Comparison of soil analytical methods for estimating plant-available potassium in highly weathered soils

Abstract

Context The NaHCO3-extractable soil K test (Colwell K) is the most frequently used method for soil K across Australia, but there is still uncertainty about the suitability of K testing guidelines for heavy-textured soils. Aims To determine whether Colwell K is suitable for predicting plant-available K on loam- and clay-textured soil types developed from highly weathered soil parent materials in Western Australia (WA). Methods Nine soil K testing methods were used to determine plant-available K on 21 soil types with a wide range of properties collected from three depths on the northern, central and southern farmlands in WA. Key results Quartz and K-feldspars were the dominant minerals in all soils, while limited white mica (1–2%) was identified in less than 10% of samples. The amount of K extracted by silver thiourea was only about 70% of the amounts extracted by NaHCO3 (Colwell K) and NH4OAc. Soil non-exchangeable K extracted by nitric acid and sodium tetraphenyl borate were from similar K pools, while aqua regia K was 1–6 times higher than these values. Conclusion There was no systemic difference or proportional bias between NH4OAc K and Colwell K, and both had good model performance (R2 > 0.86) for total K uptake by wheat in a single growing season in a wide range of soils. Implication In this diverse range of soils formed from weathered parent materials, the reliance on exchangeable K for plant K uptake suggests that plant-available K will be vulnerable to rundown due to negative K balances.