GELE Surveying (Geomatics) — Measurements and Theory of ErrorsDetailed Explanation
A detailed, step-by-step explanation of Measurements and Theory of Errors for GELE aspirants. This page goes deeper than the summary and study notes, walking through the reasoning behind each concept so you understand why Professional Regulation Commission (PRC) — Board of Geodetic Engineering tests it the way it does in the GELE Surveying (Geomatics) subtest.
Exam context
For the Geodetic Engineer Licensure Examination, Professional Regulation Commission (PRC) — Board of Geodetic Engineering tests Surveying (Geomatics) under a "Core" label, with Measurements and Theory of Errors in the 1st slot across 9 chapters. GELE candidates must clear the 70% weighted average, no sub-test below 50% cut on the 2026 paper, which draws about a meaningful share of Surveying (Geomatics) questions. Date to watch: September 2026.
Measurements and Theory of Errors - Detailed Explanation
Every measurement made in the field carries inherent uncertainty. Whether you are taping a baseline for a subdivision survey, running a level circuit across a dam site, or traversing a property boundary, your instruments, your field conditions, and even your own hands introduce small imperfections into every reading. Surveying is not the elimination of error — it is the disciplined management and quantification of it. This chapter builds the statistical and corrective framework that underpins all subsequent surveying computations: the classification of errors, the determination of the most probable value (MPV) of repeated observations, the systematic tape corrections (temperature, tension, sag, slope), and the propagation of random errors through sums of measurements. These concepts appear directly in the PRC Civil Engineer Licensure Examination under the Surveying (Geomatics) component and form the prerequisite reasoning for leveling, traversing, area computation, and curve layout covered in later chapters.
Concepts
Classification of Errors in Surveying Measurements
Survey measurements deviate from the true value due to three distinct categories of imperfection. Understanding the category determines how you handle the deviation. **1. Mistakes (Blunders)** are gross errors caused by carelessness, fatigue, or miscommunication — misreading a staff, booking a wrong digit, or losing count of tape lengths. They do not follow any statistical pattern and must be detected and eliminated before any analysis. Detection methods include repeat observations, independent checks, and closure tests. **2. Systematic Errors** follow a definite physical law and act consistently in one direction. A steel tape that is 0.02 m too short will undercount every single measurement by the same proportional amount. Temperature expansion, instrument maladjustment, and atmospheric refraction in leveling are classic systematic errors. Because they are predictable, they are **correctable** through calibration corrections applied to the raw measurement. Systematic errors do NOT average out with more repetitions — repeating a measurement with a faulty instrument just repeats the same bias. **3. Random (Accidental) Errors** are small, unpredictable residual errors remaining after mistakes and systematic errors are removed. They arise from the limitations of human senses and instrument precision — the eye cannot bisect a cross-hair perfectly every time; the wind causes a slight plumb-bob swing. Random errors follow the **Normal (Gaussian) distribution**: they are equally likely to be positive or negative, small errors are more frequent than large ones, and they tend to cancel over many repetitions. It is this third category that is treated with the probability theory in this chapter.
Examples
A single reading that is grossly inconsistent with all others signals a mistake, not a random error. The consistent cluster of ~84.35–84.37 m represents the true precision of the measurement after the blunder is removed.
Scenario
A surveying crew measures a traverse leg six times and gets values of 84.35, 84.36, 84.35, 84.37, 84.35, and 94.36 m. Classify the last reading.
Solution
The value 94.36 m is clearly a blunder (mistake) — it is approximately 10 m larger than all other consistent readings. It must be rejected and remeasured before computing the most probable value.
Systematic errors are directional and predictable. This is why calibration and environmental corrections are applied in professional surveying practice.
Scenario
A steel tape is used on a hot day when the temperature is 15°C above the standardization temperature. Each measurement reads shorter than the true distance. What type of error is this?
Solution
This is a systematic error. The tape expands due to heat, making it physically longer than its nominal length. Each full tape length laid down covers MORE ground than the tape reads — so the measured (recorded) distance is LESS than the true distance. The error always acts in the same direction and can be corrected using the temperature correction formula.
Applications
- Quality control in cadastral surveying under RA 8560 (Philippine Cadastral Survey).
- Loop closure checks in traverse surveys — large misclosures signal blunders or uncorrected systematic errors.
- Geodetic control surveys where systematic errors from atmospheric refraction and Earth curvature are explicitly corrected.
- Construction layout where a miscalibrated total station introduces systematic errors in all set-out distances.
Misconceptions
- WRONG: 'Taking more measurements eliminates systematic error.' CORRECT: More measurements only reduce random error; systematic errors must be corrected by formula or instrument calibration.
- WRONG: 'Any reading different from the mean is a blunder.' CORRECT: Small deviations from the mean are expected random errors; only large, implausible deviations are blunders.
- WRONG: 'Random errors are caused by careless observers.' CORRECT: Random errors are unavoidable even for the most careful observer — they arise from the fundamental limits of human senses and instrument precision.
Related Concepts
- Most Probable Value and Probable Error
- Normal Distribution (Statistics)
- Tape Corrections (Systematic Error Corrections)
- Closure and Adjustment of Traverses
Common Exam Questions
Example
Which type of error is caused by an improperly adjusted level bubble? Answer: Systematic error — instrument maladjustment consistently affects all readings in the same way.
Approach
Read the scenario carefully and identify (a) is it a large isolated deviation? → blunder; (b) does it consistently shift all readings in one direction due to a known physical cause? → systematic; (c) is it small and random after corrections? → random.
Question Type
Identification / Classification
Example
Statement: 'Increasing the number of measurements reduces the effect of a systematic error.' Answer: FALSE — systematic errors are directional; they accumulate, not cancel.
Approach
Recall that random errors follow a normal distribution; systematic errors do NOT average out; mistakes must be discarded.
Question Type
True/False or Multiple Choice on Properties
Key Points To Remember
- Three categories: Mistakes (blunders), Systematic errors, Random (accidental) errors.
- Mistakes must be ELIMINATED — they cannot be corrected statistically.
- Systematic errors are CORRECTABLE because they follow a physical law.
- Random errors are treated STATISTICALLY using the theory of probability.
- Systematic errors do NOT cancel with repetition; random errors tend to cancel.
- In PRC board problems, 'probable error' always refers to the random error framework.
Most Probable Value, Residuals, and Probable Error
When a quantity is measured multiple times under the same conditions, the **Most Probable Value (MPV)** is the value most likely to be closest to the true value. By the principle of least squares, the MPV of equally-weighted observations is simply the **arithmetic mean**: $$\bar{x} = \frac{\sum x_i}{n}$$ where $x_i$ are the individual measurements and $n$ is the number of observations. **Residuals** (also called corrections or deviations) are the differences between each individual observation and the MPV: $$v_i = x_i - \bar{x}$$ Residuals are NOT errors (we never know the true error), but they are our best estimate of the random variation. Their sum is always zero: $\sum v_i = 0$ — use this as an arithmetic check. **Probable Error of a Single Observation** ($E_s$): This is defined as the value such that there is a 50% probability that any single observation differs from the MPV by less than $E_s$. For the normal distribution: $$E_s = 0.6745\sqrt{\frac{\sum v_i^2}{n-1}}$$ The term $\sqrt{\dfrac{\sum v_i^2}{n-1}}$ is the **standard deviation** ($s$) of the sample, so $E_s = 0.6745\,s$. **Probable Error of the Mean** ($E_m$): Because the mean is more reliable than any single observation, its probable error is smaller: $$E_m = \frac{E_s}{\sqrt{n}} = 0.6745\sqrt{\frac{\sum v_i^2}{n(n-1)}}$$ This relationship is fundamental: **doubling the number of observations improves the precision of the mean by a factor of $\sqrt{2}$**, not by 2. To halve the probable error of the mean, you need **four times** as many measurements.
Examples
Always tabulate residuals and verify Σv = 0 before computing Σv². The probable error of the mean (±0.008 m) is smaller than that of a single observation (±0.018 m) because the mean averages out the random scatter.
Scenario
A distance is measured five times: 150.32, 150.35, 150.28, 150.34, 150.31 m. Find: (a) MPV, (b) residuals, (c) probable error of a single measurement, and (d) probable error of the mean.
Solution
**Step 1 — MPV (mean):** $$\bar{x} = \frac{150.32 + 150.35 + 150.28 + 150.34 + 150.31}{5} = \frac{751.60}{5} = 150.320 \text{ m}$$ **Step 2 — Residuals (v = x − mean):** | Obs | x (m) | v = x − 150.320 | v² | |---|---|---|---| | 1 | 150.32 | 0.000 | 0.000000 | | 2 | 150.35 | +0.030 | 0.000900 | | 3 | 150.28 | −0.040 | 0.001600 | | 4 | 150.34 | +0.020 | 0.000400 | | 5 | 150.31 | −0.010 | 0.000100 | | **Σ** | | **0.000 ✓** | **0.003000** | **Step 3 — Probable error of single obs:** $$E_s = 0.6745\sqrt{\frac{0.003000}{5-1}} = 0.6745\sqrt{0.000750} = 0.6745 \times 0.02739 = 0.01847 \approx \pm 0.018 \text{ m}$$ **Step 4 — Probable error of mean:** $$E_m = \frac{0.01847}{\sqrt{5}} = \frac{0.01847}{2.2361} = \pm 0.00826 \approx \pm 0.008 \text{ m}$$ **Result:** MPV = 150.320 ± 0.008 m
This is a direct application of E_m = E_s / √n. To reduce the error of the mean by a factor of k, you need k² observations. Boards frequently test this relationship.
Scenario
How many observations are needed to reduce the probable error of the mean to one-third of the probable error of a single observation?
Solution
We need: E_m = E_s / 3 Since E_m = E_s / √n: E_s / √n = E_s / 3 √n = 3 n = 9 observations
Applications
- Expressing the precision of bench mark elevations in geodetic leveling (e.g., 15.234 ± 0.003 m).
- Reporting angular measurements in triangulation surveys.
- Quality control — if a new measurement deviates from the MPV by more than 4× the probable error, it may be a blunder.
- Weighted observations (not equal weight) require weighted mean — an extension covered in adjustment of observations.
Misconceptions
- WRONG: 'The probable error of the mean equals the probable error of a single observation divided by n.' CORRECT: It is divided by √n, not n.
- WRONG: 'Residuals equal the true errors.' CORRECT: Residuals are deviations from the MEAN (MPV), not the true (unknown) value.
- WRONG: 'Σv² uses (n) in the denominator.' CORRECT: Use (n−1) for the probable error of a SINGLE observation (Bessel's correction for a sample). The denominator becomes n(n−1) for E_m.
- WRONG: 'The 0.6745 factor can be ignored.' CORRECT: Without it, you compute standard deviation, not probable error. The PRC exams specifically use the probable error definition.
Related Concepts
- Classification of Errors
- Error Propagation (Sum of Measurements)
- Weighted Observations and Weighted Mean
- Adjustment of Traverses by Least Squares
Common Exam Questions
Example
Five angles measured: 40°15'20", 40°15'24", 40°15'18", 40°15'22", 40°15'21". Find probable error of the mean. [Convert to seconds for arithmetic, then convert back.]
Approach
Step 1: Compute mean. Step 2: Compute residuals v. Step 3: Check Σv = 0. Step 4: Compute Σv². Step 5: Apply E_s = 0.6745√(Σv²/(n−1)). Step 6: Divide by √n for E_m.
Question Type
Compute MPV and probable error from a table of observations
Example
A line is measured 4 times giving E_m = ±0.010 m. How many total observations are needed to achieve E_m = ±0.005 m? Answer: E_m is halved, so n must increase by factor 4. New n = 4 × 4 = 16 observations.
Approach
Set E_m(new) = E_m(old)/k, solve for n(new) = k² × n(old). Or use E_m = 0.6745√(Σv²/(n(n−1))) and solve for n.
Question Type
How many more observations to achieve target precision?
Key Points To Remember
- MPV of equally-weighted observations = arithmetic mean.
- Residual: v = (observation) − (mean); sum of all residuals = 0 (arithmetic check).
- Probable error of single obs: E_s = 0.6745 × sqrt(Σv²/(n−1)).
- Probable error of mean: E_m = E_s / sqrt(n).
- 0.6745 is the z-score for 50% probability in a normal distribution.
- To halve E_m, you need 4× more observations (not 2×).
- The MPV is expressed as: MPV ± E_m (e.g., 100.10 ± 0.004 m).
Tape (Distance) Corrections
A steel surveying tape is calibrated at a standard temperature ($T_s$, typically 20°C), a standard pull ($P_s$, typically 50 N or as specified), and supported throughout its length (supported or free). Any deviation from these standard conditions introduces systematic error that must be corrected. The four principal tape corrections are: --- **1. Temperature Correction ($C_t$)** Steel expands and contracts with temperature. If the field temperature $T$ differs from the standard $T_s$: $$C_t = \alpha L(T - T_s)$$ where $\alpha$ is the coefficient of linear thermal expansion of steel ≈ $11.6 \times 10^{-6}$ /°C (sometimes given as $12 \times 10^{-6}$ /°C), $L$ is the measured length. If $T > T_s$: tape is longer → $C_t$ is positive (measured distance is short, true distance is larger). If $T < T_s$: tape is shorter → $C_t$ is negative. --- **2. Tension (Pull) Correction ($C_p$)** If the applied pull $P$ differs from the standardization pull $P_s$, the tape stretches or contracts elastically: $$C_p = \frac{(P - P_s)L}{AE}$$ where $A$ = cross-sectional area of tape (m²), $E$ = modulus of elasticity of steel ≈ $200 \times 10^9$ Pa (200 GPa). If $P > P_s$: tape is stretched → $C_p$ is positive. --- **3. Sag Correction ($C_{sag}$)** When a tape is supported only at its ends (unsupported span), it sags under its own weight into a catenary. The straight-line distance between supports is always LESS than the tape length along the curve. Sag correction is **always negative**: $$C_{sag} = -\frac{w^2 L^3}{24 P^2}$$ where $w$ = weight of tape per unit length (N/m), $L$ = length of unsupported span (m), $P$ = applied tension (N). Alternatively, using total weight $W$ of the span: $$C_{sag} = -\frac{W^2 L}{24 P^2}$$ Note: $W = wL$, so both formulas are equivalent. --- **4. Slope Correction ($C_h$)** A tape laid along a slope measures the slope distance $L$. The horizontal distance $H$ is: $$H = \sqrt{L^2 - h^2}$$ where $h$ is the height difference between the ends. For gentle slopes (small $h/L$), the approximate formula is convenient: $$C_h = -\frac{h^2}{2L} \quad \Rightarrow \quad H = L + C_h = L - \frac{h^2}{2L}$$ Slope correction is **always negative** (horizontal is always shorter than slope distance). --- **Tape Too Long / Too Short — The Most Common Board Trap** If the tape is physically **longer** than its nominal length: - When **measuring** an unknown distance, each tape length laid down covers MORE ground than the tape reads → the recorded distance is **LESS** than true → the true distance is LARGER. - **True = Measured × (Actual / Nominal)** If the tape is physically **shorter** than its nominal length: - Each tape length covers LESS ground → recorded distance is **MORE** than true → true distance is SMALLER. - **True = Measured × (Actual / Nominal)** — same formula, the ratio < 1 automatically reduces. For **laying out** (staking a known design distance): - **Stakeout distance = Design distance × (Nominal / Actual)** — you lay out less tape length if the tape is too long. --- **Sign Convention Summary:** Always compute each correction, then: True = Measured + (algebraic sum of all corrections).
Examples
T > T_s means C_t > 0, adding to measured length. Always check the sign physically: hot tape is longer, so each 50 m marked on tape actually spans more than 50 m on the ground.
Scenario
**Temperature Correction:** A 50 m steel tape (α = 11.6 × 10⁻⁶/°C, standard 20°C) is used to measure a line of 480 m at a field temperature of 38°C. What is the true length?
Solution
**Step 1 — Correction per tape length:** Not needed; apply over total measured length L = 480 m. **Step 2 — Temperature correction:** $$C_t = \alpha L(T - T_s) = (11.6 \times 10^{-6})(480)(38 - 20)$$ $$C_t = (11.6 \times 10^{-6})(480)(18) = 0.10022 \text{ m}$$ **Step 3 — True length:** $$L_{\text{true}} = 480 + 0.100 = 480.100 \text{ m}$$ The tape expanded (T > T_s), so it is longer, measured distance reads short, and the true length is greater.
Each unsupported span produces a sag correction. Since sag is always negative, the true distance is always LESS than what the tape reads under sag conditions. Increasing tension P reduces sag significantly (appears as P² in the denominator).
Scenario
**Sag Correction:** A 30 m tape weighing 0.90 N/m is supported only at its endpoints with a tension of 80 N. A distance is recorded as 270 m. Find the true horizontal distance.
Solution
**Number of tape spans:** n = 270/30 = 9 spans **Sag correction per span (L = 30 m):** $$C_{sag} = -\frac{w^2 L^3}{24 P^2} = -\frac{(0.90)^2(30)^3}{24(80)^2}$$ $$= -\frac{(0.81)(27000)}{24(6400)} = -\frac{21870}{153600} = -0.1424 \text{ m per span}$$ **Total sag correction:** $$C_{\text{total}} = 9 \times (-0.1424) = -1.282 \text{ m}$$ **True length:** $$L_{\text{true}} = 270 + (-1.282) = 268.718 \text{ m} \approx 268.72 \text{ m}$$
The ratio 29.94/30.00 = 0.9980 < 1 automatically makes the true distance less than measured — the formula handles the sign automatically. Always verify the physical logic: short tape → reads more → true is less.
Scenario
**Tape Too Short:** A nominal 30 m tape is found to be only 29.94 m when calibrated. A line is recorded as 450 m using this tape. Find the true length.
Solution
**True = Measured × (Actual / Nominal)** $$L_{\text{true}} = 450 \times \frac{29.94}{30.00} = 450 \times 0.99800 = 449.10 \text{ m}$$ The tape is short, so each 30 m recorded is only 29.94 m of ground → the true distance is LESS.
Use the exact formula when h is significant relative to L. Both give the same answer here because 8/200 = 0.04 — a very gentle slope. The slope correction is always negative (horizontal < slope).
Scenario
**Slope to Horizontal (Exact and Approximate):** A slope distance of 200 m is measured with an elevation difference of 8 m. Find horizontal distance by both methods.
Solution
**Exact method:** $$H = \sqrt{L^2 - h^2} = \sqrt{200^2 - 8^2} = \sqrt{40000 - 64} = \sqrt{39936} = 199.840 \text{ m}$$ **Approximate method:** $$C_h = -\frac{h^2}{2L} = -\frac{64}{400} = -0.160 \text{ m}$$ $$H = 200 - 0.160 = 199.840 \text{ m}$$ Both methods agree to 3 decimal places for gentle slopes (h/L < 0.10). The approximate formula is acceptable for board exams when h/L is small.
Applications
- Cadastral surveys under DENR-LMB regulations — precise boundary distances require temperature and sag corrections.
- Baseline measurement for triangulation networks — tension, temperature, and sag corrections are all applied simultaneously.
- Construction layout — laying out column grid lines with a tape requires corrections for tape condition and field temperature.
- Highway survey — measuring centerline distances on steep grades requires slope correction to obtain horizontal distances for plans.
Misconceptions
- WRONG: 'A tape too long means the measured distance is too long.' CORRECT: Tape too long → each tape span covers MORE ground → the taped count reads LESS than true → measured is SHORT. The true is MORE.
- WRONG: 'Sag correction can be positive.' CORRECT: Sag ALWAYS shortens the chord (horizontal distance between endpoints) — C_sag is always negative.
- WRONG: 'Slope correction can be positive.' CORRECT: Horizontal distance is always LESS than slope distance — C_h is always negative.
- WRONG: 'The E in C_p = (P-Ps)L/(AE) is the probable error.' CORRECT: In the tension correction formula, E = modulus of elasticity (Pa). Context determines meaning — two different quantities share the symbol E.
- WRONG: 'Temperature correction formula uses the measured length per tape span.' CORRECT: Use the total measured length L, not the nominal tape length, for the total correction.
Related Concepts
- Classification of Systematic Errors
- Error Propagation
- Horizontal Distance and Traverse Computation
- Stadia and Electronic Distance Measurement
Common Exam Questions
Example
A 50 m tape (standard at 20°C, α = 11.6×10⁻⁶/°C) is used at 35°C with a pull of 60 N (standard pull 50 N, A = 3.2 mm², E = 200 GPa). A line measures 300 m on a 5% slope. Find true horizontal distance.
Approach
Identify which corrections apply, compute each algebraically (positive/negative), sum all corrections, add to measured distance. True = Measured + ΣC.
Question Type
Compute true distance given tape corrections
Example
A 50 m tape is actually 50.05 m. A design distance of 200 m must be laid out. What tape reading should be used? Answer: Reading = 200 × (50/50.05) = 199.80 m.
Approach
Measuring: True = Measured × (Actual/Nominal). Laying out: Required tape reading = Design × (Nominal/Actual). Always check: tape too long → measuring gives short readings → multiply by ratio > 1.
Question Type
Tape too long/short — measuring vs laying out
Example
A 30 m tape (actual 30.02 m) is used at 10°C (standard 20°C, α = 11.6×10⁻⁶). A line measures 240 m. Find true length. Step 1: C_t = (11.6×10⁻⁶)(240)(10−20) = −0.0278 m. Step 2: True due to tape length = 240 × (30.02/30) = 240.16 m. Step 3: True = 240.16 − 0.0278 = 240.13 m.
Approach
Apply each correction independently, then sum. The board exam often combines temperature + tape length error in a single problem.
Question Type
Combined corrections problem
Key Points To Remember
- Temperature: C_t = αL(T − T_s); positive if T > T_s (tape longer, measured reads short).
- Tension: C_p = (P − P_s)L / (AE); positive if P > P_s.
- Sag: C_sag = −w²L³/(24P²); ALWAYS negative (always shortens).
- Slope: C_h = −h²/(2L); ALWAYS negative. Exact: H = √(L² − h²).
- Tape too long → True = Measured × (Actual/Nominal) > Measured.
- Tape too short → True = Measured × (Actual/Nominal) < Measured.
- Laying out: Stake distance = Design × (Nominal/Actual).
- All corrections: True = Measured + ΣC (algebraic sum of corrections).
Error Propagation in Survey Measurements
When a quantity of interest is computed from several independently measured components, each with its own probable error, the error in the final result is found by **error propagation**. This is essential for estimating the precision of computed distances, elevations, areas, and traverse misclosures. **Case 1: Sum or Difference of Independent Measured Quantities** If a total distance is the sum of independently measured segments $L_1, L_2, \ldots, L_n$ with probable errors $E_1, E_2, \ldots, E_n$, the probable error of the sum is: $$E_{\text{sum}} = \sqrt{E_1^2 + E_2^2 + \cdots + E_n^2}$$ Note: This applies to the **sum** (and also the **difference** — the errors still add in quadrature). The errors do NOT simply add arithmetically. **Case 2: Sum of $n$ Segments with Equal Probable Error $E$** If all segments have the same probable error $E$: $$E_{\text{sum}} = \sqrt{nE^2} = E\sqrt{n}$$ This is the **error of a series** — the total error grows as $\sqrt{n}$, not $n$. This is why a 10 km traverse has a larger expected error than a 1 km traverse, but the error grows more slowly than proportionally to the distance. **Case 3: Error of the Mean vs Error of the Sum** The mean has error $E_m = E/\sqrt{n}$ (decreases with more measurements). The sum has error $E_{\text{sum}} = E\sqrt{n}$ (increases with more measurements). These are consistent: the sum = $n \times \bar{x}$, so its error = $n \times E_m = n \times E/\sqrt{n} = E\sqrt{n}$. ✓ **Case 4: Product or General Function (Advanced)** For $Z = f(x_1, x_2, \ldots)$, the propagated error is: $$E_Z = \sqrt{\left(\frac{\partial f}{\partial x_1}\right)^2 E_1^2 + \left(\frac{\partial f}{\partial x_2}\right)^2 E_2^2 + \cdots}$$ This is the general law of error propagation. For the area of a rectangle $A = ab$: $E_A = \sqrt{(b\cdot E_a)^2 + (a\cdot E_b)^2}$. **Relative Error (Precision Ratio)** $$\text{Relative error} = \frac{E}{\bar{x}} \quad \text{or expressed as } 1:\frac{\bar{x}}{E}$$ A distance of 500 m with probable error ±0.05 m has relative error = 0.05/500 = 1:10,000.
Examples
Even though there are four errors, the combined error (0.057 m) is much less than the arithmetic sum (0.02+0.03+0.02+0.04 = 0.11 m). This is because random errors can cancel — quadrature addition accounts for this partial cancellation.
Scenario
A traverse has four sides measured with probable errors of ±0.02, ±0.03, ±0.02, and ±0.04 m. Find the probable error of the perimeter.
Solution
$$E_{\text{perimeter}} = \sqrt{(0.02)^2 + (0.03)^2 + (0.02)^2 + (0.04)^2}$$ $$= \sqrt{0.0004 + 0.0009 + 0.0004 + 0.0016}$$ $$= \sqrt{0.0033} = 0.0574 \text{ m} \approx \pm 0.057 \text{ m}$$
This is the 'equal errors, n segments' case. The error grows as √25 = 5×, not 25×. In geodetic leveling specifications, allowable closure errors are often expressed as k√K mm where K is line length in km — directly applying this principle.
Scenario
A 5 km leveling line is run in 25 sections, each measured with a probable error of ±2 mm. Find the probable error of the total line elevation difference.
Solution
$$E_{\text{total}} = E\sqrt{n} = 2\sqrt{25} = 2 \times 5 = \pm 10 \text{ mm}$$ Alternatively: $$E_{\text{total}} = \sqrt{25 \times (2)^2} = \sqrt{100} = 10 \text{ mm}$$ Relative precision = 10 mm / 5000 m = 10/(5,000,000 mm) = 1:500,000.
For products, the partial derivatives bring in the other measured dimension as a multiplier. Note that both terms contribute roughly equally — neither dimension is the sole source of error in the area.
Scenario
A rectangular lot is measured: length = 80.00 ± 0.04 m, width = 50.00 ± 0.03 m. Find the probable error of the computed area.
Solution
$$A = l \times w = 80.00 \times 50.00 = 4000.00 \text{ m}^2$$ Using the general propagation (∂A/∂l = w = 50, ∂A/∂w = l = 80): $$E_A = \sqrt{(w \cdot E_l)^2 + (l \cdot E_w)^2}$$ $$= \sqrt{(50 \times 0.04)^2 + (80 \times 0.03)^2}$$ $$= \sqrt{(2.00)^2 + (2.40)^2}$$ $$= \sqrt{4.00 + 5.76} = \sqrt{9.76} = 3.12 \text{ m}^2$$ Result: A = 4000.00 ± 3.12 m²
Applications
- Leveling networks — allowable closure = k√K mm (Philippine BRS/NAMRIA standards, where K = km length).
- Traverse closure — the expected closure error grows as E√n for n equal-precision legs.
- Area computation from measurements — combined error includes errors from all bounding dimensions.
- GPS baseline processing — error propagation from coordinate component errors to horizontal distance.
Misconceptions
- WRONG: 'If each of 4 measurements has error ±0.02 m, the total has error ±0.08 m.' CORRECT: E_total = 0.02√4 = ±0.04 m — errors do NOT add arithmetically for random (independent) errors.
- WRONG: 'The error of the mean and error of the sum are the same formula.' CORRECT: Error of mean = E/√n (gets smaller with n); error of sum = E√n (gets larger with n). These are different operations.
- WRONG: 'Error propagation for a product is the same as for a sum.' CORRECT: Products use the general differential formula with partial derivatives, not simple quadrature of the original errors.
Related Concepts
- Most Probable Value and Probable Error
- Traverse Closure and Adjustment
- Leveling — Allowable Closure
- Weighted Observations
Common Exam Questions
Example
Three distances 200±0.04, 150±0.03, 180±0.05 m — find error of total. E = √(0.04²+0.03²+0.05²) = √(0.0016+0.0009+0.0025) = √0.0050 = ±0.0707 m.
Approach
Apply E_sum = √(ΣEi²). Do NOT add errors arithmetically. Square each error, sum them, take the square root.
Question Type
Find error of a sum from individual errors
Example
What is the probable error of a 10 km level line run in 50 equal sections if each section has probable error ±1.5 mm? E = 1.5√50 = ±10.6 mm.
Approach
For equal-error segments: E_sum = E√n. Set equal to allowable closure, solve for n. Or for E_m = E/√n, solve for n = (E/E_m)².
Question Type
Find number of measurements for target precision
Key Points To Remember
- Error of sum of independent measurements: E_sum = √(E₁² + E₂² + ... + En²).
- Equal errors, n segments: E_sum = E√n (NOT n×E).
- Error of the mean: E_m = E/√n (decreases with more measurements).
- Error of a sum and error of the mean are consistent: n × (E/√n) = E√n. ✓
- Relative error = E/mean; express as 1:N ratio (higher N = more precise).
- For a product Z = ab: E_Z = √((b·Ea)² + (a·Eb)²).
- Never add probable errors arithmetically — always add in quadrature (square root of sum of squares).
Practice Problems
Working in seconds of arc simplifies arithmetic. Always verify Σv = 0 as an arithmetic check (0 ✓ here). The mean is more precise than any individual reading — E_m (0.7") is less than E_s (1.5"). This is a classic board-level five-observation problem.
Problem
**Problem 1 — MPV and Probable Error.** A surveyor measures an angle five times: 35°24'10", 35°24'14", 35°24'08", 35°24'12", 35°24'11". Find: (a) the most probable value, (b) the probable error of a single observation, and (c) the probable error of the mean.
Solution
**Convert to seconds above 35°24'00":** 10, 14, 08, 12, 11 seconds. **(a) MPV:** $$\bar{x} = \frac{10 + 14 + 8 + 12 + 11}{5} = \frac{55}{5} = 11''$$ MPV = 35°24'11" **(b) Residuals and Σv²:** | Obs | x (sec) | v = x − 11 | v² | |---|---|---|---| | 1 | 10 | −1 | 1 | | 2 | 14 | +3 | 9 | | 3 | 8 | −3 | 9 | | 4 | 12 | +1 | 1 | | 5 | 11 | 0 | 0 | | **Σ** | | **0 ✓** | **20** | $$E_s = 0.6745\sqrt{\frac{20}{5-1}} = 0.6745\sqrt{5} = 0.6745 \times 2.236 = 1.508'' \approx \pm 1.5''$$ **(c) Probable error of mean:** $$E_m = \frac{1.508}{\sqrt{5}} = \frac{1.508}{2.236} = \pm 0.674'' \approx \pm 0.7''$$ **Result:** MPV = 35°24'11" ± 0.7"
When multiple corrections act in the same direction, they compound. Here, both tape length error and temperature expansion make the tape longer, so both corrections add to the measured value. The temperature correction is applied to the full measured length (600 m), not the nominal tape length (30 m) — a common source of mistakes in board exams.
Problem
**Problem 2 — Temperature and Tape Length Corrections Combined.** A 30 m tape standardized at 20°C (α = 11.6 × 10⁻⁶/°C) is found to be 30.005 m long at standard conditions. It is used at 42°C to measure a line recorded as 600 m. Find the true length.
Solution
**Step 1 — Correction for tape being too long:** $$L_{\text{after tape corr}} = 600 \times \frac{30.005}{30.000} = 600 \times 1.000167 = 600.100 \text{ m}$$ **Step 2 — Temperature correction (applied to measured length or the corrected length; use measured for small corrections):** $$C_t = \alpha \times L_{\text{measured}} \times (T - T_s) = (11.6 \times 10^{-6})(600)(42 - 20)$$ $$C_t = (11.6 \times 10^{-6})(600)(22) = 0.15312 \text{ m}$$ **Step 3 — Total true length:** $$L_{\text{true}} = 600 + 0.100 + 0.153 = 600.253 \text{ m}$$ *(Note: Both corrections are positive — tape is long, AND temperature is above standard, so tape expanded further → measured reads even shorter → true is even longer.)*
The weight in N/m = mass/m × 9.81 = 0.04 × 9.81 = 0.3924 N/m. Always convert kg/m to N/m (multiply by g = 9.81 m/s²). The sag correction is significant here (0.693 m over 350 m) — this is why full tension support or 'normal tension' is used in precise baseline measurement.
Problem
**Problem 3 — Sag Correction.** A 50 m steel tape weighing 0.04 kg/m (0.3924 N/m) is supported only at its two endpoints with an applied tension of 90 N. A traverse leg is recorded as 350 m. Find the true length.
Solution
**Number of spans:** n = 350/50 = 7 spans **Sag correction per span (L = 50 m, w = 0.3924 N/m, P = 90 N):** $$C_{sag} = -\frac{w^2 L^3}{24 P^2} = -\frac{(0.3924)^2(50)^3}{24(90)^2}$$ $$= -\frac{(0.15398)(125000)}{24(8100)}$$ $$= -\frac{19247.5}{194400}$$ $$= -0.09901 \text{ m per span}$$ **Total sag correction:** $$C_{\text{total}} = 7 \times (-0.09901) = -0.6931 \text{ m}$$ **True length:** $$L_{\text{true}} = 350 + (-0.693) = 349.307 \text{ m} \approx 349.31 \text{ m}$$
Laying out is the REVERSE of measuring. For a short tape, you read MORE on the tape to achieve the correct ground distance. The ratio flips: measuring uses Actual/Nominal; laying out uses Nominal/Actual. A very common board exam trap: confusing these two ratios.
Problem
**Problem 4 — Laying Out with a Defective Tape.** A 50 m tape is found to be 49.96 m when calibrated. A surveyor must stake out a lot boundary of exactly 250 m. What tape reading should the crew use when laying out?
Solution
**Formula for laying out:** Stake reading = Design distance × (Nominal / Actual) $$\text{Tape reading} = 250 \times \frac{50.00}{49.96} = 250 \times 1.000801 = 250.200 \text{ m}$$ **Check logic:** The tape is short (49.96 < 50.00). Each time the crew reads 50 m on the tape, they are only spanning 49.96 m of ground. To span 250 m of ground, they need to lay out more tape than 250 — hence the reading must exceed 250 m. ✓
The √K (or √n) pattern in leveling closure specifications is a direct application of E_sum = E√n — it acknowledges that longer lines accumulate more random error proportional to the square root of distance/number of setups. This is not an arbitrary specification — it is derived from the theory of error propagation.
Problem
**Problem 5 — Error Propagation in a Leveling Loop.** A level circuit consists of 16 equal-length sections. The probable error of a single section is ±3 mm. (a) Find the probable error of the total circuit elevation difference. (b) If the allowable closure for first-order leveling is ±12√K mm where K is the total length in km, and the circuit total is 8 km, does this circuit meet first-order specifications?
Solution
**(a) Probable error of total:** $$E_{\text{total}} = E_s\sqrt{n} = 3\sqrt{16} = 3 \times 4 = \pm 12 \text{ mm}$$ **(b) Allowable closure for 8 km:** $$E_{\text{allow}} = 12\sqrt{8} = 12 \times 2.828 = \pm 33.9 \text{ mm}$$ Since 12 mm < 33.9 mm, the circuit's probable error is WITHIN the first-order specification. ✓ *(Note: First-order leveling in the Philippine context per NAMRIA standards uses allowable closure = 4√K mm for double-run; the 12√K figure used here is for illustration. The calculation principle is the same.)*
In multi-correction problems: (1) apply slope first to get a horizontal distance, (2) then apply temperature and tape length corrections to the horizontal distance. All corrections here are positive (T > Ts, tape too long) — the true horizontal distance exceeds the measured horizontal distance. This type of combined problem is common in board exams at the difficult level.
Problem
**Problem 6 — Comprehensive: Slope + Temperature + Tape Length.** A 30 m steel tape (actual length 30.03 m at 20°C, α = 11.6 × 10⁻⁶/°C) is used at 35°C to measure a slope distance of 180 m. The elevation difference between the endpoints is 6 m. Find the true horizontal distance.
Solution
**Step 1 — Slope correction to get horizontal slope distance:** $$C_h = -\frac{h^2}{2L} = -\frac{6^2}{2(180)} = -\frac{36}{360} = -0.100 \text{ m}$$ $$L_{\text{horiz}} = 180 - 0.100 = 179.900 \text{ m}$$ *(Alternatively: H = √(180² − 6²) = √(32400 − 36) = √32364 = 179.900 m ✓)* **Step 2 — Temperature correction (apply to horizontal length):** $$C_t = (11.6 \times 10^{-6})(179.900)(35 - 20) = (11.6 \times 10^{-6})(179.900)(15)$$ $$= 0.03132 \text{ m}$$ **Step 3 — Tape length correction:** $$C_l = 179.900 \times \frac{30.03}{30.00} - 179.900 = 179.900 \times 0.001 = 0.1799 \text{ m}$$ *(Or: C_l = (30.03 − 30.00)/30.00 × 179.900 = 0.001 × 179.900 = 0.180 m)* **Step 4 — True horizontal distance:** $$L_{\text{true}} = 179.900 + 0.031 + 0.180 = 180.111 \text{ m}$$
Exam Preparation Tips
- MEMORIZE THE FOUR TAPE CORRECTION FORMULAS: C_t = αL(T−Ts), C_p = (P−Ps)L/(AE), C_sag = −w²L³/(24P²), C_h = −h²/(2L). Write them out 10× until automatic. Know which variables represent what — especially that E in C_p is modulus of elasticity, NOT probable error.
- TAPE TOO LONG/SHORT: Draw a sketch. Tape too long → each span covers MORE ground → reading is SHORT → true > measured. Tape too short → each span covers LESS ground → reading is LONG → true < measured. Formula: True = Measured × (Actual/Nominal). For laying out, flip: Reading = Design × (Nominal/Actual).
- PROBABLE ERROR CHECKLIST: (1) Compute mean. (2) Compute v = obs − mean. (3) Check Σv = 0. (4) Compute Σv². (5) E_s = 0.6745√(Σv²/(n−1)). (6) E_m = E_s/√n. Never skip step 3 — it catches arithmetic errors.
- ERROR PROPAGATION: NEVER add probable errors directly. Always use E = √(ΣEi²). For equal errors, E_total = E×√n. The square-root relationship is the key formula to distinguish from incorrect arithmetic addition.
- SAG IS ALWAYS NEGATIVE, SLOPE IS ALWAYS NEGATIVE: If your computed sag or slope correction is positive, you made a sign error. Re-check.
- UNITS CONSISTENCY: In the tension correction C_p = (P−Ps)L/(AE), use: P in Newtons, L in meters, A in m² (convert mm² by dividing by 10⁶), E in Pascals (200×10⁹ Pa for steel). Cross-check units: N×m/(m²×N/m²) = m. ✓
- RELATIVE ERROR / PRECISION RATIO: Express as 1:N where N = mean/E. Higher N = more precise. NAMRIA first-order geodetic leveling requires 1:1,000,000; ordinary leveling may require 1:50,000. Know typical precision standards.
- PROBABLE ERROR vs STANDARD DEVIATION: Probable error = 0.6745 × standard deviation. In board exams, when the problem says 'probable error,' apply the 0.6745 factor. If it says 'standard deviation' or 'standard error,' do NOT apply 0.6745.
- TEMPERATURE CORRECTION SIGN CHECK: T > T_standard → tape is longer → each span covers more ground → recorded distance reads short → C_t is POSITIVE (add to measured). T < T_standard → C_t is NEGATIVE (subtract from measured).
- PRACTICE TABULAR PROBLEMS UNDER TIME PRESSURE: Board exam surveying sections are time-limited. Practice setting up the v and v² table in under 2 minutes for 5–6 observations. Speed comes from consistent table format, not different methods each time.
- CROSS-CHECK USING APPROXIMATE vs EXACT SLOPE FORMULA: If time permits, compute both H = √(L²−h²) and H = L − h²/2L and verify they agree for gentle slopes (h/L < 5%). Large discrepancy signals an arithmetic error.
- RA 544 (CIVIL ENGINEERING LAW): While not directly tested in surveying computations, RA 544 defines the scope of civil engineering practice which includes surveying. Understanding that CE licensure covers geodetic surveying context is important for professional practice questions.
In summary
Measurements and Theory of Errors is the bedrock of all surveying computations. Every distance, angle, elevation, and area derived from field observations carries an inherent uncertainty that the professional surveyor must quantify, correct, and report. The key takeaways for the PRC licensure examination are: **1. Know your error types cold.** Mistakes → eliminate. Systematic → correct with formulas. Random → treat statistically. Confusing these categories is the first trap. **2. The four tape corrections are mandatory knowledge.** C_t, C_p, C_sag, C_h — know the formula, the sign (sag and slope are ALWAYS negative), and the physical reasoning. The tape-too-long/short problem with its measuring vs. laying-out distinction is a perennial board favorite. **3. MPV = mean; probable error uses 0.6745.** Set up the v-table, verify Σv = 0, compute Σv², apply the formula. The probable error of the mean is E_s/√n — precision improves with the square root of the number of observations, not linearly. **4. Error propagation uses quadrature.** E = √(ΣEi²) for independent measurements. Never add errors arithmetically. For n equal-error measurements, E_sum = E√n and E_mean = E/√n — these are consistent and must be memorized together. As you progress through leveling, traversing, and area computation, you will apply these same principles repeatedly — the allowable closure for a level circuit (k√K), the traverse closure tolerance, and the precision of area computed from traverse coordinates all trace back to this chapter. Mastery here is not optional — it is the prerequisite for professional-level surveying practice under RA 544 and for passing the PRC Civil Engineer Licensure Examination.
Ready to practise for the GELE 2026?
Super Tutor's AI review plan adapts to your weak areas and builds a weekly practice schedule around your target GELE exam date.