In-season yield prediction using VARIwise

McCarthy, Alison and O'Keeffe, Kieran and McKay, Andrew (2019) In-season yield prediction using VARIwise. In: Australian Association of Cotton Scientists 2019 Australian Cotton Research Conference (AACS 2019), 28-30 Oct, 2019, Armidale, Australia.

[img]
Preview
Text (Presentation slides)
CottonResearchConference2019_USQ-CottonInfo_AlisonMcCarthy.pdf

Download (1MB) | Preview
[img]
Preview
Text (Accepted Version)
CottonResearch2019abstract_VARIwise.pdf

Download (47kB) | Preview

Abstract

In-season yield prediction supports improved agronomic management and planning for crop sales and insurance contracts. Yield is currently often estimated using rules of thumb and manual boll counts. Data analytics approaches have been developed using site- and season specific multispectral satellite imagery-based correlations that require significant datasets for wider scale transferability. An alternative approach is to forecast yield using known soil-plant-atmosphere interactions in crop production models and calibrated using available field data. USQ has developed software ‘VARIwise’ to provide yield prediction throughout the season combining these models with:
(i) plant parameters extracted from UAV imagery using image analysis; (ii) online soil and weather data; and (iii) on-farm management information.

In the 2017/18 and 2018/19 seasons, VARIwise was evaluated at one cotton site in Goondiwindi and 16 sites in Griffith. Management zones in the field monitored using the UAV were identified from vegetation index surveys, yield maps or satellite images. Phantom 4 UAV imagery was collected monthly at each site between January and picking for calibrating the crop model. The sites had varying levels of fruit removal, hail damage and heat stress.

In the 2017/18 Griffith trial, the percentage yield prediction errors were 10.2% in January, 6.0% in February, 2.5% in March, and 0.5% at picking, and in the 2018/19 Griffith trial the errors were 18.8% in January, 4.9% in February, 9.5% in March, and 10.1% at picking. In the 2018/19 Goondiwindi trial, the yield prediction percentage errors were 8.7% in February, 5.9% in March, 7.1% in April and 2.6% in May. The prediction errors at Griffith were higher in the 2018/19 season than the 2017/18 season because the sites experienced hail and heat stress that are not currently represented within the VARIwise crop model. The yield predictor will be evaluated in 2019/20 to improve performance under insect and hail damage.


Statistics for USQ ePrint 37859
Statistics for this ePrint Item
Item Type: Conference or Workshop Item (Commonwealth Reporting Category E) (Keynote)
Refereed: No
Item Status: Live Archive
Faculty/School / Institute/Centre: Current - Institute for Advanced Engineering and Space Sciences - Centre for Agricultural Engineering (1 Aug 2018 -)
Faculty/School / Institute/Centre: Current - Institute for Advanced Engineering and Space Sciences - Centre for Agricultural Engineering (1 Aug 2018 -)
Date Deposited: 14 Aug 2020 05:12
Last Modified: 08 Oct 2020 06:17
Uncontrolled Keywords: agronomic management and planning; crop sales; in-season yield; insurance contracts; VARIwise
Fields of Research (2008): 07 Agricultural and Veterinary Sciences > 0701 Agriculture, Land and Farm Management > 070104 Agricultural Spatial Analysis and Modelling
09 Engineering > 0906 Electrical and Electronic Engineering > 090602 Control Systems, Robotics and Automation
07 Agricultural and Veterinary Sciences > 0703 Crop and Pasture Production > 070302 Agronomy
Socio-Economic Objectives (2008): D Environment > 96 Environment > 9609 Land and Water Management > 960905 Farmland, Arable Cropland and Permanent Cropland Water Management
URI: http://eprints.usq.edu.au/id/eprint/37859

Actions (login required)

View Item Archive Repository Staff Only