Could we grow towards a landscape portfolio?
Conclusions of the perennial polycultures mini-series.
The dehesa is the Iberian agrosilvopastoral system that covers about four million hectares across the Spanish southwest, and that exists in a parallel form in Portugal under the name montado. It is the longest-continuously-functioning perennial polyculture at landscape scale anywhere in Europe. It is not preserved as a museum or a heritage site. It is an actual working agricultural region of approximately the same area as Belgium.
In the dehesa landscape, cork oaks (Quercus suber) and holm oaks (Quercus ilex) stand at roughly 40-50 trees per hectare. They are not in rows. They are widely spaced and individually shaped by centuries of pruning, casting open shade over a ground layer of grass, herbs, and occasional cereal stubble. Pigs move between the trees, fattening on the acorn fall, while sheep and cattle graze elsewhere in rotation. A few rotated parcels carry barley or oats. The trees themselves are harvested for cork on a nine-year cycle. The system produces, simultaneously: cork, acorns for the pigs, grazed pasture for the livestock, cereal grain, and firewood and charcoal from pruning. Traditionally, the system has also produced honey, wild game, and various non-timber forest products. The estate has been operating like this, more or less, since the late Middle Ages.
The dehesa is worth opening this piece with because it answers, by existing, the question this mini-series has been building toward: whether a landscape can combine multiple perennial systems with productive output and remain economically viable across long time horizons. The answer is yes — with very specific conditions. The dehesa is not a perfect template for what a contemporary landscape-portfolio agriculture would need to do — it integrates only marginal cereal cultivation with its perennial components, while the contemporary challenge is to integrate substantial annual cereal production with perennial systems. But it is the proof that the portfolio principle can sustain a working agricultural region across centuries.
The economic, ecological, and fire-resilience logic of the dehesa
The dehesa is a portfolio. No single product carries the system. The Iberian pig fattened on acorns (cerdo ibérico) produces premium pork that has become a high-value export. Cork is a substantial commodity income. The grazing animals turn the herbaceous understorey into meat, wool, and milk. The cereals provide grain for animals and people, and the rotation maintains soil fertility. Each component would be marginal on its own. Together they make the landscape economically robust.
The ecological logic is similar. The trees provide shade and microclimate buffering for animals and crops in a hot dry climate. The grazing maintains the open structure that the trees need to regenerate. The cereal rotation provides a periodic disturbance that resets the herbaceous layer. The whole landscape supports substantial biodiversity — dehesa systems host imperial eagles, Iberian lynx, black storks, and an unusually rich invertebrate community for an actively farmed European landscape.
A third logic has become more visible as Mediterranean climate change has progressed. The dehesa is among the most fire-resistant landscapes in Spain. Continuous grazing keeps the understorey low, preventing the build-up of dense shrub fuel that drives high-intensity wildfires. The wide tree spacing breaks the canopy, so a ground fire cannot easily climb into the trees. The cork oak and holm oak themselves are unusually fire-tolerant — cork bark is a thick insulating layer and both species can regenerate from the base after burning. As surrounding Mediterranean landscapes have become more fire-prone with climate warming and the abandonment of traditional grazing, the dehesa's structural fire-resilience has become one of its more consequential services. Wildfire prevention programmes in Spain and Portugal now pay shepherds to graze livestock in firebreak areas precisely because this is what extensive grazing does anyway.
What the series has shown
Before working out what a landscape-portfolio agriculture might look like, it is worth consolidating what the previous five pieces have demonstrated.
The opening piece in the series laid out the territory: most agriculture is annual monoculture. Perennial polyculture occupies a small and structurally interesting quadrant of the design space; it contains both heritage systems and modern engineered ones, and both deserve attention.
The Kernza piece showed that the change-the-crop approach — developing genuinely perennial grain species — is real yet slow. After forty years of breeding, Kernza yields about 25 percent of annual wheat. The breeding is accelerating but the timeline is measured in decades, not seasons. Kernza is a long bet on the proposition that staple grain agriculture itself can become perennial to some extent.
The alley cropping piece showed that the change-the-design approach — keeping the crops annual but integrating perennial trees into the field — has both ancient lineage (heritage systems) and modern engineered forms (temperate silvoarable). Adoption is slower than the agronomy would justify, mostly for institutional and financial reasons that the agronomy itself cannot solve.
The Restinclières piece documented this directly: thirty years of public patience produced an agroforestry experiment with paired controls revealing that the agronomy works, while the institutional environment for adoption is not generally favourable.
Shade coffee then showed what happens when a perennial polyculture has been allowed to function commercially across more than a century: it persists where the surrounding economic environment values what it produces, and contracts where these mechanisms fail.
What all five pieces converge on is a single structural finding. The agronomy of perennial polyculture is not the limiting factor. The institutional and economic environment that surrounds it is. Different systems hit different versions of this constraint, but the constraint itself is consistent.
Why a portfolio rather than a single system
If the agronomy works and the economics are the constraint, the practical question becomes: what kind of agricultural landscape would resolve the constraint? Not by abandoning annual cereals (which feed most of the world), and not by waiting for any single perennial system to scale to dominance. By integrating multiple systems at the landscape level, each playing a specific role that the others cannot.
This is the portfolio argument. It rests on three observations.
First, no single system from the series can scale alone. Kernza yields are too low to replace annual wheat at the scale humanity currently eats wheat. Alley cropping is land-intensive in ways that limit how much of any region it can occupy. Shade coffee depends on a buying public willing to pay for sustainability, which most staple crops do not have. Each system addresses part of the problem; none addresses all of it.
Second, the systems are complementary rather than competing. Kernza addresses the cereal itself. Alley cropping addresses the field. Shade coffee addresses the high-value crop. The dehesa addresses the landscape. Different climates, different intensities, different crops — but no contradiction between them.
Third, working agricultural landscapes throughout history have been portfolios. The pre-industrial European farming landscape was a mosaic of small woodlots, hedgerows, pastures, orchards, arable fields, and semi-natural areas at the scale of the parish or the watershed. The Hungarian Alföld before collectivization, the Bocage of Normandy, the German Streuobstwiesen (apple-meadow systems) — these were not single-system landscapes. They were portfolios that allocated different land to different uses according to soil, slope, climate, labour availability, and household need.
What the portfolio would contain
A working portfolio would integrate several components, each occupying a fraction of the agricultural area appropriate to local conditions:
Annual cereals would remain the backbone. Wheat, maize, rice, barley, oats, and pulses are not going away, and they should not. These are the calorie crops that feed humanity. The question is whether the annual cereal area can be reduced — not eliminated — by integrating it with perennial components that offset some of its ecological costs.
Alley cropping for the cereal land. Perennial trees integrated with annual cereals on the same parcel. Produces 20 to 40 percent more total output per hectare than separated monocultures, plus soil carbon, biodiversity, and microclimate benefits.
Perennial grains like Kernza for marginal cereal land. Sloping ground, thin soils, water-protection zones — places where annual cereals carry high erosion risk. The crop does not need to replace annual wheat; it needs a niche where its lower yield is offset by the ecosystem services it provides.
Dehesa-type systems for the dry, hilly, low-intensity land. Land currently either underused or pushed into intensive systems it cannot support. Low-intensity perennial polyculture works on such land given the right institutional framework.
High-value perennial polycultures — shade coffee, fruit polycultures, agroforestry timber — for the niches where consumer markets will pay for the system's outputs. Small fractions of any landscape, but where the economic logic is most directly visible.
Semi-natural areas — hedgerows, woodlots, riparian strips, ponds, field margins — as the connective tissue. Not strictly agricultural, but the unmonetized infrastructure that makes the whole landscape function ecologically. Pollinators move through them. Predators and parasitoids nest in them. Water moves through them.
What stands in the way
The constraints of the landscape portfolio are institutional, financial, and structural.
Land tenure is the most fundamental. Most European agricultural land is held under arrangements — short-term leases, indebted ownership, family succession crises — that punish multi-decade thinking. A farmer who plants alley cropping trees that will mature in forty years is making a bet that requires institutional and economic stability across two generations. Most contemporary land arrangements do not provide that. The dehesa exists partly because Iberian estate structures, for all their feudal origins, did provide the long time horizon the system needs. Whether comparable horizons can be built into modern land-tenure arrangements is, plausibly, the central question for the next decades of European agriculture.
Financial infrastructure follows from tenure. Banks lend against the value of an asset that can be liquidated within a credit cycle. A perennial system that takes thirty years to mature does not fit this model. Establishment costs are real; the productive phase comes later. Without specific public or institutional credit instruments designed for long-horizon investment, the financial side of portfolio agriculture is structurally hostile to its own success.
Policy frameworks (especially in the EU) are typically built around annual crops, with payments calibrated to annual production cycles and reporting requirements that perennial systems do not fit cleanly. Recent CAP reforms (eco-schemes, payments for ecosystem services, agroforestry-specific measures) have begun to address this, but the institutional weight of annual-cropping assumptions is substantial.
Consumer markets reward specific high-value outputs rather than total system output. A landscape that produces moderate amounts of many things — the dehesa, the perennial polyculture mosaic — is harder to monetize than a landscape that produces large amounts of one thing. The shade coffee piece made this point in tropical context; the same logic applies in Europe.
Each of these constraints is solvable in principle. None has been solved at scale.
What a working portfolio would need
A landscape-portfolio agriculture will not emerge from any single intervention. But three things, if they were put in place, would substantially shift the structural conditions that currently prevent it.
The first is long-horizon land tenure, possibly based on the model of community land trusts and farmland cooperatives such as Terre de Liens in France, Kulturland and BioBoden in Germany, and the Biodynamic Land Trust in Britain. These are non-profit structures that purchase farmland with community investment, then lease it to farmers on multi-generational terms. Long horizons in land tenure are the precondition for everything else. Without them, no farmer is going to plant a tree they will not live to harvest.
The second is patient credit. Public banks, agricultural cooperatives, and impact investors offering credit instruments specifically designed for perennial-system establishment. A thirty-year loan against a long-term investment is a very different financial instrument from a five-year crop loan. Cork-producing estates in Iberia have been financed in something like this way for centuries, and the basic logic can be translated to other perennial systems.
The third is policy that values portfolio outputs. Payments for ecosystem services that capture soil carbon, water protection, biodiversity, and microclimate buffering at the landscape scale rather than the field scale. Procurement frameworks that allow public bodies to buy from portfolio landscapes. Certification systems that recognise total system output rather than only the high-value cash crop.
None of this is radical. All of it has precedent. The dehesa exists because some version of these three conditions has historically been in place across the Iberian southwest. The question is not whether such conditions can support perennial-polyculture landscapes — the dehesa answers that. The question is whether the same conditions can be built, deliberately, into the institutional environment of contemporary agriculture.
If they can, the landscape we will look at in fifty years will be a portfolio — annual cereals on their best land, alley cropping where they meet trees, perennial grains on marginal slopes, dehesa-type systems on the dry hills, high-value perennial polycultures where consumer markets support them, and a connective tissue of hedgerows and field margins binding the whole landscape together. The components already exist somewhere. The system that integrates them does not yet exist. It can be built.
A note on process: see my AI Use Policy for how I work with AI in producing this website's content.