Shade, spacing and water on a coffee farm

A shade plant that learned to stand in the sun

Sourced

Coffee evolved under forest and was long considered to need shade. The physiology review reports that arabica raised from wild Ethiopian material is severely stressed and yields poorly without overhead shade. Cultivated varieties are different, and it cites one author's explanation: nearly all of them descend from plants taken from Ethiopia to Yemen and grown there for centuries in a dry climate without shade, before coffee travelled on to Asia and the Americas.

The result, on that account, is a plant that keeps the physiology of a shade species but tolerates mild drought and full sunlight. The review adds that some varieties, Typica among them, are not suited to open cultivation.

What shade trees are reported to change

Sourced

From a 2022 review of recent research on shaded coffee, and the physiology review it draws on.

The air around the coffee

Recent studies tended to confirm earlier ones: lower air temperature, wind speed and radiation under shade, fewer weeds, less water lost by the coffee plants themselves, and protection from frost. The review adds a qualification on water: shade trees transpire too, and in the study it cites the trees and the coffee together used more water than coffee in full sun.

Flowering and yield

A coffee leaf can partly make up for lower light, but shade lowers the intensity of flowering, and the review says that can reduce yield if the shade is too heavy. It calls shade a highly contentious factor for exactly this reason.

The size of the crop on each branch

Shaded trees are described as having fewer nodes per branch and a lighter fruit load, so less competition between fruits, larger beans, less alternation between heavy and light years and less branch die-back.

What ends up in the cup

Several studies the review lists found shade improving cup quality, and bean size with it. One, on Caturra in southern Colombia at about 1,440 to 1,630 metres, found the opposite, which the review puts down to shade making an already cool site too cool. It calls the effect of shade on quality highly contentious.

Pests and diseases

No single direction. The review says results vary with the site, the management, the shade species and its density: some studies found more pest or disease under shade than in full sun, others less, where there were more birds and ants or a changed microclimate. It describes many of the findings as locally specific.

The wildlife that comes with trees

Sourced

One measured example of what the animals on a farm are worth is the Costa Rican experiment described on the page about the coffee plant, in which keeping birds and bees off arabica branches cut yield by about a quarter on average. Its authors conclude that habitat for native wildlife can pay for itself in the crop. It is a single study, and it measured the animals, not shade trees as such.

Claims on packaging about shade-grown coffee are a separate matter, covered in the guide to sustainability claims. The atlas entry for Nicaragua describes how much traditional shade has been lost there, on the evidence of a published review.

How close to plant: no single answer

Sourced

Planting trees closer raises yield per hectare up to a point. The physiology review says the best density for arabica depends on the variety, on how much water and nutrition are available and on how hot and dry the air is, and that the main limit is the trees shading one another as they age, which depresses the crop strongly. It reports that closer spacing has been widely adopted where tractors cannot be used between the rows or are not needed.

Dense planting also changes the trees. The review says roots go deeper in dense stands and less water is lost, and that because the mutual shade reduces heavy flowering, it also reduces overbearing and the die-back that follows. The review quotes densities that particular authors arrived at for particular varieties; they are not repeated here, because it presents them as dependent on all of the above.

Water as a tool, where there is water

Sourced

Irrigation does two different jobs. Where there is a reliable dry season, the review describes irrigated farms in Brazil withholding water and then applying it so that the trees flower together, which concentrates the harvest. It also calls getting the timing right on a farm a difficult task and one of the most disputed questions among Brazilian growers and researchers. And because bean size is set by water in the weeks when the fruit is expanding, it says irrigation, shade or mulch in that period may give larger beans.

The review sets out the reasoning behind withholding water, and its limits. Flower buds can respond to rewetting only once they have reached a certain stage, and because buds are started at different times they do not reach it together. If the trees are kept short of water through the last part of bud development, later buds catch up while earlier ones wait, so that when water returns a larger share of the buds on each branch open at once. The difficulty is knowing how far to let the shortage go and when to end it. The review cites a water potential in the tree of about minus 1.2 megapascals as a critical value for triggering flower opening, calls the practice of fixing calendar dates for stopping and restarting irrigation an empirical way of doing it, and lists what makes any one date unreliable: the soil, temperature and humidity, the size, spacing and age of the trees, and the variety. It suggests that the stage the buds have reached may be a better guide than the severity of the stress. It also observes that flowering several times in a season spreads the filling of the seeds over time and so spares the tree's reserves, which is its reason for saying the coffee tree does not seem to have evolved towards a single flowering.

Irrigation as a plain supplement has been measured too, in older work the review summarises. In a study the review sets among Kenyan ones, irrigating at appropriate times raised yields by 30 to 40 per cent, and the irrigated trees carried many more fruit-bearing nodes with fewer fruits on each. In another experiment the review cites, which combined irrigation, mulch and nitrogen, the number of fruiting nodes was the part of yield that responded, and mulch gave a larger average response than either irrigation or nitrogen. Those are results from decades ago, reported at second hand, and no amount or timing of water is taken from them.

Whether irrigation is possible is a matter of the place. The atlas entry for Vietnam describes a robusta crop that depends on it and the pressure that puts on water; the entry for Brazil describes irrigated conilon in Espírito Santo and Bahia.

Pruning and the canopy: two studies

Sourced

Pruning decides how many stems a tree carries, how old they are and how much light reaches the inside of the bush. An experiment published in 2026 compared two systems in twelve-year-old, unirrigated Catuaí at one site in Espírito Santo, Brazil. In the traditional system a tree keeps one stem, or sometimes two, grows freely, and is cut back to a stump when it loses vigour after ten to fifteen years. In programmed cycle pruning, a method devised for conilon and adapted to arabica, each tree carries a set number of upright stems that are renewed in rotation, and side branches that have borne heavily are taken off. Over four harvests the cycle system averaged 2.25 tonnes of dry coffee per hectare against 1.59. Branch length, canopy width and flowers per branch did not differ. What differed was where the tree was active. Under traditional pruning the trees had a heavy skirt of growth low down and a weak top, a habit the authors call girdling, and leaves in the upper canopy photosynthesised at roughly half the rate of leaves lower down; under the cycle system the two levels were alike. Yield did not rise in a straight line with the number of stems.

That is one experiment, and the densities it tested are its own; none is repeated here as a recommendation. Its authors add a caution of their own: the cycle-pruned trees had been renewed in 2013 and 2014 and the traditional ones had not, so the comparison is between a renewed planting and an unrenewed one as well as between two systems. It does show the mechanism by which pruning is supposed to work, a more even canopy, being measured and not assumed.

Position in the canopy matters to the fruit as well as the leaf. A study published in 2020 took fruit from above and below 1.7 metres on the same trees, three to four metres tall, of one arabica variety on one farm in New South Wales, Australia. Green beans from the lower canopy held slightly more caffeine and trigonelline, 1.11 against 1.07 and 0.93 against 0.90 grams per hundred, and more sucrose, 8.6 against 7.7; size and weight did not differ. Roasted, they were judged to smell stronger: twelve trained panellists described the two, and 86 untrained staff and students, asked which of a pair had the more intense aroma, chose the lower-canopy coffee more often than chance. Nobody brewed or tasted it. From the activity of genes in the developing fruit the authors infer slower, longer ripening in the shaded lower canopy; they did not time it. It is a comparison within trees at one harvest, and it does not show what shading a whole farm does.

Four kinds of statement about shade

CoffeeHQ explanation

Shade is argued about with evidence of very different strengths, often in one sentence. The examples are from documents read for these pages; the sorting is CoffeeHQ's.

Observation

Something recorded on farms as they are. Shaded farms differ from unshaded ones in altitude, variety, inputs and the people who run them, so a difference in yield or quality between them cannot be put down to the trees. Much of the evidence the shade reviews gather is of this kind.

Experiment

Shade applied to some plots and not others at one site. A trial in the Ecuadorian Amazon found young robusta under about a quarter shade yielding the same as in full sun over three months of picking, and taller, with more chlorophyll in its leaves, under the two leguminous canopies, which were also the shadiest. Its authors call for longer study before the result is relied on.

Model

A projection from climate data. One study modelled the climate suitability in the 2050s of arabica, cocoa and the hundred commonest shade trees of Mesoamerica. It projects that 79 per cent of the tree species used over coffee lose much of their suitable range, among them 25 of the 30 nitrogen-fixing species, with the commonest Erythrina and Inga losing about half; and that most coffee land would still suit more than thirty other species, many of them already present on farms in small numbers. Its authors conclude that the mix of species would have to change, and say that their method looks at climate alone and leaves out pests, soil and what farmers know and want. A model of where trees could grow says nothing about what they do for the coffee under them.

Marketing

A claim on a bag that shade-grown coffee tastes better or is better for wildlife. The section above gives what was read on wildlife. On taste, CoffeeHQ read no comparison of the cup from shaded and unshaded plots of the same coffee, and treats the claim as a trade expectation. Such comparisons have been published and were not read, so that means untested by anything read here, not refuted.

Drought in a crop that was already irrigated

Sourced

The clearest account read of what water shortage does at the scale of a region is from Espírito Santo, the Brazilian state that grows most of the country's conilon. A study assembled rainfall, temperature, production and satellite records for nine municipalities over six seasons, 2010/11 to 2015/16. Average yield fell from about 37 bags per hectare in 2013/14 to about 18 in 2015/16, the worst season, and the authors attribute that partly to the trees having already been damaged by the weather of the season before.

Two of its findings cut against a simple picture of drought. In the authors' statistical model rising air temperature had a larger negative effect on production than falling annual rainfall, and the two seasons the authors single out for low yields were about one degree Celsius warmer on average than those before them. One of those two, 2012/13, had enough rain and enough water for irrigation; the authors put its poor crop down to heavy rain at flowering. And the rain that mattered was the rain of particular months, September to December and April to August, not the annual total. They add that a short dry spell from January to March, when the fruit is filling, does more harm than a longer one in winter, and say that needs testing under controlled conditions.

More than nine tenths of these plantations were irrigated, and that is the point of the heading. Yield and annual rainfall were only weakly related across the six seasons, which the authors read as irrigation doing its work in ordinary years. In 2015/16 it could not. Rainfall below the historical average from August 2014 to September 2016 emptied the rivers, streams and small earth dams the irrigation drew on, and by the authors' account most plantations lacked an adequate supply for most of that season. They remark as well that trees raised with ample water are usually the less able to withstand a shortage when one comes. All of this is an analysis of six seasons of records and not an experiment: the correlations are between the weather and the crop across a region, and six seasons are few to fit a model to. It supports a narrower statement than is often made. Irrigation protects a crop only while there is water to irrigate with, and in that hot, dry period there was not.

Pruning and feeding in the atlas, and what happens when feeding stops

Sourced

On pruning, the review makes one relevant point: robusta in Brazil is pruned on a cycle that renews its stems, and that is why its crop swings less between heavy and light years than arabica's. The atlas entries add local detail, such as the cutting back of old trees in Ethiopia and a pruning programme in Guatemala.

On nutrition, the only evidence read concerns what happens when it stops. A study of the leaf rust epidemics in Colombia and Central America found that each followed a period of low profitability in which growers cut spending on fertiliser and crop protection, and concluded that the weakened trees were more vulnerable. Nothing here says what or how much a coffee tree should be fed; that depends on soil analysis and local advice. The guide to coffee soils and plant nutrition sets out what five studies found about the soil itself.

No recommendation appears on this page

CoffeeHQ explanation

It gives no recommended shade level, spacing, pruning system or irrigation amount, and no fertiliser of any kind. It does not describe soil management, weeding or the costs of any of these practices. The effect of shade on biodiversity beyond the one experiment cited was not read and is not claimed.

What to do next

• Coffee pests and diseases — The organisms that shade, spacing and neglected trees interact with.
• Coffee and climate change — Why shade is proposed as an adaptation, and how strong the evidence for that is.
• Coffee sustainability claims — What a shade-grown or bird-friendly label does and does not establish.
• Vietnam in the atlas — A crop built on irrigation, and what that has cost in water.