Previous PageTable Of ContentsNext Page

Productive Lucerne-Crop Rotations

R.A.Latta1 and L-J.Blacklow2

1Agriculture Western Australia, Katanning, WA.
Formally University of Western Australia, Katanning, WA.


A rotation of two years lucerne (Medicago sativa) followed by wheat (Triticum aestivum) and canola (Brassica napus) was compared on the basis of soil water content, pasture production and crop yields with sub-clover (Trifolium subterranean) and annual medic (Medicago polymorpha) in the pasture phase at two sites in the Great Southern region of Western Australia. Soil water depletion during the lucerne pasture phase was up to 100 mm and 60 mm greater than the sub-clover or annual medic respectively. The lucerne rotation also had less stored soil water following the wheat and canola phases, similar or higher grain yield, grain quality and pasture production. The results indicate that incorporating lucerne in rotations can increase water use while maintaining or increasing crop and pasture production.


Lucerne, phase rotations, soil water, pasture biomass, crop yield.


Some 1.8 million hectares of agricultural land in Western Australia have been affected by salinity and this is likely to increase to more than 6 million hectares within 20 years (2). A fundamental cause of salinity is insufficient use of water in farming systems based on annual crops and pastures. One proposed solution is the development of high water-use cropping systems through the incorporation of a perennial pasture phase with lucerne being considered the most suitable pasture species.

This study investigated the production of lucerne compared to annual legume based pastures and their effect on subsequent crops. The experiments were located at sites representative of the medium rainfall wheat belt in Western Australia where lucerne has not previously been studied or widely used.


The experiments were located at Borden (34oS, 118oE) and Pingrup (33oS, 119oE), on soils classified respectively as solodic with colluvial surface soil Db 4.33, and solodized solonetz over truncated laterite Dg 2.33 (1). The pH (CaCl2) in the top 10 cm is 5.0 (Borden) and 5.8 (Pingrup).

Table 1. Pasture establishment and management of sites sown in June 1995 .




Plot size

10 m x 25 m

10 m x 25 m

Cultivars sown and
sowing rate

Lucerne cv. Sceptre @ 5 kg/ha Sub-clover cv. Dalkeith @ 10 kg/ha

Lucerne cv. Sceptre @ 5 kg/ha Annual medics cv. Serena and Santiago each @ 5 kg/ha

Fertiliser June 95 (kg/ha)

10-P, 18-K, 13-S, 22-Ca

12-P, 3-S, 12-K, 0.04-Zn

Pesticide August 96

Lemat omethoate 580g/L a.i. @ 0.1 L/ha,

Fertiliser March 97

6.8 kg/ha P, 12.3 kg/ha K, 8.6 kg/ha S, 15 kg/ha Ca

Herbicide May 97 (L/ha)

Fusilade fluazifop-p 212 g/L a.i. @ 1

Herbicide July 97 (L/ha)

Sprayseed paraquat 125 g/L and diquat 75 g/L a.i. @ 1.5

Pasture plant density and production

Lucerne and subterranean clover/annual medic plant densities were counted in May 1996 and May 1997 at both sites. The lucerne counts were repeated in February 1998, December 1998 and December 1999. Plant density was measured using 0.25-m2 quadrats randomly thrown eight times in each plot.

Herbage was harvested at approximate six weekly intervals from July 1996 to November 1997. At each harvest a 0.25-m2 quadrat was randomly placed four times within each plot and all herbage cut to 1-2 cm height. Sites were then mechanically slashed at 1-2 cm height.

Soil nitrogen

The quantities of inorganic N in soil were obtained by sampling at 0.1 m increments to 0.4 m at four random locations within each plot following the 1996, 1997, 1998 and 1999 seasonal breaks.

Establishment of wheat and canola

For the wheat phase (1998) sites were sprayed in April and again in May with glyphosate and 2,4-D to remove lucerne. Wheat cv. Cunderdin @ 78 kg/ha was sown in May with 140 kg/ha of CSBP ‘Super & Potash’ 3:1. Further pre-emergent (trifluralin, paraquat and diquat) and post-emergent (diclofop, bromoxynil and diflufenican) herbicides were applied.

In preparation for the canola phase (1999) glyphosate and atrazine were applied pre-seeding, followed by bifenthrin. A further atrazine application at Borden followed by fluazifop-p and omethoate was applied at both sites in July. Pinnacle (Borden) and Karoo (Pingrup) @ 4 kg/ha with CSBP ‘Super & Potash’ 3:1 were sown.

Grain yields were measured in December 1998 and 1999 by machine harvesting. Samples of the harvested seed were used to estimate wheat protein and oil content.

Soil water

Soil water was measured at 3-monthly intervals from October 1996 until December 1999 using a neutron moisture meter in 5 cm PVC access tubes. Readings were separated into two groups (10-33 cm and 50-150 cm) based on soil description and bulk density. Linear regression equations relating neutron probe readings to volumetric soil moisture content were obtained from field data for dry and wet soil on the perimeter of access tubes. Gravimetric soil moisture content was determined by drying the soil, calculating bulk density from the dry soil mass and known soil volume. The linear regression of soil moisture had an r2 value of 0.85 (Borden) and 0.90 (Pingrup).

Analysis of variance (ANOVA) (Genstat 5, Lawes Agricultural Trust Rothamsted) was carried out on all measurements with the neutron probe measurements analysed as a split plot design, with treatment as the main-plot factor and time as the sub-plot factor.



Rain (mm) at Borden and (Pingrup) was 341 and (295), 369 and (433), 576 and (319), and 393 and (382) in 1996, 1997, 1998 and 1999 with means of 387 and (350) mm). Lower than average rain occurred at both sites during autumn 1996 and spring 1997 and at Pingrup during spring 1998. Higher than average rain at both sites during winter 1996, and autumn 1997 and 1998. Close to, or above average rain occurred in every month of 1999, apart from February when no rain fell.

Pasture plant density and production

Lucerne plant densities in May 1996, 12 months after establishment, were 35-56 plants/m2. These declined by 20-30% by February 1998. Sub-clover densities at Borden were >400 plants/m2 in both years. The annual medic density at Pingrup was 30 plants/m2 in 1996 which increased to 255 plants/m2 in 1997. At Pingrup approximately 20% of lucerne plants (8 plants/m2) survived the chemical removal program in 1998. Three plants-m2 of lucerne remained through the canola phase at Pingrup in 1999.

The lucerne pasture at Borden produced more biomass (P<0.05) than the clover based pasture during the 1997 growing period (May-November) and the intervening (November-May) summer period of 1996/97 (Table 2). At Pingrup the lucerne produced similar biomass in 1996, less in 1997 and more during the intervening summer period. The lucerne treatments produced more total biomass at Borden and similar biomass at Pingrup as their annual pastures comparisons over the 2 year study.

Table 2. Lucerne and annual pasture production (tDM/ha) at (a) Borden and (b) Pingrup from May to November 1996, November 1996 to May 1997, and from May to November 1997.


May 1996-Nov.1996

Nov.1996-May 1997

May 1997-Nov.1997

(a) Lucerne




(a) Sub-clover




lsd. (P=0.05)




(b) Lucerne




(b) Sub-clover




lsd. (P=0.05)




Soil nitrogen

No treatment differences in inorganic soil N were measured within any 0.1 m depth increments or in the total 0.4 m profile. However, increasing inorganic N levels were measured over the 2 year pasture phase, (1996, 1997) from 40 to 90 kg N/ha at Borden and 7 to 75 kg N/ha at Pingrup.

Soil water content

At Borden (Fig. 1) there was at least 40 mm less water under the lucerne than the clover pasture from January 1997 until May 1998, with a maximum deficit of 98 mm in January 1998. This deficit was erased in August 1998 following the removal of the lucerne. The wheat following lucerne treatment re-established a 30 mm deficit by October 1998 which increased through the canola phase to 50 mm.

The soil water content in the 50-150 cm soil profile was similar to the 10-150 cm measurements. The canola following clover-wheat had 38 mm less water stored in the 50-150 cm soil profile than the potential soil water capacity measured in August 1998.

Figure 1. Stored soil water measured at Borden in response to lucerne and sub-clover in the 10-150 and 50-150 cm soil profile. Error bar I indicates least significant difference (lsd) at P=0.05.

At Pingrup (Fig. 2) soil water in the 10-150 cm profile was reduced by a maximum 63 mm in the lucerne treatment in January 1998. The deficit during the wheat and canola phases was influenced to some degree by surviving lucerne plants. Comparisons of soil water deficits in the 50-150 cm profiles in response to the lucerne and the medic-wheat-canola treatments were 50 mm versus 7 mm. The annual based treatment accessed no soil water at or near 50 cm.

Figure 2. Stored soil water measured at Pingrup in response to lucerne and medic in the 10-150 and 50-150 cm soil profile. Error bar I indicates least significant difference (lsd) at P=0.05.

Wheat and canola production

At Borden the 1998 wheat yields after lucerne (4.7 t/ha) was 18% greater than after subterranean clover (4.0 t/ha). There was no treatment effect on grain protein levels (9.3%). At Pingrup there was no effect of previous pasture species on grain yields (2.0 t/ha) but grain protein levels were higher (13.3 compared to 12%) in response to the previous lucerne. The lucerne treatment retained an average of 8 lucerne plants/m2 which survived the pre-seeding removal.

The canola yield at Pingrup was higher (1.5 t/ha compared to 1.3 t/ha) in response to the previous lucerne phase in the rotation even though an average three lucerne plants-m2 persisted through the canola phase. There were no treatment differences in canola grain yield at Borden (2.1 t/ha) or oil content at either Borden (46%) or Pingrup (41%).


The results demonstrate that a rotation of two or more years of lucerne followed by several years of crop will increase the water -use of south-western Australian grain-producing farming systems and therefore should reduce ground water recharge. Phase farming with lucerne offers an attractive option for farmers wishing to continue grain production in a sustainable manner.

The performance of lucerne in a Mediterranean climate on mildly acidic soils has been shown to be positive. The lucerne-based pasture maintained plant populations and generally increased both pasture and subsequent crop production in comparison with an improved annual legume-crop rotation. However, the wider acceptance of lucerne will rely on other economic and management issues such as effective establishment and management packages, adaptation to acid soils, ability to produce economic benefits through increased nitrogen inputs and more efficient weed control options.


The project was funded by the Grains Research and Development Corporation (GRDC). The experiments were undertaken on the properties of Richard and Nan Sounness (Borden) and Brian and Del Smith (Pingrup). John Schneider and Chris Matthews contributed excellent technical assistance.


1. Northcote, K.H. 1979. A factual key to the recognition of Australian soils. 4th Edition, (Rellim Technical Publications, Glenside, South Australia)

2. PMSEIC 1999. Dryland salinity and its impact on rural industries and the landscape. Occasional Paper No. 1, Prime Minister’s Science, Engineering and Innovation Council, Department of Industry Science and Resources, Canberra, Australia.

Previous PageTop Of PageNext Page