Fluids

The water a soil keeps from the plant

Two days after a soaking rain a field has drained all the water it is going to drain, and soil scientists call what remains its field capacity. A plant then draws water out until it can pull no harder, and wilts; what is left is the wilting point. Both are suctions, and every suction is a pore size: field capacity is the water held in pores narrower than 4.4 micrometres, the wilting point the water in pores narrower than a tenth of a micrometre. A clay soil can be visibly damp, a quarter water by volume, and give a plant nothing — because the soil it is in equilibrium with is in equilibrium with air at 98.9 per cent humidity.

Assumes: How high water will climb · The pore that fills from dry air

Soak a garden bed and leave it. For a day or two water drains downward out of it, quickly at first and then more and more slowly, and after that it effectively stops: what remains is held against gravity indefinitely, until roots or evaporation take it. Soil scientists call that state field capacity. Plant something and let the soil dry, and the plant goes on drawing water until, at some point, it can draw no more; its leaves droop and do not recover overnight. The water left then is at the permanent wilting point. The difference between the two is the water a soil can actually offer a crop, and it varies by a factor of more than ten between soils that hold, in total, much the same amount of water.

The terms come from agriculture and sound like rough categories. They are not. Each is a suction — a negative pressure in the soil water, measured in kilopascals — and a suction in water held by surface tension is, by Laplace’s relation, a pore size. Read that way, field capacity and the wilting point are two radii, a soil is a distribution of pore sizes, and the question of how much water a plant can use is the question of how much of a soil’s pore space lies between the two.

Every suction is a pore size

How high water will climb found water rising up a narrow glass tube until the weight of the column balanced the pull of surface tension at its curved top: the narrower the tube, the higher the rise, because the pressure jump across a curved meniscus is 2γ/r2\gamma/r. Turn the experiment round. Water held in a pore of radius rr stays there until the suction applied to it exceeds that same jump, and then the meniscus is pulled through and the pore empties. For water wetting a mineral surface, with γ\gamma = 0.072 N/m,

ψ=2γr,ψ in kPa≈144r in µm.\psi = \frac{2\gamma}{r}, \qquad \psi\ \text{in kPa} \approx \frac{144}{r\ \text{in µm}}.

Every suction a soil scientist names is a pore size. The suction needed to pull water out of a pore against its surface tension, 2γ/r, against the pore's radius, on logarithmic axes. Water in pores wider than 4.4 µm drains away under gravity within a day or two; what is left is called field capacity, a suction of 33 kPa, the same as hanging from a column of water 3.4 m tall. A typical crop can pull water down to about 1.5 MPa, the wilting point, which empties pores down to 0.096 µm and corresponds to a column 153 m tall. The water a plant can use is the water held in pores between those two sizes.
Fig. 1 The suction needed to pull water out of a pore against its surface tension, 2γ/r, against the pore’s radius. Field capacity, a suction of 33 kPa — the same as hanging from a column of water 3.4 m tall — is the water left in pores narrower than 4.4 µm. A typical crop can pull water down to about 1.5 MPa, the wilting point, which empties pores down to 0.096 µm and corresponds to a column 153 m tall. The water a plant can use is the water in pores between those two sizes.

The conventional field capacity, 33 kilopascals, is the pull of a hanging column of water 3.4 metres long, and it empties every pore wider than about 4.4 micrometres; the water draining out of a soaked soil in its first day is the water in the wide pores between sand grains and in cracks and wormholes, which nothing in a soil holds against gravity for long. The wilting point, 1,500 kilopascals, is the pull of a column 153 metres long, and the pores it empties are a tenth of a micrometre across — the gaps between the plate-like crystals of clay minerals. The two numbers are conventions, and the angle a liquid makes is assumed to be zero, which on clean mineral surfaces it nearly is. But the conversion between suction and size is not a convention; it is the same physics that lifts water up a capillary.

A soil as a curve

A real soil has pores of every size from millimetres down to nanometres, and how much of its volume lies in each size range is what distinguishes a sand from a clay. The measured description is a retention curve: the water content left in a soil sample against the suction applied to it. Each step along the suction axis empties the next smaller class of pores, so the curve is a cumulative distribution of pore volume over pore size, read through Laplace’s relation.

A soil is a distribution of pore sizes, read as a curve. Water content against suction for class-average sand, loam, silt loam and clay, from van Genuchten curves with Carsel and Parrish's parameters. Sand's pores are almost all wide, so it is nearly empty by field capacity, holding 0.05 of its volume. Clay's are very fine, so it is still wet at the wilting point, holding 0.27 — water a plant cannot extract. Between the two lines, field capacity at 33 kPa and wilting at 1.5 MPa, silt loam gives up 0.14 of its volume, loam 0.08, clay 0.08 and sand 0.001.
Fig. 2 Water content against suction for class-average sand, loam, silt loam and clay, from van Genuchten curves. Sand’s pores are almost all wide, so it is nearly empty by field capacity, holding 0.05 of its volume. Clay’s are very fine, so it is still wet at the wilting point, holding 0.27 — water a plant cannot extract. Between the two lines, silt loam gives up 0.14 of its volume, loam 0.08, clay 0.08 and sand 0.001.

The curves are Martinus van Genuchten’s 1980 fitting function with class-average parameters published by Carsel and Parrish for thousands of American soil samples. Sand’s curve drops almost vertically at a few kilopascals: its grains are a fraction of a millimetre across and so are its pores, and once suction passes the few kilopascals that empty them, there is almost nothing left. Clay’s curve barely falls at all across the range a plant cares about, because its pore space is overwhelmingly in gaps far finer than a tenth of a micrometre. Loam and silt loam, mixtures of sand, silt and clay particles, have pores spread across all the sizes in between, and their curves descend gently through the whole range — which is the shape that gives a plant the most.

Three fates for the water in a saturated soil

Read off at the two conventional suctions, each curve divides a saturated soil’s water into three parts. Water above field capacity drains away under gravity within a day or two. Water below the wilting point is held in pores too fine for any crop to empty. Water between is what the soil can deliver.

Three fates for the water in a saturated soil. What happens to the water in a saturated soil, by class: the part beyond field capacity (palest) drains away under gravity; the part below the wilting point (darkest) is held in pores too fine for a plant to empty; the part between is what a plant can use. Silt loam offers 0.14 of its volume, loam 0.08, clay 0.08 — while holding 0.27 out of reach — and sand 0.001, having drained almost everything. A soil can be wet to the touch and give a plant nothing.
Fig. 3 What happens to the water in a saturated soil, by class: the part beyond field capacity drains away under gravity; the part below the wilting point is held in pores too fine for a plant to empty; the part between is what a plant can use. Silt loam offers 0.14 of its volume, loam 0.08, clay 0.08 — while holding 0.27 out of reach — and sand 0.001, having drained almost everything.

The arithmetic overturns the intuition that a soil which holds more water is better for a plant. Clay holds the most water of any of them at every suction a plant can apply, and gives up no more than loam does, because most of what it holds is in pores the plant cannot empty. Sand holds almost nothing once it has drained. Silt loam wins not because it holds the most water but because its pore sizes are concentrated in the window between 0.1 and 4.4 micrometres. Every farmer who has noticed that a heavy clay can be sticky underfoot while the crop on it wilts has seen the first fact; every gardener who has watched a sandy bed need watering daily, the second.

The figure for sand needs a qualification, and it is a qualification about the description, not about the sand. The class-average curve puts sand’s usable water at a tenth of a per cent of its volume, which is too low: measured sands typically offer a few per cent. The difference is water held not in pores but in thin films on grain surfaces and in the wedges where grains touch, the same pendular rings that let the sandcastle a spoonful of water holds stand up, and van Genuchten’s curve, which treats the dry end as a constant residual water content, describes them badly. The conclusion that sand offers little stands; the precise number for sand is the least reliable on the chart.

What a fraction of a volume is worth in days

A fraction of a soil’s volume becomes useful once it is turned into a depth of water. A crop whose roots reach a metre down into silt loam has 0.14 of that metre, 140 millimetres of water, between field capacity and wilting. A cereal crop in full leaf on a warm summer day transpires about five millimetres of water, so a fully wetted silt loam root zone carries it for about four weeks without rain. The same crop in loam has a little over two weeks, and in clay the same; in sand, whose usable water by the class curve is negligible and in reality a few per cent, the reserve is a matter of days, which is why sandy soils are irrigated little and often and clay soils seldom and deeply.

The depth matters as much as the fraction, and it is set by the roots rather than the soil. Shallow-rooted vegetables drawing on thirty centimetres of loam have twenty-odd millimetres to work with and need watering every few days in summer even on a good soil; trees drawing on several metres survive droughts that kill a lawn. The pore-size window decides what fraction of the soil a plant can use; the roots decide how much soil that fraction is taken of.

The wilting point is a humidity

Laplace’s relation turns suction into pore size. A second relation turns it into humidity, and the result is the most surprising number in the subject. Water under a suction has a lower chemical potential than free water, and so a lower vapour pressure: the air in equilibrium with it is not saturated. This is Kelvin’s relation, the same one the pore that fills from dry air used to explain why fine pores fill with condensed water from air well below saturation,

ppsat=exp⁡ ⁣(−ψVmRT),\frac{p}{p_\text{sat}} = \exp\!\left(-\frac{\psi V_m}{RT}\right),

with VmV_m the molar volume of water.

A soil at the wilting point is in balance with air at 98.9 per cent humidity. The relative humidity of air in equilibrium with soil water held at a given suction, from Kelvin's relation at 20 °C. Field capacity is 99.98%; the wilting point, 1.5 MPa, is 98.9% — air that feels saturated. Soil dried on a bench in a room at 50% humidity holds its last water at about 94 MPa, sixty times the suction a wilting plant can manage. Everything a plant can use lies in the top one per cent of the humidity scale, which is why a soil that feels dry has nothing left for it.
Fig. 4 The relative humidity of air in equilibrium with soil water held at a given suction, from Kelvin’s relation at 20 °C. Field capacity is 99.98%; the wilting point, 1.5 MPa, is 98.9%. Soil dried on a bench in a room at 50% humidity holds its last water at about 94 MPa, sixty times the suction a wilting plant can manage.

The wilting point corresponds to a relative humidity of 98.9 per cent. Everything a plant can use lies in the top one per cent of the humidity scale. A soil that would feel dry to a hand — that is, a soil whose water is in equilibrium with air at seventy or eighty per cent humidity — holds its water at suctions of thirty to fifty megapascals, twenty or thirty times beyond any crop’s reach. This is why soil scientists measure the suction directly rather than the moisture. The simplest instrument, a tensiometer, is a water-filled tube ending in a porous ceramic cup buried in the soil: water passes through the cup until the tension inside matches the soil’s suction, and a gauge reads it. It stops working at about 85 kilopascals, not because the gauge runs out but because the water inside it is under tension and gives up its dissolved air as bubbles — the failure that the water columns in a tree avoid only by being free of gas nuclei. Beyond that, samples are dried on porous plates under gas pressure, or the humidity of the air in equilibrium with them is measured and turned back into suction by Kelvin’s relation. The reason for all of it is the same: two soils with the same water content can differ in what a plant can get from them by an order of magnitude, while two with the same suction offer a plant exactly the same.

The same humidity explains the limit from the plant’s side. A leaf loses water to air that is far drier than 98.9 per cent, and the water lost has to be replaced through the stem from the roots. The pull that does it is a tension in the water columns of the xylem, which the column that is pulled, not pushed followed up a tall tree; a typical crop plant can sustain a tension of about 1.5 megapascals before its leaves can no longer stay turgid, and when the soil water’s suction reaches that, there is no gradient left to drive water into the root. The leaf cells, which hold their shape by the internal pressure the stem that stands by pressure described, lose it, and the plant wilts. The figure of 1,500 kilopascals is an average for crops; plants adapted to dry ground manage four or five megapascals, and they do it by raising the solute concentration in their cells, so that the pressure that comes from counting works for them against the soil’s capillary pull. The soil’s suction and the plant’s osmotic pull are measured in the same units because they are the same quantity, the chemical potential of water, and water moves from where it is higher to where it is lower whichever mechanism set it.

Field capacity is a height

The other end of the window has a quieter surprise. Field capacity is called a capacity, as though it were a property of the soil like its porosity. It is better thought of as a height.

Field capacity is a height above the water, not a property of the soil. The water content of a soil column standing at rest above a water table, against height, for the four soils. At rest the suction at each height equals the height, so each curve is the soil's retention curve turned on its side. Sand is saturated only in a fringe a few centimetres deep and almost dry above 30 cm, at 0.05 three metres up; clay is still at 0.35 three metres up and barely knows the water table is there. "Field capacity" names one point on this curve, chosen by convention; a soil with a water table a metre down holds more than the convention says.
Fig. 5 The water content of a soil column standing at rest above a water table, against height. At rest the suction at each height equals the height, so each curve is the soil’s retention curve turned on its side. Sand is saturated only in a fringe a few centimetres deep and almost dry above 30 cm, at 0.05 three metres up; clay is still at 0.35 three metres up. “Field capacity” names one point on this curve, chosen by convention.

A column of soil standing above a water table, once the water in it has stopped moving, has at every height a suction equal to that height, in metres of water, for the same reason that pressure in a still lake equals depth. The water content at each height is then the retention curve read at that suction, and the profile of water up the column is the retention curve turned on its side. Sand is wet only in a capillary fringe a few centimetres deep and dry above that; clay three metres up has drained barely at all. The 33 kilopascals of field capacity is the profile’s value 3.4 metres above water. Where the water table is a metre down, the soil above it holds more water than its “field capacity”, permanently; where it is ten metres down, the upper soil drains further, more slowly, than the convention says. The convention was chosen because, in many field soils, drainage slows to a negligible rate near that suction within a day or two — a statement about how fast water moves through partly empty pores, not about how much the soil can hold.

The slowing has its own cause. A soil’s ability to conduct water falls steeply as it drains, because the pores still full are the narrow ones, which conduct far worse than wide ones in proportion to their size, and the water has to find its way through films and wedges. The pore that lifts highest fills slowest found the same trade going upward, in a tube: the narrow pore pulls hardest and lets water through least. Drainage stops not because the suction balances gravity everywhere but because the remaining water can no longer move fast enough for anyone to wait.

The same suction lifts frost and dissolves salt

Soil suction appears in places that do not look like gardening. The frost that lifts by drinking found ice lenses in freezing ground drawing water up from below through the same fine pores, because ice in a pore has a lower chemical potential than liquid water and pulls on it exactly as a root does; silty soils, whose pore sizes give both a strong pull and a reasonable flow, heave most, for the reason silt loam gives a plant most water. And salts that crust a soil surface dissolve once the air’s humidity rises past their deliquescence point, the humidity at which the salt that dissolves in the air’s own water takes up water from the air — a humidity that, through Kelvin’s relation, is also a suction. In each case the bookkeeping unit is the chemical potential of water, and pore size, solute and ice are three different ways of lowering it.

What the curves leave out

Every number here uses the capillary picture of a soil as a bundle of cylindrical tubes of different radii, perfectly wetted, at equilibrium. Real soil pores are irregular, connected through narrower throats than their bodies, and empty at a lower suction on drying than they fill at on wetting, so a soil’s retention curve depends on its history — the hysteresis that makes a once-dried soil harder to rewet. The fine end of the curve is not capillary at all: water on clay surfaces is held partly by adsorption and by the osmotic effect of the ions near charged mineral faces, and at suctions of hundreds of megapascals it is a few molecular layers thick. The class-average van Genuchten parameters describe a typical sample of each texture class; individual soils of the same class differ by as much as the classes differ from one another, and organic matter, structure and compaction all move the curve. And field capacity and the wilting point are conventions with real but blurred meanings: plants do not wilt at a sharp suction, and drainage does not stop at a sharp time.

Still open: how roots meet the soil

The suction in the bulk soil is not the suction at the root surface, and that is where much of the remaining uncertainty lies. As a root draws water from the soil next to it, that soil dries, its conductivity collapses, and a zone of high suction forms around the root even when the bulk soil is moist; roots exude gels that change the soil’s wetting near them, and the root’s own contact with the soil shrinks as it loses water. How much of a plant’s drought stress is set by the bulk soil’s retention curve and how much by this thin, changing interface is an active question, studied now with neutron and X-ray imaging of living roots in soil, and its answer decides how well the old pair of conventions predicts when a crop will actually wilt.

Water held by surface tension in a pore of radius r resists a suction 2γ/r, so the soil scientist’s conventions are pore sizes: field capacity, 33 kPa, is the water in pores narrower than 4.4 µm, and the wilting point, 1.5 MPa, the water in pores narrower than 0.096 µm — the equilibrium humidity of 98.9%. A soil is a distribution of pore sizes between those radii: silt loam offers 0.14 of its volume, sand almost nothing because it has drained, and clay 0.08 while holding 0.27 in pores no crop can empty. Field capacity is a height above a water table more than a property of the soil, and a damp clay can kill a plant by keeping its water.

Part 10 of 10

This essay is one argument about Capillarity. The others:

The objects named here

The third axis, after the field and the reading path: the things themselves, and every essay that touches each one.

Capillary riseContact angleKelvin equationLaplace pressureTurgor pressureWater potential