Food Forest Workshop, Osnabrück, Part 1
Last fall I was sitting at a stylish cafe in the old city center of Osnabrück Germany with my new friend, Ole Oßenbrink. He was inviting me to give a week long workshop that would leverage the tools that I’ve been building with Land Kit along with the methodology of designing landscape systems as fluid digital models. There was just one twist, we would be applying it to food forests.
Ole Oßenbrink, of Osnabrück. Photo by Chris Landau
It made perfect sense, use new advanced tools to help simplify a difficult problem. Ole was excited about the idea and it aligned nicely with some of the food forest initiatives at their school, Osnabrück University (Hochschule). These initiatives included the possibility of building a model food forest (or two) and changes to their curriculum. Coincidentally, their landscape architecture program and horticulture program share a beautiful campus, which makes it an ideal school for this kind of endeavor.
This workshop was going to be like a moon mission. Illustration by Gina Rattanakone
Even though I immediately said yes and knew I had a lot of adjacent experience, I was a bit overwhelmed by the idea. It felt like a moon mission.
First of all, I have taught many workshops in computational design for studios across the US and Canada. And I’m probably one of the most qualified computational landscape designers. But what did I know about food forests? Here is where my imposter syndrome set in, big time.
Second, were the tools ready for this? More specifically, was my software, Land Kit, ready for this? The short answer was probably, but there were a bunch of elements that I felt were maybe missing and things I hadn’t tried yet. The current tool set can be configured in a multitude of ways to tackle all kinds of design challenges and algorithmic approaches, but we hadn’t tackled the complexity of change over time.
Third, did I want to spend a week away from my family, my business, and give up a summer teaching position at Columbia? Was it worth it?
The quick answer to the last question was that I just KNEW (and still know) that food forests are a super important part of our future. Modern agriculture, while it feeds large populations, still has so many environmental issues (namely carbon and biodiversity, both closely connected to soil degradation). Food forests, on the other hand, provide many amazing benefits and advantages by building up soil and other living relationships. Unfortunately, these are harder systems for people to design and to visualize. So if I wanted help the next generation of agro-innovators with tools that matter, I knew I had to dive in.
Then I just did what I do when I’m nervous, I worked on the problem. If I was going to do this workshop and make the software updates, I was going to make it all count.
I started reading and gathering information about food forests, taking a broad view to begin and started connecting the ideas of growing food in a forest to my experiences in both landscape design and agriculture. Yes, I have some agricultural experience.
An orchard tour with my uncles. Photo by Chris Landau
My mom’s family actually grows cherries in Michigan and I had summer jobs during school to help with the harvest and a few off-season tasks. The funny thing, is that food forests were sort of something that I had dreamed about for years. What if there was more than cherries 🍒, if you grew something else in the orchard, along with the cherries? I can’t remember when I first had this idea, but now it was stuck in my head. It acted as a key element in my mental model of food forests as I gathered more information and examples from innovators like Sepp Holzer and Wouter Van Eyck.
I often thought of the orchard and the pragmatic considerations of farming as I learned new things about food forests (and forest gardens) particularly from a couple key books: Dave Jacke’s Edible Forest Gardens and Gaia’s Garden by Toby Hemenway, in addition to the obligatory YouTube videos.
A peach/plum hybrid from Ketelbroek food forest. Photo by Chris Landau
It became clear to me that food forests can come in all shapes and sizes and many implementations and approaches have yet to be explored. We are basically at the beginning of an agricultural renaissance.
As pictures of wildly differing food forests began to form in my head, I started working on tools and approaches that would help me help to solve some of the apparent problems. There were a few things that I thought would be important to have in this context. And the biggest one that I knew we needed (and that was still missing) was the element of TIME.
I had planned on using rules-based planting to help design planting locations according to the desired conditions (putting the right plant in the right place). This is the central pillar of our planting approach in Land Kit (among many other ideas). But what happens when those conditions change over time? How do you design for changing conditions? This was the first thing I needed to solve.
So, I set up a script to place a tree layer and a ground layer. It was easy enough to place the tree layer first according to a simple pattern of locations, and then to analyze the sun exposure effects. And then the bottom layer of plants would be constrained according to the mapping of sunlight hours on the ground from the trees. The challenge, though, is when your trees are younger, they cast virtually no shadows, but when they are older, they form a dense canopy above. Two completely differing conditions.
I needed a way to design for conditions that would change over time. And that’s just what we did with the new Update Conditions component. Provide an area that is already filled with plants according to one set of conditions, and then compare it to a new set of conditions. Some plants will be fine in both conditions but some will need to be replaced after the conditions change. This tool helps to make this process more direct (and iterative). For example, you could test when the trees are 2 years old vs 15 years old.
You can apply a similar logic to other kinds of conditions that might update over time. Imagine soil fertility, pH, or even topography changing over time. Any or all of these changes could be used to understand how your planting might also need to change over time.
Another challenge is always plant data. There just isn’t enough of the right kind of data out there for the sort of diversity, performance and plant characteristics we are looking for.
Pseudo species in action. Photo by Chris Landau
In the time I had available, I knew I wouldn’t be able to build the data I needed, but maybe I could generate example data that would fit the bill. So I created a generative algorithm to format the plants with the kinds of data I would need as “pseudo” species, complete with parameters no other plant database seems to have yet. This solution, perhaps imperfect for a real-world application, would work perfectly in the academic setting; because my biggest goal was to teach Ole’s students how to build and use the design systems in an adaptive way. Exact plants were secondary to the primary mission.
All the pieces were falling into place. Now I just had to buckle up for blast off.
Part 2 coming soon…
