Agriculture & Future Food Systems · Article

Sensors, irrigation and feedback

Trace a moisture reading through a control decision and a real water response, while keeping sensor limitations and crop requirements separate.

Preparation and related reading

Shared Foundations

This article helps you

Distinguish a local sensor estimate from the whole root zone.
Separate measurement calibration from choosing a control setpoint.
Explain how feedback checks the result of an irrigation action.

A reading needs a location

A probe beside one plant reports wetter soil after irrigation. Plants farther along the row may still receive less water. The reading is useful, but it does not automatically describe the whole bed.

An irrigation controller connects a measurement with a decision and an actuator, such as a valve. Feedback means checking the conditions again after action. A reliable system needs all three parts to work: relevant observations, an appropriate decision rule and actual delivery.

Identify the quantity before the threshold

Some sensors estimate volumetric water content: water volume divided by soil or substrate volume. Others measure water tension, which relates to how water is held. Their numbers have different meanings and cannot share a threshold merely because both concern moisture.

Electrical moisture estimates depend on the sensor's relationship with its surrounding medium. Soil properties, temperature and salinity can influence some devices. Calibration and checking against suitable reference observations establish how the measurement should be interpreted in its actual use.

Placement is part of measurement

Choose positions and depths that address the crop's active root zone and the variation being managed. A probe in one unusually wet spot can give a different picture from one in a representative location. Separate soil or substrate zones may need separate observations.

Good contact with the medium matters; an air gap can change the response. Keep the sensor identity, position, depth, time and calibration information with its readings. These records make an apparent moisture change distinguishable from moving the probe or changing its interpretation.

The controller needs a crop-specific rule

A bypass controller can skip a scheduled irrigation event when the monitored zone is sufficiently wet. An on-demand controller can start at a lower moisture threshold and stop at a higher one. The gap between those thresholds prevents the rule from switching at one identical value.

Choosing the thresholds is a production decision, involving crop stage, medium, root-zone conditions and delivery capacity. It is not the same as establishing sensor accuracy. Some device manuals call setting a threshold “calibration”; retain that distinction when reading the term.

A command is not delivered water

A valve command does not establish that water reached the intended plants. Supply failure, a blocked emitter or a damaged line can interrupt delivery. Compare relevant flow or delivery observations with the root-zone response.

Water movement and the measurement can take time to respond. A controller must account for that behaviour rather than repeatedly issuing commands because its first reading has not changed. The neighbouring irrigation articles explain crop and soil requirements; equipment commissioning checks this installation's response.

Keep doubtful readings visible

A stale timestamp, implausible jump or disagreement among observations calls for investigation. An unavailable sensor should not quietly become a valid “dry” reading. Define who can intervene and the installation's reviewed response when observations or delivery fail.

The invented control sequence below illustrates only how two thresholds retain a running state. Its values are not crop recommendations. Review the real crop outcome, water use and equipment behaviour together; automation is useful when it improves those outcomes under known operating conditions.

Two thresholds retain a state

Hypothetical volume fractions and rule, not recommended setpoints. Delivery and representative sensing require separate checks.

Read the diagram as text

The invented controller begins off and follows valid readings 0.24 → 0.27 → 0.31, using a start threshold below 0.26 and a stop threshold at 0.30.

Parts

0.24: start: Below 0.26; controller was off
0.27: continue: Already running; still below 0.30
0.31: stop: Stop threshold 0.30 reached

Connections

0.24: start → 0.27: continue: remeasure
0.27: continue → 0.31: stop: remeasure

Worked examples

Follow the controller's state

An invented on-demand rule starts when water content is below 0.26 and stops when it reaches 0.30. The system begins off. Valid readings occur in order: 0.24, 0.27 and 0.31. What actions follow?

0.24 is below 0.26, so the stipulated rule starts irrigation.
At 0.27 the system is already running and has not reached 0.30, so it continues.
At 0.31 it has reached the stop threshold, so it stops. The intermediate reading alone would not start a system that was still off.

Start, continue, stop. These dimensionless volume fractions and the rule are hypothetical, with valid readings stipulated; they do not establish suitable irrigation amounts.

State the uniformity assumption

A hypothetical 100 L substrate volume has a truly uniform water content of 0.25 by volume. How much water does it contain, and would one local probe prove that uniformity?

Water volume is 100 L × 0.25 = 25 L.
That calculation requires the stated uniform condition across the whole 100 L.
One probe samples a limited surrounding region; without representative evidence, its 0.25 estimate cannot establish the whole-volume condition.

25 L under the explicit uniformity assumption. A single local reading does not prove that assumption.

Check your understanding

Can a water-tension threshold be copied directly into a volumetric-content controller?
Show answer and reasoning
No.
The quantities and units differ; their relationship depends on the medium.
Does an “open valve” entry prove that every emitter delivered water?
Show answer and reasoning
No.
The command and physical delivery are separate observations.
Does selecting a start threshold demonstrate sensor accuracy?
Show answer and reasoning
No.
Control settings and measurement calibration answer different questions.

Key ideas

Keep the measurement quantity, location and interpretation attached.
Use crop-specific control rules and verify actual delivery.
Feedback needs fresh observations and a reviewed failure response.

References

Sources supporting the explanations and examples in this article.