Field NotesPlant techNo. 11
How to measure plant water stress: leaf temperature, stem diameter and sap flow
Soil sensors measure the root zone. Plant-based sensors measure the plant’s response: warmer leaves, a shrinking trunk, slower sap and a stronger pull on its water.

The short answer Measure the plant’s own response to drying: canopy temperature (often scaled into a Crop Water Stress Index), stomatal conductance, daily trunk shrinkage with a dendrometer, sap flow with heat-based sensors, and stem water potential with a pressure chamber. All of them also respond to weather, so read them alongside weather data and compare against a well-watered baseline or calibration for the same conditions.
A soil moisture sensor reports on the root zone at one spot. It can’t tell you whether those roots are working, or how hard the air is pulling on the leaves. Plant-based sensing reads the result instead. The plant is always reporting on its water supply; these instruments write the report down.
The catch, as Hamlyn Jones puts it in a review in the Journal of Experimental Botany: plant-based methods “do not usually give information on ‘how much’ irrigation to apply at any time, only whether or not irrigation is needed.”
Leaf and canopy temperature
Transpiration cools leaves. Jones notes that “leaf temperature tends to increase as plants are droughted and stomata close,” which infrared thermometers and thermal cameras detect.
José Enrique Fernández’s 2017 review in Horticulturae points out that canopy temperature also depends on radiation, air temperature, humidity and wind, so it needs normalizing. The Crop Water Stress Index (CWSI), credited to Idso and colleagues, does that:
CWSI = (Tc − Tnws) / (Tdry − Tnws)
- Tc is the measured canopy temperature.
- Tnws is the temperature of a well-watered, non-water-stressed canopy at the same vapor pressure deficit.
- Tdry is the temperature of a canopy that isn’t transpiring at all.
All three are expressed as differences from air temperature. A canopy as cool as the well-watered reference scores 0; one as warm as the non-transpiring limit scores 1.
The practical problems:
- VPD. The baseline moves with vapor pressure deficit, so log air temperature and humidity too. See what is VPD.
- Clouds and humidity. Jones says CWSI is “less useful in more humid or cloudy climates where the signal-to-noise ratio is somewhat smaller.”
- Sun, shade and soil. A thermal image mixes sunlit leaves, shaded leaves and soil; Jones points to automated image analysis for pulling out the leaf temperatures, and Fernández says the need to process the images rules out full automation.
- Emissivity. A sensor sees radiation the leaf emits plus radiation it reflects from its surroundings. Apogee Instruments lists leaf emissivity at 0.96 and says that without correction, a 25°C leaf on a sunny day could show an error of −1.5°C.
- Timing. For one image a day, Fernández cites advice to shoot at maximum stress, early afternoon in olive.
Stomatal conductance
Stomatal conductance measures how freely water vapor leaves through the stomata; a porometer reads it one leaf at a time. Fernández calls it “one of the most effective water stress indicators in terms of earliness and sensitivity,” and Jones calls the porometer the research benchmark. It rarely runs a schedule, because porometers are manual and leaf-to-leaf variation demands many readings.
Stem and trunk diameter
Trunks shrink by day and swell back. Fernández explains that the phloem and living sapwood lose water to transpiration, so the stem narrows, and the change is readily reversible. A dendrometer, such as a linear variable differential transformer (LVDT) dendrometer, logs it. Two numbers come out:
- Maximum daily shrinkage (MDS): early-morning maximum diameter minus late-afternoon minimum.
- Stem or trunk growth rate (SGR or TGR): the net change from one day to the next.
Fernández calls these the most widely used of the stem diameter indicators for scheduling irrigation. In a California almond trial by Goldhamer, Fereres and Salinas, MDS ranged from 0.1 to 0.9 mm. Since MDS also grows on hot, dry days, the researchers divided it by a baseline from fully irrigated trees at the same VPD; a signal of 1.0 meant no irrigation-related stress. Irrigation rose 10% after three consecutive days above target and fell 10% after three below.
Example: if baseline MDS for today’s VPD is 0.20 mm and your tree shrank 0.35 mm, the signal is 0.35 ÷ 0.20 = 1.75, the trial’s mild-stress target.
- Trees held at 2.75 got 20.7 inches of water through September 18 versus 33.8 inches at 1.75, a 38.8% cut, and their dry kernels were 11.5% lighter.
- MDS was more sensitive than stem water potential but noisier: mean coefficients of variation of 0.255 versus 0.127.
- Only four trees per treatment were instrumented, for one season.
In peach, lemon, grapevine and olive, Fernández notes, MDS rises as water potential falls only to a point, then declines as stress deepens, so less shrinkage isn’t always good news. Growth muddies MDS in young trees, and SGR fails when growth stops, as in grapevines after veraison. Jones notes that daily growth-rate measurements, though sensitive, are not particularly useful for controlling high-frequency irrigation systems.
Sap flow
Sap flow sensors use heat as a tracer for water moving up the stem. McElrone and Bleby’s protocol covers five heat-based techniques. Three of them:
- Heat pulse. A heater probe in the sapwood fires a short pulse, and sensors on either side track the heat. The common compensation method struggles below 2 to 10 cm per hour and can’t resolve zero or reverse flow; the heat ratio method was built for those. Drilling wounds the stem, so readings need wound corrections.
- Stem heat balance. A heater wraps the whole stem under insulation, and heat input is balanced against conduction losses and heat carried off by sap. It’s non-invasive and fits stems as small as 4 mm, but outside temperature gradients can cause errors.
- Thermal dissipation (Granier). Two needles about 2 mm across sit roughly 10 cm apart, the upper one heated at constant power; faster sap shrinks the temperature difference. The original calibration isn’t universal, and probes partly in non-conducting wood underestimate flow.
A sap flow curve follows transpiration, so it follows the weather. Jones notes that a midday dip can signal a developing deficit, but changing weather can produce the same shape, and flow lags transpiration because the stem stores water. Each tree needs calibration, and Fernández notes that large trunks need many sensors because flow varies around the trunk.
Water potential with a pressure chamber
Jones calls the pressure chamber a widely accepted reference technique, though slow, labor-intensive and unsuitable for automation. You seal a cut leaf in the chamber, stem poking out, and raise the pressure until water appears at the cut. As UC Agriculture and Natural Resources (UC ANR) puts it, “The pressure required to force water out of the stem of a severed leaf equals the water potential.”
- Readings are in bars or megapascals. Ten bars equal 1.0 MPa; one bar is 14.5 psi.
- Values are negative, and more negative means more stress: −14 bars (−1.4 MPa) is more stressed than −8 bars. On the gauge, a bigger number means more stress.
Stem water potential (SWP) uses a leaf bagged before cutting; Fernández notes it is widely recognized as potentially a better indicator than leaf water potential because changing weather sways it less. Protocols differ. UC ANR bags shaded lower-canopy leaves for about 10 minutes and measures from 12:00 to 4:00 p.m.; Fernández describes bagging 1 to 2 hours ahead and sampling at the daily low, which can come 2 to 4 hours after solar noon. Follow the protocol behind your thresholds.
Selected bands from UC ANR’s guidelines, in bars:
| Crop | Low stress, fully irrigated | Mild to moderate stress | High stress |
|---|---|---|---|
| Walnut | −2 to −4 | −6 to −8 | −10 to −12 |
| Almond | −6 to −10 | −10 to −14 (mild), −14 to −18 (moderate) | −20 to −30 |
| Prune | −6 to −8 (very low stress) | −12 to −16 | −20 to −30 (high to severe) |
These are crop-specific: almond SWP has been observed from −6 to −60 bars, far wider than walnut. Don’t borrow them for other plants.
UC also publishes weather-based baselines for well-watered trees: for almond and prune, about −5.2 bars at 75°F and 70% relative humidity, falling to −15.1 bars at 115°F and 10%. The gap between reading and baseline is the stress weather doesn’t explain.
The methods side by side
| Method | What it measures | Typical equipment | Response speed | Main limitation |
|---|---|---|---|---|
| Canopy temperature | Leaf warming as transpiration slows | Infrared thermometer, thermal camera | Continuous | VPD baseline; weak in humid, cloudy weather |
| Stomatal conductance | Vapor flow through stomata | Porometer | Early; spot readings | Manual; leaf-to-leaf variation |
| Stem diameter | Daily shrinkage and growth | LVDT dendrometer | One value per day | Growth confounds it |
| Sap flow | Water moving up the stem | Heat-based sensors | Continuous; lags transpiration | Calibration; many sensors on big trunks |
| Stem water potential | Tension on the plant’s water | Pressure chamber | Midday spot reading | Slow, manual, crop-specific |
Reading plant data in context
Soil and plant sensors answer different questions. Jones credits soil measurements with being easy to apply, with some sensors (especially capacitance and time domain types) readily automated, and with indicating “how much” water to apply. But soil heterogeneity demands many sensors, a representative spot is hard to choose, and they don’t read water status at the root surface. See what is volumetric water content for that side.
The plant, in Fernández’s phrase, is a biosensor that “integrates the soil and atmosphere water status.” So the two pair well: the plant shows when it’s short, and the soil tells you how much to put back.
Weather runs through all of it. Jones notes that the plant’s response to a given soil moisture varies with evaporative demand, and that plant sensing usually needs supplementing with information such as evaporative demand. In practice:
- Log air temperature and humidity at canopy height with every plant reading.
- Measure at the time of day your thresholds assume.
- Compare against a well-watered baseline, and act on runs of readings, as the almond trial did.
Sources
- Irrigation scheduling: advantages and pitfalls of plant-based methodsJournal of Experimental Botany (Jones, 2004)
- Plant-Based Methods for Irrigation Scheduling of Woody CropsHorticulturae (Fernández, 2017)
- Using the Pressure Chamber for Irrigation Management in Walnut, Almond, and PruneUniversity of California Agriculture and Natural Resources (ANR Publication 8503)
- Can almond trees directly dictate their irrigation needs?California Agriculture (Goldhamer, Fereres and Salinas, 2003)
- Sap flowPROMETHEUS protocols (McElrone and Bleby)
- Emissivity Correction for Infrared Radiometer SensorsApogee Instruments
General information, not professional advice for a specific plant or crop. Spot something wrong? Email info@terrathink.io.


