River Level Sensor Principles and Measurement Uncertainty: Radar, Ultrasonic and Hydrostatic

River Level Sensor Principles and Measurement Uncertainty: Radar, Ultrasonic and Hydrostatic

River stage sensors do not all measure the same physical quantity. Radar and ultrasonic instruments first measure standoff distance; a submerged hydrostatic transmitter senses liquid pressure. Without consistent elevation references and conversions, their displayed values cannot be compared directly.

For procurement and field acceptance, distinguish the sensor's distance or pressure error from errors introduced by the surveyed datum, mounting and conversion to water level.

Stage elevation, local level and depth are distinct

A bridge-mounted radar reports distance D. With surveyed sensor reference elevation Z_sensor, stage elevation is Z_water = Z_sensor − D. Water depth additionally requires a specified bed level, which may change through scour or sediment deposition.

Z_water = Z_sensor − D

Example: reference elevation 105.00 m minus 8.20 m distance gives 96.80 m stage elevation.

A 0.10 m shift in the mounting datum can create a corresponding systematic error if not corrected.

Radar distance measurement and uncertainty

Pulse radar estimates distance from propagation delay; FMCW radar uses the beat frequency between its chirp and delayed reflection. In river monitoring the tracked target must be the water surface—not a bridge pier, bank or exposed riverbed at low stage.

• Benefit: non-contact measurement without a permanently submerged sensing head; air temperature does not influence radar through sound-speed changes.

• Limitations: structural echoes, beam footprint, spray, condensation, ice and floating debris can disrupt target tracking.

• A higher carrier frequency is not proof of lower measurement error; effective bandwidth, antenna and processing matter.

Compare the METRAVON radar water-level range and confirm beam width and near-zone clearance by model.

Why ultrasonic ranging requires gas-path correction

A downward-facing ultrasonic sensor times the acoustic round trip through gas. Temperature, gas composition and gradients change sound speed; a single temperature sensor may not fully correct a strong vertical gradient.

D = c_gas × t / 2

t is round-trip time and c_gas is the effective speed of sound along the path.

Beam obstructions, heavy foam, dense vapor and near-zone blocking at maximum stage add further errors.

Ultrasound can work in suitable channels with stable gas conditions and beam clearance. Clean water is not its only application, but the atmosphere and surface must be appropriate.

Hydrostatic conversion, density and barometric compensation

A hydrostatic transmitter calculates water column height from pressure above its sensing point. Vented gauge versions commonly compensate atmospheric pressure through a vented cable; absolute-pressure versions may require separate barometric correction. Density changes with temperature, salinity and suspended solids can affect conversion.

h = P / (ρg)

P is appropriately compensated liquid-column pressure, ρ water density and g gravitational acceleration.

This is the column above the sensor; add the submerged sensor reference elevation to obtain stage elevation.

• Place and secure the probe away from persistent silt deposition; protect the cable and vent tube.

• Turbid water does not automatically prevent measurement, but burial and physical damage do matter.

Resolution, repeatability and installed accuracy

A one-millimetre display increment is not evidence of ±1 mm accuracy. Validate the full system against its model-specific specification and field reference.

Field verification and procurement FAQ

1. Record station ID, sensor reference, common elevation datum and observed reference stage.

2. Safely compare local sensor readings against a reliable staff gauge or other reference.

3. Where practical, compare several representative stages, not a single point.

4. Check units, zero offsets, output scaling and timestamps.

5. Retain acceptance records as a baseline for future shifts or reconfiguration.

Q1. Is radar always more accurate than a staff gauge?

No universal ranking is useful. Compare both observations at the same time and elevation datum, allowing for surface movement and reference-reading uncertainty. If readings disagree, check the surveyed sensor reference and target before adjusting the radar offset to match a single observation.

Q2. Does stage alone give discharge?

No. A water-surface elevation does not establish flow rate without a valid rating relation or additional hydraulic measurements. Changes to the channel can affect that relation even if the stage sensor remains stable. Keep level verification and discharge validation as separate acceptance tasks.

Q3. Is 80 GHz necessarily more accurate than 26 GHz?

Carrier frequency alone is not an accuracy specification. Compare the ordered model's distance accuracy, beam footprint, near-zone limits and installation conditions. A narrower beam may help avoid a pier, but it does not correct an incorrect mounting elevation or a moving support.

Building an installed measurement error budget

Acceptance cannot simply repeat the factory accuracy specified under ideal conditions. Include the contributions from sensing distance or pressure, surveyed datum, gas/density correction, mounting movement and quantization or software scaling.

Before combining multiple uncertainties statistically, verify independence and identify systematic bias. There is no universal formula for adding every maximum error figure.