Previous PageTable Of Contents

Simulating lucerne/crop companion farming systems in Australia

Michael Robertson1, Don Gaydon1, Roy Latta2, Mark Peoples3 and Antony Swan3

1 CSIRO Sustainable Ecosystems / APSRU, Queensland Biosciences Precinct, 306 Carmody Rd, Qld 4067, www.csiro.au Email Michael.Robertson@csiro.au, don.gaydon@csiro.au.
2
DPI, Victorian Institute for Dryland Agriculture, Mallee Research Station, PB 1 Walpeup, Vic 3507 www.dpi.vic.gov.au Email roy.latta@dpi.vic.gov.au

3 CSIRO Plant Industry, GPO Box 1600, Canberra, ACT 2601, www.csiro.au Email Mark.Peoples@csiro.au, Tony.Swan@csiro.au.

Abstract

Lucerne/crop companion cropping, whereby crops are direct drilled into an existing perennial pasture such as lucerne, is one way in which the out-of-season water use benefits of a perennial can be retained during the cropping phase. The risks and benefits associated with this new farming system need to be better defined via experimentation and simulation. There have been no previous tests of the ability of APSIM to simulate competition between a perennial and an annual crop. This paper reports the testing of APSIM against field measurements of companion farming systems made at Katanning, WA and Grogan, NSW. Long-term simulations are used to place experimental results into an historical perspective.

Media summary

Perennial species such as lucerne grown with crops show promise as a productive way to increase water use in grain growing areas and reduce the risk of rising groundwater and salinity. Simulation models are being tested on expermental measurements of this novel farming system.

Key Words

Soil water, risk, yield, wheat, canola lucerne, companion cropping.

Introduction

Perennial species such as lucerne in the pasture phase of cropping rotations have shown considerable promise as a productive way to increase whole-of-rotation water use in grain growing areas and reduce the risk of rising groundwater and salinity. The success of lucerne is dependent upon its ability to create a zone of dry soil below the normal rooting depth of annual crops that acts as a buffer against water leakage from the soil to groundwater. There is some hope that companion cropping, whereby crops are direct drilled into an existing perennial pasture such as lucerne, may be one way in which the out-of-season water use benefits of a perennial can be retained during the cropping phase. In this system a crop can potentially be sown and harvested without the costs and technical risks associated with removal and re-establishment of the perennial species. Relatively few research results are available on companion cropping for Australian conditions. Preliminary studies and anecdotal evidence suggest that this technology could provide growers with flexibility to move in and out of crop and pasture. It also allows the perennial to continue to control leakage during a cropping phase. The grain-growing areas and circumstances in which companion farming might work require better definition via simulation modelling. In this paper the Agricultural Production System Simulator (APSIM) is being tested for its ability to simulate companion farming systems against field measurements of this novel farming system.

APSIM simulates competition for light in a species mixture by taking account of the differential height and leaf areas of the different species (Carberry et al. 1996). Competing canopies are assumed to be well-mixed in the horizontal dimension. Competition for water and nutrient uptake is calculated by allowing the roots of each species to have preferential access to water and nutrients in a day-by-day rotation with the other competing species. While testing of competition between two annual species has been previously reported (Carberry et al. 1996, Robertson et al. 2000) there has been no test of the ability of APSIM to simulate competition between a perennial and an annual crop.

Methods

Datasets for model testing

One experiment (Humphries et al, in prep) was conducted at Katanning, WA (mean annual rainfall 488 mm) in the 2001 and 2002 seasons. Lucerne cultivars varying in dormancy rating (from highly winter dormant to winter active) were used. Here we report results for the medium activity cultivar, Aurora and the dormant cultivar Jindera. Treatments compared were lucerne in monoculture, lucerne companion cropped with wheat, and wheat monoculture. The wheat variety was Westonia. Wheat monoculture and companion-cropped treatments were split for with and without fertiliser N in 2002 to examine if N could influence the competition between lucerne and wheat. The duplex soil had a plant available water capacity of 62 mm to the depth of rooting of wheat (1000 mm) and 124 mm to the depth of rooting of lucerne (2000 mm).

A second experiment (M Peoples, unpubl.) was conducted at Grogan (near Temora) in southern NSW (mean annual rainfall 531 mm) over two seasons (2002 and 2003). Wheat and canola crops when grown alone, companion cropped with a medium activity cultivar of lucerne (cv. Aquarius) and lucerne was grown alone (2002 only). The wheat and canola varieties employed were Diamond Bird and Hyola 60 respectively. The clay loam soil had a plant available water capacity of 122 mm to the depth of rooting of wheat and canola (975 mm) and 140 mm to the depth of rooting of lucerne (1775 mm).

In both datasets crop and lucerne productivity was measured and detailed soil water content measurements were taken to test the ability of the model to simulate the soil water balance (and by implication leakiness) in the different systems.

Long term simulations

In order to put the experimental responses to companion cropping into a longer-term perspective, simulations for the seasons 1958 to 2003 were conducted for monocrop wheat and lucerne-wheat companion crop systems at Katanning and Temora, using the soil water parameterisation from the experiments. An established medium activity lucerne cultivar (cv. Sceptre) stand at 30 plants/m2 (250 stems/m2) was assumed in the companion cropping system and was cut at flowering, except when the wheat was present. In both monocrop and companion crop systems, wheat (cv. Kulin at Katanning and cv. Dollarbird at Temora) was sown each year at 100 plants/m2 between 1st May and 20th June when at least 15mm rain fell over 3 days, and soil nitrogen in the rooting zone was initialised (80kgN/ha at Katanning and 120 kgN/ha at Temora) at 50 days after sowing. Wheat residues were retained after harvest.

Results

Model testing

At Grogan in 2002, a very dry season, canola and wheat yields in companion crops were 0 and 12%, respectively, of that in monoculture (Table 1). Simulated yields were close to observed in absolute terms but the percent yield reduction was somewhat less than observed at 71% in both crop species. In the higher yielding 2003 season, companion crop yields were 81% and 52% of that in canola and wheat monocrops, respectively. Simulated yields were close to observed.

Table 1. Grain crop yields (t/ha) observed and simulated for monocrop and companion crop systems over two seasons at Grogan, NSW and Katanning, WA. N fertiliser treatments at Katanning were in 2002 only. Values in parentheses are companion crop grain yields as a percent of that in the corresponding monocrop.

 

Grogan, NSW

 

2002

2003

 

Observed

Simulated

Observed

Simulated

Monocrop - canola

0.87

0.79

1.52

1.95

Companion crop – canola

0.00 (0%)

0.23 (29%)

1.23 (81%)

1.54 (79%)

Monocrop – wheat

1.72

1.17

3.28

3.38

Companion crop - wheat

0.20 (12%)

0.34 (29%)

1.72 (52%)

1.60 (47%)

 

Katanning, WA

 

2001

2002

 

Observed

Simulated

Observed

Simulated

Monocrop – wheat, no N

3.30

3.47

1.50

1.41

Companion crop – wheat, Aurora lucerne, no added N

2.10 (64%)

2.20 (63%)

0.60 (40%)

0.46 (33%)

Monocrop – wheat, +22 kgN/ha

-

-

2.10

2.12

Companion crop – wheat, +22 kgN/ha, Aurora lucerne

-

-

0.90 (43%)

0.96 (45%)

Companion crop – wheat, Jindera lucerne, no added N

2.4 (73%)

2.9 (83%)

0.8 (53%)

0.64 (46%)

Companion crop – wheat, +22 kgN/ha, Jindera lucerne

-

-

1.5 (71%)

1.36 (64%)


Figure 1a. Observed and simulated wheat biomass (■) and grain (□) (left) and total profile soil water (♦) (0-2000mm) (right) and rainfall for the continuous wheat treatment at Katanning.

Figure 1b. Observed and simulated wheat biomass (■) and grain (□), lucerne biomass (*) (left) and total profile soil water (♦) (20-2000mm) and rainfall (right) for intercropped lucerne/wheat treatment at Katanning (lucerne variety ‘Aurora’)

Figure 2a. Observed and simulated canola biomass (■) and grain (□), wheat biomass (*) and grain (Δ) (left) and total profile soil water (♦) (150-1700mm) and rainfall (right) for the continuous cropping treatment at Grogan

Figure 2b. Observed and simulated canola biomass (■) and grain (□), wheat biomass (*) and grain (Δ), and lucerne biomass (▲) (left) and total profile soil water (♦) (150-1700mm) and rainfall (right) for intercropped lucerne/canola and lucerne/wheat treatments at Grogan

At Katanning in 2001, wheat yield dropped from 3.3 in the monocrop to 2.1 t/ha in companion cropping with cv. Aurora, a medium winter-activity lucerne (Table 1). With the dormant lucerne cultivar, Jindera, wheat yields were reduced less by the companion systems, and this was successfully captured in the simulation. In the lower-yielding 2002 season, wheat yields were affected more in the companion system with yields at 30-40% of those in monocrops. Again, APSIM was able to simulate the lucerne cultivar effect as well as the seasonal effect compared to 2001. N application in 2002 benefited both observed and simulated grain yield in companion and monocrops.

Seasonal dynamics of biomass and grain yield accumulation, and total profile soil water were compared with observed (Figs 1 and 2). While there were some discrepancies between simulated and observed total soil water, the simulations captured the trends in profile drying between the annual-based system and the companion system.

Long-term simulations

At Katanning observed wheat yields in companion crops were reduced by 35% and 60% relative to monocrops in 2001 and 2002 seasons. Long-term simulations reveal that median reductions are ca. 50% (Fig. 3a) (equivalent to 0.9 t/ha – Fig. 3b) with a range of 0 to 100% (i.e. crop failure) (Fig. 3a). For wheat at Grogan (Temora) the 90% reduction seen in 2002 was a 1 in 10 year occurrence (Fig. 3a). The 50% reduction for wheat in 2003 was close to the long-term median, equivalent to a yield penalty of 1.8 t/ha (Fig. 3b).

Figure 3. Cumulative probability of the penalty in wheat yield for companion cropping with lucerne, presented on (a) a percentage reduction from the wheat monocrop and (b) an absolute basis (kg/ha), for both Katanning and Temora.

Conclusion

APSIM has satisfactorily simulated wheat, canola and lucerne productivity and soil water dynamics in companion cropping systems. The response to important agronomic options of N fertiliser and lucerne cultivar used to manage competition between the perennial and the grain crop were also simulated well. Sound performance against such detailed datasets is a prerequisite to the use of APSIM to identify grain-growing areas and circumstances in which companion farming might be a viable option for farmers.

Acknowledgements

CRC for Plant-Based Management of Dryland Salinity, GRDC, CSIRO and Department of Agriculture WA provided funding for this research

References

Carberry PS, Adiku SGK, McCown RL, Keating BA (1996). Application of the APSIM cropping systems model to intercropping systems. In: C Ito, C Johansen, K Adu-Gyamfi, K Katayama, JVDK Kumar-Rao, TJ Rego (Editors). Dynamics of roots and nitrogen in cropping systems of the semi-arid tropics. (Japan International Research Centre for Agricultural Sciences), pp. 637-648.

Humphries AW, Latta RA, Auricht GC and Bellotti WD (in preparation). Overcropping lucerne with wheat: Effect of lucerne winter activity on the biomass, water-use and the yield and quality of the wheat. Australian Journal of Experimental Agriculture.

Robertson MJ, Whish J, Smith FP 2001. Simulating competition between canola and wild radish . In '12th Biennial Australian Research Assembly on Brassicas'. Victoria. pp. 106-110.

Previous PageTop Of Page