• Share post:
• Facebook
• Twitter
• WhatsApp
• LinkedIn
• E-Mail
• Copy URL zo clipboard

30.01.2019 - Peer Leithold (Send email to Peer Leithold)

Digitization in agriculture - between hype and frustration

At the moment it seems as if digitalization in agriculture is accelerating more and more. At least if you listens to the speeches of politicians, follow the agricultural trade press, count the number of scientific congresses on topics such as "Farming 4.0" or observe the corresponding activities of various agricultural technology, trade and chemical companies. Announcements of digital advances and solutions promise a better, because digital, future. If one talks to practitioners, however, it becomes clear that this topic, which is so dominant in the public eye, can sometimes cause frustration instead of a desire for this future.

 

The growing sense of disappointment

For example the mechanical engineer and farmer Michael Horsch. He gave a remarkable interview at this year's Farm&Food 4.0 in Berlin. In this interview, Mr. Horsch reports that many farm managers spend a lot of their time evaluating drones and satellite images, interpreting yield maps, digital field maps or analysis tools. He draws a disillusioning conclusion from his observations in Europe and North America as well as analyses of several working groups: a lot of effort, wasted time, little use and hardly any profit for the company. Nevertheless, Michael Horsch expects a new form of digitization of agriculture. Namely one that also makes sense. The bottom line of his interview is that only those who digitize in a meaningful way will make an economic profit out of it.

Then, of course, there is a reminder, DLG President and farmer Hubertus Paetow. In the February issue of the 'Bauernzeitung' he reports in an article about personal experiences with the digitalization of his farm and about how difficult it is to make the entire process of digital farming technically smooth and
agronomically correct or better than before. In his "self-made digitalization", Mr. Paetow clearly reached the limits of what is affordable. By recommending a cloud-based software system and a holistic approach to technology, he draws personal lessons for himself.

So there seems to be a considerable gap between the promises made by industry and the reality on the shop floor. My thanks go to both gentlemen for their honest and ruthless statements. I can only help the authors here and say, yes, that is exactly how it is.  In digitalization, we find ourselves between trial and error or between opportunity and risk.

 

In order to know what exactly is needed you should ask yourself what the core purpose of digitalization in crop production should be.

 

• In my estimation, the goal is the correct agronomic response to heterogeneity in crop production.
• Added to this is the positive effect on the environment.
• Digitization must also simplify management processes, relieve the farm manager of unnecessary office work and, above all, give him security in managing staff.

Tho, not everything that is technically possible makes sense from a farmer's point of view. After all, effort, costs, benefits and revenues are the gold standard here.

Until now, the market has been driven solely by the available technique, technology or software, rather than by agronomic issues or economic benefits. This means that we have to change the view of precision farming (digitalization in arable and crop farming) turning it upside down. We have to think from the right end. The questions we have to ask are: What can we do better? What do I need to do? How does my farm benefit from this? What does it cost?

What are the opportunities and what are the consequences?

Traditionally, we farmed on the basis of one crop choice per crop type, sometimes one choice per field. For example, if a farmer grows five crops, he has a kind of "recipe" for each of these crops. If he cultivates the fields differently, then there are perhaps 50 or 200 decisions that the farmer has to keep in mind. This can certainly be organised in a similar way with a diary, Excel and an alert mind. On uniformly farmed fields, however, there are countless different cost/income ratios. We only need to look at the yield maps to understand this relationship. This is exactly where the potential of digitization lies! In purely statistical terms, soil information varies every 50 x 50 m, plant information every 20 x 20 m.

Let us give you an example to illustrate this: A 500 ha farm with five crops must either make five decisions or, when looking at the field, perhaps 50 to 200. If we now also consider the heterogeneity of the soil, we are talking about more than 2,000 decisions (when looking at the plants we are already talking about more than 12,500).

The decisions made during the variable application of a resource must be done every second. This is neither possible in the head nor with Excel. What is correct is rather the admission that we can only access this form of precision if we are prepared to automate arable farming and crop production. However, this requires a certain degree of standardisation. Standardized processes must be controlled by agronomically sensible rules or algorithms. This process only makes sense if the result is that the revenues significantly exceed the costs of the process.

Before we turn our attention to costs and revenues, we must first assume that the digital process will function sensibly and be robustly managed. This is what most companies fail in practice. Mr Horsch and Mr Paetow have rightly put their finger on the sore spot in our industry.

 

Basic logic and success characteristics of a good digital solution

Every digital solution, whether in the personal area, in industry, trade or administration, follows the same principles: Information acquisition - decision making - application in the process. Each individual application, such as spreading lime, fertilizing nitrogen or applying fungicides, therefore requires a decision that is influenced by one or more specific pieces of information.

The right digital information

1. It must be the correct information in the sense of our question, i.e. it must be closely correlated and causally related to the intended decision (i.e. it makes no sense, to count lice on the eve of the upcoming liming).
2. The information must be available digitally and in an absolute scale (more or less of something is not enough),
3. it must be sufficiently precise in the sense of the decision and be of high resolution,
4. it must be objective and reproducible
5. it must be cost-efficient,
6. it must be easy to handle.

If one of these characteristics is not fulfilled to a large extent, the technology platform is unsuitable for practical operation. Michael Horsch makes it very clear with an example: If an operations manager spends the majority of his working time interpreting images from satellites, drones or aircraft (3rd feature), then these tools do not fulfil their purpose.

 

 

 

Agronomy
If the correct information (in the above mentioned sense) is available, a knowledge-based, agronomic algorithm implemented in software must convert this information (almost) fully automatically into a current decision. This algorithm bundles the currently best available agronomic knowledge. It represents agronomy and is the essence of digitization, it is not technology! An algorithm is good if it is proven to lead to a better economic, agronomic result. The proof of this is provided in scientifically based on-farm research experiments (OFR).

 

This decision is displayed and then automatically implemented at the right place by the application device. This applies both to online procedures (e.g. N-fertilization, plant protection) and to offline procedures (e.g. basic fertilization).

Data
The entire data management must be highly automated and "silent". In this case I absolutely share Mr Paetow's view on cloud-based data storage. This is the only way to avoid high costs, service and maintenance efforts for office software. With software, one proceeds as follows: everything that can be clearly automated, i.e. data transfer, storage, allocation, classification, routine computing work etc. is automated. Strategic decisions, user-specific procedures remain open and require the active decision of the operations manager. Scientifically developed and already proven algorithms are offered to the user for selection.

Introductory consultation
The introduction of one or more procedures of digitalization in arable and crop farming on a farm should be accompanied by external specialists in the first year. This is the only way to ensure that all available knowledge is immediately applied. Beginners' mistakes are avoided and the employees are trained in the handling. Customer service and support must also be guaranteed in the long term.

So in order for such an agronomic procedure to function correctly and cleanly, four things must come together and interact:

• robust information technology,
• the right agronomy,
• an intuitively manageable data management and
• external specialist advice.