An introduction to Syntropic Agriculture - Part 2
Syntropic agriculture reads ecological succession along two axes: systems (colonisation, accumulation, abundance) describe how much natural capital a site has built up, and within each system, life-cycle groups (placenta, secondary, climax) describe what grows there over time. Pruning is the main tool to steer the process. If you read Part 1 a long time ago, I’ve recently revised it. This post is about design; the next one will be about pruning.
1. Syntropic systems and forest gardens
If you’ve come across agroforestry in Europe or North America, you’ve probably met forest gardens, or food forests: perennial polycultures modelled on the edge of a woodland, with fruit and nut trees above and shrubs, herbs and ground covers below, designed to produce food with little work once established. Martin Crawford in England, and Dave Jacke and Eric Toensmeier in the United States, wrote the reference books.
Syntropic systems share the layers and the idea of imitating a forest, but they differ in four ways. They are far denser: a tree every metre and other plants just a few centimetres apart, where a forest garden spaces its trees for their adult size. They are actively managed: pruning is the main tool, used to keep support plants young, let light in and send growth pulses through the soil. They aim to accelerate succession, rather than keep the system at a productive, relatively stable stage, as most forest gardens do. And they try to produce something at every phase, from vegetables in the first months to timber decades later, with fertility coming mostly from biomass grown and cut on site.
The key to all of this is focus. Energy, resources, water, biomass and our own attention are never enough to spread over a whole field, so we concentrate them where they can do the most, in a row, a nest or a patch. That area regenerates first, and then spreads. Most of the design choices that follow are ways of deciding where to focus, and when.
2. Read the site first
Every design starts from the same question: which system is my land in? A compacted former arable field, an abandoned olive grove under brambles and an old orchard with dark, crumbly soil are three different starting points, and each calls for a different planting.
The best way I know to answer it is to look at what grows on its own in the abandoned corners, the verges and the strips along fences and roads (Google Street View helps more than you’d think). Gérard Ducerf’s work on bioindicator plants is the best guide I know to what they say about the soil.
Then comes the climate, and this is where the tropical model needs the most translation. Our site near Florence has two faces: summer peaks of 40–45°C with two months without rain, and winter nights down to −9°C, with around 700 mm of rain between November and March. Neither the tropical model nor the mild, coastal Mediterranean one fits it as it is.
The last thing to read is ourselves. What do we want from the system: vegetables to sell, fruit, timber, a demonstration site, food for the family? How much money, time and labour do we have? Most mistakes in agroforestry aren’t technical. They come from not being clear about the goals.
3. Three ways to start
Everything at once
The first is the one Götsch made famous: plant every life-cycle group of the site’s current system in one go, at high density. On a degraded site with no irrigation, it can look like this: a pioneer tree such as robinia or eucalyptus every half metre to one metre; cuttings of drought-tolerant aromatics like cistus, lavender and rosemary in between; annual pioneers like sorghum and millet underneath; and the seeds of fruit trees (peach, almond, chestnut) sown directly, to be grafted along the way. No shade cloth, no compost, no irrigation. This is how syntropic agriculture usually works in Mediterranean and tropical climates.
The trap, as I explained in Part 1, is to read “everything at once” as permission to plant the fruit trees of an abundance system into colonisation soil. It’s the most common mistake in conventional orchards too: apples and pears planted straight into an old pasture, still struggling after five or seven years. In nature nobody plants a fruit tree in the middle of a meadow.
Successional accumulation
The second is Scott Hall’s successional accumulation, which I described in a separate post. You start with an infestation: one or two fast-growing woody species at very high density, whose only job is to create a closed canopy, deep shade, a litter layer and biomass. Only then do you plant the long-lived trees, and later still you let birds and a sown seed bank bring in the rest. It asks for patience, but it needs almost nothing from outside.
Importing natural capital
The third is the one most of us actually use, and the debate I promised in Part 1 to come back to. No-till market gardens, food forests and hybrid systems like ours skip part of the succession by importing the natural capital the site doesn’t have yet: compost, manure and woodchip; mulch; irrigation; shade cloth and insect netting, which are imported microclimates; transplants instead of direct sowing; seeds and plants from elsewhere; infrastructure for animals. Natural capital can’t be created out of nothing, but it can be moved from one system to another. At OrtoForesta we use this route in the market-garden rows, and successional accumulation, with Amorpha fruticosa, in a marginal, experimental row.
Purist practitioners argue that a system built on imports will slip back to colonisation the moment we ease off the heavy management, while one built with plants and pruning alone stays stable in whatever phase it has reached. I think they have a point, and that the answer depends on context. On much degraded Mediterranean land, the truckloads of compost needed to plant a few hectares simply aren’t there, and building everything on site is the only realistic route. In temperate climates, with access to resources, importing some capital at the start is often the most efficient choice. Which system ends up more resilient is still an open question: as far as I know, nobody has measured it.
4. Layout and focus
Rows or nests
Syntropic systems don’t have to be planted in rows. Nests are circles 1–1.5 m across and 3–5 m apart, each with a fruit tree surrounded by its placenta and secondary species. They suit small or irregular sites, and are the best option when materials arrive a little at a time: a few plants from a nursery, a trailer of compost, a batch of cuttings. Each time, we can build a complete unit that starts to regenerate its own patch of ground, focusing everything we have in one spot instead of spreading it thin. On anything that needs to be worked efficiently, though, rows are easier to plant, prune and harvest.
Orientation
In most temperate climates, organising trees in rows on a north–south axis is a great strategy to guarantee maximum and uniform light exposure: both sides of the row get sun, and the shade moves across the alley during the day instead of sitting on the same strip. An east–west row casts permanent shade on its north side, which on a hot, dry site can be useful for shade-tolerant crops. On slopes I’d let the contour lines and the movement of water decide first, and the sun second.
Alley width
The distance between the rows is more important than it might initially seem. A wider alley gives more light, room for machinery and space for light-demanding crops or grazing, though beyond a certain width there is no further gain. Too wide, and we partly lose the “forest effect”: less shelter from wind and rain, a less stable microclimate, no shade for the alleys. If you grow vegetables, you’ll already be wondering what happens when the trees grow and shade the beds. It’s the objection I hear most often, and the answer, which is mostly about pruning, is in section 8.
In my experience, in temperate climates and at higher latitudes or altitudes, 8–15 m between rows works well for vegetables or arable crops. In hotter, drier Mediterranean climates, the rows can come down to 3–8 m, to make the most of the shade and soil protection the trees provide in summer. On very degraded land, in a colonisation system, it can make sense to start much tighter, with rows of infestation species even 1–2 m apart, to close the canopy and cover the soil as fast as possible; as the soil recovers, some of the rows are removed, a few at a time, to make space for vegetables, pasture or whatever the focus has become.
▶ IMAGE 2: 2-alley-width.png
Alley width and the height at which we keep the canopy are two sides of the same decision, which I’ll come back to in Part 3. And where the alley produces mulch for the trees, or the trees for the beds, we can size the distances so that the system is self-sufficient in biomass.
Focus
All this points to the importance of the focus of an agroforestry system: what it is meant to produce, and when. Ideally, the focus changes in space and time, to mimic succession: vegetables in the short term, small fruit in the medium term, fruit trees in the long term, timber and nuts in the very long term. Some examples:
vegetable production in the alleys for the first few years, fruit in the tree rows in the medium term, and finally timber and nuts with livestock grazing in the long term;
a more cyclical system, rotating green manures, chicken grazing, vegetables and soft fruit in the alleys, while the tallest trees in the rows are harvested for timber or biomass and pruned heavily on a rotation, to match the light and mulch needs of the alleys;
a travelling vegetable garden that rotates through the forest, inspired by indigenous swidden systems and by Götsch. An area of old planted forest is cut down, and its biomass, the growth pulse and all the natural capital it has built are reinvested in a cycle of vegetables, with new agroforestry rows planted through them. In time the patch becomes an orchard with vegetables in between, then a young woodland producing timber, nuts and mushrooms, then a mature forest. Every so many years, whichever part of the system has become old forest is cut to become the new vegetable garden. It’s the gap cycle from Part 1, run on purpose. In densely populated Europe it’s hard to achieve, but it would let us work within the landscape in a harmonious way, renewing and rejuvenating it through cycles of disturbance and accumulation.
The choice of fruit trees, rootstock included, shapes many of the distances. An apple on a dwarfing rootstock such as M9 stays at 2–3 m and can be planted close; on a seedling rootstock it grows to 6 m or more, and needs far more space. The rootstock also sets the tree’s place in the succession: a dwarf apple is productive for roughly 15–20 years, a vigorous one for decades, so the first behaves like a secondary species and the second like a climax one. On dry, unirrigated sites, vigorous rootstocks are the safer bet.
5. The support system
Support plants are the species we include in an agroforestry system not for what we harvest from them, but for what they do for the system itself: producing biomass, shading the soil, suppressing weeds, hosting pollinators and predators, cycling nutrients, opening up compacted ground, and stimulating succession through the disturbance of pruning. They are the invisible infrastructure the productive species depend on, and in a well-designed system they often outnumber them by a wide margin.
A good support plant has to thrive with no inputs at all. It should be vigorous and tough, fast-growing and generous with biomass, happy to be pruned hard and often, and easy to propagate. On top of that, each one does a specific job: fixing nitrogen, mining minerals from deep in the profile, hosting mycorrhizal fungi, or photosynthesising efficiently in heat and drought, as C4 and CAM plants do. The support system needs a time dimension too, mirroring the life-cycle groups of Part 1: cover crops and vegetables in the short term (field beans, chard, chicory), perennial herbs in the medium term (comfrey, rhubarb, horseradish), shrubs and trees in the long term (Elaeagnus, broom, hazel; poplar, willow, Paulownia, robinia).
Which layer to start from
There’s one distinction that works as a compass, and anyone who has been on one of my courses has heard it more than once. In Mediterranean climates, the design starts from the emergent layer: without the shade of the upper storeys little survives below, and without moisture, decomposition stalls, sometimes for months. In temperate climates, it starts from the low layer, which covers the soil through the winter, smothers weeds and produces plenty of biomass without stealing light that is already scarce.
Climates like ours ask for both at once, and that’s where it gets hard: a support plant has to cope with −9°C and +45°C, and many have surprisingly short windows of growth. Comfrey, chicory, chard and rhubarb are at their best in autumn, restart slowly in spring and suffer in the summer heat, while vetiver only really grows in summer. Finding species that work right through the year, or at least cover the gaps left by the others, has become the heart of my search.
An example: the banana and the eucalyptus
In the tropics, bananas and eucalyptus are the biomass engines of syntropic systems. Finding their analogues in our climates is one of the biggest challenges in adapting the method, and it helps to separate the two roles they play. The banana plays the watery role: soft, fast-growing tissues that release water and nutrients quickly when cut and laid on the ground. Rhubarb, comfrey, elecampane, horseradish, artichoke, cardoon and Jerusalem artichoke can do the same in our climates. The eucalyptus plays the woody role: height, shade, lignin, and a slower, longer-lasting input of carbon. Here that means elder, mulberry, willows, dogwood, robinia, Paulownia, poplar, Italian alder, hackberry and eucalyptus itself; in frost-free Mediterranean sites, tagasaste too.
The secret lies in management. Any species that responds well to heavy pruning, if kept in a “juvenile” phase (one year old, or cut a couple of times a year), develops fat, watery and crunchy tissues that play the role of the banana, releasing water and nutrients quickly when cut and left on the ground. These plants also keep feeding the soil food web through the growth pulse: each time they are cut, they shed part of their roots and release sugars into the soil, waking up the microbes that make nutrients available to their neighbours (I explain the mechanism in a dedicated post). A willow or an elder kept juvenile can do the banana’s job.
Finding them
To find support plants that work, I follow two paths at once: trialling species that look suited on paper, and watching the spontaneous vegetation in and around the field. At OrtoForesta, the weeds that dominate the field (couch grass, bindweed and creeping thistle) are anything but well-mannered. A few hundred metres away, though, creeping cinquefoil and lemon balm were growing well; I watched them for a couple of years and then established them in the system for good. The wild plants are the best consultants we have. After three years of trials, our low layer rests on cinquefoil, lemon balm, chives, sage, horseradish, artichoke, Jerusalem artichoke and yarrow, with field beans, chard and chicory in autumn and spring, and rhubarb in the shade of the trees. I told the whole story in a post on support plants.
A word on invasive species
Many of the best support plants are invasive somewhere in the world. Robinia, native to the eastern United States, has naturalised across Europe; tree of heaven, from China, spreads along roads and railways in Europe, North America and Australia; giant reed is a problem in the rivers of California and parts of Australia; autumn olive (Elaeagnus umbellata), a favourite nitrogen fixer in European agroforestry, is a noxious weed in much of the eastern US. The vigour, toughness and fast regrowth we look for in a support plant are the same traits that make a plant invasive.
I don’t think this should rule them out. In heavily disturbed landscapes these species are often already there, doing unasked the job of colonisation that our systems need, and the local list of invasives, tried out in the field, is a good place to start looking for infestation species. Inside a managed system, pruned hard and before flowering, they rarely get to set seed, though root suckers need watching. What matters is to know the rules where you farm (in the EU, tree of heaven and golden wreath wattle can’t be planted at all) and to keep the most mobile species, like giant reed, away from rivers and natural habitats, where nobody prunes them.
6. Putting it together: consortia
A consortium, what permaculture calls a guild, is a small community of species designed so that every plant has more than one function, and every function is covered by more than one plant. When I design one, I use a simple grid: the four layers (emergent, high, medium, low) down the side, three time horizons across the top (years 1–3, 3–8, and 10 and beyond), and in each box a production species and a support species.
Stratification
Syntropic texts give each layer a target cover: 20% for the emergent, 40% for the high, 60% for the medium and 80% for the low layer. The total, 200%, isn’t a sum. Imagine photographing the system from above at the height of each layer: at emergent height, crowns should cover a fifth of the ground; at the low layer, four fifths. A species’ layer depends on its need for light, not on its height: an olive can be kept as a medium shrub, but then it won’t fruit, and in the first summer of a young system a row of maize can be the emergent layer.
Density
Classic syntropic systems plant a tree every metre, with other plants just a few centimetres apart: a density that looks mad to a conventional eye. But nature doesn’t plant one tree at a time: it scatters seeds by the thousand, and all the mechanisms described in Part 1 evolved at high density. Density isn’t enough on its own, though. A row of tomatoes every 50 cm is dense but has no diversity: neighbouring plants should be of different species, so that their roots feed different microbial communities. A practical test: a well-made syntropic row shouldn’t let your gaze pass through it, except right after pruning. At OrtoForesta we had to triple the density of trees and shrubs to get there.
Simplified layouts
Mixing every layer in every row teaches you a great deal about how species interact, but it’s slow to prune and harvest. In systems focused on production and efficiency, two simplifications help. The first is stratified pairs: one row holds a high and a low layer (a fruit tree with an aromatic below), the next an emergent and a medium layer (a living pole of eucalyptus with currants below). The second goes further and separates support from production: rows of fruit trees with their low layer on one side, rows dedicated to support species on the other, cut to feed the alleys and the fruit rows. Productive density drops, but pruning, harvesting and machinery become far simpler, and the work can be delegated: one person can learn to prune one kind of row along its whole length.
7. A few examples
At OrtoForesta, the fruit rows are 12.9 m apart, with a narrower support row halfway, so there’s a row of trees every 6.45 m, and no-till vegetable beds (80 cm beds, 40 cm paths) in between. It’s a hybrid of the layouts above, and not a strictly syntropic one: I shared the original design, back in 2022, in a design preview. The fruit rows are mixed: each is 1.2 m wide, with a mulberry, persimmon or fig every 4 m, white willows between them at 1 m spacing for shade and biomass, elder and thornless blackberries in the medium layer, sage, lavender and broom covering the soil, and artichokes and Jerusalem artichokes along the edges. Each plant will leave the stage at a different time: the blackberries produce in the first years, the persimmons come later. The support rows, on the other hand, are simple: willow and elder every 2 m, pruned to feed the beds on either side.
8. What happens when the trees take over?
One question that many traditional and regenerative vegetable growers ask is: but what happens when the trees ta ke over? Left alone, they will, and the vegetables will end up in the shade: it’s the inhibition pathway from Part 1. The answer lies in the dynamic removal or strategic pruning of plants from the previous phase, to make space and provide a pulse to those of the next. Every year, pruning keeps the support species young and lets light into the alleys. Every few years, plants that have done their job are removed, like the blackberries once the persimmons are producing. Over decades, the focus itself moves on, from vegetables to fruit and eventually to timber, nuts and grazing. Or the vegetables move instead, as in the travelling garden, to a new patch cut from the oldest part of the system.
This is why pruning sits at the centre of a syntropic system. Each cut lets light into the lower layers and drops biomass on the soil, where it becomes mulch. Below ground, the pruned plant sheds part of its roots and releases sugars, feeding the microbes that make nutrients available to its neighbours. They grow faster, produce more biomass, and give us more to prune at the next round. So we design for pruning from the start: support species that regrow vigorously after cutting, rows we can reach with loppers and a ladder, a pruning height that matches the width of the alley. How and when to cut is the subject of Part 3.
9. What we still don’t know
Almost everything in this post comes from practice rather than from trials: of the 67 studies compiled in the most recent overview, only three come from Europe, all on young systems (Jacobi et al. 2025). Nobody has yet compared, in our climates, alley widths, establishment strategies or the stability of systems started with and without inputs, and the people best placed to do it are the farmers. If you’re keeping records of your own rows, I’d love to hear from you.
In the next post, I’ll go into detail on how to manage pruning through the year, in temperate and Mediterranean systems. Stay tuned!
References
Crawford, M. (2010) — Creating a Forest Garden, Green Books
Ducerf, G. — L’encyclopédie des plantes bio-indicatrices alimentaires et médicinales, Éditions Promonature
Jacke, D. & Toensmeier, E. (2005) — Edible Forest Gardens, Chelsea Green
Jacobi, J. et al. (2025) — The Lancet Planetary Health
Regulation (EU) No 1143/2014 — on the prevention and management of the introduction and spread of invasive alien species
Agenda Götsch — agendagotsch.com
Comments