The development and application of functions describing pasture yield responses to phosphorus, potassium and sulfur in Australia using meta-data analysis and derived soil-test calibration relationships

Abstract

An improved ability to predict pasture dry matter (DM) yield response to applied phosphorus (P), potassium (K) and sulfur (S) is a crucial step in determining the production and economic benefits of fertiliser inputs and the environmental benefits associated with efficient nutrient use. The adoption and application of soil testing can make substantial improvements to nutrient use efficiency, but soil test interpretation needs to be based on the best available and most relevant experimental data. This paper reports on the development of improved national and regionally specific soil test–pasture yield response functions and critical soil test P, K and S values for near-maximum growth of improved pastures across Australia. A comprehensive dataset of pasture yield responses to fertiliser applications was collated from field experiments conducted in all improved pasture regions of Australia. The Better Fertiliser Decisions for Pastures (BFDP) database contains data from 3032 experiment sites, 21918 yield response measures and 5548 experiment site years. These data were converted to standard measurement units and compiled within a specifically designed relational database, where the data could be explored and interpreted. Key data included soil and site descriptions, pasture type, fertiliser type and rate, nutrient application rate, DM yield measures and soil test results (i.e. Olsen P, Colwell P, P buffering, Colwell K, Skene K, exchangeable K, CPC S, KCl S). These data were analysed, and quantitative non-linear mixed effects models based upon the Mitscherlich function were developed. Where appropriate, disparate datasets were integrated to derive the most appropriate response relationships for different soil texture and P buffering index classes, as well as interpretation at the regional, state, and national scale. Overall, the fitted models provided a good fit to the large body of data, using readily interpretable coefficients, but were at times limited by patchiness of meta-data and uneven representation of different soil types and regions. The models provided improved predictions of relative pasture yield response to soil nutrient status and can be scaled to absolute yield using a specified maximal yield by the user. Importantly, the response function exhibits diminishing returns, enabling marginal economic analysis and determination of optimum fertiliser application rate to a specific situation. These derived relationships form the basis of national standards for soil test interpretation and fertiliser recommendations for Australian pastures and grazing industries, and are incorporated within the major Australian fertiliser company decision support systems. However, the utility of the national database is limited without a contemporary web-based interface, like that developed for the Better Fertiliser Decisions for Cropping (BFDC) national database. An integrated approach between the BFDP and the BFDC would facilitate the interrogation of the database by advisors and farmers to generate yield response curves relevant to the region and/or pasture system of interest and provides the capacity to accommodate new data in the future.