Measurement, Accuracy and Survey Errors

Introduction

Surveying is founded on measurement. Every plan, map, setting-out operation and level book entry depends on observations of distance, direction and elevation. Measurements are never perfectly exact: instruments have limitations, observers make mistakes, and atmospheric or ground conditions influence results. Professional surveying therefore requires more than reading an instrument. It requires selecting suitable units, recording observations correctly, assessing their quality, detecting errors and deciding whether the results satisfy the required tolerance.

This chapter introduces the principles needed by a civil draftsman to understand and use survey measurements responsibly. The emphasis is on linear and angular measurement, significant figures, accuracy and precision, classification of errors, field checks, tolerances and elementary error propagation.

Learning Objectives

After studying this chapter, you should be able to:

1. Select suitable units for distances, angles and elevations.
2. Record measurements using appropriate significant figures and decimal places.
3. Distinguish accuracy, precision, resolution and sensitivity.
4. Explain systematic, random and gross errors.
5. Identify common sources of errors in chaining, taping, levelling and angle measurement.
6. Apply checks, corrections and tolerances to field observations.
7. Use basic error-propagation relationships for sums, differences and products.
8. Judge whether survey data are suitable for drafting and construction use.

1. Measurement in Surveying

A measurement is an observed comparison between an unknown quantity and an agreed standard. In surveying, the principal quantities are:

• Distance: separation between two points, usually measured horizontally or along a slope.
• Angle: relationship between two lines or directions.
• Elevation: vertical distance of a point relative to a datum.
• Area and volume: quantities calculated from measured distances, angles and elevations.

A measured value should always be accompanied by its unit and, where appropriate, an indication of its quality. For example, 42.36 m is more informative than 42.36 alone. A survey record should also identify the instrument, date, observer, method and relevant environmental conditions.

1.1 Units of distance

The SI metre (m) is the standard unit for engineering surveying. Common conversions are:

• 1 km = 1,000 m
• 1 m = 1,000 mm
• 1 cm = 0.01 m
• 1 mm = 0.001 m

Distances on construction drawings are often expressed in metres, while small details and setting-out offsets may be recorded in millimetres. Unit conversion must be completed before arithmetic is performed. Adding 2.4 m to 350 mm requires conversion, for example: 2.4 m + 0.350 m = 2.750 m.

Distance may be horizontal, vertical, or inclined. If an inclined distance S makes an angle θ with the horizontal, the horizontal distance H is:

Here, S and H have the same distance unit and θ is an angle. The relationship assumes a straight line and a correctly defined angle measured from the horizontal. If the vertical difference is required:

where V is the vertical component. In practical surveying, instrument and target heights may also need to be included when calculating the difference in elevation between ground points.

1.2 Units of angle

Angles are commonly expressed in degrees, minutes and seconds or in decimal degrees. One degree is divided as follows:

• 1° = 60′
• 1′ = 60″
• 1° = 3,600″

The conversion from degrees, minutes and seconds to decimal degrees is:

For example, 25° 18′ 30″ equals 25 + 18/60 + 30/3,600 = 25.308333°. Decimal degrees should not be confused with degrees and decimal minutes. Thus, 25.30° is not 25° 30′; it is 25° 18′ approximately.

The radian is the coherent SI unit for mathematical work. A complete revolution is 360° or 2π radians. The conversion is:

Survey instruments may display horizontal and vertical angles in degrees, gon or mils. The unit selected must be clearly stated in field notes and software settings.

1.3 Significant figures and decimal places

A significant figure is a digit that conveys measurement information. The number of digits recorded should reflect the capability of the instrument and the quality of the observation. Recording excessive digits creates false confidence; recording too few loses useful information.

For example, a tape measurement made to the nearest centimetre may reasonably be recorded as 18.42 m, not 18.42000 m. A total station capable of resolving millimetres may justify a different format, but the displayed precision does not automatically guarantee equivalent accuracy.

Rules for practical reporting include:

• Retain guard digits during calculations and round only the final result.
• Do not report more decimal places than the observation supports.
• Match the precision of calculated results to the least precise important input.
• Record field observations, not only rounded final values.
• State whether a value is an observation, a correction or a derived quantity.

For addition and subtraction, decimal places are important. For multiplication and division, significant figures are generally important. Survey office standards may prescribe particular rounding conventions, which take priority over informal rules.

2. Accuracy, Precision and Related Concepts

Accuracy is the closeness of a measured or computed value to the accepted or true value. Precision is the closeness of repeated measurements to one another. A set of observations can be precise but inaccurate if a consistent bias affects every observation. Conversely, observations may be broadly accurate on average but poorly precise if they scatter widely.

Resolution is the smallest increment that an instrument display or scale can show. Sensitivity describes the response of an instrument to a change in the measured quantity. Resolution, precision and accuracy are related but not identical.

A useful conceptual example is a target diagram. Repeated points tightly grouped around the wrong location are precise but inaccurate. Widely scattered points around the true location have limited precision but may have acceptable mean accuracy. Survey work requires both adequate accuracy and adequate precision.

An observation may be represented as:

where e is the total measurement error. Error is not the same as mistake. An error is a deviation from the accepted value; a mistake, often called a blunder or gross error, is usually an avoidable human failure such as reading the wrong staff graduation or entering 12.60 instead of 12.06.

3. Classification of Survey Errors

3.1 Systematic errors

Systematic errors follow a consistent pattern. Their magnitude and sign may remain constant or vary according to a known relationship. Because they are repeatable, they can often be detected, estimated and corrected.

Examples include:

• a tape that is longer or shorter than its nominal length;
• incorrect prism or constant settings in an electronic distance measurement system;
• collimation error in a level;
• curvature and atmospheric refraction in long levelling sights;
• temperature, pull and sag effects in tape measurement;
• an incorrectly oriented instrument or datum error.

A general corrected observation can be written as:

where c is the correction, defined with a sign that removes the estimated error. The correction must be based on a valid calibration, formula or established procedure. Applying an arbitrary correction is not good surveying practice.

For a tape of nominal length L₀ and actual standardized length L, the correction for a measured distance D due solely to tape length is:

where C_L is the correction to the measured distance. This formula assumes the tape is used under the same conditions as the standardization and that other effects are excluded. If the actual tape is longer than nominal, the measured distance is too small and the correction is positive.

3.2 Random errors

Random errors vary in magnitude and sign in an unpredictable way. They arise from small changes in pointing, reading, centring, vibration, atmospheric conditions and other influences. Repetition does not remove random error from an individual observation, but repeated independent observations can reduce the uncertainty of their mean.

For observations x₁, x₂, …, x_n, the arithmetic mean is:

where n is the number of observations. The mean is useful when observations measure the same quantity under comparable conditions and no gross errors remain. It does not correct a systematic bias.

The sample standard deviation is:

where s estimates the scatter of the observations and n − 1 is used when estimating population variability from a sample. This elementary statistical measure is meaningful only when the observations are comparable and the data are not contaminated by blunders.

3.3 Gross errors

Gross errors are large departures caused by mistakes, equipment misuse or abnormal events. Typical examples are:

• recording a backsight as a foresight;
• reading 1.865 m as 1.685 m;
• occupying the wrong point;
• entering a bearing in the wrong quadrant;
• forgetting to enter a prism height;
• using degrees when software expects radians.

Gross errors are not normally handled by averaging. They should be detected through independent checks, comparison with field sketches, closure tests, repeated observations and review of raw data. If a value cannot be justified, it should be rejected or reobserved rather than quietly adjusted.

4. Sources of Error in Common Survey Operations

4.1 Distance measurement with a tape

Important influences include incorrect tape length, temperature, pull, sag, slope, poor alignment and incorrect marking. A tape measurement made on a slope is not the horizontal distance unless a slope correction or reduction is applied. If the slope angle is known, horizontal reduction may use H = S cos θ. For a small vertical difference h over a measured slope distance S, the horizontal distance can also be expressed as:

This relationship assumes S is the straight slope distance and h is the vertical difference between its endpoints.

Field practice should include alignment by ranging, firm and consistent tension, correct plumbing of tape ends, proper slope breaking and independent check measurements where the work is important.

4.2 Levelling

In differential levelling, the height difference between two points is obtained from staff readings. With a backsight BS and foresight FS:

If the known elevation of the first point is H_A, the elevation of the second point is:

The signs assume a conventional backsight taken on the point of known elevation and a foresight taken on the point whose elevation is required. Equal sight lengths help cancel collimation, curvature and refraction effects. The staff must be vertical, the level bubble or compensator must function correctly, and turning points must be stable.

For a closed levelling loop, the arithmetic closure error is:

For a loop returning to its starting point, H_end − H_start = 0, so e = ΣBS − ΣFS. The magnitude of permissible misclosure depends on the survey specification, route length, terrain and instrument procedure. A correction should not be distributed until the cause and acceptability of the misclosure have been considered.

4.3 Angle measurement

Angular errors may result from imperfect centring, poor pointing, eccentricity, instrument adjustment, unstable tripods, parallax or incorrect reading. Repeating an angle in both faces of a theodolite or total station can reduce or reveal some instrumental errors. Measuring angles in a closed traverse provides a geometric check.

For an n-sided plane traverse, the theoretical sum of interior angles is:

The angular misclosure is:

The formula assumes a simple plane polygon with correctly identified interior angles. If all angles are considered equally reliable and the misclosure is within the specified tolerance, an elementary adjustment may allocate equal corrections:

More advanced adjustments use observation weights and least-squares methods; equal distribution is not automatically appropriate when observations have different quality.

5. Checks, Redundancy and Tolerances

A check is an independent or semi-independent test of a result. Redundancy means collecting more observations than are strictly necessary so that inconsistencies can be detected. Examples include:

• measuring a line twice;
• closing a level run on a known benchmark;
• observing an angle on both faces;
• closing a traverse;
• checking a computed coordinate against a known control point;
• comparing total-station distances with tape or plan dimensions where appropriate.

A tolerance is the maximum permitted discrepancy under a stated specification. It is not the same as an error estimate. Acceptance requires knowing the purpose of the survey, instrument class, method, environmental conditions and applicable standard or contract requirement.

A sound tolerance procedure is:

1. Identify the quantity and required purpose.
2. Determine the governing specification before fieldwork.
3. Observe and record raw data without concealment.
4. Compute the relevant closure or discrepancy.
5. Compare it with the permitted tolerance.
6. Investigate failures before adjustment.
7. Reobserve or report limitations when necessary.

A result within tolerance is not necessarily perfect; it is simply fit for the stated purpose. A result outside tolerance should not be forced into compliance by arbitrary correction.

6. Basic Error Propagation

Error propagation estimates how uncertainty in measured quantities affects a calculated result. It is useful for planning observations and judging the reliability of derived dimensions.

For independent small errors in a sum or difference, the approximate standard uncertainty is:

where u_i are standard uncertainties of the input quantities and u_y is the standard uncertainty of y = x₁ ± x₂ ± … ± x_n. The signs do not change the variance contribution because independent random uncertainties are squared.

For a general function y = f(x₁, x₂, …, x_n), first-order propagation is:

This assumes small uncertainties, a sufficiently smooth function and independent input quantities. If inputs are correlated, covariance terms must be included.

For multiplication by a constant, y = kx, the uncertainty is:

For a product or quotient, approximate relative uncertainty for independent small uncertainties is:

for y = xz or y = x/z, provided the variables are nonzero and uncertainties are small.

Worked Example 1: Mean distance and spread

Five repeated distances, in metres, are 24.62, 24.65, 24.61, 24.64 and 24.63.

The mean is:

x̄ = (24.62 + 24.65 + 24.61 + 24.64 + 24.63)/5 = 24.63 m

Deviations from the mean are −0.01, +0.02, −0.02, +0.01 and 0.00 m. The sum of squared deviations is 0.0010 m². Therefore:

s = √(0.0010/4) = 0.0158 m ≈ 0.016 m

The observations have a mean of 24.63 m and a sample standard deviation of approximately 0.016 m. This describes repeatability; it does not prove that 24.63 m is free from systematic tape or instrument error.

Worked Example 2: Propagation in a level difference

Suppose a height difference is computed from ΔH = BS − FS. If the standard uncertainties of the backsight and foresight are each 0.003 m and are independent:

u_ΔH = √(0.003² + 0.003²) = 0.00424 m

Thus the estimated standard uncertainty of the height difference is approximately 0.004 m. The result should not be reported with meaningless millimetres beyond the quality supported by the observations.

7. Applied Case: Setting Out a Small Building Foundation

Background

This illustrative case concerns a small rectangular building foundation to be set out from two established control points. The drawing requires grid intersections to be positioned within ±10 mm, while levels must be transferred within ±5 mm relative to the site benchmark.

Problem

The initial survey team established one corner using a total station but obtained a 22 mm discrepancy when checking the diagonal and a 9 mm level misclosure over a short run. The construction crew wanted to proceed because the visible grid lines appeared close to the design.

Analysis

The diagonal discrepancy exceeded the positional tolerance. Possible causes included wrong prism height, occupation of the wrong control point, incorrect coordinate entry, poor centring or a blunder in recording. The level misclosure also exceeded the stated ±5 mm requirement. Treating both discrepancies as random and distributing them without investigation would risk transferring an incorrect foundation position and elevation.

Approach and solution

The surveyor first compared the field sketch, raw observations, point identifiers and instrument settings. The prism height was found to have been entered as 1.800 m rather than 1.680 m. The horizontal coordinates were recomputed, and the corner was reobserved from the second control point. The diagonal then differed by 4 mm, which satisfied the positional tolerance.

For the levels, the team checked the staff readings and discovered that one foresight had been copied incorrectly. A repeat run with approximately balanced sight lengths produced a 3 mm misclosure. The level was then accepted, and the corrected elevations were clearly marked in the field book and drawing register.

Outcome

The foundation grid was set out only after both independent checks met tolerance. The original observations were retained, marked as superseded, and not deleted. The final drawing included the control-point identifiers, date, instrument reference and level datum.

Learning points

1. A check is valuable only when it is independent enough to reveal a mistake.
2. Misclosure should trigger investigation before adjustment.
3. Instrument settings and point identifiers are part of measurement quality.
4. Raw observations must remain traceable for audit and future re-establishment.
5. A drawing should communicate the quality and reference framework of its survey data.

8. Professional Practice, Technology and Limitations

Digital levels, total stations, GNSS receivers, laser scanners and surveying software improve speed and data management, but they do not eliminate errors. Automated measurements can still contain incorrect coordinate systems, wrong prism constants, multipath effects, poor satellite geometry, unstable benchmarks or erroneous operator settings. Software may calculate a precise-looking answer from unsuitable input.

Good practice includes instrument calibration, field checks, regular battery and tripod inspection, independent control, metadata recording and secure file management. Surveyors should understand the datum, projection, units, geoid model and coordinate reference system used by a project. A civil draftsman must avoid importing data into a drawing at the wrong scale or coordinate system.

Ethically, survey information must not be altered to satisfy a deadline or construction preference. Uncertainty, failed checks and assumptions should be communicated to the responsible surveyor or engineer. Where legal boundaries, safety-critical work or contractual tolerances are involved, only appropriately authorized and competent personnel should certify the result.

Artificial intelligence and automated feature extraction may assist with point-cloud classification, anomaly detection and quality-control alerts. They do not replace field verification, professional judgment or responsibility for the final survey record. Any automated result should be traceable to source data and checked against known control.

Best-Practice Checklist

Before accepting survey measurements, confirm that:

• units and angular conventions are stated;
• instrument constants and heights are recorded;
• the instrument has been checked or calibrated as required;
• observations are recorded with suitable precision;
• independent checks and closures have been performed;
• gross errors have been investigated rather than averaged away;
• corrections have a documented basis and sign;
• tolerances come from the project specification;
• coordinates, datum and projection are correct;
• final drawings retain traceability to field observations.

Chapter Summary

Surveying measurements are meaningful only when their units, method and quality are understood. Distances are commonly expressed in metres and angles in degrees, minutes, seconds or decimal degrees. Significant figures should reflect the capability of the method and should not create false precision.

Accuracy describes closeness to an accepted value, whereas precision describes repeatability. Survey errors are commonly classified as systematic, random or gross. Systematic errors may be modelled and corrected; random errors are assessed statistically and reduced through appropriate repetition; gross errors require detection, investigation and usually reobservation.

Distance, angle and levelling observations should include independent checks and be compared with specified tolerances. Level and traverse closure equations provide important geometric tests. Basic error propagation shows how uncertainty in component observations affects a derived result. Modern instruments and software improve efficiency but do not remove the need for calibration, control, professional judgment and ethical reporting.

Key Terms

Accuracy: Closeness of a measurement to an accepted or true value.

Angle: The relationship between two intersecting lines or directions.

Calibration: Comparison or adjustment of an instrument against a known standard.

Closure: Agreement between a computed survey result and a known geometric or control condition.

Correction: A quantity applied to an observation to compensate for an estimated systematic error.

Datum: A reference surface or framework from which elevations or coordinates are measured.

Error: Difference between an observed value and the accepted value.

Gross error: A large, usually avoidable error caused by a mistake or abnormal event.

Levelling: Determination of differences in elevation between points.

Misclosure: The discrepancy remaining when a survey is checked against a known condition.

Precision: Closeness of repeated measurements to one another.

Random error: An error whose magnitude and sign vary unpredictably.

Resolution: The smallest increment an instrument can display or distinguish.

Significant figure: A digit that conveys meaningful measurement information.

Systematic error: A repeatable error with a consistent or modelled pattern.

Tolerance: Maximum permitted discrepancy for a specified purpose.

Uncertainty: Quantified doubt associated with a measurement or result.

Redundancy: Additional observations collected to support checking and reliability.

Review Questions

Short Answer Questions

1. Distinguish accuracy from precision in surveying.
2. Convert 18° 24′ 36″ into decimal degrees.
3. State two examples of systematic survey errors.
4. Why should gross errors not normally be removed by averaging?
5. What is the purpose of a closure check?

Long Answer Questions

1. Explain the classification of survey errors and describe practical methods for detecting or reducing each class.
2. Discuss the importance of significant figures, units and field-book conventions in professional surveying.
3. Derive and explain the relationship between backsight, foresight and change in elevation.
4. Explain how checks and tolerances should be used when accepting a traverse or level run.
5. Describe the limitations of digital surveying instruments and software in relation to measurement quality.

Analytical / Application Questions

1. A taped line is measured as 86.420 m using a tape whose actual length is 30.006 m when its nominal length is 30.000 m. Calculate the tape-length correction, state its sign, and explain the assumption used.
2. Three independent observations of a distance are 52.18 m, 52.22 m and 52.20 m. Calculate the mean and comment on whether the data alone establish accuracy.
3. A four-sided traverse has observed interior angles of 89°58′, 90°04′, 90°01′ and 89°55′. Calculate the angular misclosure and state the equal correction per angle if the misclosure is within the project tolerance.

Multiple Choice Questions with Answers

Question 1

Which statement best distinguishes accuracy from precision?

A. Accuracy is repeatability, while precision is closeness to the true value.

B. Accuracy is closeness to an accepted value, while precision is repeatability.

C. Accuracy is the instrument resolution, while precision is the unit used.

D. Accuracy and precision always mean exactly the same thing.

Correct Answer: B. Accuracy is closeness to an accepted value, while precision is repeatability.

Question 2

What is 25° 18′ 30″ expressed in decimal degrees?

A. 25.1833°

B. 25.3000°

C. 25.3083°

D. 25.5030°

Correct Answer: C. 25.3083°

Question 3

A tape has an actual length greater than its nominal length. For a measured distance, the tape-length correction is generally:

A. Positive

B. Negative

C. Always zero

D. Equal to the tape temperature

Correct Answer: A. Positive

Question 4

Which is a gross error?

A. Small random pointing variation

B. Atmospheric refraction

C. A tape length calibration bias

D. Recording a backsight as a foresight

Correct Answer: D. Recording a backsight as a foresight

Question 5

If BS = 1.845 m and FS = 2.110 m, the change in elevation is:

A. +0.265 m

B. −0.265 m

C. +3.955 m

D. −3.955 m

Correct Answer: B. −0.265 m

Question 6

For a closed four-sided plane traverse, the theoretical sum of interior angles is:

A. 180°

B. 360°

C. 540°

D. 720°

Correct Answer: B. 360°

Question 7

Two independent observations each have standard uncertainty 0.003 m. What is the approximate uncertainty of their difference?

A. 0.0015 m

B. 0.0030 m

C. 0.0042 m

D. 0.0060 m

Correct Answer: C. 0.0042 m

Question 8

Which action is most appropriate when a survey closure exceeds the specified tolerance?

A. Delete the largest observation immediately

B. Distribute the discrepancy without investigation

C. Round the result until it passes

D. Investigate the observations and reobserve if necessary

Correct Answer: D. Investigate the observations and reobserve if necessary

Question 9

For an inclined distance S and angle θ measured from the horizontal, the horizontal component is:

A. S sin θ

B. S cos θ

C. S tan θ

D. S/θ

Correct Answer: B. S cos θ

Question 10

Which statement about averaging repeated observations is correct?

A. It removes all systematic errors.

B. It can reduce the influence of random errors when observations are comparable.

C. It corrects wrong point identification automatically.

D. It makes gross errors acceptable.

Correct Answer: B. It can reduce the influence of random errors when observations are comparable.