Take home message

• Grain yield increases were recorded with increasing phosphorus rates in four of the five data sets
• Shallow phosphorus application (5-8 cm) provided the likely best economic return compared with dual placement of phosphorus shallow (5-8 cm) and deep (16-20 cm)
• The dual placement of phosphorus provided little evidence of yield advantage (only in one of the five data sets – wheat 2022) when compared to the same total phosphorus rate placed shallow
• Responses to residual dual placement of phosphorus were also recorded in the third cropping year (2022) from the initial applications in 2020
• Dual placement of phosphorus (deep P pre-sowing and shallow P at sowing) may be a useful strategy where phosphorus toxicity impacts sensitive crop species establishment and sowing equipment does not allow for seed and fertiliser separation.

Background

Wheat takes up most of its phosphorus (P) from the soil P reserve and a much smaller proportion from fertiliser P applied in the year of sowing (McBeath et al. 2012). If P in the soil reserve is concentrated in the top 8 cm, then soil surface drying can limit P uptake by wheat (Norrish 2003).

Phosphorus placement research undertaken at Mulga in QLD showed that by increasing the depth of P placement from 5-7 cm to 10-15 cm at sowing, wheat yield is increased by 0.5 t/ha (Singh et al. 2005). A further grain yield increase of 1 t/ha (valued at $320/t) occurred when the 10-15 cm P rate was increased from 10 to 40 kg P/ha (extra cost of P is $164/ha at $5.48/kg). In this research the initial 0-10 cm Colwell P was 19 mg P/kg soil. More recent research in central QLD that examined 14 site years found wheat (2) and mungbean (1) did not respond to deep P (20-25 cm) applied at 10 kg P/ha, while chickpea (6) and sorghum (5) did, with an average yield increase of 0.41 t/ha and 0.74 t/ha respectively (Sands et al. 2022). Deep P responses at one site were reported in every year for five consecutive years from an initial deep P addition of 10 kg P/ha. At all sites Colwell P was reported as low (<20 mg P/kg) in the 0-10 cm layer and <8 mg P/kg soil in the 10-30 cm layer (Sands et al. 2022). The combined evidence from central QLD clearly shows a high probability of grain yield response from deep P placement where the initial soil Colwell P is <20 mg P/kg soil and the rainfall pattern is summer dominant. The summer dominant rainfall distribution is likely to leave both winter and summer crops with limited access to immobile nutrients in the 0-10 cm layer because of high surface soil water evaporation rates (relevant to summer crops) and low amounts (mm) of in-season winter crop rainfall.

Sandral et al. (2019) highlighted possible reasons why some site year combinations may also show deep P responses in more southerly cropping regions of Australia, and this is supported in part by surface nutrient stratification (Armstrong et al. 2015). In more southern grain growing regions most paddocks have a positive partial P balance (Norton and vanderMark 2016) and therefore often have Colwell P values (0-10 cm) much greater than those reported in the central QLD research. In addition, more southerly grain growing areas receive more winter rainfall, providing a larger window of time where the soil surface moisture is adequate for crop P uptake. Consequently, deep P responses may occur at a lower frequency or not at all in southern cropping regions.

This paper reports grain yield responses from southern NSW testing the effectiveness of different combinations of dual placement of shallow (~5-8 cm) and deep (~16-20 cm) banded P against shallow (~5-8 cm) only P placement.

Materials and methods

A trial was established at French Park, southern NSW in 2020. The soil is characterised by a pH of 6.7 (pH_water 0-10 cm) and pH tends to increase with depth (pH_water 7.9 at 10-30 cm and exceeds 9 below 30 cm). The site has a Colwell P of 49 mg/kg at 0-10 cm but only 4 mg/kg at 10-30cm.

At this site, the shallow banded P treatment consisted of 0, 10, 20 and 40 kg P/ha applied at ~5-8 cm depth in bands 0.25 m apart prior to seeding but aligned with the seeding rows (shallow band). For each deep P treatment (except 0), P was applied in the subsoil (~16-20 cm depth) at 0, 20, 30 and 40 kg P/ha in bands 0.5 m apart (deep band). A very high rate of shallow banded P (80 kg P/ha) was included to estimate the severity of P deficiency. Treatment 40⁺/40⁺ (shallow/deep kg P/ha) was included to cover any possible deficiencies in zinc, sulphur and copper (+ signifies the addition of zinc, sulphur and copper). Phosphorus was applied as MAP and balanced for nitrogen (using urea) in all treatments. All treatments were disturbed to ~20 cm depths to account for any apparent ripping effect and an additional control 0/0 undisturbed (no ripping) was included. At this site all the treatments were duplicated to include a cereal and a legume each year.

In 2020 trials were sown to wheat (cv. Sunblade CL Plus) on 15 May and faba beans (cv. PBA Nasma) on 27 April and harvested on 10 and 21 December, respectively. Rainfall was slightly higher than the historical average (365 mm compared with 325 mm long-term average) but the timing of that rainfall, particularly the late-winter/early-spring rainfall allowed the plants to convert biomass to yield.

In 2021, starter P was applied below seed rows at 5 kg/ha of P as MAP to all plots on May 7 and plots were harvested on 16 December. In 2021, lentils (cv. PBA Hallmark XT) replaced faba beans and were sown over the 2020 wheat trial while the 2021 wheat was sown over the 2020 faba bean trial. Pre-sowing Colwell P for the 0-10 cm layer was 25 mg/kg. The site received well above average annual rainfall (855 mm), with 344 mm of rain during the growing season.

In 2022, the original undisturbed 0/0 treatment was treated with a shallow band of P at the maximum rate used in the initial year of the trial (i.e., 80 kg P/ha) to examine fresh vs. residual P effects. The remaining wheat plots were balanced for nitrogen only. However, for the lentil trial, the additional P was applied as triple super phosphate (TSP) to avoid a high rate of N application from MAP. Pre-sowing Colwell P for the 0-10 cm layer was 17 mg/kg. The trials were rotated and sown to wheat (cv. Sunblade CL Plus) and lentils (cv. PBA Hallmark XT) on 3 May 2022. The site received well above average annual rainfall of 1060 mm with 553 mm of rain during the April to October growing season. Although lentils had a good start, higher than average rainfall, especially during spring severely impacted the crop and this trial was lost to waterlogging stress.

Results

Wheat 2020

Wheat yield in 2020 was increased by 25% (to an average of 8.3 t/ha) for 20/40, 40/20, 40/40, 40⁺/40⁺ (shallow/deep kg P/ha) compared to the 0/0 control (6.6 t/ha) (Figure 1) and these combinations (8.3t/ha) were also significantly higher than the 0UD/0 (6.9t/ha) and 10/0 (7.4t/ha) treatments. Shallow P of 40/0 and 80/0 kg P/ha provided an average yield of 8.0 t/ha which was also significantly greater than the control 0/0 (6.6 t/ha). However, dual placed P (shallow + deep P) treatments did not outperform shallow only 40/0 or 80/0 (shallow/deep kg P/ha) and consequently there were no deep P responses.

Faba bean 2020

Faba bean yield was increased significantly when a total of 30 kg P/ha was applied, irrespective of placement. There was no difference between shallow and deep placement (Figure 1).

Lentils and wheat 2021

In 2021, phosphorus application had no effect on lentil yield (data not presented). For wheat, the 10/30, 40/40 and 40⁺/40⁺ (shallow/deep kg P/ha) treatments showed a significantly higher grain yield (average of 6.6 t/ha) than the 0/0 P (5.4 t/ha, Figure 1). Neither 10/30 nor 40/40 (shallow/deep kg P/ha) had a significantly greater yield than the corresponding shallow placement treatments of 40/0 and 80/0 (shallow/deep kg P/ha).

Wheat 2022

In 2022, the fresh application of 80 kg P/ha doubled (6 t/ha) grain yield of wheat compared to 0/0 P (3.1 t/ha, Figure 1). The dual application of 10/30 (4.5 t/ha) yielded higher than the corresponding comparison of 40/0 (3.6 t/ha) indicating an advantage for dual placement. The yield difference of 0.9 t/ha approximated $288/ha assuming a wheat price of $320/t. However, the 40/40 (4.4 t/ha) treatment did not yield significantly higher than its corresponding comparison of 80/0 (4.1 t/ha).

French Park summary

Out of five datasets from the French Park site, four showed positive responses to applied P (note: unresponsive lentil yields for 2021 are not presented). Of the four datasets that showed positive P responses, one (wheat 2022) showed a positive response to dual P placement compared with the same P rate banded shallow.

Shallow 10kg P/ha combinations

A closer examination of data examining shallow banded P at 10kg P/ha with various deep P combinations (Figure 2) showed a consistent trend of increasing grain yield across wheat in 2020, 2021 and 2022 as well as faba bean in 2020, when increasing the deep P rates from 10/0 to 10/20 (shallow/deep kg P/ha) and from 10/20 to 10/30 kg P/ha (Figure 2). In each of the four site year datasets, the 10/30 yielded significantly more than the 10/0 (shallow/deep P kg/ha). Adding additional deep P (e.g., 10/40 kg P/ha) did not increase yield in any species or year combination (Figure 2).

These findings may be useful in certain circumstances. For example, where seeding equipment does not separate P from seed and the crop species being sown is sensitive to P toxicity. The dual P strategy (e.g., deep P applied prior to sowing) allows for much safer rates of shallow P at sowing to avoid P toxicity where seed and fertiliser cannot be separated. Another practical example may include the sowing of faba beans where both the seed and fertiliser tanks in the airseeder are used for seed (e.g., no fertiliser sown in the seeding operation). In this situation yields are likely to be higher if deep P were available. In summary the results in Figure 2 show: (i) deep P has a residual value that lasts at least to year three of the cropping phase and (ii) the highest biological response was obtained with 30 kg/ha of deep P where a shallow 10 kg P/ha was used. On an annualised basis, the 10/30 treatment received an overall input of 15 kg P/ha/yr which is equivalent to an annual input of 68 kg/ha of MAP.