Donald S. Coon

I was working on a research collaboration with a friend at a large indoor plant factory facility when he asked me, “If you had to try and nail down the 4 biggest challenges to greenhouse production (mostly greenhouse, but could relate to vertical farms too) that you think need to be solved to help support greenhouse production efficiency or just overall success what do you think those would be?”

These types of questions come up often, and for some reason (likely procrastination) I took the time to write out not only a response to him, but something a bit longer for others. The results surprised me a bit! They differed slightly from my recently published review paper about exactly that subject, but utilize my perspective and experience. Below is my improved and formatted response to his question, for you.

---

For greenhouse production efficiency and overall success? Tricky question, I like it though. I think things like environmental/pest control, automation, new varieties, and proper design will always be critical to success. At this stage, I would equate greenhouses to being in the last mile of development. Getting huge increases in efficiency and success requires a lot more than it used to just because we have gotten so much better at it. This was a fun distraction to write about today! Below are what I see as the 4 biggest challenges once a greenhouse is not struggling with day to day operations.

Data streams need to be captured and used

Greenhouses started with manual controls by experts. Someone would come in and check the temperature and humidity, then decide on opening the vents, turning fans on, or whatever options they had available to control the environment. That’s great, but naturally things advanced towards automated sensing and control systems designed by engineers. These monitored the environment and turned things on/off to maintain set points (hopefully assigned by the farmer, not the engineer). Now, things are being thrust towards AI based controls where all this data is funneled to a sophisticated, trained algorithm designed to control the environment.

One commonality across all of this is a hilarious amount of data and very little need for storage. The systems are responsive, negating the need to store the data pertaining to the environment. I’ve worked in plenty of old (and even newer) systems like this where you have to double up on sensors: one for monitoring/control and one for recording. Collecting and storing daily minimums, maximums, and averages can become invaluable information. This is especially true when combined with historical yields to replicate the environments that achieved record highs or prevent excess resource consumption.

Lack of Decision Support Systems and Models in the management of Controlled Environment Agriculture

A greenhouse grower or manager is a fantastic person who knows their crop and greenhouse with intimate detail. They are certainly capable of making informed decisions regarding facility design, yield amounts, and crop timing based on their past experiences. The best growers and managers are great at these predictions. But at the end of the day, they are rough estimates, and this expert knowledge lives, moves cities, and retires with them.

Decision support systems and predictive models provide tools that not only enable the transfer of expert knowledge but also further enables growers and managers. These tools can aid in providing more precise predictions alongside uncertainty for enhanced risk management. Of course, decision support tools go way beyond yield and could include production volume, cycle planning, greenhouse design, resource use efficiency, environmental impacts, production costs, and any decisions that might help make a greenhouse successful.

Climate Change will test the limits of existing structures

Many greenhouses were designed based on old weather patterns for their locality. That still works great most of the time, and upgrades can even be made to combat the effects of climate change and its associated variations. However, the extremes beyond the capabilities of a facility that may have been once every few years may occur more frequently now. Likewise, much of the US will face increased heat and humidity, which notoriously hamper the evaporative cooling systems used by most greenhouses.

All the challenges in maintaining an ideal environment are then exacerbated by the expected increase in intensity and frequency of severe storms that could damage structures: hail punching through roofs; wind pulling at louvers; floods eroding the ground around support beams. It’s hard to plan for a catastrophic failure of the structure but should definitely be part of any plans to create a successful operation.

Disparity between farms and prospective employees –

size, money, skills, wants and needs

Coming from this place myself and seeing it repeated with other students, those graduating and trained for CEA with a bachelors or even a masters, often find themselves in a difficult position. Many large operations are looking for experienced growers/scientists to lead production/research or they need laborers. It’s common to be underqualified for the first and overqualified for the other. The few good technician roles that students match with are rare because a farm doesn’t truly need many to be successful.

Conversely, smaller operations can offer jobs, but they tend to be an all-in-one position that demands a lot, pays little, and has high risk. For agricultural engineering students, it can be even harder because they tend to want to apply that skillset at a farm designing and improving systems all while growing food, rather than working for a design/sales company that supplies greenhouse equipment. Most farms don’t want or need an engineer to be successful day to day. They need a jack-of-all trades (with a plant emphasis!) to help create stable production, no matter the circumstances.