A cost-effective approach to estimate plant available water capacity

Abstract

Accurately quantifying plant available water capacity (PAWC) of soil is important for maximising crop yields in dryland cropping systems, as PAWC positively correlates with crop yields. However, direct measurement of PAWC can be challenging due to the associated cost and complexities. We aimed to predict PAWC accurately across a wide range of soil and climatic conditions cost effectively by utilising readily available data accessible to grain growers (e.g., public soil and weather databases, crop yield history, and management records). We designed a PAWC prediction framework akin to an inverse modelling approach using the Agricultural Production Systems sIMulator Next Generation (APSIM Next Generation), using crop yields, farm management details, and weather information as inputs. In addition, we analysed the sensitivity of the prediction framework to different sources of climate and soil data. This was achieved using different combinations of national and international soil and weather data grids. The proposed PAWC prediction framework adequately predicts PAWC (R2 = 0.75, CCC = 0.75, RMSE = 47.95, bias = -19.32 mm) across Australian dryland farming systems with varying soil and climatic conditions. The sensitivity test of the framework against source soil and weather data shows that high-quality input parameters improve its predictive ability. The proposed PAWC prediction framework offers a cost-effective method of estimating the PAWC, which can be useful for making agronomic decisions.

Keywords

• APSIM next generation
• Inverse modelling
• Plant available water capacity (PAWC)
• Soil-crop interactions
• Agricultural system modelling