HERBARIUM

What the sensor thinks it knows

Cannabis grow-room sensors measuring temperature, humidity and VPD across different canopy microclimates

Placement, calibration, microclimates and the danger of one convincing number

A sensor does not measure
the room.

 

It measures one physical condition
at one sensing element
in one location
across one period of response.

The screen may say:

25.4°C.
61% relative humidity.
1.23 kPa VPD.

The decimals look complete.
The measurement is not.

  • Above the canopy may be warmer than below it.
  • Air beside a humidifier may be wetter than air beside an intake.
  • A sensor exposed to radiation may become warmer than the air it is supposed to measure.
  • Leaves may be warmer or cooler than the surrounding air.
  • A wall, lamp, wet substrate, fan or cold surface can create a local condition that the rest of the room does not share.

The sensor may report that condition accurately.
The grower may still interpret it incorrectly.

That is the first lesson:

A number can be true and still fail to represent
the place the grower thinks it represents.

The sensor has an address

Every environmental reading should be able to answer:

  • Where was the sensor?
  • How high was it?
  • Was it inside, above or below the canopy?
  • Was it exposed to direct radiation?
  • Was it beside a fan, intake, humidifier, dehumidifier or wet surface?
  • Was air moving freely around it?
  • Had it reached equilibrium?
  • What period did the reading represent?

Without those answers, “the room was 26°C and 60% RH”
is incomplete.

The room may contain
many temperatures and humidities at once.

A dense canopy modifies radiation, airflow and water-vapour distribution. Supply air, extraction, plant transpiration and equipment create further spatial gradients.

One sensor can be useful.
It cannot prove uniformity.

Better lesson:

A sensor reading has an address.
Remove the address and part of the meaning disappears.

When two identical devices disagree

“I bought two identical meters.
They do not show the same number.”

 

Good. Now the measurement has become interesting.

Two devices may disagree because:

  • they are in different microclimates;
  • one has not stabilised;
  • their response times differ;
  • one is exposed to radiation or condensation;
  • contamination or ageing has changed one sensor;
  • or both are operating within their stated tolerances.

Before deciding that one device is defective,
place them together under the same conditions.

Give them:

  • the same height;
  • the same airflow;
  • the same radiation protection;
  • the same stabilisation period;
  • and the same logging interval.

Then compare them across more than one condition.

One device may show a nearly constant offset.
Another may agree around 50% RH but diverge at higher humidity.

A one-point comparison
cannot reveal the complete response curve.

If the difference matters operationally, compare the sensors with a suitable reference or have them calibrated across the range in which they will be used.

Two matching displays do not prove accuracy.

Two disagreeing displays do not prove
that either device is useless.

Resolution is not accuracy

A display reading 24.37°C may look more trustworthy than one reading 24°C.

The extra digits prove only that the device displays smaller increments.
They do not prove that the underlying measurement is correct to the same number of decimal places.

Four terms should remain separate:

  1. Resolution
    is the smallest displayed or detectable increment.
  2. Repeatability
    describes how closely repeated measurements agree under the same conditions.
  3. Accuracy
    describes closeness to an accepted reference value.
  4. Calibration
    compares instrument response with known standards and documents the deviation.

 

Calibration is also not automatically adjustment.

A device may be calibrated, its error documented and no internal correction applied. A grower should therefore read the specification before admiring the decimal places.

The most expensive-looking number
may still carry an uncertainty
larger than the change being discussed.

Better lesson:

Precision is not accuracy.
A decimal can be uncertainty wearing jewellery.

Relative humidity borrows its meaning from temperature

Relative humidity is not a direct statement of how many grams of water vapour occupy the room.

It expresses the actual water-vapour pressure relative to the saturation vapour pressure at the same temperature.

Relative humidity can therefore change when temperature changes even if the actual amount of water vapour in the air changes very little.

Warm the air:
the same vapour content may produce a lower RH.

Cool the air:
the same vapour content may produce a higher RH.

This is why temperature error becomes humidity-interpretation error.

A combined sensor warmed by direct radiation may report:

  • air temperature too high;
  • relative humidity too low;
  • and calculated VPD too high.

The three numbers may appear to confirm one another because they were derived from the same local measurement error.

Dew point answers a related but different question:
at what temperature would the current air reach saturation?

That makes dew point useful when evaluating condensation risk on cold leaves, walls, ducts or equipment.

The moisture did not suddenly appear
when RH rose overnight.

The air may simply have cooled
towards its dew point.

The leaf is not the air

Growers often calculate VPD from:

  • air temperature
  • and relative humidity.

This produces air VPD: the difference between saturation vapour pressure at the measured air temperature and the actual vapour pressure of the air.

Water-vapour exchange from a leaf is driven more directly by the gradient between the approximately saturated internal leaf air spaces and the surrounding air.

Because saturation vapour pressure depends on temperature, a leaf warmer or cooler than the measured air can produce a different leaf-to-air vapour-pressure gradient.

  • Under strong radiation, a leaf may be warmer than the measured air.
  • Under active transpiration, it may be cooler.

Air movement, stomatal behaviour, radiation, plant water status and leaf position can all influence that difference.

This does not make an air-temperature VPD calculation useless.
It makes its definition important.

A chart calculated from air temperature and RH should not be silently presented as though leaf temperature had been measured.

A leaf-to-air VPD requires:

  • air vapour pressure
  • and a defensible leaf-temperature measurement.

A defensible infrared leaf-temperature measurement also requires attention to:

  • measurement angle;
  • field of view;
  • distance;
  • emissivity assumptions;
  • and whether the reading represents one leaf or the canopy.

VPD is not a commandment attached to one universal stage chart.

It is an atmospheric variable whose biological consequence depends on plant response, light, airflow, water supply, genotype and development.

Better lesson:

Air VPD describes the atmospheric deficit calculated at air temperature.

Leaf-to-air VPD estimates the vapour-pressure gradient
involving the temperature of the leaf.

Do not use one name
while measuring the other.

Placement is part of the measurement

A calibrated sensor in the wrong location
can control the wrong microclimate perfectly.

There is no single magic placement for every purpose.

Placement should follow the question.

  • To monitor the air surrounding active growth, place a temperature–humidity sensor near the relevant canopy zone while avoiding direct contact with leaves and local equipment streams.
  • Protect it from direct lamp or solar radiation.

Do not place it:

  • directly in a fan jet;
  • beside a humidifier outlet;
  • beside a dehumidifier intake or hot discharge;
  • against a warm or cold wall;
  • immediately above wet substrate;
  • or in a dead corner unless that corner is intentionally being investigated.

As the canopy changes height and density, the representative location may also change.

A second sensor can serve a different purpose:

  • finding a humid interior canopy zone;
  • checking vertical stratification;
  • monitoring the intake air;
  • or confirming that the control sensor is not an outlier.

The control sensor and the diagnostic sensor do not need to occupy the same place.

One helps operate the room.
The other helps reveal what the room is hiding.

The sensor also has a clock

A sensor does not respond instantly.
It has a time constant.

  • Airflow, sensor housing, protective filters and the size of the environmental change influence how quickly the displayed value approaches the new condition.
  • A humidity sensor may therefore miss, delay or smooth a short moisture spike.
  • A heavily filtered logger may show a calm environment while the crop experienced repeated brief excursions.

Condensation can create another problem.
A sensor that becomes wet may remain near saturation, recover slowly or show temporary error after the surrounding air has changed.

Logging interval matters too.
A reading stored every fifteen minutes cannot describe every one-minute event.

An hourly average can hide:

  • short heat peaks;
  • irrigation-related humidity spikes;
  • equipment cycling;
  • and rapid lights-on or lights-off transitions.

Min–max values help,
but they do not show sequence or duration.

The grower should know:

  • how often the sensor samples;
  • how often it records;
  • whether values are averaged;
  • whether the display is filtered;
  • and how quickly the sensor responds.

A number without time
is another incomplete measurement.

Barometric pressure is not room pressure

Barometric pressure describes the pressure exerted by the atmosphere at a location.

It changes with elevation and weather systems.
That can provide useful outdoor context.

It is not the same measurement as the pressure difference between:

  • a grow room and a corridor;
  • a tent and the surrounding room;
  • or the two sides of a filter, fan or duct.

Airflow through buildings and ventilation systems depends on pressure differences created by fans, wind, stack effect and flow resistance.

To determine whether a room is positive or negative relative to another space, the grower needs a differential-pressure instrument with appropriate reference points.

A single-point barometric reading cannot provide that answer.

Absolute atmospheric pressure can affect air density and precise psychrometric calculations. Ordinary weather-scale changes are usually secondary, however, to temperature, humidity, leaf temperature, airflow and water supply in a small cultivation environment.

Use barometric pressure to understand:

  • weather context;
  • elevation-related pressure;
  • or a calculation that explicitly requires absolute pressure.

Do not use it as a substitute for:

  • airflow measurement;
  • fan-performance assessment;
  • or room-to-room differential pressure.

Better lesson:

A falling barometer may describe
the weather approaching the building.

It does not tell you
whether the grow room is under negative pressure.

Automation believes the sensor before it believes the plant

A monitoring sensor can mislead the grower.
A control sensor can move equipment.

 

That difference matters.

  • Place the humidity sensor inside the humidifier plume and the controller may stop humidification before the canopy receives enough moisture.
  • Place it in a local return-air stream that does not represent the crop zone and the machine may control that stream rather than the canopy environment.
  • Expose the temperature sensor to radiation and the cooling system may fight a temperature that belongs partly to the sensor housing.

Automation does not correct a bad measurement.

It repeats the error
with electrical authority.

A stable control system needs more than a setpoint:

  • representative sensor placement;
  • documented calibration or verification;
  • suitable deadbands;
  • minimum equipment run or rest times where required;
  • alarms for missing, frozen, implausible or out-of-range values;
  • and an independent way to confirm the control sensor.

The controller should also fail visibly.

A missing sensor, frozen value or impossible reading should create an alarm—not quiet confidence.

Data should be reviewed as trends:

  • What changed first?
  • How long did the excursion last?
  • Which equipment responded?
  • Did another sensor confirm it?
  • Did the crop provide independent evidence?

The purpose of automation is not to remove judgement.

It is to apply verified decisions
more consistently.

What a serious sensor practice looks like

A serious grower does not need
the most expensive dashboard.

 

They need measurements whose limitations are understood.

Record:

  • sensor model and stated accuracy;
  • calibration or verification date;
  • exact location and height;
  • whether the sensor is shielded or aspirated;
  • sampling and logging interval;
  • offsets applied by software;
  • and any relocation as the canopy changes.

 

  • Compare co-located sensors
    before comparing distant ones.
  • Use more than one location
    when spatial variation matters.
  • Inspect trends,
    not only current values.
  • Investigate sudden agreement
    as carefully as sudden disagreement.
  • Replace or recalibrate sensors
    that drift, contaminate or recover poorly.
  • Keep control sensors away
    from the equipment streams they command.

And remember:
the plant is not living in the dashboard.

It is living in a three-dimensional environment
the dashboard is attempting to sample.

The rule

The sensor is not lying
merely because the plant disagrees.

 

The plant may be responding
to a condition the sensor never sampled.

Every reading contains:

  • an instrument;
  • a location;
  • a response time;
  • an uncertainty;
  • and an interpretation.

Remove any one of them
and the number becomes easier to misuse.

  • The sensor knows
    what touched its sensing element.
  • The grower must decide
    what that evidence represents.

A display is not the room.
A reading is not the crop.

And automation is not understanding
simply because it reacts quickly.

Factual Note

Temperature and relative-humidity sensors measure the environment at their sensing elements, not an abstract room average. Location, height, airflow, radiation exposure, nearby wet or cold surfaces and the structure of the crop can all influence the recorded value. Greenhouse research has shown that the best sensor location for estimating an overall average may differ from the locations needed to detect poorly controlled zones.

Resolution, repeatability, accuracy and calibration are different properties. A display with additional decimal places has greater displayed resolution but not necessarily greater accuracy. Calibration compares an instrument with known reference conditions and documents its response and uncertainty; it does not necessarily mean the device has been internally adjusted.

Relative humidity depends on temperature. A temperature error can therefore affect both the reported RH and any VPD calculated from that temperature–humidity pair. Sensors exposed to radiation or local equipment streams may produce internally consistent temperature, RH and VPD values that are not representative of the crop environment. Controlled-environment measurements should use suitable placement and radiation protection.

Air VPD is the difference between saturation vapour pressure at the measured air temperature and the actual vapour pressure of the air. A leaf-to-air vapour-pressure gradient instead uses saturation vapour pressure at leaf temperature and may therefore differ when the leaf is warmer or cooler than the surrounding air. These quantities should be named and interpreted separately. Plant responses to VPD also depend on species, genotype, light, water status and other environmental conditions.

Sensor response time, sampling interval and data-logging interval affect the events that become visible. Short temperature or humidity excursions may be delayed, averaged or missed. Condensation, contamination and ageing may also alter humidity-sensor behaviour, so comparison should include stabilisation time and more than one environmental condition.

Barometric pressure is atmospheric pressure at a location. It is not equivalent to the differential pressure between a cultivation room and an adjacent space. Directional airflow and building pressurisation are assessed through pressure differences associated with fans, wind, stack effect and system resistance.

Much of the detailed evidence on greenhouse sensor placement, VPD and environmental control comes from crops other than cannabis. The measurement principles are broadly relevant, but fixed cannabis thresholds and universal placement rules require crop- and facility-specific validation.

HERBARIUM

A living archive of cannabis—its cultivation in practice and its traces across science, culture, language, power and memory.

LIBERA HERBA Cannabis VADEMECUM — Early Access

Join early.

Keep the archive open.

The VADEMECUM is becoming a living archive of guides, tools, notes and practical plant knowledge.

Free member access. Join early. Keep the archive open.

LIBERA HERBA Cannabis VADEMECUM — Early Access

Join early.

Keep the
archive open.

The VADEMECUM is not just a book anymore. It is becoming a living archive of guides, tools, notes and practical plant knowledge.

Free member access. Join early. Keep the archive open.

The VADEMECUM is becoming a living archive of practical plant knowledge.

Free member access.