| Instrument | Count | Working tolerance | Typical use | Calibration |
|---|---|---|---|---|
| Robotic total station, 1″ | 4 | ±1–3 mm at 200 m | Control, setting out, monitoring | Annual UKAS + quarterly two-peg |
| Digital level, invar staff | 3 | ±0.7 mm per km double run | Precise levelling, settlement | Annual UKAS + monthly collimation |
| GNSS rover, network RTK | 6 | ±8–15 mm plan, ±15–25 mm height | Topographic detail, control extension | Annual, plus daily check into known station |
| Terrestrial laser scanner | 2 | ±2–6 mm at 50 m | Measured building, retrofit, record | Annual UKAS + per-project target check |
| Survey UAV, RTK enabled | 2 | ±25–35 mm to ground control | Volumes, faces, inaccessible ground | Camera calibration each mobilisation |
| Multibeam echosounder | 1 | ±25–50 mm at 12 m depth | Berths, scour, channel bathymetry | Patch test each mobilisation |
| Single-beam echosounder | 2 | ±30 mm + 0.1% of depth | Sections, reconnaissance, volumes | Bar check at start and end of survey |
| GPR, dual frequency | 2 | Depth ±15% of measured depth | PAS 128 utility detection | Annual, plus on-site velocity calibration |
| Tide gauge, pressure | 2 | ±10 mm | Tidal reduction for hydrographic work | Levelled to ODN on every deployment |
Calibration
Why we publish residuals
Every network we observe is adjusted by least squares, and the adjustment report goes out with the drawing whether or not anyone asked for it. It contains the residual on each observation, the standard error of each station, and the chi-squared test on the whole network.
Most clients never open it. The ones who do are usually engineers checking whether a settlement of four millimetres is real movement or is inside our noise, and that is exactly the question the report exists to answer. A survey that cannot state its own uncertainty cannot be used to decide anything marginal, and marginal decisions are the expensive ones.
The check network
We maintain eleven stations around the harbour at Doc Fictoria, observed as a closed network twice a year and levelled to a deep datum. Instruments are checked into it before any monitoring contract and after any repair, transport incident or drop. The comparison is what produces the working tolerances in the table above; they are not copied from a specification sheet.
What gets an instrument stood down
- Two-peg test outside 3″ on a total station.
- Collimation drift on a digital level between monthly checks.
- Any drop, however soft the ground, until it is re-checked.
- Scanner target residuals above 4 mm on a controlled setup.
- A patch test that will not resolve within tolerance on the multibeam.
An instrument stood down mid-contract is replaced from the fleet the same day. That is most of the reason there are four total stations for seventeen people.