Document Type

Conference Proceeding

Publisher

Grain Industry Assoc. of WA

Faculty

Faculty of Health, Engineering and Science

School

School of Computer and Security Science/ECU Security Research Institute

RAS ID

18567

Comments

This article was originally published as: Neuhaus, A. , Armstrong, L. , Leng, J. , Diepeveen, D., & Anderson, G. (2014). Integrating soil and plant tissue tests and using an artificial intelligence method for data modelling is likely to improve decisions for in-season nitrogen management. Proceedings of Agribusiness Crop Updates. (pp. 1-7). Perth, W.A. Grain Industry Assoc. of WA. Original article available here

Abstract

This paper hypothesizes that there is value in combining soil, climate and plant tissue data to give more reliable advice on nitrogen top-ups in-season when compared with models that are currently available. The benefit of soil and climate data is to factor in N mineralisation and potential yield while plant test data is a more direct approach of yield estimates when considering firstly plant N uptake from the whole soil profile and secondly biomass (important yield component). Plant test data are closer to yield in time and space than soil test data, shortening the time period for any yield prognosis by about 2-3 months, depending when plant testing occurred. A positive side-effect of plant testing is to check whether any other nutrients, apart from nitrogen, are limiting yield or an N response. Secondly, this paper explores an AI method as a comparison to the traditional modelling technique to further improve the accuracy and to turn the model into a self-calibrating model. Unlike a statistical autoregression technique, the tested AI method has dynamic functions that can be used not only on time series data but also on data such as used here.

Access Rights

free_to_read

Share

 
COinS