Simulating wheat growth response to potassium availability under field conditions in sandy soils. II. Effect of subsurface potassium on grain yield response to potassium fertiliser

Subsurface potassium (K) supply can make a significant contribution to total K uptake in wheat (Triticum aestivum L.) although its influence on grain yield response to K fertiliser is unresolved. Previous work has shown that the inclusion of subsurface (>10cm depth) soil extractable K (SEK) did not improve the prediction of relative yield (RY) compared to a prediction based on SEK in the 0-10cm soil layer only. Our understanding of the influence of subsurface SEK is constrained by the incomplete nature of the interactions between season × surface SEK × subsurface SEK directly measured in field experiments. To understand these interactions, we simulated wheat growth for two locations in a rain fed environment in south-west Western Australia (SWWA) and two soil types using the crop growth simulation model APSIM, which has been calibrated for the sandy-surfaced soils of SWWA. Sensitivity analysis of the effect of subsurface SEK on grain yield showed that the effectiveness of subsurface SEK relative to surface SEK declined exponentially as the depth of the K-enriched subsurface layer increased. We implemented a Monte Carlo simulation for a deep sand and a sand over clay soil profile for a range of surface SEK levels, subsurface SEK depths, subsurface SEK levels, locations, years, subsurface root constraint and rates of K fertiliser applied. Global sensitivity analysis showed that SEK in the 0-10cm depth was the most important factor for RY in the deep sand and sand over clay profiles followed by SEK 10-20cm and location. We used the results from the Monte Carlo simulation to develop a K fertiliser recommendation model based on SEK 0-10cm only and a recommendation model based on SEK 0-10cm together with subsurface SEK, root constraint and stored soil water at sowing. A net economic benefit (change in income exceeds extra costs) only occurred in a limited number of scenarios where SEK 0-10cm was between 40 and 60mgkg-1 for the deep sand and where SEK 0-10cm was less than 40mgkg-1 for the sand over clay. The greatest potential for improvement in profit from K fertiliser recommendation systems for soils in SWWA is for sand over clay soils where SEK 0 to 10cm is less 40mgkg-1