Improving crop productivity on sandy soils

Take home messages

  • Assessing the yield gap and the level of yield increase that the rainfall of a modified soil site can support, along with season specific effects, is an important step in assessing the risk of sand amelioration options.
  • For higher cost interventions knowing the likely longevity of effect is essential. Deep soil disturbance has shown effects for up to four years but it appears that the organic matter treatments tested to date have had most of their effect within two years of application.
  • Characterisation of sites across the sandy soils of the Southern cropping region indicated that compaction and a range of nutritional deficiencies are common issues.
  • Analysis of herbicide issues flagged glyphosate and the breakdown compound AMPA as residues of interest but their impact is still under investigation.
  • Yield responses in 2017 at Ouyen were largely driven by ripping and response to nitrogen (N) input, while at Lameroo, moderate interventions using a fertility band concept had limited impact on a high N background.
  • Economic analysis of long-term trials has assessed the return on investment for a range of treatments and highlighted the seasonal response effects on profit-risk.

Background

There are opportunities to increase production on Mallee sands by developing cost effective techniques to diagnose and overcome the primary constraints to poor crop water-use. Commonly recognised constraints that limit root growth and water extraction on sands include compaction (high penetration resistance), poor nutrient supply and low levels of biological cycling and poor crop establishment. Anecdotal evidence of herbicide related issues has been widespread in sandy soils across the low-medium rainfall region. Biological breakdown of herbicides in sandy soils may be limited due to a relatively small microbial biomass, limited organic matter to fuel microbial activity, and reduced activity due to limited soil moisture. It is thought that the issue may be arising from long-term accumulation of herbicide residues and/or inadequate plant-back periods compared to label recommendations. However, there has been little measurement of how much, and what type, of herbicide residues may have accumulated in sandy soils of the target region.

The Sandy Soils Project is aimed at increasing productivity on poor performing sands. There are a range of activities included under this umbrella of funding:

  • Analysis of chemical (nutritional, herbicide), biological (nutritional, disease) and physical constraints on transects of sand across the Southern cropping region.
  • Monitoring of the residual (up to five years) effects of a range of ripping and spading treatments with and without fertilisers, organic matter and clay.
  • Implementation of a range of new experiments at seven sites across the Southern region testing a range of approaches to overcoming constraints and increasing water use and crop productivity.
  • Machinery optimisation and soil (DEM) modelling will be used to support the implementation of new trials, to understand how machinery set-up influences profile modification, and to develop guidelines for implementation.
  • Economic and risk modelling will assess the cost-benefit outcomes of a range of project treatments, and be used to develop a framework to support decision making based on prioritising the underlying soil constraints.
  • A programme of validation and demonstration trials to extend the reach of experiments and provide local information regarding best bet treatments.

Common approaches to overcome physical constraints include techniques to fracture compacted layers (e.g. ripping) and techniques that both fracture the soil and mix and/or invert it (e.g. spading, plozza plough, mouldboard plough). Ripping is less costly to implement and the impacted soil area is far more discrete. Soil profile mixing and/or inversion are more costly to implement and create higher erosion risks. Erosion risks can be mitigated with recent developments such as the ‘spade and sow’ single pass operation. Mixing can address multiple soil constraints as it improves the physical environment for root growth, reduces issues of water repellence and enhances nutrient fertility through incorporation of amendments.

Soil modelling has been used to understand the impact of key operational parameters for spading on the extent of topsoil mixing and to compare the mixing through spading with other ploughing techniques. The modelling approach has also been used to support machinery optimisation and implementation of core research trials. An alternative to mixing approaches is the construction of permanent fertility strips; facilitating nutrition access to the crop by increasing inputs (including granulated amendments) to concentrated parts of the paddock. Permanent fertility strips have been tested elsewhere but not explored in the Mallee environment.

In this paper the most recent data from the Sandy Soils project are presented including analysis of the key constraints to production, results of new experiments implemented in 2017, residual benefits from treatments implemented up to four years ago and economic analysis of interventions to increase production on sandy soils.

Method

Analysis of constraints

Soils have been sampled from sandy soils in nine paddocks across the Southern region and analysed for chemical, biological and physical constraints to one metre depth. Herbicides assessed at 0-10, 10-20, and 20-30cm depths included glyphosate and its break-down product, AMPA, Trifluralin, Prosulfocarb, three imidazolinones (Imazapic, Imazapyr, Imazamox), 2,4-D, Triclopyr, and MCPA. Germination assays (lucerne) were also carried out and scored compared to a herbicide-free sand control.

Monitoring of residual value of sandy soil interventions

Sites established in 2015 (Bute) and 2014 (Cadgee, Brimpton Lake and Karoonda) have been monitored for ongoing yield effects in response to a range of interventions including spading and/or ripping with and without organic matter, clay and fertiliser. This long term monitoring has allowed for the assessment of the return on Investment (marginal return/total costs*100) for a range of amelioration strategies.

New Sandy Soils experiments

In 2017 new experimental sites were established at Ouyen, Lameroo and Yenda (data for Yenda not presented here). The Ouyen site (Moodie and Macdonald 2018) has deep sandy soils with high penetration resistance (>2.5 MPa) and poor sub-surface fertility and two experiments; a ripping trial and a spading trial were established. The ripping trial included shallow (20cm) and deep (30cm) ripping treatments with/without placement of N or nutrient packages (containing phosphorus (P), potassium (K), sulphur (S), zinc (Zn), copper (Cu), manganese (Mn)) at the target depths. Surface banding (~7.5cm) treatments were included as controls.

The trial design allows the impact of physical disruption to be isolated from nutrient placement. The spading trial included spading (30cm) with/without the incorporation of a range of organic amendments (vetch hay, oaten hay, vetch-oaten mixture, chicken litter, compost), and/or fertiliser. The Lameroo site is a swale-dune paddock with issues of moderate water repellence, high penetration resistance and poor sub-surface fertility (for nutrients other than N).

The first experiment, located on a non-wetting deep sand, focussed on evaluating the impact of a variety of inputs at seeding (clay, organic matter, biochar and placement depth) as options seeking to optimise the crop response to the fertility strip concept.

The second experiment focussed on evaluating the impact of an edge row sowing configuration - optimised for developing fertility strips on deep sands - but imposed on the full range of soil types in the paddock (dune, mid-slope and swale).  See Desbiolles et al. (2018) for further detail.

Results and discussion

Analysis of constraints

An important consideration before making radical changes to the soil profile is the yield that the local rainfall can support. An estimate of the physiological yield potential for wheat in an average season was determined and by subtracting farmer data (attained yield) from this value the physiological yield gap for each of our experimental sites was determined (Table 1).

Table 1. Estimates of the physiological (Phys) yield potential, unmodified soil yield potential, attained yield and physiological yield gaps for an average growing season at each of the Sandy Soils Experimental sites.

Site

GS rain
(mm)

Fallow rain
mm)

*Phys yield potential
(t/ha)

Yield potential in unmodified soil
(t/ha)

Attained yield
(t/ha)

Phys yield gap
(t/ha)

Waikerie

157

22

1.95

1.67

1.00

0.95

Carwarp

174

28

2.03

1.35

1.00

1.03

Ouyen

213

30

2.92

2.89

1.50

1.42

Karoonda

235

26

3.34

2.73

1.50

1.84

Murlong

251

21

3.54

3.28

1.50

2.04

Yenda

252

43

4.07

3.09

2.50

1.57

Lameroo

270

28

4.13

3.11

2.50

1.63

Bute

298

24

4.68

3.86

3.50

1.18

Brimpton Lake

377

21

6.33

3.12

3.00

3.33

Cadgee

410

34

7.34

3.33

1.30

6.04

*Physiological yield potential = (growing season rain + 0.25 fallow rain x 22)/1000 and yield potential in unmodified soil was estimated using APSIM set up with plant available water characterisations.

The yield gap highlights the limits of any potential yield gains that we want to capture and places it in the context of the environment, and local knowledge of yield gains that can be made from simple interventions like change in crop sequence and fertiliser management. Because the sites have been characterised for plant available water capacity we were also able to use APSIM to estimate the water and nitrogen driven yield potential of unmodified soil. Where the gap between the physiological yield potential and the unmodified soil yield potential is relatively small (eg. Waikerie) it suggests that gains from soil modification may also be small (Table 1). While these estimates will be refined with further characterisation, they do moderate expectations with respect to anticipated yield gains and the level of investment that should be made in amelioration strategies, particularly in low rainfall regions.

Penetration resistance is an assessment of the potential for physical constraints to limit root growth. Measurements were commonly moderate (>1.5 MPa) within the top 20cm and high (>2.5 MPa) within the top 30cm (Figure 1). Nutrients commonly measured at marginal levels included N, P, Zn, Cu, and Mn (data not shown).

Diagram of penetration resistance (Mpa) in response to depth at key Sandy Soils Research sites. Black line represents the average, with the shaded grey indicating the range at the site.

Figure 1. Penetration resistance (Mpa) in response to depth at key Sandy Soils Research sites. Black line represents the average, with the shaded grey indicating the range at the site.

A survey of herbicide residues in sandy soils in the southern low rainfall region did not find Prosulfocarb, imidazolinones (Imazapic, Imazapyr, Imazamox), Triclopyr, MCPA Trifluralin or 2,4-D at a significant level. Glyphosate and its breakdown product AMPA were measured at all nine sites, where the combined residue load (glyphosate plus AMPA) represented between 0.7 and 6.1 typical applications. The majority (~85%) of the herbicide residue was found in the top 10cm, and was predominantly AMPA (~80%) rather than glyphosate. Little is known of the toxic effects of AMPA and how it may affect root growth and function. However, the concentrations measured here had no impact on the germination of lucerne seeds in a lab bioassay.

Monitoring of residual value of Sandy Soil interventions

An experiment established at Bute by Trengove Consulting in 2015 compared the effects of all combinations of increased inputs of annual fertiliser, ripping and 5 and 20t/ha chicken litter (20t/ha chicken litter not presented here (Trengove et al. 2018)). Compared to a nil control yield of 4.4t/ha/3 yrs, there was a yield gain of 3t/ha with relatively high inputs of annual fertiliser (N, P, S, K, Zn, Cu, Mn at a cost of $430/ha), 2t/ha extra grain yield with chicken litter at 5t/ha (cost $180/ha) and 2.4t/ha response to ripping (cost $60/ha). The cumulative yields for combinations of two factors were quite similar (8.2-8.9t/ha), while the combination of the three treatments yielded 9.6t/ha. There were strong seasonal responses to the treatments with barley in 2016 highly responsive to nutrition and lentils in 2017 relatively more responsive to ripping than the other treatments. The lowest cost intervention was ripping and this generated the most substantive return on investment at 1342% while chicken litter with ripping generated 521%.

Cumulative (2015-2017) yield response (left hand side graph) and return on investment (%) (right hand side graph) to fertiliser, chicken litter and ripping at the Bute site implemented and managed by Trengove Consulting.

Figure 2. Cumulative (2015-2017) yield response and return on investment (%) to fertiliser, chicken litter and ripping at the Bute site implemented and managed by Trengove Consulting.

Long-term spading trials developed in the New Horizons Program have demonstrated cumulative yield gains at Karoonda and Brimpton Lake of an additional 2t/ha grain over four years, with a smaller gain (1.3-1.6t/ha) from incorporation of high rates of lucerne based organic matter. Gains from the incorporation of organic matter appear to be short-lived (~two years) at these sites. Yield gains at Cadgee (+1.6t/ha) were driven by the incorporation of organic matter, and not the physical impact of spading, with gains appearing to increase over time (Fraser et al. 2016). Site specific responses at these long-term sites highlight the need to identify the underlying constraints and to understand the components of the response in economic terms before investing in modifications.

New Sandy Soils experiments

In the relatively low rainfall season of 2017 at Ouyen, deep ripping reduced penetration resistance to ~40cm and resulted in a wheat yield gain of 0.9t/hawhile shallow ripping (20cm) had little impact on penetration resistance or yield. Deep placement of nutrients had no effect on production above the physical impact of ripping. Spading also reduced penetration resistance to the target depth (30cm) but resulted in a relatively small yield gain (<0.4t/ha) above the control (1.4t/ha). The smaller relative gain from spading compared to ripping may be due to lower plant establishment caused by poorer seeding depth control in the spader-seeder approach used.

All spaded treatments outperformed the non-spaded control, with the exception of the spaded oaten hay which likely immobilised N (Figure 3). Incorporation of N rich organic matter (vetch hay, chicken litter, compost) significantly increased yield (0.6 and 1t/ha), grain protein, and harvest index. Ongoing monitoring will evaluate the continuing effects on penetration resistance, nutritional legacy effects, rates of organic matter decomposition, and crop water-use.

In 2017, the Lameroo site had good conditions for mineral N supply (>100 kg N/ha/m depth at sowing) from a preceding legume pasture phase and above average growing season rainfall. As a result, most of the relatively low cost/ low input interventions of the fertility strip concept did not have a significant impact in the first year of inputs. Their effects, if important, are expected to increase with time. While the amendments in fertility strip treatments banded at 10cm furrow depth did not yield more than the zero-amendment baseline, 100kg/ha clay + 100kg/ha organic matter applied at a 20cm furrow depth was the only fertility strip treatment to show an early effect (4.0t/ha versus 3.5t/ha wheat, Desbiolles et al. 2017).

Column bar graph illustrating wheat grain yield response to spading and spading with organic matter inputs at Ouyen in 2017 (site established and managed by Moodie Agronomy).

Figure 3. Wheat grain yield response to spading and spading with organic matter inputs at Ouyen in 2017 (site established and managed by Moodie Agronomy).

Conclusions

Substantial opportunities to increase yield on poor sandy soils have been demonstrated in recent trials. However understanding the rainfall-limited yield potential and season specific effects is important for assessing the likely scope of yield gains and the associated investment risk. Of critical importance to the higher cost interventions is the longevity of effect.

While effects of deep soil disturbance have proven measurable after four years, organic matter inputs appear to often lose effect after two years. Determining the most economic treatment options for growers by diagnosing key constraints, optimising treatments (both machinery and inputs) and understanding longevity in the system is a large and ongoing effort in the Sandy Soils Project.

References

Desbiolles et al. (2018) Testing the concept of fertility strips to increase productivity on deep sand.

Fraser et al. (2016) Overcoming constraints on sandy soils amelioration strategies to boost crop production

Moodie and Macdonald (2018) Increasing water extraction production mallee sands enhanced nutrient supply root zone

Trengove et al. (2018) Amelioration of sandy soils opportunities for long term improvement

Acknowledgements

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 author would like to thank them for their continued support.

This work is funded under the GRDC project CSP00203; a collaboration between the CSIRO, the University of South Australia, the SA state government through Primary Industries and Regions SA, Mallee Sustainable Farming Inc. and AgGrow Agronomy and Research P/L.

We would like to acknowledge the inputs of the wider project team, Bill Davoren, Willie Shoobridge, Stuart Sheriff, Chris Saunders, Mustafa Ugcul, David Davenport, Nigel Wilhelm, Tanja Morgan, Barry Haskins and Rachael Whitworth.

Contact details

Therese McBeath
08 8303 8455
therese.mcbeath@csiro.au

GRDC Project code: CSP00203