When does potassium application pay? Three seasons of WMG trial results

By Simon Kruger, WMG Project Communications Officer

Potassium management has been a recurring question for growers across the West Midlands region, particularly on the coarse-textured sandy soils that dominate much of the cropping area. Growers and advisers have long suspected that fertiliser applications were falling short of what crops remove at harvest, but surface soil testing alone was not providing a reliable picture of availability across the full profile.

Between 2023 and 2025, WMG investigated potassium management across the region through a combination of paddock-scale nutrient budgeting, replicated field trials and a demonstration site, supported by GRDC investment WMG2304-001SAX. This article summarises the key findings and what they mean for fertiliser decisions heading into the 2026 season.

What the 2023 survey found

In 2023, potassium budgets were completed across 10 paddocks spanning the West Midlands region from Regans Ford to Coorow, covering a range of soil types from deep yellow sand through to heavy clay.

Plant tissue testing at GS30 found 70% of crops were at sufficient potassium status at the time of sampling. However, in-season fertiliser applications fell short of whole-crop potassium requirements at all 10 sites. Using a crop requirement of 16 kg of potassium per tonne of grain removed, the gap between applied potassium and total crop requirement was present across every soil type and location surveyed. Growers were relying on potassium recycled through stubble and drawn from subsoil reserves to bridge that gap.

Soil testing to 90 cm confirmed substantial subsoil reserves at most sites, with cumulative potassium in the top 90 cm ranging from 206 kg/ha on a deep yellow sand at Regans Ford to 1,059 kg/ha on a dark brown sand at Dandaragan. The majority of that potassium was located below 30 cm, a finding that surprised many participating growers who had rarely tested beyond the surface layer.

Compaction was identified as a significant constraint to root access, with penetrometer readings exceeding the threshold associated with restricted root growth at 80% of sites, generally between 30 and 50 cm depth. Where compaction limits rooting depth, crop access to subsoil potassium reserves is restricted regardless of how much potassium is present at depth.

Dandaragan 2024: does deep ripping pay?

A replicated trial at Lachie Brown’s property near Dandaragan tested whether deep ripping to 60 cm could improve crop access to subsoil potassium reserves across five species: mid-season wheat (Tomahawk), short-season wheat (Anvil CL), canola (Blazer TT), lupin (Jurien) and serradella (Cadiz). Ripping was conducted post-emergent, approximately two to three days after seeding.

Ripping substantially reduced soil strength, with the depth at which resistance became limiting to root growth shifting from approximately 225 mm in non-ripped plots to 650 mm in ripped plots. A strong relationship between biomass production and potassium uptake was confirmed across all species and treatments (R² = 0.81), confirming that maximising biomass is the primary pathway to cycling potassium from depth. Ripped treatments accumulated approximately 20 kg/ha more potassium in plant biomass than non-ripped treatments.

Grain yield was significantly higher in ripped plots for both wheat types. Mid-season wheat yielded 3.35 t/ha ripped versus 2.85 t/ha non-ripped. Short-season wheat yielded 2.76 t/ha ripped versus 2.28 t/ha non-ripped. Canola, lupin and serradella did not show statistically significant yield differences between treatments, though ripped plots trended higher for most species.

Crop	No-rip (t/ha)	Ripped (t/ha)	Difference (t/ha)
Mid-season wheat (Tomahawk)	2.85	3.35	+0.49*
Short-season wheat (Anvil CL)	2.28	2.76	+0.48*
Canola (Blazer TT)	0.90	1.17	+0.28
Lupin (Jurien)	2.28	2.33	+0.05
Serradella (Cadiz)	0.35	0.31	-0.03

* Statistically significant. Dandaragan 2024. Bird damage noted on first range of ripped plots.

The economics of deep ripping were favourable for wheat and canola under all grain price scenarios tested. For mid-season wheat, the 0.49 t/ha yield benefit returned a net single-season gain of $58–108/ha depending on grain price, after deducting the $80/ha ripping cost. When spread across three years, the annual cost burden falls to around $27/ha and net returns range from $111–161/ha per year. Canola returned a positive single-season result at $650/t. Lupin did not recover the ripping cost in year one.

A few qualifications are worth noting. The post-emergent ripping technique reduced plant establishment across several species. In the 2024 season, crops compensated and still produced higher yields. In a difficult establishment year, that compensation may not occur. The amortisation assumption of three years may also not hold on highly dynamic sandy soils under wheeled traffic, where re-compaction can occur sooner.

Badgingarra 2024: rate, placement and product

A replicated trial at Rohan Marriott’s property near Badgingarra tested potassium fertiliser rates of 15, 30, 45 and 75 kg/ha of product, applied either banded below the seed or with the seed, using both muriate of potash (MOP) and sulphate of potash (SOP) in a wheat crop.

K30 MOP banded below seed was the economically optimal treatment, returning $122/ha above the nil K treatment at $330/t APW1. This finding was consistent across the original 2024 harvest prices and all three price scenarios tested.

Higher rates did not improve the outcome. K45 returned a slightly lower net return than K30 despite similar yield, because the additional fertiliser cost was not recovered through yield gain. K75 consistently returned below the nil K treatment at all grain price scenarios after fertiliser cost was accounted for.

No statistically significant effect of placement or product was detected at any rate. Given SOP costs approximately $400/t more than MOP, there is no economic basis for using SOP at this site unless specific sulphur or chloride sensitivities exist. The lowest-cost delivery method, MOP banded below seed, was also the most economically effective.

With-seed application reduced plant establishment at higher rates. Crops compensated and final yields were not significantly affected in 2024, but banded below placement avoids this risk entirely and is recommended where agronomy allows.

Ballidu 2025: broadcast K on contrasting soils

A paddock-scale demonstration at Corey Mincherton’s property near Ballidu compared broadcast MOP rates across two adjacent paddocks with contrasting soil types: a yellow sand and a clay loam. All treatments also received 15 kg/ha MOP banded at seeding.

The two sites produced markedly different responses.

On the yellow sand, yield increased from 2.98 t/ha in the control to 4.03 t/ha at 75 kg/ha broadcast MOP, a lift of 1.05 t/ha. Most of that response was achieved at moderate rates. Yields at 37.5 and 50 kg/ha broadcast MOP were only marginally below the highest rate, indicating the response curve flattens above approximately 37.5–50 kg/ha. On a net return basis at $330/t, the 37.5 and 50 kg/ha rates returned $276 and $274/ha more than the control respectively, compared with $284/ha at 75 kg/ha.

On the clay loam, the control produced the highest yield at 3.92 t/ha. All broadcast K treatments returned lower yields, and broadcast K reduced net return at every rate and price scenario tested. At $330/t, the control returned $1,294/ha net compared with $1,053–1,234/ha for treated plots.

These results are from a single season at one demonstration site. Yield figures were derived from precision harvest map strips rather than replicated small plots, and treatment-level differences should be treated as indicative of directional response rather than statistically confirmed results.

The divergent response between soil types within the same property reinforces the case for variable-rate K application on paddocks with mixed texture. Applying a uniform broadcast rate across both soil types in the same season would have either under-applied on the yellow sand or over-applied on the clay loam, at economic cost either way.

Cross-site summary and practical implications

Site	Strategy tested	Best treatment	Net return vs baseline	Key qualification
Dandaragan	Deep ripping to 60 cm	Ripped mid-season wheat	+$83/ha single yr; +$136/ha amortised (at $330/t)	Post-emergent ripping applied. Multi-year data needed.
Badgingarra	K rate, placement, product	K30 MOP banded below	+$122/ha vs Nil K (at $330/t)	K75 returns below Nil K at all prices.
Ballidu — Yellow sand	Broadcast K rate	T4 by yield; T3/T5 similar economics	+$284/ha vs T1 (mid price)	Economics favour 37.5–50 kg/ha on risk-adjusted basis.
Ballidu — Clay loam	Broadcast K rate	T1 (control) — no K response	All K treatments below T1	No yield benefit from broadcast K in 2025.

Three practical messages emerge from across the three sites.

Soil type is the most important variable in potassium response. Coarse-textured sands are most likely to benefit from targeted potassium investment. Heavier soils may not require additional broadcast inputs in seasons where existing reserves meet crop demand. Testing to depth, rather than relying on surface Colwell K alone, provides a better basis for decision-making.

Moderate rates captured most of the available yield response at responsive sites, while high rates consistently failed to recover their cost. At Badgingarra, K75 returned below the nil K treatment at all price scenarios. At Ballidu, the difference in net return between 37.5 and 75 kg/ha broadcast MOP on the yellow sand was around $8/ha.

Physical constraints matter as much as fertiliser rates on many West Midlands soils. Where compaction is restricting root access below 30–50 cm, investing in higher fertiliser rates may not deliver a return until the physical constraint is addressed.

The economic analysis used a MOP price of $820/t. Unlike urea and DAP, MOP supply is not heavily concentrated through Middle Eastern export routes, and potash pricing has been less directly affected by the current conflict. Growers should confirm current prices with their supplier before applying these figures to their own operations, as local pricing varies by supplier, location and purchase timing.

Acknowledgements

This article draws on research conducted under GRDC investment WMG2304-001SAX – the K Extension Project. WMG acknowledges Lachie Brown (Dandaragan), Rohan Marriott (Badgingarra) and Corey Mincherton (Ballidu) for hosting trial and demonstration sites and for their ongoing involvement in the K Extension Project.