Planting potatoes in a field at Van den Borne precision agriculture

Increase understanding of soil to make strides in precision ag

12 June 2024

Jacob van den Borne is Mr Precision Agriculture. Based in Reusel, he uses drones, sensors, satellites and robots to collect all sorts of data. And meanwhile, he also runs a large arable farming business. “If you don’t keep records, you can’t learn.” His farm is an important experimental location, both for WUR and the international smart farming and agri-tech community.

The little red book. Not Mao’s – it’s Jacob van den Borne’s grandfather’s. In it, he wrote in beautiful cursive handwriting what he had done every day on his land, along with anything else that had caught his eye. Which part of the land he had ploughed, how high the grain was or how much he had harvested, and how many millimetres of rainfall there had been that day. It was intended as something to learn from, says Van den Borne. “If a year later he noticed something unusual about a crop, he could look back and see what he had done in the field before that, or what had happened there.” You could think of it as old-school data. “If you walk across your field every day, you start seeing how things grow, why they’re failing to grow or why they’re not growing quite as you’d expected. You want to know why, and how you could improve things. If you don’t keep records, you can’t learn.”

Precision agriculture is not about technological tools

Most farmers don’t learn from what they have done in the past. They often do the same thing all their lives. That’s because they aren’t looking properly, and because they don’t want to learn, says Van den Borne. And that’s what lies at the heart of precision agriculture. Whenever he talks about precision agriculture, people immediately associate it with drones, sensors, robots and scans. But precision agriculture isn’t about technological tools. “You want to know what to do, and when and where to do it, in order to get a better crop and higher yields. I don’t use a little red book for that. I use data, obtained through technology.”

Dutch smart farmer Jacob van den Borne flying drone on arable field

Jacob van den Borne is ‘Mr. Precision Agriculture’

Van den Borne processes all that data in the Dacom management programme, using the Cloudfarm optional extra. Using those two tools, he can collate all his field locations, plot sizes, crop records and the work carried out. And if something goes wrong, he can review his records to avoid that happening in future. Data on fuel use, yields, crop varieties, and application maps are all collected and displayed directly in Cloudfarm. “That enables me to immediately create a prescription map for fertilising or spraying a plot.”

Van den Borne is Reusel’s Mr Precision Agriculture, and he sees that as something he just had to do. To understand why, you need to zoom in on a map of the Netherlands. Reusel is located in the Kempen region of Brabant. It’s a region of sandy soils, and in the past was home to farmers who barely had a pot to piss in. They grew some grain and crops grown for the canning/preserving industry , and that was it. Just like Van den Borne’s grandfather. Then, 70 years ago, the area began to be used for cattle and pig farming.

Chip factory

His father sensed a new opportunity. He heard that just across the border in Lommel, Belgium, there were plans to open a chip factory. They would be needing potatoes, and he would supply them. But there was a problem: you can only grow potatoes in the same soil once every four years. If you only have 20 hectares at your disposal, you can only use five of them each year. That’s not a great business model, so his father paid a visit to other farmers in the area to ask if he could plant potatoes on their land. By doing that, he expanded his potato empire from 20 to 300 hectares.

In 2006, Van den Borne took over his father’s farm. He had to make a choice: continue to scale up like his father had done, or increase the quality of his potatoes at a lower cost of production. He realised it would be difficult to expand any further in the local area, so he decided to focus on quality. That decision was based on what he had learned while studying at the HAS University of Applied Sciences in Den Bosch, which was that there was room for improvement in potato cultivation. To do it, he would need precision technology. It was nice for his father to have had all those small plots of about three hectares each, adding up to 180 fields spread across 120 farmers, with 80% of them in Belgium. But they weren’t his. And unlike his grandfather, he couldn’t go on a daily round of the fields to see for himself how his crops were growing and flourishing. “I needed some electronic vision for that,” says Van den Borne. I had no other choice.”

Management at the farm level

Four years after taking over, Van den Borne started practising precision agriculture. It was nothing new, he says. Scrutinising plants is something people were doing even in the very first agricultural settlements around 9000 BC. Which is logical, because if your crop isn’t okay, you’ve got nothing to eat and you’ll starve. “Over time, you can see how farmers increasingly drifted away from the actual plants themselves.” Today, almost 90% of all farmers manage their farms at the farm level. They have no idea how much they earn per plot or per zone, let alone per crop. Nor how much time and money they put into it. And that’s something he does want to know: not just at the plot level, but preferably also per zone or even more specifically, at the plant level. “We already have the technology to do that, but as yet, we don’t have the knowledge.”

Van den Borne has divided precision agriculture into 14 steps. From soil scanning, tillage and crop sensing to irrigation, variable fertilisation and storage. It starts in the winter when he prepares his field plans, he says. He inputs all the field boundaries into the on-board computer of his Fendt tractors. These are full of gadgets and can automatically detect any field when they enter it a few months later. While it’s great to have such a proactive computer, there’s a clear difference between that and his grandfather’s little red book: the human factor. His grandfather didn’t just note what he’d sprayed that day, he also looked around him. He would also note that just after he’d done his spraying, there had been 15 minutes of heavy rain. His tractor doesn’t know that, which is why Van den Borne has added a few more functions to the standard computer programme. When you’re driving off the field, the tractor asks: have you finished? Yes? Did you complete your task successfully, and what grade would you give the result? One, seven or 10 out of 10? Or should the work actually not have been done because the ground was too wet? “Using that assessment, I can decide whether or not to work with the dataset that’s been produced. Is the data actually representative of the task that’s been performed?” In other words: “When you start working with big data, filtering is the challenge. How do you know which data is good? You need to collect metadata from your data. Rating that data – adding the metadata to it – is the most important factor, because that’s what you really want to know.”

He shows a map that displays how much fuel a tractor has used for ploughing, turning the soil and sowing a particular potato plot. It needed between 9.4 and 13.2 litres of fuel to do this. Van den Borne can see on the map exactly where the tractor used more or less fuel. In one spot, it shows an unusually high amount. “So then you really need to ask the driver why exactly that is, before you draw any of your own conclusions.” In this case, it wasn’t a soil compaction problem, but something else. Someone else had mown a ditch and left the muck on the land, and it’s harder work for an engine to drive through that. “Data is good, but don’t forget the human factor. Grandfather’s notebook .”

Soil conductivity

Soil quality is the most important factor that determines his yields, says Van den Borne. He has tried every available technology to analyse his soil as best he can. His preferred tool is Dualem’s soil conductivity sensor. He performs his soil scans on a quad bike with a sled fixed to the rear. At the front, a Gamma-Ray sensor measures the level of clay and sedimentary rocks in the soil, and on the back the Dualem sensor takes EC measurements. Together, these tools provide a good picture of the conductivity of the soil. “You can compare soil conductivity to a battery,” says Van den Borne. The greater the conductivity, the greater the soil battery, and the more nutrients it can hold. And, in turn, the more that can be grown on that land. Many farmers take a soil sample to test this. ‘My potatoes need 200kg of nitrogen. There’s still 100kg left in the soil, so we’re going to sprinkle the whole plot with 100kg.’ But hold on: if you don’t know the soil conductivity of your land, then a section with a small battery could easily flood. That’s how you get emissions, the nitrogen leaching. “Precision agriculture enables you take a tailored approach. You can apply less to areas of soil that have a small battery, and more where there’s a large battery.”

Van den Borne tests available technology to analyse the soil, such as Dualem's soil conductivity sensor

Van den Borne performs soil scans on a quad bike with a sled fixed to the rear. Source: Van den Borne Aardappelen

Testing prototypes

Van den Borne maintains good relationships with equipment manufacturers. He tests their innovations, providing feedback on things like GeJo’s Smart Grader Reader, Soilmaster’s prototypes and Dualem’s soil scanners. He also used the first sprayer released by BB Leap. He felt the prototype was too focused on the plant level. His top tip is to start at the beginning. “It’s nice to do things at the plant level, but if you don’t already understand it at the plot level, you shouldn’t start at the plant level. You have to go from macro to micro.”

That’s important when it comes to precision agriculture. It’s not something farmers should just suddenly start doing. They need to first come up with an answer to the question of why they do what they do. “80% of all farmers can’t answer that question,” says Van den Borne. “They do things because they’ve always done them that way. To get what they have always gotten.” He knows people don’t like to change, and that it’s difficult. But the world is changing. Nature is changing. So as a farmer, you have to change too. “People often only start to change when they really have to. Then there’s the question of how long they can keep it up.”

According to Van den Borne, the further development of precision agriculture requires a better understanding of soils and plants. “You want to know how to make the best use of the soil.” A lot of research was done into this at Wageningen after WWII, with Sicco Mansholt at the helm. “He poured millions into ‘Wageningen’ in order to reform the European agricultural system.” At that time, it was all about growth models, land consolidation, mass mechanisation, fertiliser systems and, above all, production increases. The latter remains important, but it needs to be done sustainably. Everyone understands that. We need to move on from what is almost a chemical way of farming, with its excessive use of fertilisers and crop protection agents. “That means starting your research all over again,” says Van den Borne. “To see how you can change the agricultural system.”

Sampling and measuring

Van den Borne has already made a start on this, alongside the running of his enormous potato business. He uses every tool he can get his hands on to measure anything and everything, from the weather to the soil, from crop growth to photosynthesis. He uses sensors, tractors, drones and satellites. How big is the plant? How many leaves and stems does it have? What’s the colour and spectral colour? The temperature? Yields, size grade, dry matter and nitrate content? He really measures everything. All of that data gets entered into databases. “Since 2012, we’ve had a sampling programme like Wageningen had after the war. When we take samples, we gather a group of students and go into the field. Always at an average spot, so it’s most representative of the whole plot.” He now wants those average spots to be selected by AI, and also to let AI figure out exactly what the students should measure there. “I want to take all our data and get AI to train a model which farmers can then use to scan the soil themselves. The aim is to enable the farmer to maximise both product quality and sustainability. Products have to be sustainable, healthy and economically viable. It has to be a business model.”

Regional crop plans

As part of this, it’s also essential that farmers start working together. “Farmers all have their own crop plans,” says Van den Borne. “They’re all tinkering away on their own little plots. It’s much smarter for them to develop a collective crop plan, based on a clear objective, and for that to be measurable.” A livestock farmer with 50 hectares just uses the manure from his own cows to fertilise that land. In the past, it would have been fertilised by maybe 20 herds. “If farmers start swapping their land, and if they make a regional crop plan, they can apply different types of manure at different times in the right quantities. That increases soil utility.”

Harvesting potatoes at Van den Borne farm using AVR machinery

Potato harvest at Van den Borne. Source: Van den Borne Aardappelen

Van den Borne’s advice is that if you want to use precision agriculture to optimise your production, you need to first know where you currently stand. For him, yield measurement was an eye opener, especially the variation across different plots. He used his potato harvesting machines to figure this out. The aim was to find out how many potatoes were being harvested from specific parts of the plot, what their quality was, and how much tare soil they came with. “That gave me a yield map for each plot. I was able to compare those with other data maps from other surveys we do.” Those maps have enabled him to determine the yield potential of the potatoes. “We also use those yield maps to evaluate and plan site-specific cultivation interventions.”

Hyperspectral cameras

To assess the quality of his potatoes, Van den Borne uses not only weight sensors but also hyperspectral cameras. These enable you to assess your tare soil in real time. The cameras also measure the percentage of stones/clods/soil in the mass flow, and enable you to see the grading sizes of the potatoes in it. The cameras communicate this data to AVR Connect, which translates the information into site-specific data. “With those cameras you can use weighing sensors to determine your gross yield. Of course, you’re mainly interested in the net yield, and you can calculate that if you know the amount of tare soil.”

His three pieces of advice to fellow farmers: “If you want to improve your business operations, you first need to know where you stand. How are you performing?” If you’re already achieving a good return, then going hi-tech isn’t going to give you much more. If your yields are just average or well below average, then it will. But who are the ones investing in hi-tech? It’s farmers who are already doing well, and want to do even better. So they just end up being disappointed, and that’s why the technology isn’t gaining traction and more buyers. “It’s because the group that has tried it are saying that it doesn’t deliver,” says Van den Borne. “But it’s actually worthwhile for the farmers at the lower end. They’re the ones with potential to grow, but they don’t have money.”

Revenue model as a starting point

His second piece of advice is to think about what you’re selling as a farmer, and what your revenue model is. Most don’t even know. They’ll say: I sell potatoes. Okay, but are you selling them per kilo, per metre or per row. “The definition of what you sell and where you make your money is very important. Many farmers are just focused on maximising crop yields. I did the same. But that’s not the smartest choice from the perspective of technical production costs.” Put simply: “If you can achieve 50 tonnes of potatoes through a normal level of investment, and 80 tonnes through a high level of investment, then the 80 tonnes will probably be more expensive per kilo than the 50 tonnes. Your starting point should be your net yield, your revenue model, rather than maximising your yield.”

His third piece of advice is to make sure you share your knowledge rather than keeping it to yourself. And to keep learning from each other. Every week, Van den Borne receives three busloads of people. “They all say wow, wow, and go home all excited. But a few days later they think, no, it’ll probably be too difficult for me.” That’s why he’s shifting his focus to the government. “I need to start by getting the government on board, because they’re the ones who will determine the future of farmers.”

Measurement isn’t the problem. You can measure everything. The more important thing, he says, is what the data means. “If a sensor says 10, does that mean 12 is good too or is 26 better? Who’s going to tell me what it should be?” That’s a job for knowledge institutes. Hence his appeal to WUR: provide more computing capacity and people to validate data. So that it can be fed into AI and enable farmers to take the next step with it. This will improve our understanding of soil and of all the other steps within precision agriculture. “We only understand two per cent of what happens in the soil. If we learn to understand soil better, we can also make faster strides in agricultural production.” It’s not about taking soil samples. That’s an analysis based on nutrients and coming up with a balance. Add this, remove that. “That has nothing to do with soil life and all its spores, fungi and bacteria. We all need to look at biology much more than chemistry.”

Strip cropping

Van den Borne doesn’t believe in strip farming as a revenue model. “The whole agricultural system is based on capacity and scale. Doing as much as possible with as few people as possible. If you have to do all that on strips of land, it’s not really efficient. We can do it if we make progress with robotisation and have a trillion robots to do all the work on those strips. But that’s a long way off.”

He has a strip cropping field himself too, with wheat, potatoes, sugar beet, parsnips, maize and onions. He’s certainly learnt something from that project, particularly about soil fertility and insects. “In a lot of potato plots I’ve turned spraying paths into flower strips, because I can clearly see the added value of having a ‘highway’ in your field for biological pest control.” To be clear: he’s not saying that strip cropping is bad, but that it’s too expensive. It’s the same reason he’s creating a food forest in Reusel. “It’s all very nice, but  it’s not a business model. It’ll triple the price of chips. We need to be willing to say that out loud too.”

Erik Pekkeriet Vision Robotics

ing. EJ (Erik) Pekkeriet

Programme Manager Vision+Robotics

Contact ing. EJ (Erik) Pekkeriet