

sulted in bulk densities usually <1.7 Mg m~3 and afairly uniform bulk density with depth (Table 1). Plotsreceiving traffic at harvest had high bulk densities inthe surface of up to 1.9 Mg m-3. Bulk density de-creased with depth and most of the compaction wasrestricted to the top 0.25 m (Tables 2 and 3).

No Traffic Treatment (NWT)Although a low bulk density of about 1.45 Mg m-3

was created by tillage, the soil consolidated with time(Table 1). Most of the bulk density increase was causedby the first flood irrigation in April 1983 although therewas a small amount of settling from the sprinkle ir-rigations and rain (November 1982-March 1983). Thelargest increase in bulk density occurred during thefirst flood irrigation at the 0.45- and 0.55-m depths(change of 0.22 Mg m-3). Subsequent changes werenegligible at all depths. The final bulk densities (1986)at 0.05 and 0.15 m were lower than the rest of theprofile.

Preplant Treatment (PR)These plots increased in bulk density only slightly

after the first flood irrigation because they had beencompacted by traffic previously (Table 1). Traffic be-fore planting resulted in final 1986 bulk densities at

MEEK ET AL.: SOIL COMPACTION EFFECTS ON ALFALFA 235

Table 2. Bulk density at selected times and depths for therepeat (RE) treatment.

Table 3. Bulk density of selected times and depths for the GRtreatment (average of GR-M and GR-H).

Depth1982fFall

1983 1983

March May 1984 1985 1986

m0.050.150.250.350.450.550.65

Mgm-'1.631.641.741.651.541.611.70

1.671.741.681.571.611.70

(18)t (18)LSD (P = 0.05) = 0.049

1.641.641.741.681.571.561.66(36)

1.841.811.821.741.641.641.72

(54)

1.801.811.791.731.621.621.70(18)

1.831.821.811.761.641.631.72(18)

Depth

m0.050.150.250.350.450.550.65

June

1.66bt1.68b1.70c1.64c1.60b1.62b1.61c

September

1.88a1.86a1.74b1.68b1.63b1.66b1.66b

1984

Mg m J

1.91a1.88al.SOa1.74a1.71a1.73a1.74a

1985

1.87a1.83a1.79a1.73a1.67ab1.67b1.68ab

1986

1.90a1.84a1.83a1.77a1.71a1.70ab1.71ab

t Values are an average of all measurements taken that year, except for1983 when measurements were taken before the first flood irrigation(March), and the May values are an average of measurements made afterthe first and second flood irrigations.

t Number of measurements averaged for each value in column.

the 0.45- and 0.55-m depths, which were significantlylower than the NWT treatment. This probably re-sulted from preplant traffic decreasing water intake rateduring flood irrigation and having the soil wet with alower soil-water potential. There were only smallchanges in bulk density from May of 1983 to 1986.

Repeated Traffic Treatment (RE)Traffic over all of the soil surface resulted in a high

bulk density at the 0- to 0.30-m depth with values to1.84 Mg m-3 (Table 2). The only significant changesin bulk density occurred between May 1983 (beforeharvest traffic was applied) and 1984. The changes inbulk density were largest at the soil surface (0.2 Mgm~3 at 0.05 m) and decreased with depth (0.06 Mgm-3 at 0.65 m).

Bulk density increases occurred during the first five2.0 -,

t Row values followed by the same letter are not significantly different(Duncan's multiple range test, 5% level).

passes of traffic at the 0.5- and 0.15-m depths, but atthe deeper depths additional passes were necessary toreach a stable bulk density (Fig. 1).

Grower Treatment (GR)This treatment resulted in a wide range of bulk

densities for the various traffic zones up to a high of1.9 Mg m-3 for the GR-M and GR-H zones (Table 3).In July of 1984, the bulk density values at the 0.15-m depth were 1.65, 1.77, 1.89, and 1.84 Mg m-3 forthe GR-O, GR-L, GR-M, and GR-H zones, respec-tively. There was no significant difference in bulk den-sity between the GR-M and GR-H lanes. The GR-Mand GR-H lanes resulted in compaction below the til-lage zone at the 0.65-m depth.

Alfalfa YieldsThe yields for 3 yr were reduced 10% by the GR

treatment and 17% by the RE treatment comparedwith the NWT treatment (Table 4). If crown damagehad been the only factor contributing to yield reduc-

Tlmes TrafficRpplied

ab

0.00.05 0.15 0.25 0.35 0.45 0.55 0.65

DEPTH ( m )Fig. 1. Bulk density measured at depths between 0.05 and 0.65 m as affected by the number of traffic passes. Within each depth, values

followed by the same letter are not significantly different (Duncan's multiple range test, 5% level).

236 SOIL SCI. SOC. AM. J., VOL. 52, 1988

Table 4. Alfalfa yields for 1983, 1984, and 1985 for thefour treatments.

Treatment

NonePreplan!RepeatGrower

1983

19.7aJ19.5a15.9b18.8a

1984

———— Mg ha-26.2a26.0a21.5c22.6b

1985

't —————2S.2a24.5b21.8c22.5c

Total

71.170.059.263.9

t Weights are oven dry basis.t Column values followed by the same letter are not significantly different

(Duncan's multiple range test, 5% level).

tion, the grower treatment (which covered 48% of thesoil surface with wheel traffic) should have reducedyield (0.48 X 17% [yield reduction from 100% traffic])by 8.2%. In this experiment we could not separate theeffects of soil compaction and crown damage fromharvest traffic since they both relate to yield, but an-other experiment has begun that will separate thesetwo factors. A more detailed discussion of the rela-tionship between wheel traffic and growth of alfalfa(taken from this experiment) is presented in Rechelet al., 1987.

DISCUSSIONTraffic resulted in high bulk densities in this sandy

loam soil. Values to 1.9 Mg m~3 were obtained evenwhen care was taken to avoid trafficking when wet andwheel weights were <2100 kg. Gupta et al. (1985)tested 87 soils and found that a Wasco sandy loamsoil had the highest bulk density of any soil when com-pacted under standard conditions. Heavy traffic re-sulted in compaction below the tilled zone at the 0.65-m depth. More passes of traffic were required to com-pact the soil at the deeper depths.

The tilled soil without traffic settled to a bulk den-sity of about 1.65 Mg m-3 when flood irrigated. It waspossible to maintain a bulk density of 1.60 Mg m~3

or less at the 0.45- and 0.55-m depth if soil was lightlytrafficked before the first flood irrigation. Sprinkler ir-rigation resulted in less settling than flood irrigation.

Yields were improved 8% by applying all traffic inlanes compared with covering 100% of the soil surface.Yield decreases may have been greater if the soil hadbeen compacted before the alfalfa was planted; by thetime harvest traffic was applied, alfalfa roots had al-ready grown through the compacted layer.

Bulk density values that cause problems are differ-ent depending on soil texture and other soil properties.Erikspn et al. (1974) have suggested the comparisonof soils using a standard degree of compaction, whichis the bulk density of the soil in the field divided bythe bulk density when compacted under standard con-ditions (200 kPa). They obtained the maximum yieldat about 85 to 90% of the standard degree of compac-tion. The soil in this study settled without traffic toabout 100% of the standard degree of compaction andto 110% of the standard degree of compaction whencompacted by wheel traffic.

The zone production system with traffic lanes sep-arated by plant production zones would provide ad-vantages for alfalfa production. Traffic lanes wouldprovide better traction and allow harvests when the

soil is at a higher moisture content, whereas plantgrowth in the production zone would be optimum be-cause crown damage and soil compaction are elimi-nated.