Sensors used in agricultural and landscape automation differ significantly in measurement principle, response time, calibration requirement and communication interface. Selection considers soil texture, salinity, temperature range and long-term stability.
Soil moisture sensors
Soil water content is most often inferred from dielectric permittivity: water has a relative permittivity (ε ≈ 80) that is far higher than the soil matrix (ε ≈ 4), which makes electrical sensing sensitive. Three core technologies are common:
- TDR (Time-Domain Reflectometry): a voltage pulse propagates along the probe; permittivity is computed from the round-trip time of the reflection. The Topp equation (1980) converts the result into volumetric water content (θv, %VWC). It is the method least affected by salinity, with accuracy near ±2% VWC.
- FDR (Frequency-Domain Reflectometry): measures capacitance around the probe at a fixed frequency (typically 70–150 MHz). Cheaper than TDR; needs temperature and salinity correction.
- Capacitive probes: typically operate in the 10–100 MHz range; low-cost but the most dependent on soil-specific calibration.
Tensiometers and granular-matrix sensors (Watermark) instead measure matric potential (kPa), which directly reflects how hard the plant has to work to extract water. The irrigation threshold for most turf falls between −30 and −50 kPa.
Hydraulic and flow sensors
- Turbine / paddle-wheel flow meters: pulse output, low cost, ±2–5% accuracy.
- Electromagnetic flow meters: based on Faraday's law, no moving parts, ±0.5% accuracy; suitable for saline water or fertigation streams.
- Ultrasonic flow meters: transit-time principle, can be installed clamp-on without piercing the pipe.
- Pressure sensors: piezoresistive or ceramic; typical range 0–10 bar, output 4–20 mA or 0.5–4.5 V ratiometric.
- Water level sensors: hydrostatic (submersible), ultrasonic and radar; used to track tank fill.
Micrometeorology sensors
The FAO-56 ETo calculation requires four variables: ambient temperature, relative humidity, wind speed and solar radiation. These sensors are typically grouped in an integrated weather station. A site may add a tipping-bucket rain gauge (resolution 0.2 mm), a leaf-wetness sensor (for fungal-disease models) and a PAR sensor (for greenhouse use).
Communication protocols
Bringing dissimilar field sensors into one system is possible thanks to standard interfaces:
- SDI-12 (Serial Digital Interface, 1200 baud): an asymmetric digital protocol designed for agriculture and hydrology, released by USGS in 1988. Supports addressing of up to 62 devices on a single cable; very low power.
- Modbus RTU (over RS-485, 1979): the industrial de facto standard. Used to integrate multi-vendor flow meters and PLCs; typical baud 9600–115200.
- 4–20 mA current loop: a noise-immune analog standard for long runs; 4 mA = zero reading, 20 mA = full scale. A broken loop (≤3.6 mA) is detectable.
- 0–10 V analog: a simple actuator and sensor interface for short distances.
- I²C / SPI: intra-device interfaces for integrated digital sensors.
- Bluetooth Low Energy (BLE 4.0+) and LoRa: preferred in wireless soil-moisture nodes; LoRa trades a low data rate (0.3–50 kbps) for kilometre-scale range.
Calibration and drift management
All electrochemical and dielectric sensors drift over time. Manufacturers typically declare an annual drift of ±1–3%. Calibration is verified by a two-point method (dry-soil and saturated-soil readings) or by laboratory gravimetric comparison. EC (electrical conductivity) sensors are additionally normalised to 25 °C with temperature compensation.