Ms Si Yang Han1, Prof. Thomas Bishop1, Dr Patrick Filippi1
1The University of Sydney
Biography:
Si Yang is a Postdoctoral Research Associate at the Precision Agriculture Lab at the University of Sydney. Research interests include validating existing spatial products and modelling on-farm management and yield at large extents.
Abstract:
Frost damage in winter grains significantly impacts grower profit and is an obstacle to global food security. Although frost events cannot be prevented, the severity or risk of damage can be substantially reduced with management decisions around crop types, crop varieties and the timing of sowing. While most growers have a general idea of the frost prone areas on their farm, a map of frost damage to crop yield that clearly identifies the location and magnitude can be a valuable guide at sowing. Frost damage can be observed in end-of-season yield maps. Given the widespread adoption of yield monitors in Australian winter grains and the abundance of publicly available spatial data, this study showcases the opportunity to use data-driven yield models to create historical maps of yield lost to frost damage as a pre-season decision-support tool.
Two approaches are outlined to model frost damage: 1) predicting yield pre-frost then post-frost, and 2) manipulating frost variables to have an actual yield prediction (with frost) and an artificial yield prediction (without frost). The differences between these pairs of yield predictions reflects the effect of frost in the model, which in essence is frost damage. Validating exact frost damage is difficult, however, the first approach better reflected the grower’s estimate in terms of location and magnitude. Future directions include collecting more frost validation data and using downscaled land surface temperature from satellites to improve grain yield and frost damage predictions at within-field level.