Narrow orchard systems for future climates (AS22002)

Video transcript

The second dot point there is the objective for our Narrow Orchard Systems project broken down into five key points there. And and it's all really, focused on making sure our orchards are safer to attract a labour force, which of course has been a real big problem the last few years more, more profitable, uniform and there's accelerated adoption of robotic and sensing technology. So I've always sort of seen this as being part and parcel with the the work we're doing with respect to narrow orchard systems. So it's not just about tree training, but it's also about the technology that is appropriate for what we call narrow orchard systems. And of course, the final point there is about climate resilience. and sustainability.

It's a five year project that commenced in June this year and it's actually more than five years, about 5. 5. The usual sort of carry over into a final, an extra financial year in 2028 for the delivery of the final report, which is great having a project of this sort of duration. We hope to be able to at least get some pretty good bits of work done.

The project team, so our little partners and team, so yeah, as Marguerite said, AgVic leads the project. DPIRD Western Australia is a partner, University of Queensland, New South Wales DPI, SARDI, and Plant and Food Research are all partners in the project and I've listed the people are in each one of the agencies that are working on this project.

And there's obviously people in this webinar whose names appear there. But I thought this photo captures pretty much all of the key people involved in the project. And most of you, if you, I'll go from left to right, that's Ken Breen from Plant and Food. He was our host when we did a recent trip to the site with doing that the FOPS project. Next to Ken is Mark O'Connell, Steele Jacob, Alessio Scalisi, myself, Darren Graetz from SARDI. And that's Richard Oliver. Richard Oliver's in Sensing Guy based in Hamilton in New Zealand. Is that Liqi there? Yeah, and Liqi, the second last guy is Liqi so, Liqi Han, he's University of Queensland, so they're a partner in the project, and Kevin Dodds on the far right from New South Wales DPI. So yeah, I think that's quite a nice photo to show, some of the key players in the project. And in the background, of course, is the narrow row, I shouldn't call it that. It's the FOPS experiment that Plant and Food Research have been doing for quite a number of years.

So yeah, I thought to, at the start of this presentation, I'd just give a bit of a definition of what we mean by narrow orchard systems. And it's really our starting point for the work that we will be doing. So, I've got there four key points. So, first of all, narrow orchard systems is about 2D, narrow hedgerow canopies, what are often referred to as a fruiting wall. It's about narrow road spacings. In other words, two metre row spacings. It's about widely spaced trees within a row. So, not a high density, ultra high density down the row but actually having trees spaced at two to three metres between the trees. And then the what I'd call the training systems. The training system is about multiple vertical leaders arising what is referred to, arising from what is referred to as a cordon, and that expression is quite commonly used in viticulture, as a training system. We're basically laying a horizontal structure, limb, in other words, from which vertical leaders arise, and you're looking at, six to ten vertical leaders arising from that cordon.

So that's basically the starting point of what our Narrow Orchard Systems is all about, and we'll be doing work to look at things like rootstocks and how you can actually train to a cordon system. We'll be looking at obviously different crops and of course in different environments as well.

And with respect to a bit more background, there's already obviously work being done, not only in Australia, but overseas. So this is, shows some of the background work that's already been done with respect to Cordon systems in Australia. So, we've been working on using Cordon on, in our in our pear experiment here at Tatura. Those photos there on the left hand side, the two photos there are firstly the Tatura trellis or open Tatura trellis and a vertical system using a six, well, it's an eight leader on the open Tatura and a six litre on the vertical training system. Directly underneath them is a block of plums from a commercial guy at Swan Hill who's been using a cordon system for plums. Next to that is a photo of cherries. A cordon system on a traditional v, traditional Tatura trellis V system where there's, somewhere between six and eight vertical leaders arising from the cordon. And then above that it's just showing the, the, basically the a V system of nectarine and a vertical system of nectarine where I think Mark O'Connell was playing around a bit with using a cordon system on nectarines.

So there's a bit of background already been, in, in terms of a cordon system in Australia. In the United States, there's often referred to as a UFO system, which has been around for quite a while and again, it's, obviously a planar type canopy using a cordon system where the trees actually leaned over and, the uprights arise from that main stem where the, where the cordons laid at a bit of an angle to prevent the end, to prevent the dominance of the vertical leaders arising from close to the the trunk. But yeah, again, that's on cherries. That's Greg Lang, the guy there, who's obviously done a lot of work and along with Matt Whiting on this system of training trees into planar systems using a cordon. And then probably the, the most similar work to what, we originally what we are planning is doing as part of this project is, work being done in Italy on what they refer to as the Guyot system I've added narrow road pedestrian Guyot because, the photo there on the right is a two metre row spacing and trees at two metre height, and the Guyot training system is where you actually have trees from a nursery that have already got feathers on them. As this, as the middle photo there shows a guy, training the the vertical leaders. That's a tree that's just been planted. So, the Guyot system is a registered trademark, by the way of having trees from a nursery that were already feathered in the right position, so you can just basically have a those vertical leaders straight off the shelf, so to speak. The photo there on the left too is an interesting one, because ,and Dario will be doing a bit of this work in Western Australia where the cordons are actually arched over each other to stop them from, when you're trying to bend over cordons, you can often crack them or break them. So, that method there is being used to try and avoid that from happening. That, I think, is referred to as a double Guyot system. And, of course, the work that's been done in New Zealand as part of their future production systems project. I've, I've added a few more words to the description of it because I think it's a narrow row, tall planar cordon system.

So, where they landed is and there's a photo there of Mark O'Connell by the way with his arms outstretched in the, one of the treatments that was a 1. 5 metre row spacing. But I think where they landed is, the preference for a 2 meter row spacing , and the critical thing from the bit of work that was done in New Zealand, and this was probably the impetus to, for us to, do the work in Australia and obviously get the funding from Hort Innovation, was that yield response function that's shown there with light as per, Stuart Tustin's publication in Scientia Horticulturae, you can see that, you're getting yields some of those dots are getting up around, 200 tonne per hectare. And that is well, that's exceptionally high yields. And a light, corresponding to a light interception of, getting up towards 80%. Obviously with a narrow row, tall trees the amount of light that can be intercepted is obviously getting very high and associated that with the work that I've done is, very high yields.

So what are we going to do? This slide here is lists of sites where we're doing both setting up experiments as well as some demonstration sites. Here in Tatura, we'll have both a reasonably large experiment as well as a demonstration block. So we'll be working on pear, apple, plum, nectarine and cherry. The table there just gives you, I won't go through them, but it gives some detail of the different rootstocks that we'll be doing in this experiment. So basically every crop has got three different rootstocks, and I think rootstocks are going to be critical for narrow orchard systems to work in Australia, we're probably more concerned about lack of vigour with some of these dwarfing rootstocks and not being able to fill the canopy, as opposed to, environments like in New Zealand and even the United States, where they've got much deeper soils, they still have to try to contain the vegetative vigour on some of these dwarfing rootstocks.

So yeah, it, we're thinking that we might actually not be able to fill the actual allocated canopy space with in our environment with some of these rootstocks. Manjimup, focusing on apples, Loxton, apricot and looking at the combination of a few different cultivars as well as different rootstocks. The guys at SARDI will also, have also set up an experiment in Adelaide Hills on a, commercial property, Global orchards on cherry, where they'll be looking at not necessarily, it won't be a, a two meter row spacing. I think it's a three meter row spacing in the orchard, but there's a comparison between a vertical UFO system versus a V system, but again, no planar, very narrow canopies, which is, of course, a significant component of this project, component of this project as well. And Kevin Dodds has been trying to establish a demonstration site at, in Batlow and he's got a guy up there who's just planting an orchard at the moment, three metre row spacing on a cordon system in other words, 2. 4 metres between the trees with eight leaders on each one of those trees, eight vertical leaders and aiming to get to a 3. 5 metre canopy height. So, Kevin's quite excited about this project, as we all are, but, Kevin in particular is, he's an extension officer and pretty keen to make sure that they have a significant contribution to make for their growers in this narrow orchard systems.

UQ are going to do studies on digital twin simulations, University of Queensland that is. Liqi is obviously the main person to do this work. So, he's really got two main components of work. The first one here is on orchard design to maximise lighter deception. So he's I'll just hit the down button to show a, a LiDAR scan of, trees that we actually did a few years ago at Tatura, but what, Liqi is going to be doing is first of all, calibrate and validate his light simulator model that he's got, which basically picks up on a digital twin that you create from a LiDAR image. Our first bit of work that we'll do is to make sure that Existing light simulator is is appropriate and works for, stone and palm fruit.

He's obviously done a lot of work in tropical fruits to develop the light simulator, and our job, first of all, is to make sure it works for our crops. So we'll be comparing what might, what gets generated from his simulation model using LiDAR scans of canopies of pome and stone fruit trees compared with our physical measures of light interception with the, for example, the light trolley shown in the photo there on the right. So that's the first sort of cab off the rank. And then the plan is to build a tool that can be used by industry to help best design an orchard for a particular environment so that it can maximize light interception, and of course, what I mean by light interception is a combination of total light interception by the tree, but also light distribution down through the canopy, because that's obviously critical.

And the third dot point there is that Liqi will do a bit of work on exploring the relationships between what gets simulated by the model in terms of light incidence on at a particular point within the canopy and fruit colour and floral initiation. So I'm quite excited about that last dot point because I think, I've got an opportunity here to use the technology, i. e. LiDAR and optical images to automatically capture this really large and rich data set that then we can look at establishing relationships between that light incident at any point in the canopy and what it's doing in terms of the productivity of the crop.

The other bit of work that Liqi's group will do is on sprayer designed for narrow orchard systems to maximize obviously spray efficiencies. So again, he's already got a model that he developed for tropical fruits and he'll validate that for pome and stone fruit, temperate crops in other words. Then the second dot point there is about designing a sprayer for narrow orchard systems that actually maximises the efficiency of application. So, the photos there are just giving you some idea of the different options of sprayers that, firstly, what we currently use on the far left there, that's our air blast sprayer we've got at Tatura. Secondly, that those over the row type sprayers that are common, well, not common, but they are used in viticulture. Thirdly, a very, simple sprayer that isn't an air blast. It's just basically using pressure to apply the spray. If we're talking about a thin canopy, maybe that's all we need. And the last sprayer there on the right it's a tower sprayer. That one there is only 2 metres high, but you can get them that are, I think, are about 3. 5 metre high tower sprayer. It's got a , I forgot the name of the fan on it, but it's like a vertical fan as a I think they're quite, that to me will be just off the top of my head, I think that's one of the preferences I'd be using for narrow orchard systems, a sprayer like that. The other as I said earlier, it's really critical that we bring along the ag tech and sensors that are appropriate and for narrow orchard systems, not only appropriate, but it actually could make narrow orchard systems work better, perform better, be more profitable, in other words. During the progress of this project will be testing and evaluating, demonstrating various bits of AgTech. And I've got in brackets there in that first dot point about it's going to be governed by the Project Reference Group. So that was Hort Innovations idea was to bring along the Project Reference Group with what AgTech might be available, but also have their inputs in turn, where the gaps in equipment might lie, and of course, one of the first things that the growers, when we've had discussions with them about is two metre row spacing. They're going to struggle to get a bin down the row. Well, so, you know, think about, well, what ag tech can we actually use to overcome a problem that would have been the lack of manoeuvrability of a bin being carted up and down a row? So anyway, here's some, some of the things that we've bounced around and got, ideas of what we might use, whether it's autonomous, something like a Burro which is shown in that photo on the left hand side there. There's a autonomous tractor as well with a tower sprayer. That's the tower sprayer I was mentioning before with the higher boom on the back. In the middle there is, it's doesn't, it's really hard to sort of see from that photo what it actually is, but that's a robotic harvester that's being developed by a company called Ripe Robotics in the Goulburn Valley, and their idea is to make the robotic harvester compact and small, which would be appropriate and suitable for a two meter row space orchard, narrow orchard systems in other words. And have, making them small enough and cheap enough that, of course, you can have multiple robotic harvesters. And the photo there on the right is the Smart Apply variable rate sprayer. That particular photo is the one we've got up at the Mildura Mid Area Smart Farm.

Ian, can I have a question? I guess that's Dave, is it? Yes, it is. Thank you. In in rainy Mildura, unfortunately. It's normally stinking hot, but it's not today. Are you guys going, do you have the equipment you need to manage your narrow rows? Or will you have to simulate some of the expected costs associated with this? Do you have narrow tractors and all the equipment you need?

Yeah, we do have narrow tractors, and the, even the air blast sprayer we've got will go down a two metre row. I'm pretty keen to progress the spray component of work of Liqi's because I want to purchase one of those, a sprayer that's more suitable to a narrow orchard system, right? In terms of, yeah, we've got, small enough slasher to go down the row. So we've got the equipment already to be able to manage the block, Dave. Oh yeah, thank you.

But it and this sort of leads on to this next up point in some ways, what you just brought up, Dave, because this one says, can existing orchard management technology actually be used in a narrow orchard system, right? And, and so we'll be looking at obviously we've, as I said, we've got the equipment already at Tatura to be able to, I'm not, I shouldn't answer for the others, other sites, by the way, maybe Tim can answer that question later, and I'm not sure if, Dario is out of the country at the moment, but, but it's not only that sort of obvious orchard equipment, but it's also things like, I've got examples here of a leaf blower and a Darwin flower thinner, and a hedger. So with those, you could easily argue, well you shouldn't need a lot of this equipment like a leaf blower and a hedger in a well managed narrow orchard system, a planar canopy, but growers are still cognizant as that mightn't be the case. They still might have to go up and down rows doing this sort of stuff, right? And particularly, I think of leaf blowing to allow better light through the canopy, and maybe if we do have more vigour than we anticipate, and we don't have enough light going through the canopy onto the adjoining row, that a leaf blower technology might have actually had to be used. So, the question I'm proposing here is, can this equipment be used in narrow orchard systems?

And of course, the technology to capture data. Here I've got, I think I gave a presentation about Cartographer a few years ago, but, we obviously still heavily rely on Cartographer to take a lot of our measures within our experiments, and we'll be continuing to do that. But of course, it also has a lot of applications for commercial orchard management as well. And other sensors as well, I've got there a trunk dendrometer. At the moment they're either going through a LoRaWAN system of communication or a proprietary type communication into the mobile system. So, there's various options we're currently using. But I think Dario in Western Australia is pretty keen to progress this a bit more to come up with a more universal, I'll call it that, system of capturing, some of this sensor data and also in the interpretation of the data too. And I've got there a graph of, that's a trunk dendrometer showing it's shrinkage and swelling over a period of time of about, what is that, approximately a week during November and there's still a necessary human intervention there in interpreting that data. And I think what Dario's got in mind is trying to use AI to interpret a graph like that to determine whether trees need to be irrigated or not.

And the last slide with, it's not the last slide I think, but it's the last slide on the ag tech is to look at a couple of other ag technologies, one being the ability to actually size fruit in the orchard and automated retractable netting. We're definitely going to look at automated retractable netting at Tatura and this particular image here is video at least. It shows it's actually an orchard in Ardmona close by here where they put in a bit of a trial to see whether they could use automated retractable netting systems because, with these netting systems, you will start losing out in colour and even potentially crop load and fruit size. So, the idea is with the retractable netting, automated retractable netting, is you can draw it over when the conditions are appropriate for excessive amount of radiation or a hail event, etc. The other photo on the left there is, we've been here at Tatura just trying to develop a, on our platform harvester, you can see that there's a box on one of the platform harvester conveyor belts, which has actually got sensors in it to measure fruit quality, including, fruit size and colour. And the proposal in this project is, Dario's group are going to pursue something similar to that, whether it's feasible to be able to do it in a narrow orchard systems, because that platform harvester there, as you can see, is a bit wide for going down a two metre row spacing.

The profitable NOS concept represents an innovative, tech-ready design that will be resilient to climate and market uncertainty. The project will provide knowledge that supports an industry transition to:

* Safer orchards that attract a labour force.

* More profitable orchards due to lower operating cost (e.g., energy, labour).

* Uniform orchards that consistently produce high yields of quality fruit.

* Accelerated adoption of robotic and sensing technology.

* Climate resilient and environmentally sustainable orchards.

Stone and pome fruit orchard design is at a crossroad. Dwarfing rootstocks have enabled orchards to be planted at high density but there is still debate over canopy architecture (i.e., a narrow wall 2D canopy or a wider 3D structure like a spindle) despite growing evidence from Europe that there are substantial yield and quality benefits from narrow 2D canopies. Furthermore, row width and resultant tree height is dictated by existing machinery and fruit bin size such that canopies must be high to get comparable yields. The alternative is to narrow the row spacing to 2 metres and keep the tree canopy at a height that can be managed from the ground without ladders or platforms. Such orchard design, referred to as a narrow orchard system (NOS), is attractive to field workers and amenable to robotics, automation, sensing and crop protection, and has the potential to use less labour, energy and increase the efficiency of inputs.

The overall objective of this project is to provide orchard designs to maximise yield, fruit quality, climate resilience and the efficient use of labour and resources for pome and stone fruits. We propose that a narrow row two-dimensional (2D) 'pedestrian' orchard will meet these requirements. Planting trees in narrow 2 metre rows with trees spaced at 2 metres along the row and maintaining a narrow 2D canopy will increase the evenness of light as it passes through the foliage to maximise fruit quality whilst keeping tree costs equivalent to current high-density planting systems. Restricting the tree height to approximately 2 metres by grafting to the best dwarfing rootstocks will enable easy access for labour and remove the use of ladders, which represent a hazard for orchard operation. In addition, gains in the accuracy and efficiency of sensing and robot orchard operations (crop load estimation, pruning, thinning, pest and disease monitoring, fruit picking, etc.) are anticipated in narrow row 2D pedestrian orchards compared to existing tree training systems. In addition to field experiments to test this hypothesis, the project will undertake an economic study to determine the profitability and payback period of NOS and compare the benefits and costs to current standard practice.

The project will establish narrow row 2D orchards at the Tatura SmartFarm for cherry, nectarine, plum, apple and pear. Part of the orchard will be dedicated to comparing the performance of different dwarfing rootstocks and tree canopy heights in designed experiments. The remainder of the orchard will be established to demonstrate resource use efficiency, robotics, sensors and data integration (e.g., robotic harvester, autonomous bin pickups, low horsepower electric machinery, variable rate spray technology, automated retractable netting, spatial monitoring of yield and fruit quality, plug-and-play internet-based wireless (i.e., IoT) sensors for irrigation and pest detection, fruit tracking) as well as existing machinery (e.g., string flower thinners, pneumatic defoliators, hedger).

The design of NOS in future climate, the investigation of light distribution in the canopy and its impact on fruit quality, the innovation of a new sprayer (featuring low horsepower, good adaptability to narrow environment and high precision) as well as other measures of orchard uniformity and productivity will be supported by a digital-twin research component involving LiDAR scanning, artificial intelligence and computing experts from the University of Queensland.

Two satellite NOS sites will be established in Western Australia and South Australia. A satellite experimental narrow row 2D apple orchard will be established at the Manjimup research station in WA. Appropriate rootstocks for WA and different options for training trees as a planar cordon will be investigated. The site in WA will explore autonomous vehicles, mechanical pruning and auto field grading technology. Similarly, an experimental narrow row 2D apricot orchard will be established at the Loxton Research Centre in SA. Cultivar and rootstock will be investigated. An additional experiment in a commercial 2D cherry orchard in Adelaide Hills SA will be established to investigate the effects of training system and cultivar. All sites will be used to communicate results from the project and to demonstrate sensors, tracking technologies and data integration.

The Tatura SmartFarm (Vic), Batlow (NSW), Applethorpe (Qld), Adelaide Hills (SA) and the Manjimup (WA) and Loxton (SA) research stations will be used for field walks and events to communicate the NOS concept and associated technology to the pome and stone fruit industries as well as to service providers, policy makers, scientists and students.

Video transcript

Hi, my name's Ian Goodwin. I'm from Agriculture Victoria. We're at the Tatura Smart Farm today in the Narrow Orchard Systems experimental block.

I'm in front of the cherries. It's the 18th of March today, and what I'll quickly demonstrate is the laying down of the cordon. These trees, a bit of background. These trees were planted in September as a rod. We cut them back at planting. Subsequently, we went through in late November, early December, removed the shoots that were coming out of the top of the tree because they're extremely vigorous and come out a very steep angle, and the shoots that are remaining here are what grew subsequently. So, to lay it down, I've already laid down this one on this side, as you can see. So I'm just going to lay down this shoot on the other side here, and just put a couple of clips on the tree to hold the cordon in position. I'll probably put in three because what might happen is that the end of this shoot will probably turn and start growing, growing up like that, which is fine because that could be our, the outer vertical leader from the cordon. There's a couple of shoots here too that I'll just take out that we don't want anymore, and there's a root stock that's coming up there, which I'll chop out as well, and that's it.

These are the nectarines. As you can see, this tree has got, multiple shoots coming off where we actually cut it back to the cordon wire after planting. We've already been through previously and cut some of these out above the cordon wire. And as you can see, now in March, they're strong enough and large enough to lay down. As I said, multiple shoots coming out here and we only want to pick two of them to lay down, which I'll show in a second. And also, compared with cherry, we've got a lot of these, what are called selectic side shoots coming off the main shoot going up here. But yeah, pretty simple to just bend this over this time of year without it breaking, and clipping it onto the wire, and I'll just put on a couple of clips.

and then yeah, there's quite a number of those selectics that I, I want just stub them back down the cordon, and then of course we select the vertical leaders from this cordon that'll be spaced at about 50 centimetres apart because of the nature of the nectarines, fruiting on lateral wood. And of course, towards the end of that those laterals. So, we'll space, these vertical leaders at 50 centimetres apart along this wire.

We are in the plum in the narrow orchard systems experiment. In this particular tree here again, we purchased them as a rod. They were cut off just above the cordon wire, and obviously multiple shoots have arisen. We did go through like the other crops in December and removed the shoots that were, above or close to that cordon wire. I've already laid down a couple of cordons here, and so these are the remaining shoots that we want to take out. So, I'll quickly, take these out without cutting off the cordon by mistake. There's a sucker there. Which we'll cut back. You can rip these out. The ones up the top there. Pretty simple. And of course, cutting back these shoots selectics that have risen off the cordon. I'm just stubbing them back for now. And next year of course we'll create the structure of the vertical leaders and space them out on the plums.

Now we're in the Apple in the Narrow Orchards experiment and I'm going to demonstrate, laying down the cordon. In this particular tree here it obviously again was, cut off when we when they were planted, and any shoots coming out of above that wire been broken off here, ripped out in other words, which is fine. But what I wanted to show here is how this particular, strong shoot coming up here, rather than try and lay it down that way, I'll actually lay it to avoid it cracking or splitting away from the main trunk. I'm going to lower it, lay it down so it's crossing over arching it like that so that so that it doesn't break away from the trunk of the tree. And again, just clipping it on. And then it's not long enough, but I'll just leave that shoot there. I'll lay down this one for you. It's pretty simple. This one here, just laying it down. Putting a second clip on here a fair way along towards it because in the next month or so, this will probably turn up a bit. If we had it, if we clipped it back here, it'd come start coming up too far back, which is not what we want. And then yeah, just taking out that other shoot here and just stubbing that back, and that's it.