Mapping our food from space will soon be the norm. Scientists from the University of New England (UNE) are testing satellite technologies as a way to track the health and growth of tropical tree crops, including avocados, mangoes, macadamia and bananas.
The scientists – Andrew Robson, Moshiur Rahman and Jasmine Muir, from the Agricultural Remote Sensing Team (ARST) within the university’s Precision Agricultural Research Group (PARG) – are investigating whether satellite-based remote sensing technology can provide accurate measures of crop yield, fruit size and quality.
If successful, growers will have access to up-to-date parameter-specific maps within a growing season to help them identify areas performing poorly because of inadequate nutrition, pests, diseases, too much moisture or not enough management. This will allow them to better manage crop inputs and make more informed decisions regarding harvest scheduling and forward selling.
At present, yield forecasting of tree crops such as avocado is undertaken by counting the fruit of a small number of trees, then extrapolated across the entire farm – a process both labour-intensive and inaccurate. An initial evaluation of satellite imagery coupled with targeted field sampling has indicated this approach to be more accurate for both avocado and macadamia. Using satellite imagery allows differences in individual tree health to be clearly seen across an orchard. When calibrated with actual fruit yield parameters, the imagery can be converted into surrogate fruit yield, size and quality maps.
The UNE team, in collaboration with the University of Queensland, University of Sydney, Central Queensland University and Queensland Department of Agriculture and Fisheries, has been sampling mango, banana, macadamia and avocado orchards across four Australian states, as part of an ongoing project funded by the Federal Rural Research and Development For Profit scheme and Horticulture Innovation Australia.
The scientists use satellite imagery, as well as a number of other ground and airborne sensors, to measure the health or vigour of individual tree canopies via their spectral characteristics. From this information, measures such as the Normalised Difference Vegetation Index, a scale commonly used to determine the amount of live green vegetation in a given area, is used to select specific trees for targeted field sampling. Once sampled, the varying yield parameters are correlated with additional vegetation indices to identify that which produces the strongest relationship.
For avocados in Bundaberg, Queensland, the team identified a correlation between a number of vegetation indices and fruit weight, both as tree yield and for individual fruit. These results are being validated across other growing regions and across seasons.
“We are still investigating this relationship,” says Robson, “but the research indicates there is an association between how big or healthy a tree is and its ability to set and then fill fruit.”
Satellite sensing might also enable farmers to better determine the quality and maturity of fruit across an orchard. This will lead to greater efficiencies at harvest time.
At present, fruit-pickers undertake selective harvesting at the start of the harvest season to pick fruits of optimal size and quality, which attract highest prices. This is achieved by the pickers inspecting each piece of fruit and choosing only those ripe and ready. In some cases, entire regions of an orchard may only produce lower grade fruit, the harvest of which demands high costs of wages and fuel, but results in low returns.
Strong correlations between satellite imagery and fruit size, achieved over three seasons, indicate that “fruit size” maps can be derived. This information allows growers to adopt targeted harvesting to intensively pick only those areas of the orchard that are bearing large fruit.
Although there is still a great deal of research to validate these results, if confirmed the technologies investigated through the entire project have the potential to revolutionise the Australian tree cropping industry and potentially other agricultural sectors as well.