Soil-water tension.

Moisture percentage tells you how much water is in the medium. Tension tells you how hard the plant has to pull to get it. That effort-to-extract is what the root actually feels, and it is the number that lets you compare a peat pot, a coco slab, and a field bed on the same scale. The reading travels between media. The setpoint you water at does not, and that distinction is the whole art of using it.

Water in a substrate is held by the particles. A little water clings loosely and the plant takes it easily; the last of it is gripped so tightly the root can't win it back. Matric potential, measured as tension, is that grip: the suction, in kilopascals (kPa) or centibars (cbar), that the root has to overcome. Saturated medium is near zero; as it dries, tension climbs. In a mineral soil, 10 cbar is wet and easy, 50 cbar is getting hard, and past that, sensitive crops start to throttle growth to conserve water long before they wilt. In a soilless container mix the whole working range sits far wetter than that: a peat or coco substrate can be past its comfortable range at 5 cbar. The unit is the same; the numbers a crop lives at are not.

Content tells you how much; tension tells you how hard.

This is the distinction almost every cheap "moisture meter" misses. Volumetric water content (VWC) is a quantity: liters of water per liter of medium, a percentage. It is useful, but it is not what the plant responds to. The plant responds to availability, the energy it must spend to pull water free, and two media at the identical VWC can present wildly different availability. Forty percent moisture in a peat mix and forty percent in a sandy mix are not the same offer to the root: one holds it loosely, one grips it hard. The plant feels the grip, not the percentage.

What travels between media, and what does not.

Because tension measures the physical pull rather than the volume, the measurement is comparable across media: 15 cbar is the same pull on the root in coco, rockwool, soil, or a raised bed. "40% VWC" is not comparable, because the same percentage is a different grip in every substrate. That is the real advantage of tension, and it is why a tension reading can be written down, handed to someone else, and put in a recipe.

The setpoint is another matter. The tension a crop should be irrigated at depends on how the medium holds and moves water, and that changes with the substrate, the container size and depth, the crop and its stage, where the sensor sits, and how the water is applied. A field crop in a loam might be irrigated at 30 to 40 cbar; a tomato on a rockwool slab or a seedling in a peat plug lives at a fraction of that, because a soilless substrate stops moving water to the root at a tension a soil would call barely damp. Container substrates are usually managed inside roughly 1 to 5 kPa (1 to 5 cbar) of available water; mineral soils are managed across tens of centibars. Same instrument, same unit, a different band. Move the sensor to a new medium and you keep the meter and re-learn the number.

What tension makes possible is irrigation by demand instead of by clock: a trigger that fires when the plant is actually working to get water, set for that crop in that medium, and checked against the plant.

How to read it: tensiometers and matrix sensors.

Two instruments measure tension directly. A tensiometer is a water-filled tube capped with a porous ceramic tip and a vacuum gauge; as the medium dries and pulls water out through the tip, the gauge reads the suction. It reads true matric potential and is accurate in the wet range growers care most about (roughly 0–80 cbar), but it needs refilling and maintenance and stops working if it dries out. A granular matrix sensor (the Watermark type) is a gypsum-and-matrix block whose electrical resistance changes with how wet it is; it is low-maintenance and reads a wider range, but it measures resistance, not pressure, so it must be driven with alternating current: feeding it steady DC polarizes and corrodes the electrodes and ruins the reading. A small node that supplies the AC excitation and reads the result is the DIY build; pro probes hand you the value over SDI-12 or an analog line.

Either way, the reading becomes oat-ods the same as any other input: a tensiometer with a pressure transducer arrives as an analog or 4–20 mA signal, a matrix block as a resistance you convert, an SDI-12 probe as a finished number. The gatherer reads all of them and lands the tension in your own data, tied to the bed, not the sensor.

Irrigate by what the plant feels.

Tension is most powerful as a pair of thresholds, not a single line: a wet limit you don't want to sit at for long (a medium held near saturation has its pores full of water and short of air, the failure mode the dissolved-oxygen input warns about) and a dry trigger that starts the next irrigation. Sitting inside that band is the wet-then-partly-dry cycle that keeps roots both fed and breathing. Where the two lines fall is a property of the medium and the crop, not a constant: in a mineral soil the wet limit is somewhere around 8 to 10 cbar and the trigger for many crops lands between 20 and 40; in a peat, coco, or rockwool container the whole band sits inside the first few centibars, and the wet limit is governed by how long the substrate stays saturated after an irrigation event rather than by a tension number. Set the band by crop, stage, and medium, and when the medium changes, re-set the band. Our soil-water tension reference lays out starting bands by crop and medium.

The trap: chasing a percentage that changes meaning by medium.

The common mistake is steering by a moisture percentage or, worse, a clock, and then wondering why the same "40%" that worked last cycle is too wet or too dry this time. It's because the medium, the pot size, or the root mass changed, and VWC quietly changed meaning with it. Tension reads the plant's actual experience in a unit that means the same thing everywhere. The second trap is the mirror image: carrying a tension setpoint from a soil bed to a soilless slab as if the number travelled with the unit. Measure the pull, set the band for the medium in front of you, check it against the plant, and let the percentage be whatever it needs to be.