Take home messages

1. One-time tillage in continuous no-tillage (NT) system reduced weed populations and improved crop productivity and profitability in the first year after tillage, with no impact in subsequent four years.
2. Soil properties were not impacted in Vertosols, however Sodosols and Dermosols suffered short-term negative soil health impacts. A Sodosol and Dermosol also posed higher risks of runoff, and associated loss of nutrients and sediment during intense rainfall after tillage.
3. Tillage reduced plant available water in the short-term which could result in unreliable sowing opportunities for the following crop especially in semi-arid climate.
4. Generally there were no significant differences in crop productivity and soil health between tillage implements and tillage frequencies between tillage and NT.
5. The study suggests that occasional tillage can be a viable strategy to manage constraints of NT systems, with few short-term soil and environmental costs and some benefits such as short-term farm productivity and profitability and reduced reliance on herbicides.

Introduction

No-tillage (NT;seeding with low soil disturbance and no prior tillage) is a key component of conservation agricultural systems which has provided tangible, economic, environmental and social benefits as compared to conventional tillage which involves intensive disturbance of soil prior to crop sowing. The adoption of NT has progressed globally and in Australia, particularly within north-eastern cropping regions. However, there are concerns regarding long-term sustainability of such systems due to build-up of herbicide-resistant weed populations, increased incidence of soil and stubble-borne diseases, and stratification of nutrients and organic carbon in the top soil. There is an increased interest in the use of an occasional tillage to combat both biotic and abiotic constraints in NT system. The impact of tillage in an otherwise NT farming system on agronomic, soil and environmental factors has either been shown to be inconsistent or studied for a relatively short period. Only a few studies have examined the impact of occassional tillage for periods of four to five years, and these yielded inconsistent results.

Materials and methods

A total of 14 field experiments were established in four different phases in northern NSW and Queensland during 2012-15 on sites with a long-term history of continuous NT and controlled-traffic farming:

(i) five sites with or without tillage treatments in otherwise NT fields in winter 2012;

(ii) three sites with treatments involving different timing, frequency and type of tillage implements with each treatment factor (timing or, frequency or type of tillage) in a separate experiment in a complete randomised block design in winter 2013,

(iii) two sites with different types of tillage implements (strip tillage, narrow chisel, and disc) and one experiment to quantify soil water loss and recovery with one-time tillage in summer 2013, and

(iv) three sites with or without tillage treatments to quantify runoff and loss of soluble nutrients immediately following tillage under simulated rainfall in winter 2015.  
All tillage operations were shallow to an approximate soil depth of 0-15 cm.

A randomised, complete-block design was used for all sites except at Warwick. At Warwick, a long-term experiment consisting of a factorial combination of tillage practice (NT vs conventional tillage), crop residue management (burnt or retained) and nitrogen (N) ferilizer rates (0, 30 and 90 kg N/ha) was longitudinally split into two, with half of the NT section receiving chisel and the remaining half left un-tilled as a control. In the present study, we reported results for no-till-stubble retained-N90 treatments only and analysed as randomised complete block design. All sites had a minimum of three replicates. A detailed description of each site, and tillage implement used, timing and frequency is given in Table 1. The most commonly tillage implements used included chisel, disc, offset disc, scarifier and ‘Kelly’ prickle chain. We determined short-term soil water loss due to evaporation on a Vertosol (100 m x 12 m plots with 9 m buffer space between plots) at Felton A site (Table 1) on day 1, 2, 3, 4, 7, 9, 17, 30, 44, 60, and 65 post tillage using an EM38 (Geonics MKII) in both vertical and horizontal modes. Volumetric water content was estimated using a pre-calibrated linear relationship between volumetric water content and EM38 readings recorded at the site.

At rainfall simulation sites during 2015; runoff was generated for at least 30 min at a rainfall intensity of 70 mm h^-1 from four plots of NT and tillage on three soil types. Samples of runoff from rainfall simulations were analysed for volume, sediment and nutrient contents. Nitrous oxide (N₂O), carbon dioxide (CO₂) and methane (CH₄) fluxes were measured to quantify short-term greenhouse gases (GHG) emissions. Gases were extracted manually from chambers at 0, 25, 50 and 75 min after enclosure once per day before rainfall, within 3 h post-rainfall, and then 1, 2 and 3 days after rainfall. Soil samples were obtained yearly prior to sowing and analysed for physical, chemical and biological properties.

Results and discussion

Soil impacts

Soil water storage and crop water supply are the major factors affecting grain production in the semi-arid region of north-eastern Australia. Tillage in a NT farming system would potentially increase soil evaporation. In the present study, tillage initially reduced the soil water content, due to evaporation. Within 4 weeks of tillage, however, sufficient rainfall replenished soil water loss and helped to recover the soil moisture to pre-till moisture status (Fig. 1). Evaporation losses due to tillage can be as high as 20 to 30 mm. Evaporation can be affected by climatic conditions above the soil surface, depth of the tilled layer, time and nature of tillage, the nature of induced surface structure and the pore geometry of the tilled layer.

Following one-time tillage, soil water content in the 0-0.1 m layer was not significantly impacted at any site or time; except at Warwick 3 months after tillage (Table 2). Soil water recovery was due to substantial rain (80 - 99 mm) received in the first 3 months after tillage operations during 2012 and 2013 at most sites. The exception was a significantly negative impact at the Warwick site, despite 118 mm rainfall between tillage and the time of soil sampling after 3 months. High clay content soils, like that at Warwick (65%), require more rain to refill the soil profile. No significant differences in soil water content due to tillage treatments were observed prior to seeding in subsequent years i.e. after 12, and 24 months after the tillage operation. Given that crop production in north-eastern Australia heavily relies on stored soil moisture during the fallow period, the initial loss of soil water in extreme cases could present either unreliable sowing opportunities or a poor crop establishment. No-till farming systems, especially in high clay soils, usually have a much wider sowing window due to retention of adequate water for crop establishment. In most circumstances, the occurrence of rain after the tillage operation would determine the success of tillage in NT farming systems. This raises the importance of timing of tillage in the farming systems context in relation to the seasonal forecast. Overall, the climatic conditions throughout the season will influence soil water, aeration and temperature and thus will have a marked influence on crop responses and yields in different seasons.

Immediately following a tillage operation, there was significantly higher cumulative CO₂ emissions over 2 days on the Sodosol (67 mg C m^-2 and 37 µg C m^-2 from tillage and NT plots, respectively) but not on the Vertosol (209 µg C m^-2 and 215 µg C m^-2 from tillage and NT plots, respectively). However, the impact of one-time tillage on SOC was not significant (P<0.05) after 3, 12 or 24 months tillage on any of the soil types studied (Table 2). There are a number studies that indicate large amounts of CO₂ are lost from the soil immediately following tillage due to an increase in microbial respiration, typically occurring with a release of trapped CO₂.  However, in the long-term, there is no consistent effect of one-time tillage on soil TOC in long term NT systems. Particulate organic carbon (POC) is considered to be more easily decomposed and is preferentially degraded over humic TOC by tillage. In the present study, POC ranged from 10% of TOC in the Black Vertosol at Warwick to 30% of TOC in the brown Sodosol at Condamine. The POC content was significantly decreased in the Black Vertosol at Biloela 3 months after tillage. There was no significant impact on POC at other sites. The decrease in POC with tillage may be due to changes in incorporation of residue into soil, redistribution and decomposition, or where aggregate breakdown results in increased mineralisation. In the present study, surprisingly there was an increasing trend in POC in black Vertosol at Warwick with tillage after 3 months. This could be due sampling anomalies and/or the high spatial variability associated with carbon distribution in field conditions and/or high stubble load prior to tillage leading to incorporation. Available soil P tended to be lower in the surface soil (0 – 0.1 m) at all sites 3 months after tillage, however, the reduction was significant only on the Vertosol at Biloela A. At 12 and 24 months after tillage, available P was similar at all sites. Most studies have reported a general decrease in available P in the surface soil due to tillage. However, in the present study, use of shallow tillage and low P mobility did not result in redistribution of nutrients into the nutrient-poor subsoil.

Total microbial activity (TMA) in the top 0.1 m soil depth was quite variable and not significantly affected at any site three months after one-time tillage operation. However, there was a significant decrease in TMA 12 months after tillage as compared to 3 months after tillage in the Black Vertosol at Warwick (Table 2). Differences may be associated to changes in seasons between the collection times. Generally, FDA as a measure of TMA is considered to be a ’broad-scale’ measurement for enzymatic activities and may not be sensitive enough to detect changes in specific processes due to functional redundancy of soil microbial communities. Alternatively, the result suggests the soil communities were functionally stable in these soils. Furthermore, one-time tillage did not affect the mycorrhizal associations or Pratylenchus thornei nematode populations measured in Black Vertosols at Warwick and Jimbour 12 months after tillage. There was no incidence of crown rot at most sites. However, at the Jimbour site with crown rot infection, one-time tillage did not result in a significant decrease in crown rot (results not shown).

Environmental impacts

On the Dermosol at Moonie B and the Sodosol at Billa Billa there was significantly higher runoff from tillage plots than from NT plots. Runoff volume was highly variable and similar between tillage and NT treatments on the Vertosol at Felton B (Fig. 2). Consistent with these effects on runoff volume, infiltration rates were significantly higher on the NT plots than the tillage plots on the Dermosol and Sodosol (Fig. 3). Erosion and total N loads were highest after tillage on the Sodosol however there were no significant differences due to tillage on the Dermosol or Vertosol. Total P loads in runoff were also significantly higher from tillage than from NT on both the Sodosol and Dermosol. The impact of tillage on runoff and nutrient loads was largely attributed to removal of groundcover by tillage and an increased vulnerability to erosion. Soluble P losses were low overall (<12% of total P) and rather than reducing surface soil enrichment with P and associated soluble P concentrations in runoff, TILLAGE increased soluble P concentrations in runoff at two sites (Fig. 3).

As well as the higher cumulative CO₂ emissions over 2 days after tillage on the Sodosol, approximately three and four times more CH₄ was absorbed over the sampling periods by the NT treatment than the tillage treatment in the Vertosol and Sodosol, respectively. There were no significant impacts on cumulative N₂O fluxes due to tillage, with 111 µg N m^-2 and 100 µg N m^-2 measured from the Vertosol and 61 N m^-2 and 56 µg N m^-2 measured from the Sodosol, from tillage and NT plots, respectively. The N₂O fluxes were low relative to higher input cropping systems, and emissions from both the tillage and NT treatments may have been limited by soil mineral N availability. An increased risk of runoff, erosion and nutrient loss from Sodosols and Dermosols after tillage for weed control in cropping systems typical of the NGR are trade-offs that need consideration in tillage decisions. However, the low GHG emissions measured, despite intense rainfall, suggests that tillage practices in these farm systems have a low global warming potential.

Agronomic impacts

Weed populations on all the sites were significantly decreased 3 months after one-time tillage in 2012 and a trend toward decreased weed density at two sites established in 2013 (Table 2). Twelve months after tillage, weed populations were significantly lower on the black Vertosol at Biloela and the grey Dermosol at Moonie. On the brown Sodosol at Condamine, there was an increase in weed density, in particular African turnip weed (Sysibrium thellungii). Twenty-four and 36 months’ post tillage, there were indications of lower weed populations in TILLAGE as compared to NT but results were not significant. Most studies suggest a positive impact of tillage on reducing weed density. However, tillage has the potential to move buried weed seed to the surface soil, thus providing a more favourable environment for germination by breaking seed dormancy.

The first year after tillage operation generally provided higher productivity at all the sites compared to NT, however these results were not significant. On average, one-time tillage resulted in 0.1 t ha^-1 higher yield as compared to NT. The Brown Sodosol at Condamine recorded a marginally significant increase in chickpea yield (1.07 – 1.16 t ha^-1) after a single chisel treatment (P = 0.08). In the second year of tillage operation, slight positive trends were observed on the Black Vertosol at Biloela, Grey Dermosol at Moonie and the Black Vertosol at Warwick. The Brown Sodosol at Condamine recorded a decrease in yield when compared to NT, likely resulting from a significant increase in the weed population. 24 and 36 months after the tillage operation, no significant differences were observed between NT and tillage treatments. It appears that reduced weed population in the tillage treatments resulted in improved grain yield, as also observed elsewhere. Net return with one-time tillage was generally positive at most sites ($10-$35 ha^-1) except from Vertosol at Biloela (-$3 ha^-1), Jimbour (-$17 ha^-1) and Moree (-$5 ha^-1). However, overall total net return over the 4 years was positive for all sites. Most studies conducted in North America (USA and Canada) and Europe suggests that introducing occasional tillage in continuous NT systems could improve productivity and profitability in the short term. However in the long-term, the impact is negligible or even negative.

*The Billa Billa and Moonie B sites were strategically cultivated to shallow depths (≤ 150 mm) for weed control or pupae busting after cotton cropping once or twice, respectively, in the four years prior to the experiment

↓ or ↑ indicate a significant decrease or increase, respectively at p<0.05; NS, non-significant  (+) NS increase; (-) NS decrease; (~) no result; 3m, 3 months after tillage; 12m, 12 months after tillage; 24m, 24 months after tillage; TMA, total microbial activity (µg/mL FDA/g soil/hour); MBC, microbial biomass (μg C g^-1 soil); BD, bulk density (g/cc); SW, soil water (mm); TOC, total organic carbon (t/ha); POC, particulate organic carbon (t/ha); P, available P (Colwell-P) mg/kg.