Will new approaches lead to a breakthrough?

The idea of site-specific management is now more than a quarter of a century old. It is based on the heterogeneity of a field in terms of soil type and relief. This results in differences in water supply, nutrient availability and in the risk of leaching of fertilisers and pesticides.

While many applications in the field of automation have become a matter of course in practice (automatic steering systems, section control), site-specific management is not really catching on in practice. One of the reasons for this is that technical developments have often been one step ahead of agronomic knowledge and advice. However, due to the new legal framework conditions (keyword fertiliser ordinance), the importance of precision farming will presumably increase. Because in the future, the focus will be less on maximising yield and more on the economically sensible use of operating resources.

So far, the most experience with site-specific management has been gained in the area of N fertilisation.For a long time, sensor technologies were primarily used for this purpose. However, the high initial costs of such a sensor system deterred many (especially medium-sized and small farms). In the meantime, however, more and more solutions based on satellite data are entering the market.

A nitrogen sensor has already been in use for 20 years on the land of the agricultural community Groß Germersleben on the western edge of the Magdeburger Börde region. "Our aim was to better understand yield differences, to use fertiliser nitrogen as efficiently as possible and to establish high-performance crops with uniform qualities," says farm manager Sven Borchert. The total of 1,700 ha of the community are structured differently. The field sizes range up to 100 ha. With these dimensions, digital helpers are a welcome support. "Of course, every farmer knows his/her land best. But technical tools like sensors offer the possibility to substantiate our experience and professional gut feeling with data that supports our decisions or challenges them," says Borchert.

Karsten Twietmeyer also confirms this: "A sensor is like a pair of strong glasses. People can only detect about 20% yield differences in a field. A sensor similarly registers differences in the low percentage range." This agricultural engineer ought to know as he himself wrote the algorithms for a nitrogen sensor processing and developed an app for biomass analysis. Since 2017, he has been managing the Wichmannsdorf Agricultural Producers' Association in the Uckermark region (Brandenburg) with 2,500 ha and 1,300 beef cattle together with his brother. The site is characterised by comparatively low acreage numbers, which also fluctuate quite strongly with values between 17 and 50. In addition, water is a limiting factor. And this is exactly where people come into play again: "Many farmers expect what has definitely not worked until today: You can't just put an inexperienced employee on a tractor with a nitrogen sensor and let him manage the fertiliser application ," says Twietmeyer. "There are too many uncertainties in the entire fertiliser system for that: What can the soil compensate for in terms of drought/precipitation? Are there any other nutrients that are deficient? How great is the replenishment potential of the soil? Are there other influencing factors such as diseases or soil compaction? All these questions play a decisive role in the right fertiliser strategy. A sensor can't know that. It regulates rigidly according to programmed algorithms."

Dr Hans Georg Brunn has also had this experience.He is a farmer in Pabstorf and farms a total of 770 ha in eastern Lower Saxony and western Saxony-Anhalt, which also vary in soil quality. Over the past 12 years, he has already worked with several nitrogen sensors from different suppliers. "The decision on the right fertiliser strategy for the site must always be made by the farmer. The technology can only improve the distribution of the fertiliser," says Brunn. There are two variants here: Either one wants to homogenise the stocks or to differentiate them more. This means that one tries to bring the weak areas of a crop closer to the stronger ones, or to promote the strong ones even further. "Here, the knowledge of one's own soil and the weather plays a supporting role. A sandy soil needs to be fertilised differently than clay. The same measured value can therefore have completely different consequences at two locations. That's why a sensor can only be as good as its user," is the farmer's experience.

With an average field size of 10 ha and fairly homogeneous soil quality, Christoph von Breitenbuch does not see the need to invest in a nitrogen sensor yet. Von Breitenbuch is the managing director of the Leine-Solling farming community in southern Lower Saxony. He attaches great importance to precise arable farming and has extensively digitised the farm community's land. With the help of a service provider, so-called farmyard soil maps have been developed in recent years, which serve as the basis for application maps for the Isobus fertiliser spreader. "The public's view of agriculture is becoming increasingly critical. Therefore, we need reliable, objective data on which to base our decisions. And last but not least, the Fertiliser Ordinance requires us farmers to become even better in the area of nitrogen efficiency. In this respect, every technical solution is a welcome support," says the farmer from Lower Saxony. For him, documentation is also an important factor in this context. "Of course, the fertiliser technology must also be able to write the data for spreading back to the computer." In practice, however, it is not uncommon to find 20-year-old spreaders that are generally unable to do this. In addition, site-specific fertilisation with centrifugal spreaders is physically difficult anyway. 

Satellites can also take a look at the crops. Especially since the European Space Agency (ESA) launched its Sentinel satellites into space, new service offers for site-specific N fertilisation have mushroomed.These offers are therefore in direct competition with nitrogen sensors.Sven Borchert, Karsten Twietmeyer and

Hans Georg Brunn already have experience with these kinds of offers. "We currently use an app that divides our fields into different growth potential zones based on satellite biomass maps from the past six years. The app does not give a concrete fertiliser recommendation in kg per ha, but instead distributes the nitrogen by application map according to the site-specific average fertiliser level for the field, which I specify myself," explains Sven Borchert. The application maps are transferred to the tractor terminal via USB stick, e-mail or cloud.

The handling of such satellite-based digital services is comparatively simple:Farmers and advisors can register online. The respective system then automatically records the field boundaries using the shape files from the area application program.Application maps can then be created on the basis of biomass maps.

The fact that the satellite services only work in a grid of 10 x 10 m, and therefore have a lower measuring accuracy than the sensors close to the ground, does not bother the farmers. "It doesn't come down to the last centimetre," says Dr Brunn. "The spreading technology is not (yet) able to meter out very small-scale differences in demand at the usual driving speeds today anyway." Nevertheless, he still regularly scans his fields with the sensor. However, only in order to have a "sparring partner for his intuition" and to record the data. However, he does not leave the regulation of fertilisation to the automatic system, but instead creates the application maps for each individual field himself on the PC. However, the biomass maps of the satellite software would also suffice as a basis for this. "We have also achieved very good results with this method," says the farmer. Sven Borchert also confirms this: "The bottom line is that both systems achieve the same effect."

In terms of accuracy, Karsten Twietmeyer also sees no disadvantage compared to the sensors close to the ground. "The beauty of satellite systems is that you get a two-dimensional picture of the entire crop. So it's a complete recording of the field, so to speak. A sensor system always records only part of the crop and then interpolates for the entire working width. In addition, a calibration run in a representative track is required before starting work," he says.

Furthermore, farmers do not see any problems from atmospheric influences in satellite services. "Even a sensor does not always run trouble-free," says Dr Hans Georg Brunn. "Changing conditions during a crossing, e.g. due to dust or dew, can also cause measurement errors here."

Data availability. The Sentinel-2 satellites provide up-to-date images every two to three days. However, as only cloud-free images are used if possible, in extreme cases there can be a gap of three weeks between two biomass images. As a rule, however, this is not the case. Nevertheless, one is more flexible in this regard with a sensor system. The faster the growth (especially in April and May), the more important it is to have up-to-date images. "However, our app is, for example, not based on current biomass images at all, but on a blend of biomass images from previous years," says Sven Borchert. "In addition, it should be borne in mind that we are also dependent on the driveability of the areas with the sensor. And you have to take care of the hardware as well," he adds.

The bottom line is that the high cost of a sensor system is still a barrier for many farm managers.The new satellite services offer significantly cheaper alternatives here. Above all, farmers who are new to site-specific management can use satellite services to slowly "get to grips" with the topic. With the right approach, they can achieve the same results as with a sensor system.

Katrin Rutt, DLG-Mitteilungen (Announcements)