CIV3403 ENGINEERING SURVEYING 1

LECTURE IV
ERRORS IN LEVELING

Sources of Error in Leveling

All measurements have error. In the case of leveling, these errors will be instrumental,
observational and natural;

Instrument Errors:

i. The main source of instrumental error is residual collimation error. As already

indicated, keeping the horizontal lengths of the backsights and foresights at each
instrument position equal will cancel this error. Where the observational distances are
unequal, the error will be proportional to the difference in distances. The easiest
approach to equalizing the sight distances is to pace from backsight to instrument and
then set up the foresight change point the same number of paces away from the
instrument.
ii. Parallax error.
iii. Staff graduation errors may result from wear and tear or repairs and the staffs should
be checked against a steel tape. Zero error of the staff, caused by excessive wear of the
base, will cancel out on backsight and foresight differences. However, if two staffs are
used, errors will result unless calibration corrections are applied.
iv. In the case of the tripod, loose fixings will cause twisting and movement of the tripod
head. Overtight fixings make it difficult to open out the tripod correctly. Loose tripod
shoes will also result in unstable set-ups.

1
Observational Errors:
i. Levelling involves vertical measurements relative to a horizontal plane so it is important
to ensure that the staff is held strictly vertical. It is often suggested that one should rock the
staff back and forth in the direction of the line of sight and accept the minimum reading as
the truly vertical one. However, as shown in Figure 1a, this concept is incorrect when using
a flat-bottomed staff on flat ground, due to the fact that the staff is not being tilted about its
face. Thus it is preferable to use a staff bubble, which should be checked frequently with
the aid of a plumb-bob

Figure 1. a. The effect of rocking the staff on staff reading b. Levelling Plate

ii. There may be errors in reading the staff, particularly when using a tilting level which gives
an inverted image. These errors may result from inexperience, poor observation conditions
or overlong sights. Limit the length of sight to about 25 – 30m, to ensure the graduations
are clearly defined.
iii. Ensure that the staff is correctly extended or assembled. In the case of extending staffs,
listen for the click of the spring joint and check the face of the staff to ensure continuity of
readings. This also applies to jointed staffs.
iv. Do not move the staff off the CP position, particularly when turning it to face the new
instrument position. Always use a well-defined ground (Figure 1b).
v. Avoid settlement of the tripod, which may alter the height of collimation between sights or
tilt the line of sight. Set up on firm ground, with the tripod feet firmly thrust well into
ground. On pavements, locate the tripod shoes in existing cracks or joints. In precise
levelling, the use of two staffs helps to reduce this effect. Observers should also refrain
from touching or leaning on the tripod during observation.

2
 vi. Booking errors can, of course, ruin good field work. Neat, clear, correct booking of field
data is essential in any surveying operation. Typical booking errors in leveling are entering
the values in wrong columns or on the wrong lines, transposing figures such as 3.538 to
3.583 and making arithmetical errors in the reduction process. Very often, the use of pocket
calculators simply enables booker to make the errors quicker. To avoid this error source,
use neat, legible figures; read the booked value back to the observer and have them check
the staff reading again; reduce the data as it is recorded.

vii. When using a tilting level remember to level the tubular bubble with the tilting screw prior
to each new staff reading. With the automatic level, carefully center the circular bubble
and make sure the compensator is not sticking. Residual compensator errors are
counteracted by centering the circular bubble with the instrument pointing backwards at
the first instrument set-up and forward at the next.
This procedure is continued throughout the leveling.

Natural Errors:
i. Curvature and refraction have already been dealt with. Their effects are minimized by equal
observation distances to backsight and foresight at each set-up and readings more than 0.5
m above the ground.
ii. Wind can cause instrument vibration and make the staff difficult to hold in a steady
position. Precise leveling is impossible in strong winds. In tertiary leveling keep the staffs
to its shortest length and use a wind break to shelter the instrument.
iii. Heat shimmer can make the staff reading difficult if not impossible and may make it
necessary to delay the work to an overcast day. In hot sunny climes, carry out the work
early in the morning or in the evening. Careful consideration of the above error sources,
combined with regularly calibrated equipment, will ensure the best possible results but will
never preclude random errors of observation.

Instrument Errors:
For equipment to give the best possible results it should be frequently tested and, if necessary,
adjusted. Surveying equipment receives continuous and often brutal use on construction sites. In
all such cases a calibration base should be established to permit weekly checks on the equipment

3
Two Peg Test for Checking and Re-calibrating

Figure 2 Two peg test

This method is either for an optical or digital level. The two-peg test is very simple, but provides
a way to test the accuracy of a level, and if you know which screw to turn (for analog instruments)
or menu to follow (for the digital level), you can adjust it to remove the error. See specific
instrument instructions for making adjustments. Collimation error occurs if the line of sight is not
truly horizontal when the tubular bubble is centered, i.e. the line of sight is inclined up or down
from the horizontal. The procedure being as follows (Figure 2):

4
a. Set up the instrument midway between two pegs A and B set, say, 20 m apart and note the
staff readings,a1 and b1, equal to, say, 1.500 m and 0.500 m respectively.
Let us assume that the line of sight is inclined up by an angle of α; as the lengths of the
sights are equal (10 m), the error in each staff reading will be equal and so cancel out,
resulting in a ‘true’ difference in level between A and B.

ΔHTRUE = (a1− b1) = (1.500 − 0.500) = 1.000 m

Thus we know that A is truly lower than B by 1.000 m. We do not at this stage know that
collimation error is present.
b. Move the instrument to C, which is 10 m from B and in the line AB and observe the staff
readings a and b equal to, say, 3.500 m and 2.000 m respectively. Then

ΔH = (a2− b2) = (3.500 − 2.000) = 1.500 m

Now as 1.500 ≠ the ‘true’ value of 1.000, it must be ‘false’.

ΔH FALSE = 1.500 m and it is obvious that the instrument possesses a collimation error the
amount and direction of which is as yet still unknown, but which has been revealed by the
use of unequal sight lengths CB (10 m) and CA (30 m). Had the two values for ΔH been
equal, then there would be no collimation error present in the instrument.
c. Imagine a horizontal line from reading b2 (2.000 m) cutting the staff at A at reading a3.
Because A is truly 1.000 m below B, the reading at a3 must be 2.000 + 1.000 = 3.000 m.
However, the actual reading was 3.500 m, and therefore the line of sight of the instrument
was too high by 0.500 m in 20 m (the distance between the two pegs). This is the amount
and direction of collimation error.
d. Without moving the instrument from C, the line of sight must be adjusted down until it is
horizontal. To do this one must compute the reading (a4) on staff A that a horizontal sight
from C, distance 30 m away, would give. By simple proportion, as the error in 20 m is
0.500, the error in 30 m = (0.500 ×30)/20 = 0.750 m.
Therefore the required reading at a4 is 3.500 − 0.750 = 2.750 m.

5
 e. If titled level is used, tilt the telescope until it reads 2.750 m on the staff by using tilting
screw. This movement will cause the tubular bubble to go off center. Re-center it by means
of its adjusting screws, which will permit the raising or lowering of one end of the bubble.
f. For other analog instruments, you can adjust the cross hairs by loosening the top adjusting
screw that allows cross hair to move up, or the bottom capstan screw that allows it to move
down then tightening to other screw.
g. For digital instruments, the adjustments can be carried out by following the user manual of
the equipment.

Closure Tolerances and Error Distribution

Leveling should always be checked by running closed circuits or loops. This can be done either by
returning to the starting benchmark, or by ending the circuit at another benchmark of equal or
higher reliability. The final elevation should agree with the starting elevation if returning to the
initial benchmark. The amount by which they differ is the loop misclosure. If closure is made to
another benchmark, the section misclosure is the difference between the closing benchmark’s
given elevation and its elevation obtained after leveling through the section. Specifications, or
purpose of the survey, fix permissible misclosures. If the allowable misclosure is exceeded, one or
more additional runs must be made. When acceptable misclosures achieved, final elevations are
obtained by making an adjustment.
When the misclosure is assessed, one must then decide if it is acceptable or not. In many cases the
engineer may make the decision based on his knowledge of the project and the tolerances required.
Alternatively the permissible criteria may be based on the distance levelled or the number of set-
ups involved. A common criterion used to assess the misclosure (E) is:

E = m k (0.5)
where K =distance leveled in kilometers,
m =a constant with units of millimeters, and
E =the allowable misclosure in millimeters.

6
The value of m may vary from 2 mm for precise levelling to 12 mm or more for large engineering
leveling. In many cases in engineering, the distance involved is quite short but the number of set-
ups quite high, in which case the following criterion may be used:

E = m n (0.5)

where n = the number of set-ups,

and m = a constant in millimeters.
As this criterion would tend to be used only for construction leveling, the value for m may be a
matter of professional judgment. A value frequently used is ±5 mm.

In the case of a leveling circuit, a simple method of distribution is to allocate the error in proportion
to the distance leveled. For instance, consider a leveling circuit commencing from a BM at A, to
establish other BMs at B, C, D and E (Figure 3) for which the heights have been computed without
taking the misclosure into account.

Figure 3 Example of Leveling Circuit

7
On completing the circuit the observed value for the BM at A is 20.018 m compared, with its
known and hence starting value of 20.000 m, so the misclosure is 0.018 m. The distance leveled is
5.7 km. Considering the purpose of the work, the terrain and observational conditions, it is decided
to adopt a value for m of 12 mm. Hence the acceptable misclosure is 12(5.7)0.5= 29 mm, so the
leveling is acceptable. The difference in heights is corrected by (0.018/5.7) × distance in kilometers
travelled. Therefore correction to AB =−0.005 m, to BC =−0.002 m, to CD =−0.003 m, to DE
=−0.006 m and to EA =−0.002 m. The values of the BMs will then be B = 28.561 m, C = 35.003
m, D = 30.640 m, E = 22.829 m and A = 20.000 m. In many instances, a closing loop with known
distances is not the method used and each reduced level is adjusted in proportion to the cumulative
number of set-ups to that point

from the start. Consider the table below which shows the observations for a short section of
leveling between two bench marks of known height:

Table 1. Example of Closure, Tolerance and Error Distribution

i. There are four set-ups, and therefore E = 5(4)0.5 = 0.010 m. As the misclosure is only
0.008 m, the leveling is acceptable.
ii. The correction per set-up is (0.008/4) = −0.002 m and is cumulative as shown in the table.