e-Gro

Silke Hemming: "The time has come for data-driven cultivation"

Silke Hemming
Silke Hemming

As a researcher at Wageningen University & Research and co-organiser of the Autonomous Greenhouse Challenge, Silke Hemming closely follows developments around data-driven cultivation. "There are still many bumps in the road, but the whole world is watching and cooperating to take remote cultivation further. Fifteen years ago, no one saw it coming. That is clearly different now."

Much has already been said and written about the first edition of the Autonomous Greenhouse Challenge.  For the second edition, which starts in mid-December 2019 and will be even more challenging, interest is overwhelming. More than 100 teams of mixed composition (students, researchers, programmers and crop specialists) from all parts of the world had signed up for the challenging competition. Grodan is one of the sponsors facilitating this competition, providing GroSens sensors and substrate slabs for the participating teams.

Not new, but innovative
Data-driven cultivation in itself is not new. After all, growers have long been guided in their decisions by climate data measured inside and outside Greenhouses. They compare these with the past or other growers. So why is data-driven growing now so suddenly in the spotlight?

"That is because there is currently a profound evolution in technology that allows us to zoom in on the plant and in mathematical models that can accurately predict how the plant will respond to set points. With that, you can also predict the future and support the horticulturist in their decisions," says Silke Hemming. "Tech companies have been working on data platforms for food for some time. With new business models emerging in vegetable production - think vertical farming concepts and intensive cultivation with LED / light - the need for control options is increasing and the development of remote cultivation is gaining momentum. The challenge is a fun way to gain new knowledge and experience, try out new technology and learn from it."

From reactive to proactive
Current crop management is still largely based on personal observations and feeling. According to the researcher, even conservative growers are starting to realise that while green fingers remain useful, they are not blissful. "crop management still has a strong reactive character," Hemming continues. "You change setpoints when the crop develops too vegetatively or too generatively, or when deficiency symptoms become visible. Properly considered, the plant has then performed sub-optimally for quite some time. Sensors are now becoming available that allow us to monitor in real time the root environment, the microclimate around the plant and even photosynthesis. That makes it possible to react much faster. Combined with advanced growth models, you can then anticipate to avoid sub-optimal performance. That will lead to higher yields, improvement in product quality, higher resilience and more efficient use of energy, CO2, water and nutrients."

Challenges
The researcher notes that the supplying industry is busy developing platforms that facilitate data-driven growers and make them accessible to practitioners. "Your e-Gro concept is and nice example of that," she says. "I am sure this kind of initiative will help horticulture move forward. I see it as 'work in progress', because there are still a lot of challenges before there can be truly autonomous systems covering the full spectrum of observation, data processing and crop optimisation." One such challenge is the fact that data on crops, growing conditions, climate and water/nutrients comes from different data sources, from manually written down to digitally, some data per minute (climate), others per day (harvest), still others per fortnight (nutrient analysis). Analysing, linking and interpreting all that data is still a tall order. Moreover, there continue to be observations that are not (yet) recorded digitally, such as grower observations as he walks through the Greenhouse.

Relevance
Hemming: "We also have to ask ourselves whether everything we currently measure and can measure is relevant to what we actually want to know, namely how the plant feels and whether it performs to its capacity. The answer to that question is obviously in the negative. In addition, not all the data currently being measured is easy to interpret or translate into concrete actions. We are still nowhere near being able to look as deeply into the plant as we would like, but there is movement in that too."

Once a grower has the desired relevant data, what can he do with it? Steering production, quality and harvest timing, of course, but for maximum financial results, he should also include market information.

New generation
Eventually, platforms will emerge that aggregate all relevant information and support growers in their decision-making, the Wageningen researcher expects. They will certainly not make growers redundant, but will greatly increase their coping capacity. This is also necessary, as the number of good employable growers is steadily decreasing worldwide. Moreover, the great thing about new technology such as e-Gro is that it can lead to renewed interest in the green sector among young people, who are brought up on interactive games and apps. New growing concepts such as vertical farming, for which there is great interest in many countries, also contribute to this. So are other technology- and artificial intelligence-based developments, such as robotisation and remote sensing, with or without drones.

"You can safely say that the time for data-driven growers is ripe," concludes Hemming. "That was not yet the case in 2006, when my colleagues in our greenhouse in Naaldwijk had a pepper crop guided by computer models with reasonable success. Back then, few people got excited about it. My expectation is that in the next five years development will undergo a huge acceleration and autonomous cultivation platforms will become part of the standard equipment of a new generation of cultivation managers."