Key messages

• Pasture with late-season weed growth tended to dry the soil out at depth more than a grain crop. Early spray-out would be required to conserve water for following crops.
• The root system of wheat and barley reached ~150–160cm soil depth, although most activity was <110cm.
• Wheat following fallow had a deeper root system than following a grain crop, utilising soil water from the full 160cm of the measured profile.
• Canola had a deeper root system than wheat and barley and, in some seasons, dried the soil out more, particularly deeper in the profile (70–160cm).
• Legumes had a similar rooting depth to wheat. On occasion albus lupin used less water than wheat at ~70–90cm depth. Chickpea water use varied and in some seasons chickpeas used a greater proportion of available water/rainfall late in the season (Oct/Nov), perhaps reflecting the crop’s slower maturity than wheat.
• Fallow had a marked benefit on following wheat yield after a dry summer/autumn and no effect after a wet summer.

Aim

The aim was to determine the effect of different rotations on crop yield and water use.

Introduction

The core principles of conservation agriculture (no-tillage) include full residue retention, minimum soil disturbance and diverse rotations. A survey in 2005 (Derpsch 2005) identified the main issues of Western Australian no-tillage systems as being a lack of cover on the soil, inadequate diversity in the rotation, herbicide resistance and weed control. A long term no-tillage project was started in 2007 with the overall aim of testing if current cropping systems could be improved by adherence to these key conservation agriculture principles. Measurements included soil carbon sequestration, diseases, insects, weeds, crop water use, as well as crop yield and profitability. The three main collaborators were WANTFA, UWA and CSIRO.

This paper reports on patterns of water use by crops in five different rotations/treatments from 2013 to 2019, when similar rotations were used. Wheat yields in two contrasting rainfall years of 2017 and 2019 are also presented.

Method

This 12-year experiment was started in 2007 at the College of Agriculture Cunderdin. The soil was an alkaline red, sandy clay-loam. Plots were 36m x 80m, with a 2m wide buffer along each side of the plots, providing a 4m guard between plots. The treatments included permanent pasture, monoculture wheat (mono.) and three 3-year rotations: cereal rotation (cereal/cereal/cereal); diverse rotation (wheat/legume/brassica); and ‘farmer’ rotation (cereal/cereal/legume or fallow) (Table 1). All phases (crops) in the rotation were grown every year. Similar crops were maintained for a three-year period (until all plots had completed the same rotations), then varieties or crop types (within the broad definitions mentioned – cereal, legume etc.) could change, to maintain relevance of the treatments. Wheat was grown over the whole site in 2019 (year 13), to compare the long-term effects of the different rotations on wheat yield. The trial design was a randomised complete block with three replications. The intention was to retain maximum residues so all crop treatments were seeded with a low soil disturbance ‘disc opener’, except the farmer rotation where a tyne and knife-pint seeder was used.

In 2010, plots were split into spread residue and windrow burning, except for the farmer rotation where whole plots were windrow burnt. However, there were few differences in soil water between the two subplots, so whole plot data is reported, i.e. for the rotations/crops.

Varieties used and soil water measurements

For the pasture, Dalkeith sub-clover (Trifolium subterraneum ssp. subterraneum) was sown in 2008, which was reseeded, following a knockdown herbicide, in 2009 with Scimitar medic (Medicago polymorpha), and again in 2016 with a Dalkeith sub-clover/Scimitar medic mix. Over time, the pasture became weedy and it was mown occasionally after anthesis to reduce weed seed set, however, sometimes thistles re-grew and remained green into December. From 2013 to 2015 the varieties were Mace wheat (Triticum aestivum L.), Striker chickpea (Cicer arietinum L.), Scope barley (Hordeum vulgare L.) in 2013/2014 and Buloke in 2015, with Sturt TT canola (Brassica napus L.) in 2013 and IH30RR canola in 2014/2015. From 2016 to 2018, Impress CL (2016/17) then Chief CL wheat, La Trobe barley, Amira albus lupin (Lupinus albus L.) and IH30RR canola were grown (Table 1).

Table 1. Crop rotations and sequences at Cunderdin from 2007 to 2019

One access tube was placed in each sub plot and soil water measured with a neutron moisture meter (NMM) between 0-160cm in 20cm increments, starting at 10cm. This was done on a monthly basis throughout the year for the 12 years. Only data from the last six years of the experiment is presented, as there were consistent rotations over this period (Table 1). Crop lower limits (CLL) were estimated using soil water data from the end of the growing season (mostly December), in years which were relatively dry in Oct–Dec (2015, 2016 and 2018), although there was about 28mm in November in 2015. The fallow was only introduced in 2013, so the effect on the following wheat crop could only be determined in following years, therefore 2013 was excluded from the CLL estimates. The drained upper limit (DUL – field capacity) for the topsoil was estimated using NMM measurements during August 2011 and 2016, after recent rainfall and allowing for drainage. For the other depths, DUL was determined from the fallow plots at the times of CLL determination, as it was unlikely that drainage was still occurring. The rooting depth of crops was determined in the same years as the crop lower limit, however 2013 was included for chickpea as this was only grown from 2013–2015. The depth of the roots was estimated from the change in soil water down the profile between August (wettest month) and December (driest month). The rainfall was obtained from the SILO patched point database for Cunderdin (Jeffrey et al 2001). To illustrate the effect of rainfall on yield over the trial period, site-average barley yields are presented, except for 2019, when only wheat was grown. This was because frost damage reduced the yield of wheat in some years.

Results

Crop yield and rainfall varied over the 12 years, with highest barley yields in 2016–2018, the final three years of the rotations (Figure 1). The lowest rainfall years were 2010, 2015 and 2019, with annual rainfall <250mm. The importance of good pre-season rainfall on these soils was shown in 2016 and 2017, which had two of the three highest yielding years.

Figure 1. Rainfall (2007–2019) and site average barley yield (2007–2018) and wheat yield in 2019 at the Cunderdin site. Bars show annual rainfall comprising pre-season rainfall (Jan-April), growing season rainfall (GSR - May-Oct) and Nov-Dec.

Drained upper limit and crop lower limit

The drained upper limit was ~426mm and the site average crop lower limit 245mm, giving an estimated plant available water of 181mm (Figure 2). However, the crop lower limits varied between crops in any one year and across years.

Figure 2. Drained upper limit (DUL) and site average crop lower limit (CLL) for the Cunderdin trial site.

Apparent crop lower limits (0-160cm)

Pasture had the lowest crop lower limit in December with a mean of 222mm, presumably due to continued growth of weeds in most years. This suggests that crops grown after long-season pasture may start with less soil water compared with after a grain crop (Figure 3). Canola in the diverse rotation also appeared to dry the soil out more than wheat and barley in the cereal rotation, with an average of 254mm remaining compared 292mm, respectively. However, it should be noted that the drying effect varied across years, depending on rainfall and canola growth. The relatively dry soil following canola appeared to carry over to the following wheat, were an average of 267mm remained after wheat. This CLL was not significantly different from crops in the cereal rotation, but was significantly less than wheat monoculture, which had the highest CLL of 311mm (Figure 3). As expected, fallow had the most soil water at the end of the season.

Figure 3. Crop lower limits (at end of growing season) for different crops in the cereal, diverse and farmer rotations and monoculture wheat and pasture (average 2015, 2016 and 2018). Error bars show ±SE (n=6 except for wheat mono. and pasture where n=3). Different letters above bars show significant differences at P≤0.05.

Rooting depth

The rooting depth of wheat and barley was similar at ~150–160cm, with most activity above 110cm (Figure 4a, b). However, there was greater water extraction with wheat following fallow in the farmer rotation, especially from 50–160cm, indicating deeper root activity (Figure 4a).

Figure 4. Change in soil water with depth from August to December (average 2015, 2016 and 2018) for: a) wheat monoculture and wheat after cereal (cereal rotation), after canola (diverse rotation) and after fallow (farmer rotation); and b) wheat after cereal (cereal rotation), barley after cereal (cereal rotation) and barley after wheat (farmer rotation). Error bars show ±SE (n=6 except for wheat mono. where n=3). Asterisk above lines show significant differences at P≤0.05.

By comparison, canola had greater root activity than wheat for the full 160cm, with the roots extending beyond 160cm (Figure 5a). Averaged across 2016 and 2018, albus lupin had a similar rooting depth to wheat and there appeared to be less water use at 70–110cm depths (Figure 5a), although differences were relatively small, and this only occurred in one of the two years (2018). Chickpea was grown from 2013–2015 and showed a similar water use pattern to wheat in the cereal rotation (Figure 5b). Interestingly, chickpea appeared to have more water remaining deeper in the profile in Aug–Oct, with a larger proportion of this water used at the end of the season compared with wheat (data not shown). This perhaps reflected the longer maturity of the Striker chickpea compared with Mace wheat.

Figure 5. Change in soil water with depth from August to December for: a) wheat after cereal (cereal rotation), canola after legume (diverse rotation) and albus lupin after wheat (diverse rotation) (average 2016 and 2018); and b) wheat after cereal (cereal rotation) and chickpea after wheat (diverse rotation) (average 2013 and 2015). Error bars show ±SE (n=6). Asterisk above lines show significant differences at P≤0.05.

Water loss over summer/autumn and the fallow effect

Differences in the fallow effect are shown by comparing 2016/17 (Figure 6a) and 2018/19 (Figure 6b). At the end of the 2016 growing season, in December 2016, the fallow in the farmer rotation had stored an additional 120mm compared with after a wheat crop in the cereal rotation (Figure 6a). However, there was 223mm of rainfall between December and May the following year and the soil ended up with a similar amount of soil water following both these rotations, with seemingly no stored water benefit from the fallow. This was borne out as the yield of wheat in 2017 following fallow was similar to that following wheat in the cereal rotation (Figure 7a). The yield of wheat following canola in the diverse rotation was slightly lower than following fallow in the farmer rotation and crops in the cereal rotation (Figure 7a). This was thought to be due to lower wheat establishment following canola. In 2017, despite the high summer rainfall, little rain fell between the end of March and early June and there were low levels of crop residue following canola, resulting in a dry topsoil at seeding.

Figure 6. Soil water remaining in December and the following May for two seasons, after fallow in the farmer rotation and wheat in the cereal rotation: a) December 2016 and May 2017 where a total of 223mm of rainfall fell in the period between the two measurements; and b) December 2018 and May 2019 where a total of 27mm of rainfall fell in the period. Error bars show ±SE (n=6).

Figure 7. Yield of wheat in a) 2017 and b) in 2019, following monoculture wheat, pasture (2019 only) and different crops in the cereal, diverse and farmer rotations. Error bars show ±SE (n=6 except for wheat mono. and pasture where n=3). Different letters above bars show significant differences at P≤0.05.

By contrast, there was an additional 132mm stored in December 2018 after the fallow compared with after wheat in the cereal rotation (Figure 6b). There appeared to be little change in soil water at 0–30cm, presumably because about 24mm of the total 27mm rainfall (between December 2018 and May 2019) fell in April/May. The fallow ended up losing ~17mm of soil water over summer/autumn and the wheat plots ~11mm over this period, giving a total loss (including 27mm rainfall) of 44mm and 38mm, respectively. Most of the water loss occurred between 30–70cm following wheat and from 30–130cm in the fallow (Figure 6b).

At seeding (May) of the following 2019 wheat crops, fallow had an additional 126mm of stored water, compared with the cereal rotation, where wheat was grown the previous year (Figure 6b). The additional stored soil water following fallow in the farmer rotation resulted in a 70-100% increase in yield (3.4t/ha) compared with wheat grown after other grain crops (Figure 7b). The second highest yield was following pasture the previous year (2.2t/ha), although this was not significantly different from wheat grown following crops in the diverse rotation (~2t/ha). Wheat in the cereal rotation was slightly lower yielding (~1.7t/ha) again and this was similar to wheat following wheat in the farmer rotation. Lowest yield occurred in monoculture wheat (1.5t/ha), although this was not significantly different from the two lowest yielding wheat crops in the cereal rotation (Figure 7b).

Conclusion

In several seasons there was less soil water remaining after pasture compared with growing a crop, particularly deeper in the profile. Pastures could be sprayed out by end of August to early September to increase the stored soil water deeper in the profile for the following crop. Canola also appeared to dry the soil out more than wheat and barley, especially between ~70–160cm. The drying effect carried over to the next wheat crop, although that crop mostly yielded more than wheat after a cereal crop, suggesting other benefits from canola such as reduced disease or increased nitrogen availability (Kirkegaard et al 1994).

The rooting depth of wheat and barley was similar at ~150–160cm, with most activity above 110cm. As expected, there was greater water extraction down to 160cm in wheat following fallow, indicating deeper root activity. Canola had greater root activity than wheat for the full 160cm, with the roots extending beyond 160cm. Albus lupin had a similar rooting depth to wheat and there appeared to be less water use at 70–110cm depths. Chickpea showed a similar water use pattern to wheat, although this crop appeared to have more water remaining deeper in the profile in Aug–Oct, with a larger proportion of water used at the end of the season compared with wheat. This, perhaps, reflected the longer maturity of the Striker chickpea compared with Mace wheat.

In some seasons a substantial amount of soil water was stored after winter fallow (up to 126mm), while in other seasons, with high summer/autumn rain, there was little additional soil water compared with after a grain crop. So, the yield benefit from the fallow on the following wheat crop varied from 0 to 100%, compared with wheat after a grain crop. The loss of soil water from the fallow over summer/autumn varied with season and rainfall but in 2018/19 was estimated to be ~44mm in the fallow and ~38mm after a wheat crop (change in soil water plus 27mm rainfall). In 2019, wheat yield after pasture (2.2t/ha) was similar to wheat after crops in the diverse rotation (~2t/ha), which was higher than wheat grown in the cereal rotation (~1.7t/ha) and monoculture (1.5t/ha). Wheat after fallow was significantly higher than wheat after the other rotations/crops in 2019 (3.4t/ha).

Acknowledgments

The research undertaken as part of this project is made possible by the significant contributions of growers through both trial cooperation and the support of the GRDC, the authors would like to thank them for their continued support.

The authors would like to thank the Western Australian College of Agriculture at Cunderdin for hosting the trial and their support over the 13 years. Thanks also to Dr Geoff Anderson for reviewing the paper.

References

Derpsch, R. (2005). Final Report - Situational analysis of no-tillage systems in WA and recommendations for the way forward. Report on consultancy to Western Australia and South Australia and WANTFA. GRDC – Australian Government Department of Agriculture, Fisheries and Forestry Sustainable Industries Initiative of the National Landcare Program.

Jeffrey, S.J., Carter, J.O., Moodie, K.B. and Beswick, A.R. (2001). Using spatial interpolation to construct a comprehensive archive of Australian climate data, Environmental Modelling and Software, Vol 16/4, pp 309-330. DOI: 10.1016/S1364-8152(01)00008-1

Kirkegaard, J.A., Gardner, P.A., Angus, J.F. and Koetz, E. (1994). Effect of Brassica break crops on the growth and yield of wheat. Australian Journal of Agricultural research. 45: 529-545.

Contact details

Ken Flower
The University of Western Australia, 35 Stirling Highway, Perth, Western Australia
Phone: 0417952080, Email: ken.flower@uwa.edu.au