Restinclières: What Thirty Years of Public Patience Made Visible

The temperate deep-dive in the Perennial Polycultures mini-series.

Restinclières is the longest-running silvoarable trial in Europe, designed by Christian Dupraz in 1995 near the French city of Montpellier. It is one of the most comprehensively monitored alley cropping sites in the world. It has produced thirty years of paired data on tree growth, crop yield, soil carbon, root architecture, light competition, water dynamics, and microclimate — all measured against monoculture and forestry control plots that share the same soil, slope, and weather. There is nothing comparable in temperate agroforestry. Most other long-term trials are younger, smaller, or lack proper controls. The empirical case for European temperate alley cropping rests, to a substantial degree, on what Restinclières has shown.

This piece is about what the trial has shown, what it has not yet been able to answer, and why the unanswered question is the one that matters now.

What Restinclières is

The site is a working agricultural estate on the southern edge of the Languedoc plain. The soil is a calcareous clay-loam over deep alluvium, the climate is Mediterranean with hot dry summers and cool wet winters, and the rainfall is around 800 millimetres a year — enough for cereals, not enough to make trees easy. The Département de l'Hérault, the regional public authority, owns the land and has supported the trial financially and politically — alongside other public bodies — since its establishment. Without that long-term institutional backing the trial would not exist. Thirty years of monitoring is not something that fits inside a normal research grant — what Restinclières has shown depends on the patience of public authorities willing to commit beyond electoral and budgetary cycles.

Christian Dupraz designed the original layout in 1995, working through INRA (now INRAE), the French national agricultural research institute. The trial parcels follow a strict experimental geometry. Tree rows are spaced 13 metres apart. Trees within the row are spaced about 4 metres in early years, thinned to roughly 9 metres at maturity, giving an adult density of about 85 trees per hectare. The alleys carry annual crops — mainly durum wheat and barley in the early decades, with rotations later including alfalfa, peas, and chickpea. Each agroforestry parcel is paired with a monoculture control plot growing the same crop without trees, and a forestry control plot growing the same trees without crops. This three-way comparison is what makes the data interpretable. Most agroforestry trials worldwide compare alley cropping against "the conventional system" in vague terms; Restinclières measures it against two specific, neighbouring alternatives.

The headline tree species is hybrid walnut (Juglans nigra × Juglans regia), mostly the cultivar NG23. This was chosen for timber value rather than fruit — the trial was set up to test whether high-value temperate timber could be grown in alley configuration without sacrificing arable productivity. Smaller experimental parcels at Restinclières have since added wild cherry, sorb, and grapevine, but the walnut–cereal system is the spine of the experiment and the source of most of the long-term data.

Restinclières in the European research network

Restinclières is the empirical anchor of European agroforestry research, but it is not isolated. From 2001 to 2005, the EU-funded SAFE project (Silvoarable Agroforestry For Europe) brought together trial sites in France, Spain, the Netherlands, and the UK to build a comparative dataset and to develop the modelling tools that have since become standard in the field. The Yield-SAFE biophysical model, which simulates tree–crop competition for light and water, and the Farm-SAFE economic model, which projects long-term financial returns, both came out of that project. Both lean heavily on Restinclières data for calibration. When agroforestry researchers in Germany, Belgium, or northern Italy run economic projections for proposed alley cropping systems today, they are running models that were trained on what Restinclières measured. The trial's influence on the field extends well beyond its own boundary.

What thirty years of data actually show

Five findings stand out. Some are confirmations of what was assumed; others were somewhat unexpected.

1. Land Equivalent Ratio

The Land Equivalent Ratio holds up. Across the agroforestry parcels, the combined output of trees and crops, expressed as a fraction of what would be needed in separate monocultures to deliver the same total, has consistently come in above 1. The figures sit in the 1.2 to 1.4 range across most parcels and most years. This is the headline agronomic finding and it has held through droughts, heat waves, and the maturation of the trees. A hectare of alley cropping at Restinclières produces what it would take 1.2 to 1.4 hectares of separated monocultures to match. The system is, in the strict sense of the metric, more land-efficient than the alternatives.

2. Light competition

The light competition curve is now mapped. As the trees grow, they shade the alleys, and at some point this begins to suppress alley crop yields. The work of Lydie Dufour, Dupraz, and their colleagues has documented exactly when and how this happens. The threshold rule that has come out of the trial — that crop yield is barely affected until the ratio of tree height to alley width reaches roughly 0.8 — is one of the most useful design rules in temperate alley cropping, because it tells designers how to size alleys so that competition kicks in late in the rotation rather than early. C3 cereals like wheat and barley tolerate the gradual increase in shade reasonably well. C4 crops like maize and sunflower would not, which is part of why they are not the alley crops of choice in walnut systems.

3. Tree root depth

Tree roots go much deeper than expected, and competition with the crop drives them even deeper. Work by Rémi Cardinael and colleagues has documented walnut roots active well into the deep soil profile at Restinclières, in some parcels reaching the bedrock at 9.4 metres. Interestingly: the roots of trees grown with crops in the alleys go deeper than the roots of trees grown alone in the forestry control. The crop, by occupying the upper soil layers, forces the tree to develop its functional rooting zone below the crop. This is counterintuitive — the intuition is that the trees and crops would compete in the same soil layer. The trial documented the opposite: vertical separation of the rooting zones, partly induced by the crop itself, is one of the reasons the alley cropping system works as well as it does.

4. Carbon accumulation

Carbon accumulates in the agroforestry parcels in two pools, and the distinction matters. Cardinael and colleagues have documented soil organic carbon accumulation under the French silvoarable systems including Restinclières at roughly 0.25 tonnes of carbon per hectare per year in the upper 30 centimetres, with additional storage in deeper layers bringing the figure to about 0.35 tonnes per hectare per year over the upper metre at Restinclières specifically. On top of this, the trees themselves accumulate carbon in their above- and belowground biomass at roughly 0.65 tonnes per hectare per year. The figures are point estimates with substantial uncertainty in each component, but they sum to approximately 1 tonne of carbon per hectare per year for the whole system.

This is several times slower than what the most productive tropical multistrata agroforestry systems can sequester — figures of 5 to 10 tonnes per hectare per year are reported for tropical systems with fast-growing trees and dense canopies — but the comparison needs to be read carefully. The Restinclières system stores most of its carbon in the soil pool; tropical multistrata systems store most of theirs in aboveground biomass. These two pools have different fates. Biomass carbon returns to the atmosphere when the tree is harvested or dies, while stable deep soil carbon persists on much longer timescales. The tropical advantage in raw accumulation rate is real, but so is the temperate trade-off: slower accumulation, with a larger fraction of the total ending up in the more durable pool.

Multiplied over decades and over the landscape area that European policy might plausibly bring under alley cropping, the carbon mathematics begin to matter — both in absolute terms and in the durability terms that climate policy is increasingly being asked to weigh.

5. Microclimate buffering

The microclimate buffering effect is real, increasingly important, and probably the finding with the largest implications going forward. The buffering does not work mainly through direct cooling of the air in the alleys — that effect is small and inconsistent across alley cropping studies. It works through three mechanisms acting together. The trees slow wind, which lowers the rate at which water is pulled from the crop's leaves and from the soil surface. They shade portions of the alley during peak summer hours, lowering leaf-surface temperatures on the hottest days. And they reduce the vapour pressure deficit experienced by the crop in the immediate vicinity of the tree rows, which means the air around the crop is less aggressively pulling water from it. Each effect on its own is moderate. Together they substantially reduce the cumulative heat and drought stress the crop experiences over a growing season.

The Restinclières group has quantified the buffering through a combination of field measurement and biophysical modelling. In a recent study using long-term Restinclières data calibrated against projected mid-century Mediterranean climates, heat, drought, and nitrogen stresses on durum wheat in agroforestry parcels were reduced by 20 to 35 percent compared with monoculture controls, even with medium-sized trees in the alleys. The reduction is an alley-wide average and the underlying picture is a gradient: shade and humidity effects are strongest in the strips next to the tree rows and weaken toward the centre, while wind reduction reaches across the entire alley because mature tree rows shelter the full thirteen metres between them. The crop in the middle of the alley never receives direct shade, but it does experience slower wind and lower evaporative demand, and that alone is a substantial part of the buffering. The effect strengthens as the trees mature. By the late phase of the rotation, the agroforestry crop experiences a meaningfully gentler climate than the open-field crop on the same soil under the same weather.

As Mediterranean and continental European agriculture moves into a climate where 40-degree summers are no longer rare, the value of this buffering effect is rising. It may turn out to be the most economically and ecologically consequential finding of the entire experiment, not because it was the most surprising, but because the world around the trial has changed in a way that makes it matter more.

What the trial has not resolved

Three things are worth being honest about.

The economics at scale are still partly modelled rather than observed. The Farm-SAFE projections, calibrated on Restinclières data, suggest that walnut–cereal alley cropping is profitable over a full rotation under a range of grant regimes, and that under some scenarios it outperforms separated monocultures financially as well as ecologically. But the trial is a publicly supported research site. It is not a commercial farm carrying its own land costs, financing its own establishment from farm credit, or selling its timber on a commodity market at the end of a 60-year rotation. The economic case rests on modelling and on the partial evidence from French commercial walnut alley cropping farms, where the rotations are still incomplete. The strongest economic claim that can honestly be made today is that the modelling is consistent with profitability under reasonable assumptions, and that what limited commercial-farm evidence exists is broadly consistent with the modelling. That is not the same thing as a completed empirical demonstration.

The trial is still only halfway through a tree rotation. Restinclières is now thirty years in. The hybrid walnuts were planted to be harvested for timber at sixty to seventy years of age. The data on establishment and middle-rotation phases is rich; the data on late rotation, timber harvest, and what happens to the alleys when the canopy peaks and then is removed, is still ahead. We are reading findings from a half-finished experiment and projecting them forward. Most of what we think we know about the late phase of a temperate walnut alley cropping system is inference rather than measurement. Restinclières will eventually answer this, if the institutional support holds. It cannot answer it yet.

Transferability across European climates is partly answered and partly not. Restinclières is Mediterranean. The walnut species, the soil profile, the rainfall pattern, the depth to bedrock, and the temperature regime are all specific. The SAFE project sites and subsequent trials in Atlantic France, the Netherlands, the UK, and Germany have shown that the broad principles — light competition curves, root vertical separation, microclimate buffering — carry across climates with adjustments. Whether the LER figures, the carbon sequestration rates, and the economic projections transfer with the same fidelity is less certain. A working farmer in Schleswig-Holstein or in Wallonia cannot read the Restinclières data straight onto her own field. The principles transfer; the numbers need recalibration.

The case for persistence – why should we keep on with the Restinclières trial?

Two arguments hold up.

Data this rich does not exist for any other temperate alley cropping system, and it cannot be created retroactively. If European policy is going to make evidence-based decisions about agroforestry over the coming decades, Restinclières is much of the evidence base. Continuing the trial is not just a research investment; it is an investment in the policy infrastructure that surrounds the practice. The marginal cost of running the trial at least until the rotation completes is small compared to the marginal value of having empirically anchored answers to questions that will keep arriving as the climate changes.

The microclimate buffering findings are increasingly important. The trial was set up in 1995 to answer agronomic questions about productivity and ecosystem services. Restinclières now sits inside a world where summer heat extremes are common, where vineyard and orchard yields are increasingly limited by heat stress, and where the value of in-field buffering has risen sharply. Restinclières happens to be one of the few sites in the world that can document this with thirty years of paired control data. The trial's relevance has grown beyond its original design, and that growth is likely to continue.

What Restinclières cannot answer

What the trial cannot answer is the question of how to translate that agronomic case into adoption on commercial farms operating under annual credit cycles, short land tenure, and policy frameworks that were built around annual crops. Restinclières is publicly supported, monitored by full-time researchers, and protected from the financial pressures that would test whether a working farmer could absorb the establishment years and finance the wait. The trial is, in this sense, an ideal-conditions experiment. It shows what the system can do when the institutional environment is favourable. It does not show what it would take to build a favourable institutional environment in the first place.

That is not a criticism of the trial. It is a recognition of what a research site can and cannot do. The agronomic question was the right question to ask in 1995. The financial-frameworks question is the right question for now.

Whether it gets answered in the coming decades will determine whether the agronomy of temperate alley cropping translates into landscape-scale change, or remains what it has been so far: a beautiful, well-documented exception.

A note on process: see my AI Use Policy for how I work with AI in producing this website's content.