Australia’s largest horticultural producer, the Costa Group, has contracted UNE’s Remote Sensing Centre to provide yield forecasts for its extensive citrus and avocado operations.
UNE’s Applied Agricultural Remote Sensing Centre (AARSC) was chosen by Costa after a three-year evaluation of varied technologies and approaches to yield forecasting.
“We recognise the depth and strength of AARSC’s data science team,” says Jesse Reader, Costa’s Commercial Manager – Ag Technology & Innovation.
“I’ve worked with lots of different people, and it stands out that the AARSC team are very practical, they understand what we do and what we need, and they give us the confidence to proceed based on empirical evidence.”
Mr Reader said AARSC’s approach also has other qualities that stood out, not least the quality of the Centre’s data. “The forecasting accuracy that AARSC was able to achieve was a step-change better than our existing methods.”
“And AARSC’s process allows us to be much less hands-on about yield forecasting. We don’t have to go out, sample trees and measure pieces of fruit with callipers. We ask for information and we get a quick, accurate response.”
Costa’s confidence in AARSC is a milestone for the Centre too, observes AARSC Director, Professor Andrew Robson.
“We’ve developed our tree crop mapping and yield forecasting methodologies in collaboration with Australian tree crop industry bodies, but the Costa contract shows that our work is being recognised by commercial operators as an adoptable and beneficial tool,” Prof. Robson says.
“After spending years of field sampling across the country to build our forecasting approaches, it is gratifying that the Centre’s yield forecasting accuracies have stood up when compared alongside other commercial solutions”.
“There has been increasing momentum in recent times to ensure that peer-reviewed research is transitioned into an adoptable tool, and this a really nice example of how this can be achieved. We at the Centre are very proud of this commercial engagement with Costa.”
For Costa, AARSC’s technologies change expectations of what’s needed to resource an important part of the operation.
Until now, Costa has had to produce citrus and avocado yield forecasts by measuring portions of each crop, and then scaling up the results to reflect an estimate of overall yield.
When dealing with thousands of hectares of tree crop, this manual process – involving, Mr Reader says, “looking at hundreds of blocks and putting callipers on thousands of pieces of fruit” – is money- and resource-intensive and lacks the spatial resolution required for the desired accuracy.
AARSC’s approach, using calibrated satellite imagery and historical data, all piped through some clever algorithms, requires few people and minimal time, yet delivers a much better result than the manual method.
Mr Reader says that yield forecasting is necessary not just to manage supply chain logistics during harvest. For a publicly listed business, as Costa is, accurate reporting to the market is vital.
“But beyond that reporting function, accurate yield forecasting is essential for good agronomic practice, and understanding what we're growing. Knowing the volume of the crop, and when it needs to come off, is critical to any horticultural business.”
Prof. Robson will manage the Costa project for AARSC, while Dr Moshiur Rahman will lead the avocado component and Dr Angelica Suarez will lead the citrus.