Differential Leveling Loop Closure Calculator

Differential leveling (also called spirit leveling) is the most common method of establishing precise vertical control. A surveyor sets up a level instrument at an arbitrary point between two rod stations and reads the height of the staff at each station through the telescope.

The Height of Instrument (HI) is the key intermediate quantity. It equals the known elevation at the backsight point plus the backsight rod reading:

Once the HI is known, the elevation at any foresight (turning point) is found by subtracting the foresight reading:

A level loop starts and ends at the same benchmark. A BM-to-BM run starts at one benchmark and closes on a different one with a known elevation. In either case, the computed final elevation is compared with the known value to check the quality of the work.

For a closed loop where the start and closing benchmarks are the same point, the sum of all backsight readings minus the sum of all foresight readings should equal zero:

In practice, random errors in reading the staff cause a small deviation. The closure error is:

A positive closure error means the computed elevation is too high — the foresight readings were collectively too small or the backsight readings too large. A negative closure error means the computed elevation is too low.

The closure error is always expressed in millimetres for comparison with tolerance standards, regardless of whether your field readings were in metres or feet.

When a loop closure passes the tolerance test, the closure error is systematically distributed among all turning points so the final elevation exactly matches the known benchmark. The simplest method — and the one used by this calculator — is equal distribution (sometimes called the Bowditch rule applied to leveling).

The correction per setup is:

This correction is applied cumulatively to each successive turning point:

• TP1 adjusted elevation = raw elevation + 1c
• TP2 adjusted elevation = raw elevation + 2c
• TPn (closing BM) adjusted elevation = raw elevation + nc = known elevation exactly

Equal distribution assumes all setups are of equal quality and equal sight length. When setups vary significantly in length, a distance-weighted distribution is more rigorous — but for routine site work, equal distribution is the accepted standard.

If the loop fails the tolerance test, do not apply the Bowditch correction. Instead, re-run the loop and look for the setup that produced an abnormally large change — that is likely where the blunder occurred.

A blunder is a gross error — a mistake large enough that the loop closure will exceed even the most lenient tolerance. Unlike random errors, blunders cannot be removed by distributing corrections. The loop must be re-run.

• Mis-reading the staff: Confusing a 1.423 with 1.243 is a classic transposition error. If your loop fails by a large margin, look for a setup where the BS or FS value seems inconsistent with the others.
• Unstable turning point: If the staff is held on soft ground between the BS and FS readings, the staff can sink. The BS and FS are recorded at different heights, creating a systematic error in that setup.
• Collimation error: If the instrument's line of sight is not truly horizontal, long sights introduce a systematic error proportional to sight length. Balance backsight and foresight distances at each setup to cancel this effect.
• Curvature and refraction: Over long sights (beyond 100 m), the Earth's curvature and atmospheric refraction cause the line of sight to curve upward. Balanced sight lengths cancel most of this effect automatically.

Blunder isolation: If the loop fails, re-run from the starting benchmark to the midpoint of the loop. If that section closes correctly, the blunder is in the second half. Halve iteratively until the bad setup is found.