Hhhhoooo boy. I have been working on my experimental designs for about a year now, and tell you what. It's hard. Even the plant selection for the experimental plots is hard, despite it being the part I most looked forward to! For science, I have collected and recorded site factors and plant traits in excruciating detail, all to make the selection process credible and my results more applicable… But that will be a post for another day. Before that, I thought I'd describe how I approach planting design in my projects.
First, some planting design theory (in progress)!
My planting design approach is largely based on ideas on designed plant communities, drawing on a variety of conceptual design tools and visual examples developed within the paradigm. Much of the paradigm rests on ideas of combining ecology with artistry, which both demands application of detailed site- and plant knowledge as well as the creation of intentional aesthetic effects. Additionally, the paradigm stresses functionality. “Function” can of course mean almost any number of things in planting design: enhancing site atmosphere through pleasant appearance that suits the site, ensuring high canopy cover to minimize the need for weeding, providing food and shelter for a variety of insects to maintain biodiversity, having high evapotranspiration rates to aid in water management, and so on. Each project will prioritize different functions. Together, the site ecology, creative ambitions and specified functions combine to make project-specific design intentions.
As I see it, applying the design intentions into realized plantings requires a reasonable grasp on three different “domains” [*1]: Plant assemblage (the combination of plants within a defined space at any given timeframe), the spatio-temporal configuration of vegetation (how plants occupy space and are distributed in four different dimensions: Height, width, depth and time), and plant ecology (how plants interact with each other, how they interact with humans and other organisms on site, as well as how they interact with abiotic site factors).
Note the word “grasp”, as opposed to “control”. Each of these domains contains a multitude of factors, not to mention that all of the domains are interconnected and interdependent. As such, it is impossible (or at the very least highly impractical) to even try and have absolute control over the domains. On the other hand, the designer does inevitably influence all of the domains, and the domains in turn influence if, how, and when the design intentions are fulfilled. As such, it is necessary to be mindful of the domains in all project phases, from site analysis to prioritizing intended functions, species selection to individual plant placement, and finally vegetation development and management.
A design intention may fall neatly under any given domain. Maybe the brief is to specifically build a specimen planting consisting only of roses (Rosa spp.). In other cases, getting the spatial qualities right in the long-term might be the most important thing: in traffic environments you have to limit the height of your planting to ensure traffic safety, but on the other hand you can't well play hide-and-seek in a Sedum-based low dry meadow mix. Or maybe the project is purely about creating ideal habitat for a specific butterfly species. But even these types of projects are concerned with all of the domains.
What does the theoretical approach above mean in design practice? Here’s an example.
A new tree is to be planted into a park. The abiotic site conditions are suitable for a number of different species, but a silver birch (Betula pendula) is chosen for its rapid growth, which will quickly provide shade for people and habitat for birds. Its tall and narrow shape will contrast with the surrounding existing trees, enhancing people’s experience of the park. The shade of the silver birch will cause a change in the understorey vegetation, as the most shade-intolerant species will be outcompeted. The silver birch is also known for its capacity to sharply decrease water availability within its root zone, thus further modifying the understorey species assemblage. Here in the Nordics birches also tend to work as nurse trees for Norway spruce (Picea abies), which will mean that in time the site will become even more shaded, but as the birches are replaced by spruces the water balance will change as well. There will also be an increase in organisms living on the spruce, and a change in the soil pH as the spruce needles slowly make the soil more acidic… You get the idea.
Of the three domains, plant assemblage and spatio-temporal vegetation configuration are the ones that the planting designer has the most direct control over. Many designers thus start the design process by defining the desired spatial qualities within suitable project timespans, and use these to guide most of the plant selection. Others might place non-spatial functions or creative intentions first, even if it might require changing the abiotic site conditions permanently or intermittently through management actions. Neither of the other abovementioned approaches are wrong, though, and I could see myself changing my process for specific projects.
Generally speaking, my plant selection process isn't very rigid, and I don't have a huge amount of rules of thumb. I don't use any specific tools rigorously, either, although I do apply many prevalent German planting design concepts selectively. I try to switch between broadening and narrowing the plant selection throughout the process, as I feel that this keeps me from getting into a rut in my species choices and -combinations. I’m also acutely aware of the lackluster evidence base available for planting design, both in terms of objective but specifically applicable plant knowledge, as well as in terms of best practice for site analysis. As such, I have accepted that my design decisions are educated guesses at best, and wishful thinking at worst. But that doesn’t stop me from doing my best to reach the project-specific design intentions.
Phase 1: Initial plant selection
Personally, no matter the project priorities, design intentions and domain focus, I tend to start from the subdomain of interactions between plants and their abiotic environment, and use these to make the initial plant selection. Of course, I do take care to familiarize myself with the overall assignment and intended functions/ goals for the plantings first, but these tend to be used as background information rather than the starting point in the early phases. Similarly, while I do try to jot down my immediate inspiration (often a combination of individual plants and a rough theme, like "woodland" or "industrial setting") when visiting or studying a site, formulating and applying explicit selection filters based on these tends to come later in my process. It is usually more important for me to lay out the basic site context (e.g. cultural and/or natural context, spatial attributes) the relevant garden habitat [*2] and possible other decisive site conditions (such as high quantities of salt, windiness, frequent or intense wear) early in the process. Otherwise the initial selection is bound to take a very long time, and to produce too many duds that have to be cut from the list later anyways.
To make the initial plant selection, I often consult German plant nursery catalogues [*3]. To make the process faster, I might go through plant lists that match the garden habitats [*2] (courtesy of Bruns Pflanzen) I’ve deemed relevant for the project, and pick species that feel intuitively right from these. Or then I just flip through other German catalogues that classify plants based on their garden habitats. Most of the intuitive choices I make in this phase can be traced back to the initial theme and site context, but they’re also motivated by my prior experience and knowledge on each taxa's behavior. Naturally, my personal preferences and biases come into play as well in this phase.
While I’m making the initial list, I tend to also document various attributes of the plants that I find interesting. The extent and contents of the documentation has varied during my career, as my plant knowledge has increased, and my design process has become more streamlined. Usually, I write down the scientific name, garden habitat(s), light requirements, and blooming time as well as possible attributes relevant to specific desired functions or site limitations. Before, I used to write down many more attributes, including size, sociability classification, expected longevity and spreading attributes, but now I only do this to plants I don't know so well.
Phase 2: Plant selection- The Editor's Cut
My initial lists are usually quite long at this point, averaging probably between 20 and 40 taxa. This is because I rarely have a super clear or strict theme, as well as because I rarely start with the most limiting selection factor. That will be rectified at this point in the process, where I start removing plants from the long list. Partially this is an artistic editing process, where I simply start noticing patterns in the plant selection (phenology, sizes, colors, shapes, textures, overall feeling) and remove what doesn't feel like it "belongs" or deciding between two plants with overlapping attributes. Partially this is a technical process, where I read through the detailed descriptions of each plant, and discard those whose behavior or suitability to the site is questionable, or that don't contribute enough to the intended function. I also check what local (Swedish or Finnish) or broadly respected (e.g. Missouri Botanical garden) sources say about the taxa, and factor that in to adjust my idea of the plant's expected size, blooming time, and winterhardiness. Usually, the list shortens to about half or so during this process.
In the third phase of the process I take a look at the selection, and compare it with the size of the areas that I have. How many species are actually reasonable to have cohabiting this space? I decide on a rough number. Sometimes it's above, sometimes below the actual length of the list. To reach my target number, I double-check my lists. Which ecological functions are under-or overrepresented? Do I have enough flowers or other visual interest during different seasons (and does it matter if some seasons don't have special interest)? Do we reach an acceptable vertical structure (heights, layering, volume, intended spatial function)? Are there enough plants with different spreading strategies and vigor? And so on. If there are too many plants, I remove functional redundancy where needed, if too few, I go back to the catalogues and search for plants that fill the missing visual, ecological, or technical roles. I also double-check the availability of the taxa I want at 1-3 local plant nurseries and drop or swap out unavailable taxa. If I have the time and budget, I add a picture of each plant at this point to my records. This is both good for me, my colleagues, and for possible clients. I tended to do this earlier in the process when I was less confident in my plant knowledge, but now I leave it to the first phase where I need to share the selection with others.
Phase 3: Which plant goes where?
Then comes the spatial configuration phase. I consider possible microhabitats or specific areas in a planting that might require specific types of species mixes. I also think about how I want the overall look and dynamics to be, and decide roughly on the planting pattern strategy I take (i.e. where on the small-scale intermingled- large-scale monospecific-continuum I want to be, and how many different patterns and sub-mixes I need). My one personal (and completely arbitrary!) rule of thumb for planting design is that each taxa should take up a minimum of 5 m2 within a project, and no more than 20 m2 [*4]. This rule of thumb serves two purposes: For one, I'm trying to limit the species count and ensure enough repetition to prevent the planting of becoming an insufferable mishmash, and two, I'm also trying to ensure enough biodiversity to prevent monotony and large-scale plant failures. BUT, it is good to remember that there is no proof that biodiversity in itself ensures the resilience of a planting, and as such adding diversity is no substitute for knowing your plants.
When it comes to deciding planting distances, I have two simple guidelines, courtesy of the Grand Old Man of the office: 1) ensure good, quick coverage and 2) make it easy to draw and count. Because of this, I pretty much only use 2 separate planting distances for perennials: 35 cm and 50 cm. 35 cm yields about 8 plants/ sqm, which ensures good coverage for most plants within two growing seasons, excepting the smallest groundcovers like Cerastium tomentosum or Ajuga reptans. 50 cm is reasonable for bigger plants when they either grow fast or are combined with other bigger plants. Sure, dense plantings encourage competition and may diminish the size of individual plants, but as long as you’re considering the overall competitive balance of the plant combinations, don’t use especially weak taxa, and can afford to make the initial investment, it will not make that much of a difference. Nobody is going to space the plants exactly according to the specification, anyway.
Then I start setting out the plants on the drawing. This is the last phase where I can swap out, add, or drop a few more taxa, based on both functional, ecological and aesthetic criteria. Oftentimes I work with modules, i.e. repeating patterns of specific species combinations. Sometimes I draw all of the plants out, one by one. Yet at other times I do draw blocks and drifts. It depends on the project size and the intended visual effect, as well as the ecological attributes and behavior of the plants. Whichever drawing strategy I take, I have to consider proportions between different plants first. I don't use any of the role-setting systems available [*5], but rather I go by project- and place-specific aims and my gut feeling. The "gut feeling" is largely based on my knowledge of the species' habit, reproductive behavior, phenology and longevity. A small part of it is just aesthetic preference, but I tend to consider it less at this point of the process than I probably should.
Whenever I intermingle plants or otherwise place different taxa in close quarters, I try to imagine how they will live together and how it will look. This is especially important when creating intermingled modules. How do the phenologies go together? Habits? Blooming time? Competitive strategies? If I'm satisfied I make a pattern that satisfies both the aesthetic intentions and that I assume to give a good, reliable species composition and groundcover. Then I repeat this for each module, and consider how the different modules play off of each other in time and space. Similarly, if I draw the plants out one by one I think about how the neighbouring plants will interact with each other, and how the different functions become distributed in space and time. I wish I could learn to work with placement to heighten spatial and experiential qualities, because that would help me to add more value and layers to the designs. I rarely do blocks and drifts nowadays, because I actually think they're harder to draw and count. But when I do, I try to systematize the pattern somehow, so that I know exactly how many plants fit in each with a set planting distance. Otherwise the combos will be considered on same terms as when drawing everything out.
Phase 4: And it's... done?
The planting plan is done! After this, I write the order specification documents that include plant sizes (this is sometimes important to consider well beforehand, e.g. to set the right planting distance, or to make sure that the plant root system fits into the planting area at the time of planting). Then the documents go to the client, who takes it to the company who will build the planting, who will contact the nurseries [*6]. And then it will be built! Time to see the design in action.
How will the final planting develop? It depends. I usually try to consider as many factors as relevant for the vegetation development in the design process, and imagine the most likely scenarios for development as it pertains to the three domains. Still, the success of my process in ensuring the realization of the design intentions is largely unproven. This year, a few of my bigger non-scientific projects have been built, and my planting design experiments should also start. Will they develop as I intended, or foresaw? If they fail, how will I know if the problem was due to my design, or due to an uncontrollable environmental or human variable? I don’t know yet. And while both my practical planting design process and theory on planting design are still sub-optimal and largely subjective, I trust that there is a good chance of improving both – and thus improving the chances of learning how to make reliably functional designed plant communities.
[*1] not dissimilar to Rainer & West's (2015) 3 essential relationships, although these ideas weren't directly inspired by them.
[*2] Short explanation about the Garden habitat-system: The GH-system is a way of describing different combinations of site conditions in garden settings as they correspond to rough natural habitat archetypes. The basic Garden habitats are woodland, woodland edge, open ground (with subtypes), Rock gardens (with subtypes), border, water's edge and water. These habitat archetypes give a general indication of the light-, soil-, and water conditions. While the Garden habitats were first defined by Hansen & Stahl in the early 1980's, it is Sieber's version from the early 90's that is commonly used by German plant nurseries and practitioners. In Sieber's version each of the habitat archetypes is complemented by a number that expresses moisture conditions. I'll make an in-depth post about the Garden habitat system at a later point in the future.
[*3] Because the Germans have their shit together.
[*4] Most taxa can’t reliably cover even this much space by themselves.
[*5] Once again, a subject for a different day.
[*6] At this point there may be an additional need for swapping out plants due to poor availability; based on my limited experience, this is the phase where all of the previous hard work can go down the drain, if the designer is not involved in deciding the replacements.
Comments