Smart Orchard 2.0

An ongoing project in Washington State puts technology to the test. By Anna Mouton.

The explosion of new technologies has been a mixed blessing for fruit growers. Never before have growers had access to so much data — and never before have they been so overwhelmed by options. How can they decide which tools will help them to increase the productivity and profitability of their orchards?

Tree-fruit physiologist Prof. Lee Kalcsits of the Washington State University Tree Fruit Research and Extension Center is part of a team that has been testing new technologies in apple orchards. “The project started in 2019 in response to industry demand,” said Kalcsits. “Growers wanted an unbiased evaluation that would improve the successful adoption of these technologies.”

The first trials were conducted in 2020 in a Gala orchard. Kalcsits came on board in 2021 when a medium-density Honeycrisp block was added — hence Smart Orchard 2.0.

Smart Orchard research is an industry-university partnership between the Washington Tree Fruit Research Commission and Washington State University. Two commercial growers provided access to their orchards. In addition to Kalcsits, the team includes extension specialists Bernardita Sallato and Dr Jenny Bolivar-Medina, irrigation engineer Prof. Troy Peters, and remote-sensing engineer Prof. Lav Khot, all of Washington State University.

Specialist agricultural data-science company innov8.ag led by Steve Mantle supply data-processing and machine-learning capabilities. “They look at data integration, so working across multiple software platforms and bringing it all together in one place on a platform where growers can view that data,” explained Kalcsits.

“A grower that uses a universal platform can use their data much easier than if they had five different software platforms and they’re moving between five different apps to try and decide what they want to do or how they want to manage their water.”

Mantle is also the connectivity specialist. “We have areas in the state where we have very poor cell phone and data connectivity,” said Kalcsits. This can be a challenge when sensors need to deliver data to smartphones or computer networks.

The big-data toolbox

“There are lots of tools and some of them are better than others,” commented Kalcsits. The advantage of the Smart Orchard project is that it allows side-by-side comparison of different technologies in the same orchard. “We’ve partnered with a lot of the sensor companies. If a company wants to put their product against things that are tried and tested, this is the place to do it.”

Products for soil-mapping technology are among those under evaluation. Sallato and Khot have been combining drone and satellite imaging with electroconductivity mapping and SoilOptix. SoilOptix uses a gamma radiation sensor to map individual nutrients in the soil while electroconductivity mapping provides information about soil texture and drainage.

Soil data generated by different sensors were assessed in relation to tree vigour. Maps of soil electroconductivity showed how trees were excessively vigorous in some wetter areas of the block. But Kalcsits cautioned that sometimes certain sensors also identified differences that had no measurable impact on plant response.

Sensors that assist with irrigation management have also proliferated. Remote sensing can gather data on temperature and evapotranspiration while soil-based sensors record moisture and plant-based sensors monitor stem water potential. Whereas satellite imagery shows crop water use at coarse scales — several trees per pixel — drone-based imagery records leaf-level variation in crop water use.

Detailed maps combining soil water-holding capacity and tree water use to reflect spatial variability in water use can change how growers view irrigation management. However, Kalcsits reminded the audience to focus on practicalities when placing sensors. “If your management unit is your orchard block, there’s no point in putting in more than one sensor. If you want to split the block into different units, then you need multiple sensors, and it becomes much more expensive.”

Kalcsits and postdoctoral fellow Dr Victor Blanco are particularly interested in plant-based sensors such as dendrometers and microtensiometers. Microtensiometers are installed on the tree and measure stem water potential continuously. Older methods for measuring stem water potential involved pressure chambers and could only generate discrete data points.

Stem water potential is correlated with plant water status and fruit growth. Kalcsits believes that microtensiometers hold promise for improving water management. “It’s not the cheapest technology but the cost is in line with other irrigation management sensors that are already on the market.”

Be sensible about sensors

The number of sensors on the market increases daily. Kalcsits had some advice for growers trying to decide whether to adopt a new product. “You want to make sure you’re getting something that’s tested and peer-reviewed.” That includes establishing whether the underlying methodology of the sensor is grounded in good science.

Kalcsits adds that growers should be wary of data ownership. “Long-term knowledge of your blocks and how those trees are performing is powerful. Some companies out there are subscription-based — you could get stuck because you don’t want to let go of your data.” Sensors should ideally also be compatible with different software platforms.

It seems that talk about data is everywhere these days. But data is only useful when it is converted into information that supports decision-making. The Smart Orchard project shows how the data deluge can be channelled into supporting better productivity and profitability for tree-fruit growers.

“When you adopt something into your operations, it comes down to thinking about how the data lead to a decision that improves fruit quality and productivity long-term,” said Kalcsits.

Some of the Smart Orchard trials are coming to an end and Kalcsits hopes that new technologies will continue to be tested in future. “A big part of what we’re moving toward now is integrating layers of data — we have all this data on the soil, the environment, the spatial variability within the block, and what’s going on in the plant. We’re working with computer scientists that do a lot of artificial intelligence work to see if we can take all these really intensive layers of data and develop more useful and reliable models for the industry.”

Image: Dr Lee Kalcsits, Washington State University Tree Fruit Research and Extension Center. Supplied by Dr Lee Kalcsits.