\[LEDE — the first paragraph IS the answer an AI engine will extract\]
01Why this lesson matters.
Every monitoring decision starts with what to measure. A grower who buys sensors without thinking carefully about that question ends up with a dashboard full of numbers that are technically accurate and practically useless. Sensors that measure the wrong thing at the wrong place at the wrong time do not produce insight — they produce noise.
The gap between a sensor reporting a number and a grower making a better decision is bigger than it seems. A temperature sensor reports the temperature where the sensor happens to be. That may or may not be the temperature the plants are experiencing. A soil moisture sensor reports the moisture in a specific volume of substrate at a specific depth. That may or may not be what the roots care about. A light sensor reports the light that hits its measurement surface at a specific moment. That may or may not represent what the crop received over the course of the day. Turning a reading into an actionable number requires understanding what the sensor is actually sensing, and whether that matches the decision the grower is trying to make.
This lesson organizes sensors by what they measure, not by what the sensor is called. Growers think in terms of questions — what is the temperature, how wet is the soil, is there enough light, is the CO2 where it should be. Sensor categories follow that structure. Inside each category, the lesson covers the common sensor technologies, the quality tiers, the pitfalls, and the match between sensor type and growing context.
A grower who finishes this lesson will know what to ask for when they want to measure something specific, what quality level fits their situation, and what traps to avoid on each sensor type. That is the practical goal. The philosophical goal is that the grower learns to think about sensors the way an experienced monitoring designer does: not as products to buy, but as answers to specific questions the grower needs to ask.
02Sensor basics.
Before the category-by-category tour, a few concepts apply to every sensor.
Accuracy.
How close the sensor's reading is to the true value. A temperature sensor with plus-or-minus half a degree Celsius accuracy reads within half a degree of the actual temperature. Accuracy is always stated as a range around the true value, often with conditions (at 25°C, at sea level, etc.). Accuracy usually degrades at the extremes of the sensor's operating range and in conditions the sensor was not designed for.
Precision.
How consistent the sensor's readings are, separate from whether they are correct. A sensor can be precise (the same reading every time under the same conditions) without being accurate (the readings are all consistently off by two degrees). A sensor can be accurate on average but imprecise (the readings bounce around the true value). Good sensors are both. Precision matters when you care about change over time; accuracy matters when you care about the absolute number.
Resolution.
The smallest change the sensor can detect. A temperature sensor with 0.1°C resolution reports in tenths of a degree; one with 1°C resolution only reports whole degrees. Resolution does not imply accuracy — a sensor can report to 0.01°C while being accurate to plus-or-minus 2°C. The three concepts (accuracy, precision, resolution) are often confused. A complete sensor specification includes all three.
Range.
The span of values the sensor can measure. A temperature sensor rated from -40°C to \+85°C will not give accurate readings outside that range, even if it produces numbers. Matching range to the application matters — a temperature sensor for a freezer needs to work below 0°C; a sensor for a compost pile needs to work above 60°C.
Response time.
How long the sensor takes to reflect a change in the environment. A fast-response temperature sensor reads the new temperature within seconds of exposure; a slow-response sensor may take minutes. For environmental monitoring at typical reporting intervals, response time rarely matters. For automation triggering off sudden changes, response time can be critical.
Calibration.
The process of comparing a sensor's readings against a known reference and adjusting as needed. Some sensors hold calibration for years; others drift and need periodic checking. Understanding which category a sensor falls into is essential for long-term accuracy. A sensor that drifts one degree per year and is never recalibrated produces usable data for a year and misleading data for a decade.
Drift.
Gradual change in a sensor's readings over time, separate from actual environmental changes. A CO2 sensor that was accurate when installed may read 200 parts per million too high after two years of continuous operation. Drift is the silent killer of long-term agricultural monitoring — the data looks fine, the sensor looks fine, but the readings no longer mean what they used to mean.
Environmental tolerance.
A sensor's ability to work correctly in the conditions it will actually face. A sensor designed for indoor use may fail in high humidity. A sensor designed for a clean laboratory may corrode in ammonia-rich greenhouse air. A sensor designed for room temperature may give unreliable readings at freezing. Matching the sensor's environmental tolerance to the actual deployment environment is often the difference between a sensor that lasts five years and one that fails in five months.
03Temperature.
The most basic measurement in agriculture and the starting point for almost every monitoring system. Temperature affects plant growth, metabolic rate, transpiration, flowering, dormancy, disease pressure, pest activity, harvest timing, and storage life. A single degree of difference can mean a good harvest or a lost one.
Air temperature.
The most common measurement. Small sensors based on semiconductor technology (often combined with humidity sensors in the same package) cost a few dollars and give accurate readings across the ranges that matter for agriculture. The common technologies: thermistors (resistance changes with temperature, cheap, good accuracy over moderate ranges), resistance temperature detectors (RTDs, more expensive, more accurate, better for extreme temperatures), and digital sensors with onboard processing (what most modern low-cost sensors use).
For agricultural air temperature, a good digital sensor with plus-or-minus 0.5°C accuracy is plenty for almost any decision a grower makes. Paying for plus-or-minus 0.1°C accuracy rarely changes outcomes. Most common sensors include both temperature and humidity measurement in a single package, which matters because humidity and temperature together determine vapor pressure deficit (VPD), a more actionable measurement than either alone.
Placement matters more than precision. A sensor in direct sunlight will read five to ten degrees higher than the actual air temperature because it is measuring the sensor's warming, not the air. A sensor above a heat source reads that source, not the room. A sensor near a cold wall reads cooler than the plant canopy. Place sensors where the plants actually are, shield them from direct radiation, and accept that a few sensors placed thoughtfully are more useful than many sensors placed carelessly.
Soil temperature.
Often more important than air temperature for certain decisions. Soil temperature determines germination timing, root activity, nutrient uptake, and disease pressure in the root zone. Soil temperature lags air temperature by hours to days depending on soil type, moisture, and depth. A cold soil in warm air can check root growth even while the grower thinks everything looks fine from above.
Soil temperature sensors are usually waterproof probes that go underground at specified depths. Common measurement depths: 2 inches (5 cm) for surface effects, 4 to 6 inches (10-15 cm) for root zone, and 12 inches (30 cm) or deeper for seasonal baseline. Multi-depth probes (like the four-thermocouple probe common in CEA and row crops) give a temperature profile that reveals how heat moves through soil over a day.
Surface temperature.
Infrared sensors measure the surface temperature of an object without touching it. This matters for plant canopy temperature (an indicator of stress and transpiration), greenhouse glazing temperature (for heat loss analysis), and cold chain monitoring where probing food is impractical. Infrared sensors are tricky because they measure the surface they are pointed at, and that surface's emissivity (how efficiently it radiates heat) affects the reading. Plant canopies are close enough to an emissivity of 1.0 for practical use, but metal surfaces and glass require calibration.
What to watch for.
Sensor placement in direct sun causing overreading. Condensation on sensors causing erratic behavior. Sensors mounted too close to heat sources (lights, heaters, equipment) reading that source rather than the environment. Drift over time with cheap sensors (though modern digital sensors drift very little if at all). Incorrect calibration when a sensor reads consistently one or two degrees off — check against a reference thermometer before assuming the sensor is broken.
05Light.
Plants respond to specific wavelengths of light (photosynthetically active radiation, or PAR), at specific intensities, for specific durations. A light sensor's value depends on whether it measures what the plants actually use. Most consumer light sensors do not — they measure what the human eye sees, which is not the same thing.
Units of light measurement.
The units are confusing because they come from different fields. Lux and foot-candles measure light as the human eye perceives it — not what plants use. PPFD (photosynthetic photon flux density, measured in micromoles per square meter per second, or μmol/m²/s) measures the photosynthetically useful photons hitting a surface. DLI (daily light integral, measured in moles per square meter per day, or mol/m²/d) measures the total photosynthetic light received over a full day. For horticulture, PPFD and DLI are what matter. Lux is a hobbyist's unit.
PAR sensors.
Sensors calibrated to measure photosynthetically active radiation — the 400 to 700 nanometer range that drives photosynthesis. Quality PAR sensors (often called quantum sensors) cost a moderate amount for a decent one, more for laboratory-grade. A PAR sensor is necessary for any grower making serious lighting decisions — supplemental lighting calculations, DLI tracking, comparing one light fixture to another. Consumer lux meters cannot do this job.
DLI tracking.
Daily light integral is the PPFD integrated over the whole day. Many crops have specific DLI requirements — lettuce does well at 12-17 mol/m²/d, tomatoes prefer 22-30, cannabis often targets 35-55. Tracking DLI over days and weeks reveals whether the crop is getting the light it needs across changing weather. DLI is one of the most actionable measurements in controlled environment agriculture.
Spectrum sensors.
More specialized sensors that report the distribution of light across different wavelengths. Useful for evaluating grow lights, comparing sunlight to supplemental lighting, and research applications. Most growers do not need spectrum sensors — PAR and DLI are sufficient for operational decisions.
What to watch for.
Using lux meters for horticultural decisions (the results will mislead). Placement of light sensors matters — a sensor in shade reports shade, not what the canopy receives. PAR sensors need to be kept clean and oriented correctly; a horizontal sensor reads differently than a vertical one. Outdoor PAR sensors need UV-stable domes; the domes yellow over time and reduce readings, causing apparent (but false) declines in light over the years.
06Soil and substrate sensors.
What is happening in the root zone is often invisible without sensors. A grower can see the plant and feel the soil at the surface, but what the roots are actually experiencing — moisture, temperature, salts, oxygen — is underground. Soil and substrate sensors are how a monitoring system sees underground.
Volumetric water content (soil moisture).
How much water is in a given volume of soil or substrate, usually expressed as a percentage (15 percent VWC means 15 milliliters of water per 100 cubic centimeters of soil). Most modern soil moisture sensors measure this capacitively — they use the dielectric properties of water to estimate how much of it is in the soil around the probe. Capacitive sensors are more reliable than older resistive sensors, which corroded and drifted.
The common mistake is trusting cheap soil moisture sensors. A five-dollar capacitive soil moisture probe gives readings that change in the right direction but may be off by factor of two from true VWC, particularly in substrates the sensor was not designed for. A twenty-dollar probe is usually better. A forty-dollar probe with proper calibration is reliable. Professional-grade probes (Meter Group's Teros series, Campbell Scientific's CS650) cost a moderate amount and give research-quality readings but the investment is justified when the decisions depend on getting the number right.
A practical compromise: use cheap sensors to see trends and changes, use at least one quality sensor to anchor the absolute numbers. A grower watching twenty cheap sensors can still manage irrigation by observing how they change over time, even if the absolute percentages are not perfectly calibrated.
Electrical conductivity (EC).
How well the soil or substrate conducts electricity, which correlates with dissolved salts — fertilizer, accumulated minerals, sometimes salinity. EC is a proxy for nutrient levels in hydroponic and substrate growing, where controlling fertigation is central. Too low and plants are underfed; too high and they are stressed by salt. In soil, EC drifts over seasons with fertilizer applications and rainfall leaching.
EC sensors are usually built into multi-function soil probes that also measure moisture and sometimes temperature. Measuring EC accurately is harder than measuring moisture, because the result depends on soil moisture (wet soil conducts better than dry soil even at the same salt concentration). Quality probes account for this; cheap probes may not.
pH.
How acidic or basic the soil or substrate is, on a scale of 0 to 14 where 7 is neutral. pH affects nutrient availability — most nutrients are most available between 6.0 and 7.0 in soil, slightly lower in soilless substrates. Soil pH changes slowly; substrate pH can change fast. pH is almost always measured with a glass electrode probe that requires regular calibration and has a limited lifespan (one to three years of continuous use).
Continuous pH monitoring with a probe in the substrate is possible but problematic — the probes foul, drift, and eventually die. Most growers measure pH manually (handheld meter, dipping a probe in a drain sample) rather than continuously. For hydroponic nutrient solutions, in-line pH probes are more practical and last longer because the solution keeps them clean.
Soil temperature (covered under Temperature above).
Most multi-function soil probes measure temperature as well as moisture. A single probe can deliver volumetric water content, EC, and temperature from the same depth, which is often the right compromise between cost and data quality.
Water potential.
A different way to measure soil water — how hard the plant has to work to pull water out of the soil, measured in kilopascals or bars. Water potential sensors (often called tensiometers or matric potential sensors) measure what plants experience directly, not how much water is present. For irrigation scheduling, water potential is often more actionable than volumetric content because plants respond to potential, not percentage. These sensors cost a substantial amount depending on type and are more common in research and precision agriculture.
What to watch for.
Cheap probes giving plausible-looking numbers that are systematically wrong. Probes installed in air pockets rather than in contact with roots. Probes that have dried out and no longer make contact with the substrate. Cable damage where the probe enters the soil — this is the most common failure mode for soil sensors and is often invisible. Sensor drift in salty environments. Incorrect substrate calibration (a soil sensor calibrated for mineral soil will read differently in peat-based potting mix).
07Water sensors.
Water is central to every growing operation. Monitoring it — flow rates, pressure, tank levels, quality — prevents the failures that destroy crops. A well pump that fails unnoticed, a tank that runs dry, an irrigation zone that stopped working — these are the events that cost harvests. Water sensors catch them before they become disasters.
Flow meters.
Measure how much water is moving through a pipe. Pulse-output flow meters send a signal for each unit of water (one pulse per liter, or per gallon, depending on the meter). Totalizing flow meters show cumulative volume. The common technologies: paddle-wheel (cheap, moderate accuracy), turbine (more accurate, slight pressure drop), electromagnetic (accurate, no moving parts, requires conductive water), and ultrasonic (strap-on, no plumbing modifications, moderate accuracy). For most agricultural monitoring, paddle-wheel or turbine meters with pulse output integrated into the monitoring system work well.
Pressure sensors.
Measure the pressure of water in a pipe, usually in pounds per square inch (psi) or bar. Low pressure indicates a leak, a closed valve, a failed pump, or a filter that needs cleaning. High pressure indicates a blockage downstream. Pressure monitoring at the point of use catches irrigation problems immediately — a zone that is not pressurizing is a zone not watering. Pressure transducers that output a standard signal (4-20 mA or 0-10 V) cost a moderate amount and integrate easily with monitoring systems.
Tank level sensors.
Measure how full a water tank or reservoir is. Common technologies: float switches (simple, discrete on/off at a set level), ultrasonic (reads distance to water surface, works without contact), pressure (reads pressure at the bottom of the tank to calculate height), and capacitive (a strip inside the tank). For alerting that a tank is low, a float switch at the critical level is cheap and reliable. For continuous level monitoring, ultrasonic or pressure sensors provide the full range.
Water quality.
pH and EC in water are measured the same way as in substrate, with in-line probes that sit in the flow. Dissolved oxygen, oxidation-reduction potential (ORP), and specific ion probes exist for specialized applications. For most agricultural operations, in-line pH and EC in the fertigation line are the most useful water quality measurements.
Rain.
Rainfall matters for outdoor operations and for understanding supplemental irrigation needs. Tipping-bucket rain gauges are the standard — a bucket tips for each 0.2 mm (or similar increment) of rain, sending a pulse the monitoring system counts. Consumer weather stations include rain gauges that integrate well with modern monitoring systems. Accuracy depends on sitting, calibration, and freedom from debris. Placement matters — a rain gauge under a roof overhang misses rain; one surrounded by obstructions catches wind-driven rain inaccurately.
What to watch for.
Flow meters that clog, especially on well water with debris. Pressure transducers damaged by water hammer (sudden pressure spikes from fast valve closures). Level sensors fouled by biofilm or mineral deposits. pH probes that have died but still produce numbers. Freeze damage to any outdoor water sensor. The usual advice: a single failure in a water sensor is fixable; an undetected failure is catastrophic. Build in redundancy for the measurements whose failure costs the most.
08Air and atmosphere sensors.
Beyond temperature and humidity, the composition of the air matters for many growing operations. CO2 for photosynthesis, oxygen for roots, ammonia from ventilation problems, ethylene from ripening fruit, volatile compounds from stress or disease — the invisible side of air affects plants as much as temperature does.
Carbon dioxide (CO2).
Plants need CO2 for photosynthesis. In a closed greenhouse or indoor grow, CO2 can drop below atmospheric levels (400 ppm) during active photosynthesis, limiting growth. In a controlled environment with CO2 enrichment, typical targets are 800 to 1,500 ppm. CO2 sensors use non-dispersive infrared (NDIR) technology — infrared light passes through a sample of air, and CO2 absorbs specific wavelengths. The amount absorbed indicates concentration.
Quality CO2 sensors cost a moderate amount for a reliable one; laboratory-grade versions cost more. The pitfalls: CO2 sensors drift over time and need recalibration, many older low-cost CO2 sensors are single-beam and drift significantly, and the cheapest CO2 sensors are often inaccurate by hundreds of ppm. Newer technology CO2 Sensors are starting to show up with better accuracy and lower cost. For CO2 enrichment control, invest in a sensor with auto-calibration or plan regular manual recalibration.
Oxygen.
Atmospheric oxygen is stable at about 20.9 percent and rarely needs monitoring in plant growing environments. Exceptions: root-zone oxygen in hydroponic systems where low dissolved oxygen causes root death, enclosed storage environments where reduced oxygen is part of preservation, and any enclosed space where humans work (safety monitoring). Dissolved oxygen in water is measured with specific probes; atmospheric oxygen is measured with electrochemical or paramagnetic sensors.
Ammonia.
A byproduct of some fertilizer programs, livestock operations, and decomposition. High ammonia damages plants and is toxic to workers. Electrochemical ammonia sensors cost a moderate amount and measure in parts per million; they are common in livestock facilities and sometimes in intensive greenhouse operations.
Ethylene.
A plant hormone that accelerates ripening. In storage rooms and packing facilities, ethylene from one batch of ripening fruit accelerates all surrounding fruit. Ethylene sensors are specialized and expensive; handheld gas detectors are more common for spot-checking.
Volatile organic compounds (VOCs).
A broad category of gases that plants release under stress or disease. VOC sensors detect the total concentration, not specific compounds, but changes can indicate plant stress before other symptoms appear. Low-cost VOC sensors are available but interpretation requires experience; they are more useful for research than for routine operations.
What to watch for.
CO2 sensor drift — a sensor that read correctly two years ago may read 200-400 ppm high today. Cross-sensitivity — many gas sensors respond to multiple gases, so a reading may reflect something other than the target gas. Temperature and humidity effects — most gas sensors are affected by environmental conditions and need temperature/humidity compensation. The usual rule: for any gas sensor, plan for periodic calibration or replacement.
09Presence, motion, and position sensors.
Not every sensor measures a physical property of the environment. Some measure whether something is happening — a door opened, a pump running, water flowing, motion detected, equipment activated. These event-based sensors are the operational monitoring side, as important as the environmental side for running a real operation.
Door and gate sensors.
Magnetic reed switches or Hall effect sensors that detect whether a door or gate is open or closed. A sensor on the greenhouse door tells the grower when it was opened and for how long — useful for security, for correlating climate disruptions with entries, and for intrusion alerts. These sensors cost little and run for years on small batteries. Consumer smart-home door sensors work fine in most agricultural contexts.
Motion sensors.
Passive infrared (PIR) sensors detect the heat movement of a warm body. Useful for security, for detecting people in storage rooms, and for automation (turning on lights when someone enters). Ultrasonic and microwave motion sensors use different principles and have different strengths. For agricultural use, PIR is the most common and cheapest.
Contact sensors.
Similar to door sensors but applied to other applications — detecting whether a lid is closed, whether a cover is in place, whether a latch is engaged. Any binary state that matters can be monitored with a simple switch and a contact input.
Pump and equipment status.
Current transformers around a power wire report when equipment is drawing current. A current transformer on the pump circuit tells the monitoring system whether the pump is running, regardless of whether the control signal said it should be. This catches the disaster case where the controller thinks the pump is running but it is not. Current transformers cost a modest amount and install without breaking the circuit.
What to watch for.
Motion sensors triggering on animals, which may or may not be what the grower wants. Door sensors that fail to close completely and report as open. Current transformers that need correct calibration for the specific circuit. False alarms from sensors placed where environmental factors trigger them (sunlight warming a PIR sensor, for example).
10Cameras and imaging.
Cameras are sensors too, and visual monitoring has become affordable enough that most operations benefit from at least a few. A camera in the greenhouse provides context that numerical sensors cannot — did the plants wilt before the alert fired, is the condensation really that bad, did someone leave a door open. Cameras also enable more specialized applications: plant counting, stress detection, pest spotting, growth tracking.
IP cameras.
Network-connected cameras that stream video over Ethernet or WiFi. Consumer IP cameras cost a moderate amount and work well for greenhouse monitoring. Features to look for: day/night capability, weatherproof rating (IP65 or better for outdoor use), POE support, ONVIF compliance (so the camera works with standard software), and local recording to SD card or network storage.
Wildlife cameras.
Battery-powered cameras with motion triggers, originally designed for hunting and wildlife monitoring. Useful in remote locations where IP cameras cannot be powered. They record to SD cards and do not stream, so real-time monitoring is limited, but they catch what happens overnight.
Thermal cameras.
Infrared imaging that shows temperature patterns rather than visible light. Useful for finding heat loss in greenhouses, detecting hot spots in electrical equipment, and seeing plants in darkness. Consumer thermal cameras for smartphones cost a moderate amount; dedicated thermal cameras cost more. Specialized agricultural applications (canopy temperature mapping for stress detection) are research-grade rather than operational for most growers.
Specialized imaging.
Multispectral and hyperspectral cameras see wavelengths beyond human vision and can reveal plant stress, disease, and nutrient deficiency before they show visually. Drone-based multispectral imaging is operational in larger field operations. For small growers, the technology is still emerging and cost-prohibitive for most applications.
What to watch for.
Network bandwidth consumed by video streams. Privacy considerations in areas where people work. Storage requirements for continuous recording — a single camera at HD resolution can consume a terabyte per month. Outdoor camera housings that fog up in high-humidity environments. Cheap cameras with security vulnerabilities that put the whole network at risk.
11Quality tiers and what to spend.
Sensors of every type exist at multiple price points. A framework for thinking about what to spend:
Consumer grade.
Designed for home use. Adequate accuracy for many agricultural applications. Often lacks the environmental tolerance of purpose-built agricultural sensors. Best for protected environments (indoor grows, greenhouses) with moderate precision requirements. Examples: BLE temperature and humidity sensors, capacitive soil moisture probes, consumer weather stations.
Prosumer and light industrial.
More rugged, better calibration, longer-lasting. Works in outdoor and harsher environments. Better accuracy than consumer grade. The sweet spot for serious small-and-mid-scale agricultural monitoring. Examples: industrial temperature probes, mid-tier soil moisture probes, pressure transducers, flow meters, IP cameras.
Industrial and agricultural-specific.
Purpose-built for agricultural environments. High accuracy, long calibration intervals, multi-year warranties. Integrates with industrial control systems and standard protocols (Modbus, 4-20 mA). Worth the cost when the measurement really matters and the environment is harsh. Examples: quality soil probes (Teros, Campbell Scientific), PAR sensors (Apogee, LiCor), industrial CO2 sensors, weather stations.
Research and laboratory.
Reference-grade accuracy, traceable calibration, specialized measurement capabilities. Usually not needed for operational agriculture. Appropriate for research, regulated applications, or anchoring a sensor network with a ground-truth reference. Examples: laboratory grade pH probes, reference PAR sensors, calibrated gas analyzers.
The honest rule: start with consumer or prosumer grade for most measurements, upgrade specific sensors to industrial grade when the data actually drives expensive decisions. A CO2 sensor for a cannabis operation where CO2 directly affects yield justifies an industrial sensor. A humidity sensor for a hobby greenhouse does not. Match the quality to the decision.
12Placement and sampling.
A sensor's value depends entirely on where it is placed. The most accurate sensor in the wrong location produces misleading data. The cheapest sensor in the right location can be transformative.
Measure where decisions matter.
Place sensors near the plants, not at the door where it is convenient to install. Air temperature at the canopy level, not at the ceiling. Soil moisture in the root zone, not at the surface. Light at the plant, not on top of the fixture. This seems obvious and yet most installations violate it to save wire runs.
Representative sampling.
A single sensor reports a single point. A greenhouse is not uniform — the south end is warmer, the north end cooler, the middle may have its own climate. For meaningful monitoring, place multiple sensors in representative locations. A five-zone greenhouse might have five temperature sensors, one per zone. A field might have sensors in each soil type. Redundancy also helps detect failing sensors — one sensor reading wildly different from its neighbors is probably broken, not revealing a hidden microclimate.
Height and depth matter.
Temperature and humidity vary with height in a greenhouse. Soil temperature varies with depth. Choose the height or depth that matches the decision. For seedling germination, surface conditions matter. For established crops, canopy conditions matter. For irrigation decisions, root zone conditions matter. For seasonal planning, deeper measurements that represent long-term averages matter.
Edge effects.
Sensors near walls, doors, vents, or equipment will read differently than sensors in the center. Sometimes this is useful (monitoring the cold spot near a vulnerable wall); sometimes it is misleading (drawing conclusions about the overall climate from a sensor in a drafty corner). Know what each sensor is really measuring.
Reporting frequency.
How often a sensor should report depends on how fast conditions change and how much battery life matters. Temperature and humidity that affect crops can be sampled every 1 to 5 minutes without missing anything actionable. Soil moisture that changes over hours or days can be sampled every 15 to 60 minutes. Flow rates and pressures for alerting need shorter intervals (every 10 to 30 seconds) because failures can be catastrophic quickly. Balance this against battery life — more frequent reporting means more power consumed.
13Common sensor failure modes.
Knowing how sensors typically fail helps diagnose problems and design systems that catch failures before they cause bad decisions.
Silent drift.
The most dangerous failure mode. Sensor produces numbers that look fine but have drifted far from reality. The data dashboard looks good; decisions made from it are wrong. Prevention: periodic calibration checks against known references, track long-term patterns and look for unexplained shifts, replace sensors at the end of their calibration-stable lifespan rather than waiting for obvious failures.
Sudden failure.
Sensor stops reporting entirely or reports nonsensical values. Easy to detect — the absence of data or impossible readings are obvious. Cause: dead battery, broken wire, component failure, physical damage. Prevention: every sensor should report into a monitoring system that alerts on data absence or impossible values. A sensor that reads -99°C or \+150°C has failed; the system should know this.
Physical damage.
Cables chewed by rodents, probes knocked by equipment, sensors flooded by overflow, enclosures cracked by weather. The dominant failure mode in rough environments. Prevention: install sensors with mechanical protection, route cables through conduit, choose sensor locations that are not in the path of operations, inspect periodically.
Contamination.
Dust, salts, biofilm, chemical residue accumulating on the sensor element. Changes behavior without obvious external damage. Prevention: design for periodic cleaning, choose sensors appropriate for the environment, install in locations that minimize contamination (shield from fertilizer mist, etc.).
Wrong placement.
Not really a failure of the sensor, but a failure of the installation. Sensor works perfectly and reports accurately — about a location that does not represent what the grower needs to know. Prevention: think carefully about placement before installation, move sensors if initial placement proves wrong, add sensors in genuinely representative locations rather than convenient ones.
Communication failure.
Sensor works but its data does not reach the monitoring system. Covered in Understanding Communications — radio problems, network outages, hub failures. Relevant here because the grower sees the same symptom (no data) and may misdiagnose the sensor.
14Rules of thumb for sensor decisions.
A practical summary:
Start with the decisions you need to make, not the sensors you want to buy. A sensor that does not support a decision is clutter.
Match sensor quality to decision consequence. High-consequence decisions justify industrial sensors; low-consequence decisions do not.
Measure at the plant, not at the wall. Placement matters more than precision.
Use multiple cheap sensors where redundancy and spatial coverage matter more than absolute accuracy. Use one good sensor as an anchor for the cheap ones.
Consumer-grade BLE sensors are often the appropriate technology for small-scale environmental monitoring. A greenhouse does not need industrial sensors for every variable.
For any measurement that drives expensive decisions, invest in a quality sensor with known calibration intervals. VPD for controlled environment growing, EC for hydroponic, PAR for supplemental lighting, soil moisture for irrigation scheduling.
Include sensor battery voltage and health as part of the reported data. A failing sensor reports its own deterioration; use that information.
Calibrate periodically or plan to replace. Drift is invisible; scheduled maintenance prevents it from becoming harmful.
Every sensor needs a plan for what happens when it fails. Alert on data absence. Cross-check against nearby sensors. Build redundancy for critical measurements.
When in doubt, more sensors in thoughtful locations beats fewer sensors in careful locations beats one great sensor in a bad location.
\[FAQ — JSON-LD to be generated from these Q&A pairs. Organized beginner → technical → specific measurements → placement → troubleshooting → cost → myths.\]
Frequently asked questions.
The honest version.
What is a sensor?
A sensor is a device that measures a physical property — temperature, humidity, light, moisture, pressure, flow, motion — and converts that measurement into an electrical signal a computer can read. Agricultural sensors range from simple thermometers to sophisticated probes that measure multiple properties at once. The common function is turning the physical world into numbers the grower can act on.
What does accuracy mean for a sensor?
Accuracy is how close the sensor's reading is to the true value, usually expressed as a plus-or-minus range. A temperature sensor rated at plus-or-minus 0.5°C reads within half a degree of the actual temperature. Accuracy is different from precision (how consistent the readings are) and resolution (the smallest change the sensor can detect). A complete specification includes all three.
What is sensor drift?
Drift is the gradual change in a sensor's readings over time, separate from actual changes in what it is measuring. A CO2 sensor that reads correctly when new may read 200 parts per million too high after two years of use. Drift is dangerous because the data looks fine but no longer represents reality. Preventing drift requires periodic calibration or scheduled replacement.
What is VPD?
VPD stands for vapor pressure deficit. It measures the difference between the amount of water vapor the air is currently holding and the maximum it could hold at the current temperature, expressed in kilopascals (kPa). VPD is what plants actually respond to — it determines how fast they transpire. Most crops do best with VPD between 0.8 and 1.2 kPa. VPD is calculated from temperature and humidity measurements.
What is DLI?
DLI stands for daily light integral. It is the total amount of photosynthetically useful light a plant receives over a full day, measured in moles of photons per square meter (mol/m²/d). Different crops have different DLI requirements — lettuce needs 12 to 17 mol/m²/d, tomatoes prefer 22 to 30, cannabis often targets 35 to 55. DLI is calculated by integrating PPFD measurements over the whole daylight period.
What is PPFD?
PPFD stands for photosynthetic photon flux density. It measures the rate of photosynthetically useful photons hitting a surface, in micromoles per square meter per second (μmol/m²/s). PPFD is the instantaneous measurement of usable light at a point — think of it as the intensity of light that drives photosynthesis right now. DLI is PPFD integrated over a day.
What is PAR?
PAR stands for photosynthetically active radiation. It is the range of light wavelengths (400 to 700 nanometers) that plants use for photosynthesis. PAR sensors are calibrated specifically to measure light in this range. Consumer lux meters measure what the human eye sees, which is different from what plants use — they do not work correctly for horticultural decisions.
What is EC?
EC stands for electrical conductivity. It is a measurement of how well a liquid or moist substrate conducts electricity, which correlates with dissolved salt content. In agriculture, EC is used as a proxy for nutrient concentration in hydroponic solutions and substrates, and for salinity in soils. Too low and plants are underfed; too high and they are stressed by excess salts.
What is volumetric water content?
Volumetric water content (VWC) is the percentage of a given volume of soil or substrate that is water. A soil with 20 percent VWC contains 20 milliliters of water in every 100 cubic centimeters of soil. VWC is what most modern soil moisture sensors measure, using capacitance to estimate how much water is present in the soil around the probe.
What is pH?
pH is a measurement of how acidic or basic a substance is, on a scale of 0 to 14 where 7 is neutral, values below 7 are acidic, and values above 7 are basic. Soil and substrate pH affects which nutrients are available to plants — most nutrients are most available between pH 6.0 and 7.0 in soil, slightly lower in soilless substrates. pH is measured with a glass electrode probe that requires regular calibration.
What is the best temperature sensor for a greenhouse?
For most greenhouse applications, a combined temperature and humidity sensor based on a digital sensor chip (like Sensirion SHT3x or Bosch BME280) is the right choice. Consumer BLE sensors using these chips cost little and give plus-or-minus 0.5°C accuracy. Multiple such sensors placed at canopy level throughout the greenhouse give better information than one industrial sensor in the wrong spot.
Where should I place a temperature sensor?
At the height of the plants, not at the ceiling. Shielded from direct sunlight (sensors in sun read five to ten degrees higher than the actual air temperature). Away from heat sources (heaters, equipment, lights). In locations representative of the zone you want to monitor, not the most convenient wire run. In larger greenhouses, use multiple sensors at representative points rather than one sensor in the corner.
Why does my temperature sensor read higher than my thermostat?
Common causes: the sensor is in direct or reflected sunlight, the sensor is near a heat source, the sensor has a radio or other heat-generating electronics packaged with it that warms the enclosure, or the thermostat is actually in a cooler location than the sensor. Compare the sensor's reading to a reference thermometer placed next to it to determine whether the sensor is accurate — often the issue is placement, not calibration.
What is the difference between relative humidity and absolute humidity?
Relative humidity (RH) is the percentage of water vapor the air is holding compared to the maximum it could hold at the current temperature. Absolute humidity is the actual amount of water vapor in the air (grams per cubic meter). The same RH means different things at different temperatures — 70 percent RH at 20°C contains much less water than 70 percent RH at 30°C. For actionable decisions, VPD (calculated from RH and temperature) is usually more useful than either humidity measurement alone.
Can I use a lux meter to measure light for plants?
Lux meters measure light as the human eye perceives it, which is different from what plants use for photosynthesis. Readings from a lux meter do not translate reliably to the PAR or PPFD values plants respond to. For any serious lighting decision — supplemental lighting calculations, DLI tracking, comparing grow lights — use a PAR sensor, not a lux meter. A quality PAR sensor costs a moderate amount.
How do I measure DLI?
DLI is measured by integrating PPFD over a full day. A PAR sensor reports PPFD continuously or at regular intervals; software integrates those readings into DLI. Modern monitoring systems (including Home Assistant setups with PAR sensors) automatically calculate DLI from PPFD data. The resulting DLI value reports the total photosynthetically useful light the plants received that day.
Does my grow light give off PAR?
LED grow lights, fluorescent grow lights, and high-pressure sodium lights all emit PAR, but the efficiency and spectrum varies widely. The useful measurement is how much PAR the plants actually receive, which depends on the light's output, the distance to the plants, and the shape of the light's distribution. Measure PPFD at the plant canopy with a PAR sensor to know what the plants are really getting.
Why are cheap soil moisture sensors unreliable?
Cheap capacitive soil moisture probes use a single-frequency measurement that is strongly affected by soil type, salinity, and temperature. The readings may move in the right direction — wetter when water is added, drier as soil dries — but absolute values can be far from true volumetric water content. A more expensive probe uses better electronics and calibration to give accurate absolute readings. For watching trends, cheap probes work. For making irrigation decisions that depend on specific moisture levels, better probes are worth the cost.
What is the best soil moisture sensor?
For research-grade measurements, probes like the Meter Group Teros 12 or Campbell Scientific CS650 give accurate VWC plus EC and temperature for a moderate amount per probe. For general monitoring, mid-tier capacitive probes work well enough to manage irrigation. The right choice depends on what the data will be used for — watering decisions based on specific thresholds need accurate probes; general monitoring can use cheaper ones.
How often should I calibrate a pH probe?
Continuous-immersion pH probes should be calibrated at least monthly, and many operations calibrate weekly. Handheld pH meters should be calibrated before any important measurement. pH probes have a limited operating life — typically one to three years of continuous use — after which the electrode material degrades and readings become unreliable regardless of calibration. When a probe's calibration offsets become large or unstable, it is time to replace the probe.
How do I measure water flow for irrigation monitoring?
Pulse-output flow meters are the standard choice. The meter sends an electrical pulse for each unit of water (commonly one pulse per gallon or per liter). A monitoring system counts the pulses and reports flow rate and total volume. Turbine and paddle-wheel meters are most common; electromagnetic meters are more expensive but have no moving parts. Ultrasonic strap-on meters install without cutting pipe. Costs range from a moderate amount depending on size and accuracy.
What pressure should my irrigation system run at?
Typical micro-irrigation (drip, low-volume) runs at 15 to 40 psi (1 to 2.7 bar). Typical impact and pop-up sprinklers run at 40 to 60 psi (2.7 to 4 bar). Specific requirements depend on equipment; consult the specifications of your sprinklers or emitters. A pressure transducer installed in each zone's supply line reveals when pressure is out of specification — low pressure indicates a leak, a closed valve, or pump problems; high pressure indicates a blockage.
How do I monitor a water tank level?
Simple: install a float switch at the critical low level (tank almost empty) — this gives a discrete alert when the tank needs filling. Better: install a continuous level sensor (ultrasonic, pressure, or capacitive) that reports the actual level in real time. Continuous monitoring reveals consumption patterns, detects leaks, and gives advance warning before the tank runs dry. Sensor cost spans a wide range — a basic float switch is inexpensive, a quality ultrasonic sensor moderately priced.
How accurate are cheap CO2 sensors?
CO2 sensors are typically inaccurate by hundreds of parts per million and drift significantly over months. For general awareness of CO2 levels in a room, they are acceptable. For CO2 enrichment control where the target is 800 to 1,500 ppm, a cheap sensor can be off by a quarter of the target value. A quality NDIR CO2 sensor with auto-calibration is worth the cost for any application where CO2 levels drive decisions.
Why do CO2 sensors drift?
CO2 sensors measure how much infrared light is absorbed by air passing through a sample chamber. Over time, the light source ages, the detector sensitivity changes, and the optical surfaces can become contaminated. Cheap single-beam sensors drift faster because they lack a reference beam for comparison. Better sensors use dual-beam or automatic baseline calibration (assuming the lowest reading over a period represents the outdoor atmospheric CO2 level, then adjusting the readings accordingly).
Do I need to monitor oxygen in a greenhouse?
Usually not. Atmospheric oxygen is very stable at around 20.9 percent. Only in very specific circumstances — tightly sealed environments with high plant respiration, confined workspaces with combustion equipment, or controlled atmosphere storage — does atmospheric oxygen need monitoring. For most growing operations, oxygen monitoring is unnecessary. Dissolved oxygen in hydroponic solutions is a different matter and can be worth monitoring.
Can I monitor whether my greenhouse doors were opened overnight?
Yes, a simple door contact sensor (magnetic reed switch) costs little and reports opened/closed status to the monitoring system. Combined with timestamps, this gives a complete log of every door opening. Consumer smart-home door sensors work well for this purpose in most agricultural contexts. Combined with motion sensors or cameras, the monitoring can indicate whether the opening was authorized entry, an animal, or wind.
How do I know if my pump is actually running?
Install a current transformer (CT) around the pump's power wire. The CT reports when current is flowing through the wire, which means the pump is drawing power. This catches the failure case where the control signal says the pump should be on but the pump is not actually running (failed motor, tripped breaker, disconnected wire). Current transformers cost a modest amount and install without breaking the circuit.
How do I know if a sensor is worth the money?
Match the sensor's quality to the decision it supports. A temperature sensor for general greenhouse awareness can be cheap. A temperature sensor controlling a critical crop's automated climate response should be better. A CO2 sensor for rough awareness can be cheap. A CO2 sensor controlling CO2 enrichment in a production cannabis facility should be industrial-grade. The test: what is the cost of a bad measurement? If the answer is significant, invest in the sensor. If not, do not.
How much should I spend on sensors for a small greenhouse?
A functional small-greenhouse monitoring system with temperature, humidity, light, and soil moisture sensors can be assembled for a moderate amount in sensors plus the hub computer and communication hardware. More comprehensive monitoring (CO2, VPD, multi-zone coverage, imaging) brings the sensor investment to a substantial amount. For most small operations, the sweet spot is a moderate amount in sensors covering the key measurements.
Why does my sensor show impossible values like -40°C or 999 percent humidity?
These are failure indicators. Common causes: disconnected wire or connector, bad power to the sensor, broken sensor element, or a communication error corrupting the data. Good monitoring software detects impossible values and flags them as sensor failures rather than treating them as real readings. If a single sensor shows impossible values, check its physical connection and power; if multiple sensors fail simultaneously, the issue is likely in the hub or network.
Why do my two sensors read different values in the same place?
Differences of a small amount (less than 1°C, less than 5 percent RH) are normal — every sensor has tolerance, and small variations in placement affect readings. Larger differences indicate a problem: one sensor may be failing, one may be in a slightly different microclimate (different sun exposure, different airflow), or one may be miscalibrated. Compare both sensors to a reference (quality handheld thermometer/hygrometer) to determine which is correct.
How long do sensors typically last?
Quality sensors in protected environments often last 5 to 10 years. Consumer-grade sensors in the same environments may last 2 to 5 years before drift becomes significant. Sensors in harsh environments (outdoor weather, chemical exposure, high humidity) last less. Sensors with moving parts or consumable elements (pH probes, chemical gas sensors) have shorter lifespans, often 1 to 3 years. Plan for replacement as part of the operational budget.
Is more expensive always more accurate?
Not always. A well-designed consumer-grade sensor can match or exceed the accuracy of a poorly-designed industrial sensor. What expensive sensors usually offer is better calibration stability (less drift), wider operating range, better environmental tolerance, and longer warranties. For accuracy at installation, an inexpensive sensor can match a moderately priced sensor. For accuracy five years later, the moderately priced sensor usually wins. Match the price to the longevity and conditions needed, not just the accuracy claim.
Do I really need to calibrate sensors?
It depends on the sensor type. Temperature sensors based on digital chips rarely need field calibration and hold their accuracy for years. Humidity sensors drift slowly and benefit from occasional calibration. pH probes need frequent calibration (weekly to monthly). CO2 sensors drift and need periodic recalibration. In general: the less stable the measurement principle, the more calibration matters. Plan for calibration on probe-based chemical sensors; less so on solid-state temperature and humidity sensors.
Can I measure everything with one multi-function sensor?
Some multi-function probes are excellent (soil probes that measure VWC, EC, and temperature together; environmental sensors that measure temperature, humidity, and pressure in one chip). But trying to cover everything with one sensor location rarely works — the right place for a soil moisture measurement is not the same as the right place for an air temperature measurement. Use multi-function sensors where the measurements genuinely belong together, and separate sensors where they do not.