Benchmark in Surveying: How a BM Carries Levels to Your Site

What is a benchmark in surveying?

A benchmark is a stable reference mark, cut into a pillar, culvert, rock or building plinth, whose reduced level (RL) has been established by careful levelling. It gives the starting height for a level survey. A national benchmark refers heights to mean sea level; a site benchmark may carry a borrowed national value or an agreed arbitrary one, such as 100.000 m.

The traditional mark is a chiselled broad arrow pointing up to a horizontal bar; the bar marks the level.

Types of benchmark in surveying

If services are buried near your new levels, work out your GPR survey cost for a scan as well.

How are levels carried from a benchmark?

Every benchmark in surveying is used the same way, by a closed levelling run. The run goes from the BM to the new points and back to the BM, or on to a second known mark. The return gives a closing error. With an auto level we work to about ±2.5 mm per km of run, and many specifications set an allowable closure that grows with the square root of the distance.

A worked example

The BM reads RL 198.450 m. A backsight on its staff reads 1.235 m, so the height of instrument is 199.685 m. A foresight on a new plinth reads 0.870 m, giving RL 198.815 m. If the run returns to the BM at 198.452 m, the 2 mm misclosure is shared out along the run.

A field scenario: bringing a level onto a new plot

The nearest reliable mark is often a kilometre or more away, so the level is carried in and brought back. Suppose a benchmark on a culvert sits 1.8 km down the road from a new warehouse plot. The crew runs a line of levels to the plot, sets two temporary benchmarks on its boundary wall footings and returns to the culvert, a loop of about 3.6 km. A specification written as 12√K mm, with K in kilometres, would allow about 23 mm on that loop. A misclosure of 8 mm passes comfortably and is shared out; a misclosure of 60 mm sends the crew back to find a misread staff.

Where levelling with benchmarks goes wrong

Levelling errors are seldom about the instrument and almost always about the marks and the bookkeeping.

• A disturbed mark. Road resurfacing, culvert repairs and new kerbs bury or shift old benchmarks.
• Mixed datums. One drawing at an arbitrary 100.000 m and another at mean sea level cannot be overlaid directly.
• Staff errors. Inverted staff readings on soffits and ceilings carry a minus sign that is easy to drop.
• Open runs. A level line that never closes has no check at all.
• Temporary marks on movable objects. A painted mark on a manhole cover or a loose kerb stone is not a benchmark.

How benchmarks are recorded on Mezux level sheets

Each sheet says where the height came from and how well it closed. We describe the starting benchmark, its value and its source, then list every TBM set on site with a sketch, a photograph and its reduced level. The closing error and the adjustment method are printed with the results, so a contractor using the TBMs later knows what they rest on. Levels themselves can be checked by the ISO 17123-2 field procedure, the part of that series written for levelling instruments.

Why not use GNSS heights instead?

GNSS heights are measured above an ellipsoid, not above mean sea level, and are less precise than positions: 2–3 cm against 1–2 cm. That holds for every form of satellite-based positioning, from a single GPS receiver to a base station and rover pair. Even where DGPS works from a network of ground-based reference stations, the GPS signals yield heights that need a geoid model before they can be compared with a benchmark. For drains, sewers and floor levels a few centimetres matter, so we either level from a benchmark or check GNSS heights against one. Horizontal position is a separate question, handled in grid coordinates; read Easting and Northing. Where levels and positions must be carried through built-up streets together, a short guide on Total Station Traverse shows the method.

Before hiring a team, read the questions to ask before a total station survey, or ask for a quotation for your project.

Frequently asked questions

What is the difference between a BM and a TBM?

A benchmark is permanent and its value is established by authority or careful levelling. A TBM is set by the site team for convenience and is checked back to a benchmark.

Can I trust an old benchmark on a culvert?

Only after checking it. Structures settle or are rebuilt, so we level between two marks before relying on either.

Does a DGPS survey also need a benchmark?

Yes in most cases, if heights matter: a differential global positioning system survey fixes position well, but its heights still need a level check. See the DGPS survey process step by step.

How far apart should temporary benchmarks be?

Close enough that every part of the site can be reached from one set-up, typically within a hundred metres or so. Always set at least two, so each can check the other.

Can a TBM be marked on a building wall?

Yes, if the wall is stable and will not be demolished during the works. A chiselled or drilled mark with a painted reference is common.

Why do two nearby benchmarks sometimes disagree?

Settlement, rebuilding or a different original datum. Levelling between them shows which one still holds its value.

Does a GNSS height replace a benchmark on a large site?

It can supplement one. With a good geoid model and a check on at least one levelled mark, GNSS heights suit earthwork and contour work; drains and floors are still levelled.

What is the difference between a benchmark and a control pillar?

A benchmark gives height; a control pillar, often used as a reference station for GNSS, normally gives horizontal coordinates and may carry a level too. Many project pillars serve both purposes.

Written by Saniya Parveen · Technically reviewed by Rahul Rai · Updated 8 November 2026