Levelling Principles and Practice

Introduction

Levelling is the surveying operation used to determine the elevations of points relative to a chosen datum and to establish points at specified elevations. It is fundamental to construction because roads, drainage systems, foundations, pipelines, canals, pavements and earthworks must be positioned not only horizontally but also vertically. A small error in level may cause ponding, inadequate drainage, excessive excavation or structural difficulties.

The principal instrument used in ordinary levelling is the level, which establishes a horizontal line of sight. A levelling staff provides a graduated scale from which staff readings are taken. The essential observation is the difference between two staff readings: when the same horizontal line of sight is used, the difference in readings gives the difference in elevation between the two staff stations.

This chapter develops the principles and field procedures for differential, profile, cross-section, reciprocal, fly and precise levelling. It also explains the height-of-instrument and rise-and-fall methods, curvature and refraction corrections, closure checks, field booking and sources of error.

Learning Objectives

After studying this chapter, a student should be able to:

• explain the datum, benchmark, reduced level, backsight, foresight and intermediate sight;
• carry out differential, profile, cross-section, reciprocal, fly and precise levelling;
• reduce observations by the height-of-instrument and rise-and-fall methods;
• apply curvature and refraction corrections when required;
• perform arithmetic and geometric checks and assess levelling closure;
• identify major sources of error and select appropriate field controls;
• interpret levelling results for construction and infrastructure work.

1. Fundamental Concepts

1.1 Datum and reduced level

A datum is a reference surface or reference elevation from which levels are measured. Mean sea level is commonly used as a national datum, although a project may use an assumed datum for convenience. The elevation of a point above or below the datum is its reduced level (RL).

A benchmark (BM) is a permanent or temporary point whose RL is known or established. Common types include a national benchmark, permanent benchmark, temporary benchmark and arbitrary benchmark. A benchmark should be stable, identifiable and protected from disturbance.

1.2 Line of collimation and line of sight

The line of collimation is the line joining the intersection of the cross-hairs and the optical centre of the objective lens. When the level is correctly adjusted, this line is horizontal. The staff reading is the vertical distance from the point on which the staff stands to the line of sight.

If a staff reading is large, the staff station is lower relative to the instrument line of sight. If the reading is small, the station is higher. Thus, for two points observed from one instrument position:

where ΔRL is the elevation of A relative to B, and R_A and R_B are the staff readings at A and B in metres. Equivalently, the point with the smaller reading has the greater RL.

1.3 Staff readings

A backsight (BS) is the first staff reading taken after setting up the instrument, normally on a point of known RL or a change point. A foresight (FS) is the last reading taken before the instrument is moved, normally on a change point or closing benchmark. An intermediate sight (IS) is any other reading taken without moving the instrument.

A change point (CP), also called a turning point, is a firm point on which a foresight is taken before moving the instrument and a backsight is taken after moving it. The staff must not move between the two observations. A poor change point, such as loose soil or a stone that can shift, produces a false level transfer.

The height of instrument (HI), or height of collimation, is the RL of the line of sight. For a point of known RL:

For any point observed from that setup:

These equations assume that the staff reading is correctly observed and that the line of sight is horizontal at the instrument.

2. Instruments and Field Practice

A conventional automatic level contains a telescope, cross-hairs, objective lens, focusing system, circular bubble and compensator. The compensator automatically makes the line of sight approximately horizontal within its operating range. A digital level reads a coded staff electronically and can reduce reading and booking errors. A precise level has a high-quality telescope and compensator and is used with precise staffs and controlled procedures.

Before observing, the surveyor should inspect the tripod, clamps, compensator, eyepiece, objective lens and staff. The instrument is centred approximately over the setup point only when the method or project requires it; ordinary differential levelling does not require centring over a known point. The tripod should be firm, the instrument levelled, the eyepiece focused, and parallax eliminated. Parallax exists when the staff image appears to move relative to the cross-hairs as the observer's eye moves; it is removed by separately focusing the eyepiece and staff image.

The staff should be held vertical. A staff bubble, rocking technique or electronic staff aid may be used. Readings should be recorded immediately, with units and station descriptions. Distances to backsight and foresight should be kept approximately equal where possible because this reduces the effect of collimation error and partially balances atmospheric effects.

3. Differential Levelling

Differential levelling transfers elevation between points that cannot be observed conveniently from one instrument position. The procedure is:

1. Set up the level at a position from which the benchmark and a suitable change point are visible.
2. Take a BS on the benchmark.
3. Take an FS on the change point.
4. Move the instrument beyond or to the other side of the change point.
5. Take a BS on the same change point.
6. Continue until the destination point or closing benchmark is reached.

The instrument should not be set directly on a change point. A stable change point and balanced sight lengths are essential.

Height-of-instrument method

In this method, each instrument setup is reduced independently. The arithmetic check is:

The check follows by adding the successive equations for HI and RL. Intermediate sights do not appear in the check because they are readings from points observed and reduced within a setup.

Worked Example 1: Height-of-instrument method

A benchmark has RL 100.000 m. The following readings are observed:

First setup: HI₁ = 100.000 + 1.245 = 101.245 m. Therefore, RL_P1 = 101.245 − 1.860 = 99.385 m and RL_CP1 = 101.245 − 2.115 = 99.130 m.

Second setup: HI₂ = 99.130 + 0.955 = 100.085 m. Therefore, RL_P2 = 100.085 − 1.430 = 98.655 m and RL_BM2 = 100.085 − 1.775 = 98.310 m.

Check: ΣBS − ΣFS = (1.245 + 0.955) − (2.115 + 1.775) = −1.690 m. Also, RL_BM2 − RL_BM1 = 98.310 − 100.000 = −1.690 m. The computation therefore closes arithmetically.

Rise-and-fall method

The rise-and-fall method compares consecutive staff readings. If the next reading is smaller, the ground has risen; if it is larger, the ground has fallen.

The RL is then obtained by adding a rise or subtracting a fall:

The complete check is:

The rise-and-fall method gives a valuable check at each consecutive pair of readings, but it requires careful identification of whether a reading is associated with a new point or a repeated change point. At a change point, the FS and subsequent BS refer to the same elevation and must not be treated as a rise or fall between different ground points.

4. Profile and Cross-Section Levelling

4.1 Profile levelling

Profile levelling determines elevations along a selected alignment, such as a road centreline, railway, pipeline or canal. Stations are usually marked at regular chainages and at significant changes in slope, ground condition, drainage or proposed design grade. The resulting longitudinal section shows existing ground RL against chainage and supports grade design, cut-and-fill estimation and drainage analysis.

A typical procedure begins with a benchmark near the alignment. BS, IS and FS readings are taken as the instrument is advanced. The chainage, point description and RL are recorded. Instrument stations should be selected so that the line of sight is not obstructed and the backsight and foresight distances remain reasonably balanced.

4.2 Cross-section levelling

Cross-sections are taken at right angles to the centreline at selected chainages. They establish the transverse ground shape required for road width, side slopes, drainage channels, quantities and construction control. Readings are identified by offset and side, for example, 5 m left, centreline, 5 m right and 10 m right.

Cross-sections may be taken using an ordinary level, a laser level, a total station or a GNSS receiver. The chosen method must provide accuracy appropriate to the design. A section should include all points where the transverse slope changes, not merely points at fixed offsets.

5. Fly Levelling and Precise Levelling

5.1 Fly levelling

Fly levelling is the rapid transfer of RL from one benchmark to another. Only BS and FS readings are normally taken; intermediate details are omitted. It is used to establish temporary benchmarks around a project or to connect a work site to a control network. Speed does not justify poor practice: change points must be stable, sight lengths should be balanced, and the line should close on a known point whenever possible.

5.2 Precise levelling

Precise levelling is used where small vertical errors are significant, such as settlement monitoring, deformation surveys, major structures and high-order control. It requires:

• calibrated precise levels and approved staffs;
• short and approximately equal BS and FS sight lengths;
• careful staff verticality and firm turning points;
• repeated or double-run observations where specified;
• controlled observation timing to limit heat shimmer and refraction;
• a documented allowable misclosure.

For a levelling line of length K kilometres, a commonly used project tolerance may be expressed as:

where e is the closure error in millimetres, K is the line length in kilometres, and C is a specification-dependent constant in millimetres per square root kilometre. The value of C must come from the governing survey standard or project specification; it is not universal. A tolerance is an acceptance criterion, not a correction formula.

6. Curvature and Refraction

A level instrument produces a horizontal line of sight, whereas the earth's level surface curves away from that tangent line. At long sight distances, the staff reading is affected by both curvature and atmospheric refraction.

The curvature correction, for distance d in kilometres, is approximately:

This represents the amount by which the curved earth surface lies below the tangent line. In a staff-reading convention, the correction to the observed reading due to curvature is negative because the observed reading is too large relative to the level surface:

Atmospheric refraction bends the line of sight slightly downward, partly cancelling curvature. Under a commonly adopted standard atmosphere, the combined correction is approximately:

where d is in kilometres. The numerical coefficients are empirical approximations and may vary with atmospheric conditions and the adopted convention. For ordinary short levelling with balanced sights, the effects largely cancel and are usually neglected. For long, unequal sights, they should be considered explicitly or controlled by observation design.

The safest practical strategy is to keep backsight and foresight distances equal. Then the curvature and refraction effects on the two readings are nearly equal and cancel when their difference is used to determine elevation difference.

7. Reciprocal Levelling

Reciprocal levelling is used when two points are separated by an obstacle, such as a river, valley or busy highway, and equal sight lengths cannot be obtained from a single setup. Observations are taken with the instrument near A and then near B.

Let:

• a = reading on A when the instrument is near A;
• b = reading on B when the instrument is near A;
• c = reading on A when the instrument is near B;
• d = reading on B when the instrument is near B.

The true difference in elevation of A relative to B is:

The apparent error in the first setup is isolated by comparison of the two observed differences. The collimation, curvature and refraction effects are substantially removed because they reverse in relation to the two points when the instrument is moved.

The difference between the observed elevation differences is also a diagnostic. If the first setup gives b−a and the second gives d−c, a large disagreement indicates collimation error, unstable staff stations, reading mistakes or rapidly changing atmospheric refraction.

8. Closure and Error Assessment

A levelling run should close on a benchmark of known RL, either directly or through a loop. The closure error is:

If the run begins and ends on the same benchmark, the theoretical difference is zero and e = RL_observed − RL_start. Closure error must be distinguished from arithmetic error. The BS–FS check may pass even when the entire survey has a systematic error, because it checks internal arithmetic rather than independent accuracy.

When a run is accepted but has non-zero closure error, an adjustment may be distributed according to the project specification. A simple length-weighted correction for a point at distance l along a total line length L is:

This assumes error accumulates approximately with distance. Equal corrections per setup may be used where the specification assumes setup-based accumulation. Corrections should never be applied to conceal a failed tolerance test; a failed run should normally be investigated and reobserved.

9. Illustrative Case: Drainage-Line Control Survey

Background

A construction team is preparing a 600 m stormwater drain. The site benchmark has RL 52.840 m, but the work area contains a low-lying road crossing that prevents a direct line of sight.

Problem

The team must transfer the benchmark RL to the far end, establish temporary benchmarks, and obtain a centreline profile. A direct one-way run across the road could be affected by unequal sight lengths and unstable ground.

Analysis and approach

The surveyor first carries out a fly levelling run using stable change points. At the road crossing, reciprocal levelling is used: one setup is made on each side, and readings are taken to both staffs. This reduces collimation and earth-curvature effects. The far temporary benchmark is then used for profile levelling along the proposed drain. Cross-sections are taken at 20 m chainages and at every visible break in transverse slope.

The field book distinguishes BS, IS and FS readings and records chainage, offset, point description and staff position. A height-of-instrument reduction is used for rapid checking in the field, while a rise-and-fall reduction is completed independently in the office. The run closes on a second known benchmark.

Outcome

The independent reduction agrees, and the observed closure is within the project tolerance. The profile identifies a section where the proposed drain grade would be too flat to maintain the required hydraulic fall. The design team lowers the downstream invert slightly and adds a maintenance access point. Cross-sections also reveal that the initial earthwork estimate omitted a side drain on the left bank.

Learning points

The case demonstrates that levelling is not merely the recording of staff readings. Instrument setup, sight balance, benchmark stability, method selection, independent checks and communication with designers all affect construction quality. A reliable vertical control network can prevent expensive rework and drainage failure.

10. Errors, Limitations and Best Practice

Important errors include collimation error, staff not held vertical, parallax, incorrect booking, wrong change-point identification, settlement of the tripod or staff, temperature effects, vibration, wind, heat shimmer and refraction. Blunders are large mistakes such as transposing digits or reading the wrong staff; they must be detected through independent checks rather than statistically averaged away.

Best practice includes:

• use firm, clearly marked benchmarks and change points;
• balance backsight and foresight distances;
• keep sight lengths appropriate to visibility and required precision;
• remove parallax and check the compensator;
• hold the staff vertical and read at eye level where practical;
• record observations immediately and describe every point;
• perform both arithmetic and independent closure checks;
• repeat questionable observations rather than silently adjusting them;
• protect benchmarks and document their locations;
• use digital records only with verification, since automation does not eliminate setup or identification errors.

Modern digital levels can store observations, calculate RLs and export data to surveying software. Robotic total stations and GNSS receivers may support vertical control, but GNSS heights depend on geoid models and are not automatically interchangeable with conventional level elevations. Artificial intelligence may assist with anomaly detection and field-data validation, but responsibility for instrument checks, datum selection, acceptance tolerances and final engineering decisions remains with the competent surveyor.

Chapter Summary

Levelling determines elevations relative to a datum and is essential for construction control, drainage, earthworks and infrastructure alignment. A level establishes a horizontal line of sight, while staff readings provide the vertical differences between points. Differential levelling transfers RLs through change points; profile levelling establishes longitudinal ground form; cross-section levelling describes transverse form; fly levelling rapidly transfers control; reciprocal levelling reduces errors across obstacles; and precise levelling provides high-accuracy vertical control.

The height-of-instrument method calculates the line-of-sight elevation and subtracts each staff reading. The rise-and-fall method compares consecutive readings and provides a stronger local arithmetic check. The fundamental check is ΣBS−ΣFS = RL_last−RL_first. Curvature and refraction become significant on long sights, but balanced sights largely cancel their effects. Closure must be checked against an independent known RL and a project-specific tolerance. Accurate levelling depends equally on theory, stable field procedures, careful booking, independent verification and professional judgement.

Key Terms

Datum: Reference surface or elevation system from which RLs are measured.

Reduced level (RL): Elevation of a point relative to a datum.

Benchmark: Fixed point with a known or established RL.

Temporary benchmark: Stable project point established for short-term control.

Line of collimation: Line through the cross-hairs and optical centre of the instrument.

Backsight (BS): First staff reading after an instrument setup.

Foresight (FS): Final reading before the instrument is moved.

Intermediate sight (IS): Reading taken without moving the instrument.

Change point: Stable point used for transferring elevation between setups.

Height of instrument: RL of the horizontal line of sight.

Differential levelling: Transfer of elevation using successive instrument setups.

Profile levelling: Levelling along an alignment to obtain a longitudinal section.

Cross-section levelling: Levelling transversely across an alignment.

Fly levelling: Rapid transfer of RL using mainly BS and FS observations.

Reciprocal levelling: Levelling from both sides of an obstacle to reduce systematic error.

Precise levelling: High-accuracy levelling performed under controlled procedures.

Rise: Increase in elevation between consecutive points.

Fall: Decrease in elevation between consecutive points.

Closure error: Difference between observed and known closing elevations.

Parallax: Apparent movement of the staff image relative to the cross-hairs.

Review Questions

Short Answer Questions

1. Define reduced level, benchmark and change point.
2. Distinguish between a backsight, intermediate sight and foresight.
3. State the principal arithmetic check for the height-of-instrument method.
4. Why are backsight and foresight distances kept approximately equal?
5. What is the purpose of reciprocal levelling?

Long Answer Questions

1. Explain the complete field procedure for differential levelling, including the use of change points.
2. Describe profile and cross-section levelling and explain their applications in road construction.
3. Explain the height-of-instrument and rise-and-fall methods, including their checks.
4. Discuss curvature and refraction corrections and explain how balanced sights control them.
5. Describe the requirements and field controls for precise levelling.

Analytical / Application Questions

1. A benchmark has RL 210.500 m. Readings are BS 1.325 m, IS 2.010 m, FS 2.455 m, BS 0.875 m, IS 1.640 m and FS 1.920 m. Reduce the levels by the height-of-instrument method and perform the BS–FS check.
2. A level line crosses a river where equal sight lengths cannot be obtained. Explain how reciprocal levelling should be organised and identify the observations needed to determine the true difference in elevation.
3. A profile run closes 4 mm high over a 1.6 km line, while the project tolerance is 5√K mm. Decide whether it passes and explain how any accepted closure error could be distributed along the line.

Multiple Choice Questions with Answers

Question 1

A benchmark has RL 100.000 m and the backsight is 1.250 m. What is the height of instrument?

A. 98.750 m

B. 100.000 m

C. 101.250 m

D. 102.500 m

Correct Answer: C. 101.250 m

Question 2

If the next staff reading is smaller than the previous reading, the ground has experienced a:

A. Fall

B. Rise

C. Collimation error

D. Closure error

Correct Answer: B. Rise

Question 3

Which reading is normally taken immediately before moving the instrument?

A. Backsight

B. Intermediate sight

C. Foresight

D. Check sight

Correct Answer: C. Foresight

Question 4

The principal purpose of a change point is to:

A. Measure horizontal angles

B. Transfer elevation between instrument setups

C. Eliminate the need for a benchmark

D. Determine magnetic bearing

Correct Answer: B. Transfer elevation between instrument setups

Question 5

For a staff reading distance d in kilometres, the approximate curvature correction magnitude is:

A. 0.00785d m

B. 0.0785d² m

C. 7.85d² m

D. 0.785/d² m

Correct Answer: B. 0.0785d² m

Question 6

Reciprocal levelling is especially useful when:

A. The ground is perfectly level

B. An obstacle prevents balanced direct sights

C. Only one staff is available

D. Horizontal angles are required

Correct Answer: B. An obstacle prevents balanced direct sights

Question 7

Which equation is the standard levelling arithmetic check?

A. ΣBS + ΣFS = RLfirst + RLlast

B. ΣBS − ΣFS = RLlast − RLfirst

C. ΣBS × ΣFS = RLlast

D. ΣRise + ΣFall = RLfirst

Correct Answer: B. ΣBS − ΣFS = RLlast − RLfirst

Question 8

A point has HI 75.430 m and staff reading 2.115 m. Its RL is:

A. 73.315 m

B. 77.545 m

C. 75.430 m

D. 2.115 m

Correct Answer: A. 73.315 m

Question 9

Which practice most effectively reduces curvature and refraction effects in ordinary differential levelling?

A. Using unequal long sights

B. Balancing backsight and foresight distances

C. Taking only intermediate sights

D. Holding the staff at an angle

Correct Answer: B. Balancing backsight and foresight distances

Question 10

If a 1.6 km line has tolerance 5√K mm, its allowable magnitude of closure error is approximately:

A. 2.0 mm

B. 4.0 mm

C. 6.3 mm

D. 8.0 mm

Correct Answer: C. 6.3 mm