Leveling

Differential leveling carries elevations from a benchmark using a level and a rod. The bookkeeping is simple; the discipline is in the checks.

Fig. 1 The level sits between two rods. Its line of sight is horizontal, so the backsight reading minus the foresight reading is the difference in elevation. 3D illustration (sample).
BM TP1 line of sight = HI BSFS

The three readings

Field-note checks

For the points that carry the run, ΣBS − ΣFS must equal the last elevation minus the first. This proves the arithmetic — not the readings. The real check is to close: end on a second benchmark, or loop back to the first. The difference between the computed and known elevation is the misclosure.

Adjusting a level run

Errors build up with each setup (or each unit of distance). The simplest adjustment gives a point reached on setup k a correction of −e·k/n for n setups. With balanced sight lengths this is the same as distributing by distance.

Accuracy is usually expressed as an allowable misclosure proportional to the square root of the run length. FGCS (1984) vertical-control classes range from 3 mm·√K (first order) to 12 mm·√K (third order), K in kilometres; construction and agency specs often use their own value in feet × √miles.

Main error sources

Trigonometric leveling

A total station gives elevation differences from slope distance and zenith angle: Δh = S·cos z + instrument height − rod (prism) height. Over long sights add the curvature-and-refraction correction (about 0.0206 ft × (distance in thousands of feet)²).

Try it in the calculators

When not to use: this guide teaches the method. It is not a survey standard and does not replace your state's minimum technical standards for surveys or the project specifications.

Original explanations written for this site. Plane-surveying fundamentals only — not exam questions, and not a substitute for a surveying text or your agency's survey manual.