GEODESY AND CARTOGRAPHY ⃝c Polish Academy of Sciences

Vol. 68, No. 2, 2019, pp. 361–373 DOI: https://doi.org/10.24425/gac.2019.128463

Research paper

Evaluation of some levelling techniques in surveying application

Atınç Pırtı1∗ , Ramazan Gürsel Hoşbaş2

Yildiz Technical University,

Department of Surveying Engineering
Davutpasa, 34220 Esenler, Istanbul – Turkiye
1
e-mail: [email protected], ORCID: http://orcid.org/0000-0001-9197-3411
2
e-mail: [email protected], ORCID: http://orcid.org/0000-0002-3189-7696

∗
Corresponding author: Atınç Pırtı

Received: 31 May 2019 / Accepted: 1 August 2019

Abstract: Applications in geodesy and engineering surveying require the determination of

the heights of the vertical control points in the national and local networks using different
techniques. These techniques can be classified as geometric, trigonometric, barometric and
Global Positioning System (GPS) levelling. The aim of this study is to analyse height dif-
ferences obtained from these three techniques using precise digital level and digital level,
total station (trigonometric levelling) and GPS which collects phase and code observations
(GPS levelling). The accuracies of these methods are analysed. The results obtained show
that the precise digital levelling is more stable and reliable than the other two methods.
The results of the three levelling methods agree with each other within a few millimetres.
The different levelling methods are compared. Geometric levelling is usually accepted as
being more accurate than the other methods. The discrepancy between geometric levelling
and short range trigonometric levelling is at the level of 8 millimetres. The accuracy of the
short range trigonometric levelling is due the reciprocal and simultaneous observations of
the zenith angles and slope distances over relative short distances of 250 m. The difference
between the ellipsoidal height differences obtained from the GPS levelling used without
geoid and the orthometric height differences obtained from precise geometric levelling is
4 millimetres. The geoid model which is obtained from a fifth order polynomial fit of the
project area is good enough in this study. The discrepancy between the precise geometric
and GPS levelling (with geoid corrections) is 4 millimetres over 5 km.

Keywords: levelling, accuracy, height, GPS, precision

1. Introduction

Levelling is the general term applied to any process by which elevations of points or
differences in elevation are determined. It is a vital operation in producing the necessary
data for mapping, engineering design and construction. Differences in elevation have
⃝c 2019 by the Author(s). Submitted for possible open access publication under
the terms and conditions of the Creative Commons Attribution (CC BY-NC) license
0 (http://creativecommons.org/licenses/by/4.0/).
362 Atınç Pırtı, Ramazan Gürsel Hoşbaş

traditionally been determined by digital levelling, trigonometric levelling, barometric

levelling and GPS levelling. The method selected depends primarily on the accuracy
required, although the type of terrain over which the levelling is done is also a fac-
tor. Geometric levelling can produce the highest order of accuracy. In spite of the fact
that trigonometric levelling produces a somewhat lower order of accuracy than digi-
tal levelling, the method is still suitable for many projects such as establishing verti-
cal control for topographic mapping or for lower order construction setting-out. It is
particularly convenient in hilly or mountainous terrain where large differences in el-
evations are encountered. Both methods are subject to systematic and random errors.
The primary systematic errors include earth curvature, atmospheric refraction and cal-
ibration of the surveying tools. The Global Positioning System (GPS) can be used for
low-order vertical control surveys. To get accurate elevations using this method, the
geoidal undulations in the area must be known and applied. The ellipsoidal heights
obtained from GPS cannot be used directly for practical surveying. The ellipsoidal
heights have to be transformed to orthometric heights, being the distance measured
along the plumb line between the geoid and a point on the Earth’s surface and taken
positive upward from the geoid. The difference between the ellipsoidal and the ortho-
metric height is defined as the geoid height or geoid undulation. With the knowledge
of the geoid (quasigeoid), it is possible to derive orthometric or normal heights from
ellipsoidal heights. The most recent and advanced levelling approach is GPS levelling
which is the combination of the GPS derived ellipsoidal heights and the geoid infor-
mation in order to determine orthometric heights. The studies relating to the GPS tech-
nique need ellipsoidal heights with centimetre or sub-centimetre accuracy which are
obtained from the GPS solution. This will only be possible with an interconnected net-
work or reference stations (Ayan, 2001; Ceylan and Baykal, 2008; Schofield, 2001)
assuming the Precise Digital Levelling (PDL) data as minimal error for the time be-
ing and equal weight of all sections (all sections assumed to be 250 m long). In this
study, accuracy of the height determination techniques has been compared the twenty
vertical control points were marked in the project area. In order to determine of height
differences between these 20 points, three methods were used and analysed, namely,
GPS levelling, reciprocal and simultaneous short range trigonometric levelling, and dig-
ital and precise digital levelling. The procedures and obtained results are explained
in this paper.

2. Digital levelling

Recent advances in electronics now enable surveyors to perform digital levelling with
an electronic (digital) level. The development of this type of level has become pos-
sible due to the advances in microchip technology and image processing. The at-
tributes of self-levelling instrumentation coupled with digital array photography and
electronic image processing have generated a digital level, which is very much close
to being truly automatic. The digital level processes an electronic image of a bar
coded staff for the determination of heights and distances with automatic recording of
 Evaluation of some levelling techniques in surveying application 363

the data for future processing on a computer. The digital level is an automatic level
(with a pendulum compensator) capable of normal optical levelling with normal grad-
uated staffs. The level can also be used with the bar coded staff and rod readings
obtained digitally with output to a display on the instrument. The measuring system
of the digital levelling consists of a level comprising optics and compensator, a bar
code scale mostly on an invar band fixed into a rod frame, a couple-charged device
(CCD) linear array and software controlling all operations, procedures and processes
of the digital level (Anderson and Mikhail, 1998; Ingesand, 1999; Kahmen, 2000;
Rüeger and Brunner, 1982; 2000; Wolf and Ghilani, 2002; Woschitz and Brunner, 2002;
Schofield, 2001).

3. Precise digital levelling

Precise digital levelling may be used in certain instances in construction such as in

deformation monitoring, the provision of precise height control for large engineering
projects like long-span bridges, dams and hydroelectric schemes and in mining sub-
sidence measurements. Precise digital levelling staffs have an invar bar code kept un-
der constant tension on the face. Invar has a very low coefficient of expansion, and
the staffs are calibrated to an exact length. The maximum sight distances are (50 m,
60 m, 90 m) and allowable differences between backsight and foresight lengths (2–
4 m, 5–10 m, 10–10 m) for first-second, and third-order levelling respectively. Rod
persons can pace or count rail lengths or highway slab joints to set sight distances.
Precise levelling demands good-quality turning points. Lines of sight should not pass
closer than about 0.5 m from any surface, e.g., the ground, to minimise refraction.
Readings at any set up must be completed in rapid succession; otherwise, changes
in atmospheric conditions might significantly alter refraction characteristics between
them (Assuming the PDL data as error free for the time being and equal weight of
all sections (all sections assumed to be 250 m long), (Featherstone et al., 1998) as-
suming the PDL data as error free for the time being and equal weight of all sec-
tions (all sections assumed to be 250 m long) (Federal Geodetic Control Committee,
1984; 1998).

4. Short range simultaneous reciprocal trigonometric levelling

Trigonometric levelling is the method of obtaining height differences using measured

slope distances and zenith angles. The targets in trigonometric levelling can always be
placed at the same height above the ground. Thus, the sighting distances are not lim-
ited by the inclination of the terrain and systematic errors, e.g. refraction, because the
back and foresight lines pass through similar layers of air. By extending the sighting
distances to a few hundred metres, the number of set-ups per kilometre is minimised.
The accuracy of trigonometric levelling mostly depends on the earth’s curvature and re-
fraction which directly affect the zenith angle and distance observations. Usually, the
364 Atınç Pırtı, Ramazan Gürsel Hoşbaş

surveyor measures the reciprocal zenith angles and slope distances from both ends of
the baseline. Besides, the surveyor should then use the mean value of the computed
height difference in order to correct for earth curvature and refraction. However, this is
not always practical and warranted. It should be remembered that in order to minimise
the errors introduced by curvature and refraction, the distances between the instrument
set-ups should be shorter than 250 m. The influence of the earth’s curvature and refrac-
tion is given by c&r = 0.0675 K2 metres, where K is the distance in kilometres. For
0.5 km the effect of c&r is 16.8 mm, for 0.2 km it is 2.7 mm, and so on (Anderson and
Mikhail 1998). Figure 1 illustrates the short range trigonometric levelling with recipro-
cal and simultaneous measurement of√zenith angles and slope distances. Field tests show
that standard deviations of ≤ 2 mm / km are achievable, even along inclined terrain, at
speed of 10 km/day when using sights no greater than 250 m for the reciprocal method
(Ceylan and Baykal, 2008) assuming the PDL data as error free for the time being and
equal weight of all sections (all sections assumed to be 250 m long), (Chrzanowski et
al., 1985). The electronic total stations were set up on the 20 marked points of the 5 km
run. The instruments were set up on the marks and the heights of instrument (iA and iB)
were measured three times from the centre of the telescope (horizontal axis) above the
station mark in the field, see Figure 1. The height difference between two points, namely,
A and B can be written as,
B
∆HAB = ∑ ∆hij + tA − tB (1)
A

where ∆HAB is the height difference between the terminals A and B, ∆hij are the indi-
vidual height differences, tA and tB are the height of the target at A and B from ground
to the total station. The observations have been made reciprocally and simultaneously
between two points (Figure 1). The zenith angles (Z′AB and Z′BA ) and slope distances
(CAB and CBA ) were measured for the test scenario. The height difference between
A and B points (∆HAB ) are computed using the following section (Anderson, 1998;
Ceylan and Baykal, 2008; Wolf and Ghilani, 2002).

Fig. 1. Model of the reciprocal and simultaneous short range trigonometric levelling

In our method, the average height difference is free from the impact of the refraction
and curvature of the Earth. However, this method has one disadvantage:
 Evaluation of some levelling techniques in surveying application 365

– the measurement of altitude iA and iB is not very precise. Another method of

trigonometric leveling based on differential measurement from the inside can be
used (Chrzanowski et al., 1985):
• the instrument stands between the points: (SAB /2) = 125 m – measurement in
2 face poles with supports, the poles are mounted with reflectors on the same
height,
• the instrument is not set on the point, so you do not need to measure its height,
– differential measurement eliminates the residual impact of refraction and curva-
ture of the Earth,
– the measurement is carried out in the main and return direction, therefore it is
possible to evaluate the internal accuracy as well as in geometric leveling.

5. GPS levelling

The geoid is defined as the gravity equipotential surface which best approximates mean
sea level over the entire Earth. It has been defined as the datum for the orthometric
height system. The irregular shape of the geoid, however, does not allow for an easy
computation of the horizontal positions of points. Therefore, a reference surface of reg-
ular shape, usually a biaxial ellipsoid, is selected to best approximate the geoid either
locally or globally. The geometric relationship between the geoid and the reference el-
lipsoid surface can be fully described by their separation and the slope of the geoid with
respect to the reference ellipsoid. The former is known as the geoidal height (N), and
the latter is known as the deviation of the vertical (θ ). The deviation of the vertical is
defined as the spatial angle between the ellipsoid normal and the actual gravity vector.
With a large number of monuments where both the GPS ellipsoidal heights and the or-
thometric heights from digital levelling have been observed, the geoidal heights at these
points can be approximated by using the following simple relation

N = h−H (2)

where N is the geoidal height, h is the ellipsoidal height from GPS surveys, and H is the
orthometric height from digital levelling (Kuang et al., 1996; Ayan, 2001; Featherstone
et al., 1998; Ceylan and Baykal, 2008).

6. The United States Federal Geodetic Control Subcommittee (US FGCS) accuracy
standards

The FGCS established accuracy standards and specifications for various orders of lev-
elling. The (US) Federal Geodetic Control Subcommittee (FGCS) established accuracy
standards and specifications for various orders of geometrical levelling. The FGCS rec-
ommends the following formula to compute the allowable misclosures (tolerances):
√
TMisc = ±m K mm (3)
366 Atınç Pırtı, Ramazan Gürsel Hoşbaş

where TMisc is the allowable loop or section misclosure, in millimetres; mis a constant;
and K is the total length levelled in kilometres. For loops (circuits that begin and end
on the same bench mark), K is the total perimeter distance. The FGCS specifies the
constants (m) of 4 (first-order class I), 5 (first-order class II), 6 (second-order class I),
8 (second-order class II), and 12 mm (third-order) for five classes of levelling. It is im-
portant to point out that meeting the FGCS misclosure criterion alone does not guarantee
that a certain order of accuracy has been met (Assuming the PDL data as error free for
the time being and equal weight of all sections (all sections assumed to be 250 m long;
Featherstone et al, 1998; Federal Geodetic Control Committee, 1984; 1988).
The accuracy parameters of 1st and 2nd order levelling network for Turkey are
T [mm] ≤ 12 K (km).

7. Methods and data

The experiment was conducted in the Samandira region of Istanbul, see Figure 2.
The GPS and terrestrial surveys (geometric and short range trigonometric levelling) were
performed on a levelling route of about 5 kilometres length. In order to minimise the
errors introduced by earth curvature and refraction, distances between the tests points
had to be restricted to about 250 m. Figure 3 illustrates the distribution of the selected

a) b)

c)

Fig. 2. Project Area (a) and GPS Network (b) and quasi-geoid isolines on the Turkey map (c)
 Evaluation of some levelling techniques in surveying application 367

Fig. 3. The distribution of the test points in the project area

points in the project area. The digital levelling (both digital and precise digital level-
ling) was carried out in order to assess the accuracy of the short range trigonometric
levelling and the GPS levelling (Telci et al., 2006). The height differences between the
20 points were determined by geometric levelling performed as double run levelling.
The instrument used for the precise digital levelling was a Leica DNA 03 precise dig-
ital level, together with two bar coded invar staffs of three metres length and the staffs
were stabilised with struts. The Leica DNA 03 precise digital level provides rapid, ac-
curate solutions for a wide range of levelling applications, from topographic and con-
struction surveys to first-order levelling and monitoring. It provides 0.3 mm accuracy
on a 1 km double run line with an invar staff and is ideal for first and second order
levelling and high precision measurements. The levelling routine was performed dou-
ble observance (BFFB, aBFFB) to increase the reliability of the measurement and to
reduce possible errors caused by the staff sinking. Applying alternating observations
procedures (aBFFB = BFFB FBBF) to eliminate horizontal tilt (residual error of the
automatic compensator) (Assuming the PDL data as error free for the time being and
equal weight of all sections (all sections assumed to be 250 m long). The instrument
used for the digital levelling was a Topcon DL 102 digital level (0.8 mm/km with fi-
bre glass staff) and with the two bar-coded aluminium rods of five metres length. The
levelling routine was performed observance (BFFB). Minimum ground clearance of
0.5 m required to refractionary influences of ground proximity. Limit target distance,
< 30 m. Levelling staffs with adjustable brace poles provide good stability. The staffs
were stabilised with struts. Include equal backsights and foresights, maintaining a line
of sight > 0.5 m above the ground and levelling the instrument to minimise any errors
368 Atınç Pırtı, Ramazan Gürsel Hoşbaş

due to the obliquity of horizon problem. All these precautions were taken during the
survey. Maximum allowable staff reading is 3 m. Four persons were performed in dou-
ble run digital levelling. It took 6.5 hours on 5 km double run levelling. In the short
range trigonometric levelling, the distances were measured by using two Nikon DTM
330 total stations with the (3 mm + 2 ppm) distance specification and 4.5′′ zenith an-
gle accuracy. The zenith angles and slope distances were reciprocally and simultane-
ously measured by using the same instruments four times in two faces. The heights
of instruments, prisms, and targets for all points were measured three times to ob-
tain mmlevel accuracy. The instruments were only set up on the 20 mark points. The
instrument uses dual-axis compensation, and electronic level sensors, and it applies col-
limation, vertical index, and trunnion axis corrections automatically. Auto dual-axis
compensation can ensure the accurate levelling of total station (Automatic dual-axis
compensator with working range ±3′ (±55 mgon)). In the compensation range 3′ , al-
though the instrument is tilt, the horizontal and zenith angle can be measured pre-
cisely.
The GPS surveys were carried out in order to determine the ellipsoidal heights
of the 20 points along a levelling route (~5 km). The GPS measurements were taken
with four Ashtech Z Max GPS receivers using the static method. The GPS data were
recorded in seven sessions. These sessions were measured on 10 May 2006 between
8:00–18:30 h (Local time). During this period, the satellite visibility varied between
6 and 9 satellites, and the PDOP values between 2.8 and 1.5. The GPS data were
collected in 10 seconds epoch intervals. The station occupation time was ten hours
and thirty minutes for the reference point (P1) and about sixty minutes for the re-
maining points. The reference station (P1) suffered no sky obstructions and was set
up on Ortadag Hill which is the highest point in the project area, see Figure 3. The
GPS data processing and adjustment for the reference point (P1) was conducted us-
ing the Bernese Software 4.2. In the adjustment procedure, the ITRF 2000 coordi-
nates of ISTA (IGS (International GPS Service) Station) were held fixed. The GPS
data for the rest of the points were processed by the Ashtech Solutions 2.60 Software
using the reference point (P1) as fixed. The horizontal and vertical (height) position-
ing precision of the points is obtained, on average, as 1 mm and 5 mm, respectively.
These results show that the GPS measurements are quite accurate and consistent (Telci
et al., 2006).

8. Results and discussion

In order to compare the results of the levelling methods, the height differences between
the points were separately determined, as are shown in Table 1. In these tables, ∆HBack is
backward levelling; ∆HFore is the forward levelling; ∆HMean is the mean of the backward
and forward levelling; ∆HTrig is the height difference from reciprocal and simultaneous
short range trigonometric levelling and ∆HGPS is the GPS levelling height differences
between the points.
 Evaluation of some levelling techniques in surveying application 369

Table 1. Comparison of height differences obtained from different methods (Note: GPS data comparison of
the height differences derived from digital, GPS and reciprocal and simultaneous short range trigonometric
levelling method

dPL−DL dPL−TL dPL−GPS

∆HPL ∆HDL ∆HTL ∆hGPS
[mm] [mm] [mm]
1–2 −40.56827 −40.5695 −40.566 −40.565 1.23 −2.27 −3.27
2–3 −17.79039 −17.7892 −17.788 −17.787 −1.19 −2.39 −3.39
3–4 −18.65703 −18.6574 −18.658 −18.660 0.37 0.97 2.97
4–5 −0.34028 −0.3403 −0.339 −0.344 0.02 −1.28 3.72
5–6 −0.93113 −0.9311 −0.933 −0.939 −0.03 1.87 7.87
6–7 −0.74409 −0.7447 −0.742 −0.741 0.61 −2.09 −3.09
7–8 −1.12021 −1.1204 −1.120 −1.122 0.19 −0.21 1.79
8–9 −1.57150 −1.5714 −1.573 −1.575 −0.10 1.50 3.50
9–10 −3.82522 −3.8254 −3.822 −3.820 0.18 −3.22 −5.22
10–11 −1.79897 −1.7994 −1.796 −1.795 0.43 −2.97 −3.97
11–12 −13.39333 −13.3934 −13.391 −13.391 0.07 −2.33 −2.33
12–13 −9.63998 −9.6397 −9.641 −9.642 −0.28 1.02 2.00
13–14 −5.35411 −5.3538 −5.352 −5.347 −0.31 −2.11 −7.11
14–15 2.77112 2.7711 pmi 2.768 2.767 0.02 3.12 4.12
15–16 −1.42499 −1.4249 −1.425 −1.426 −0.09 0.01 1.01
16–17 −0.17022 −0.1706 −0.170 −0.172 0.38 −0.22 1.78
17–18 2.59111 2.5912 2.594 2.588 −0.09 −2.89 3.11
18–19 4.36295 4.3632 4.361 4.369 −0.25 1.95 −6.05
19–20 1.71840 1.7184 1.715 1.720 0 3.40 −1.60
−105.88614 −105.8873 −105.878 −105.882
∑ Pd2 = 14.43 362.22 1144.63
σ1 KM = ±0.92 mm ±4.28 mm ±8.02 mm
σ5 KM = ±2.06 mm ±9.57 mm ±17.93 mm

9. Accuracy analysis of height determination techniques

This section covers the basics of the statistical theory used to determine the level of ac-
curacy for a survey. The standard deviations and the differences between paired observa-
tions for precise digital levelling and ordinary digital levelling are illustrated in Table 2.
The precise digital levelling measurements were processed for the 5 km double run lev-
elling; the standard deviation of a single run measurement of 250 m is ±0.25 mm/km;
the standard deviations of double run levelling of 1 km is ±0.17 mm. The ordinary digi-
tal levelling measurements were being processed for about 5 km back and fore levelling
range; the standard deviation of one measurement in 250 m is ±1.20 mm/km; the stan-
dard deviation of double run levelling in 1 km distance levelled is ±0.85 mm/km (these
370 Atınç Pırtı, Ramazan Gürsel Hoşbaş

values were calculated according to the Levallois formula based on the differences from
the main and the return measurement). The height differences determined by precise dig-
ital levelling were assumed as true values for the the comparison of the height differences
derived from digital, GPS and reciprocal and simultaneous short range trigonometric lev-
elling method. In Table 1, the d values could be considered the corrections (residuals) of
the DL, TL, and GPS (corr.) data to the PDL data taken a reference.

Table 2. Accuracy analysis of two height determination methods (precise digital levelling with Leica
DNA03 and (ordinary) digital levelling with TOPCON DL–102) and accuracy analysis and testing all height
determination methods (Note: GPS data corrected with geoid), (Telci et al., 2006)

Precise (Ordinary) Trigonometric GPS

Digital Levelling Digital Levelling Levelling Levelling
∆HBack ∆HBack
∆HFore ∆HMean ∆HFore ∆HMean ∆HTrig ∆HGPS
−105.88519 4 −105.8903
∑ ∆H
105.88706 −105.8861 105.8837 −105.8873 −105.878 −105.882
Misclosure = 1.87 mm Misclosure = 6 mm
√
TMisc = 4 K = 9 mm
(FGCS First Order Class I)
√
TMisc = 6 K = 13 mm
(FGCS Second Order Class I)
Line Length Weight Precise Digital Levelling (Ordinary) Digital Levelling
K [km] P [1/km] [1/K] D [mm] d [mm]
∑= 4.934 km ∑ Pd2 = 2.31 [mm2 /km] 54.09 [mm2 /km]
Single run 1 km precision ±0.25 mm ±1.2 mm
Double run 1 km precision ±0.17 mm ±0.85 mm
Double run 5 km precision ±0.55 mm ±2.65 mm
0.85 8.15 4.15
t= √ = 0.314 t= √ = 1.180 t= √ = 0.340
0.552 + 2.652 0.552 + 6.862 0.552 + 12.222
T {α = 0.05, f = 38} 2.024

In Table 1, the d values could be considered the corrections (residuals) of the DL,
TL, and GPS (corr.) data to the PDL data taken a reference deviation of the differences
1
between precise digital levelling value and digital levelling value. P value is for precise
S
1
and digital levelling, however P value is 2 for precise and digital levelling. After that,
S
we calculated the absolute t-value.
Assuming the PDL data as error free for the time being and equal weight of all
sections (all sections assumed to be 250 m long), then the standard deviation of the
 Evaluation of some levelling techniques in surveying application 371

DL, TL and GPS (corr.) √ height differences (over 250 m) can be calculated approx-
[dd]
imately as σ250 m = (n = 19). For Digital Levelling σDL(250 m) = ±0.46 mm,
n
for short range trigonometric Levelling σTL(250 m) = ±2.14 mm, for GPS Levelling
σGPS(250 m) = ±4.01 mm are calculated in Table 1. To get the approximate 1 km pre-
√
cision multiplying 4, that is 2. For DL σDL = ±0.92 mm, for short range trigonomet-
ric levelling σTL = ±4.28 mm, for GPS Levelling σGPS = ±8.02 mm is calculated in
Table 1.
In this study the t distribution is used to compare two different data test results. The
following sections provide the mathematical equations to calculate the t–test statistic.
We calculated the difference (dPL − dDL ) between the precise digital levelling value and
the ordinary digital levelling value for each set of test number, see Table 1. Then, we
calculated the arithmetic average (mean, d) of the differences between the precise digital
levelling value and the digital levelling value (last column) and the standard deviation of
the differences between precise digital levelling value and digital levelling value. After
that, we calculated the absolute t-value using Equation (4).
The sample size for observed differences is n = 19.

d
|t| = (4)
σd

where | | – absolute value (disregard sign), d – average of the differences between test
results, σd – standard deviation of the differences between test results, n – total number
of sections (sample size).
The comparison among the precise digital levelling method, digital levelling method,
trigonometric levelling and GPS levelling method has indicated no significant differ-
ences (the 95% confidence level) between accuracies the four techniques; refer to Table 2
(in the last rows) (Colorado Procedure, 2007).
The height differences can be determined with an accuracy of a few mm/km using
the levelling method named “Precision Trigonometric Levelling”. In Ceylan and Baykal
(2008) and Chrzanowski et al. (1985)√ demonstrate that accuracy can be achieved trigono-
metric levelling better than 1 mm K.
The effects of errors in digital levelling can be reduced by taking equal backward
and forward observation range, the round trip surveying, following BFFB or FBBF ob-
servation order or surveying calibration in lab and surveying additional parameters such
as pressure, temperature and time at the survey moment. Trigonometric levelling is “low
order” compared with conventional differential levelling. The reason for this is that dur-
ing a 3D traverse, the height of every single instrument and target set-up must be mea-
sured. These heights are usually measured somewhat crudely with a folding ruler, and
the accuracy of the results suffers directly from these imprecise height measurements.
The total station used in trigonometric levelling had to be controlled and calibrated be-
fore the measurements. Short range trigonometric levelling methods use two total sta-
tions mounted on points observing simultaneous reciprocal angles, requiring numerous
temperature and pressure measurements. In order to exploit the long range the modern
372 Atınç Pırtı, Ramazan Gürsel Hoşbaş

instruments, maintaining the desired accuracy, reciprocal and simultaneous observations

should be used so that errors due to the curvature and atmospheric refraction would
be minimised. The effect of the vertical on the zenith angles should, also, be investi-
gated in all cases. While the EDM Height Traversing technique is by no means new, the
advancement of equipment technologies now makes the use of Total station (The Au-
tomatic Target Recognition (ATR)) demonstrates its full benefits during routine repeat
measurements, e.g. monitoring, set measurements and measurement at two telescope
faces. Using the sight, the observer aligns the telescope roughly with the target point and
triggers a distance measurement. The total station automatically moves the telescope to
the centre of the prism measures the distance and corrects the angle 1 mm with deviation
to the centre of the prism (High constant accuracy independent of the observer, fatigue
free and quick, no focusing necessary, measures using any standard prism) a practicable
replacement for the normal spirit levelling. The Total station Levelling technique has
a number of benefits over normal spirit levelling. The elimination of collimation errors
and staff calibration errors and minimisation of refraction errors make the technique at-
tractive to those undertaking first order levelling. The use of significantly longer sight
lengths makes it attractive to everybody else. The technique does require slightly longer
observation periods per standpoint; however this is offset by fewer instrument stand-
points for the same length of run. The other significant advantage is that the Total station
already exists on the surveyor’s equipment list. There is no need for extra equipment
to complete highly accurate level runs. It may be said that GPS-derived height determi-
nation techniques give effective solution and accurate results which will meet the most
engineering projects’ requirements. However, this method requires geoid undulation or
geoid undulation differences.

10. Conclusions

In this study, four different levelling methods are compared. Geometric levelling is usu-
ally accepted as being more accurate than the other methods. The discrepancy between
geometric levelling and short range trigonometric levelling is at the level of 8 millime-
tres. The accuracy of the short range trigonometric levelling is due the reciprocal and
simultaneous observations of the zenith angles and slope distances over relative short
distances of 250 m. The difference between the ellipsoidal height differences obtained
from the GPS levelling used without geoid and the orthometric height differences ob-
tained from precise geometric levelling is 4 millimetres. The geoid model which is ob-
tained from a fifth order polynomial fit of the project area is good enough in this study.
The discrepancy between the precise geometric and GPS levelling (with geoid correc-
tions) is 4 millimetres over 5 km. This shows the necessity of an appropriate geoid model
which for the study area. It was seen that the short range trigonometric levelling and the
GPS levelling techniques give sufficiently accurate results when compared to geomet-
ric levelling. This study presented some practical solutions towards determination of the
heights of vertical control points in engineering surveying applications using different
techniques.
 Evaluation of some levelling techniques in surveying application 373

Acknowledgement

We thank Prof. Dr. Halil Erkaya (Okan University, İstanbul) and Prof. Dr. D. Uğur
Şanli (Yıldız Technical University, İstanbul) for comments that greatly improved the
manuscript. The research does not have external founds.

References
Anderson, J.M. and Mikhail, E.M. (1998). Surveying, Theory and Practice, 7th edition WCB/McGraw-
Hill, Boston.
Ayan, T. (2001). Geodetic Network Densification in Istanbul, IGNA (Istanbul GPS Network), IV. Turkish–
German Joint Geodetic Days, Berlin.
Ceylan, A. and Baykal, O. (2008). Precise height determination using simultaneous-reciprocal trigono-
metric levelling. Survey Review, 40 (308), 195–205. DOI: 10.1179/003962608X290997.
Chrzanowski, A., Greening, T., Kornacki, W., Second, J., Vamosi, S. and Chen, Y.Q. (1985). Applications
and limitations of precise trigonometric height traversing. Third International Symposium on the
North American vertical Datum, Rockville, Maryland, USA.
Colorado Procedure 14-03, Standard Practice for F and t-test Statistical Method for HBP Voids Ac-
ceptance, (2007). http://www.dot.state.co.us/designsupport/Field%20Materials%20Manual/2007/
CP%2014.pdf.
Featherstone, W.E., Dentith, M.C. and Kirby, J.F. (1998). Strategies for the Accurate Determination of Or-
thometric Height from GPS. Survey Review, 34(267), 278–296. DOI: 10.1179/sre.1998.34.267.278.
Federal Geodetic Control Committee (1998). Geometric Geodetic Accuracy Standards and Specifications
for Using GPS Relative Positioning Techniques, Version 5.0, U.S. Department of Commerce, May
11, 1988, reprinted with corrections, August 1.
Federal Geodetic Control Committee (1984). Standards and Specifications for Geodetic Control Net-
works, U.S. Department of Commerce, September.
Ingensand, H. (1999). The evolution of digital leveling techniques – limitations and new solutions. In
Lilje, M. (ed.) The Importance of Heights. FIG, Gävle, Sweden, 59–68.
Kahmen, H. (2000). Vermessungskunde, 20. völlig neu bearbeitete Auflage, Walter de Gruyter, 2000.
Kuang, S., Fidis, C. and Thomas, F. (1996). Modelling of local geoid with GPS and levelling: a case study.
Surveying and Land Information Systems, 56(2), 75–88.
Rüeger, J.M. and Brunner, F.K. (2000). On System Calibration and Type Testing of Digital Levels.
Zeitschrift für Vermessungswesen 4.
Rüeger, J.M. and Brunner, F.K. (1982). EDM Height Traversing Versus Geodetic Levelling. The Canadian
Surveyor, 36, 69–81.
Schofield, W. (2001). Engineering Surveying Theory and examination problems for students, 5th Edition,
Butterworth–Heinemann, Linacre House, Jordan Hill, Oxford, 521 pp.
Telci, G., Gürel, N., Bahşi, N.M., Hames, F.A., Uzun, I.A. and Inel, A. (2006). GPS Heightening (GPS ile
Yükseklik Tayini), Thesis, Yildiz Technical University, Istanbul – Turkey, 57 pp. (in Turkish).
Wolf, P.R. and Ghilani, C.D. (2002). Elementary Surveying, An Introduction to Geomatics, 10th Edition,
Prentice Hall Upper Saddle River, New Jersey, 887 pp.
Woschitz, H. and Brunner, F.K. (2002). Calibration of Digital Levels – Experimental Results of Systematic
Effects. Reprint of paper published in: INGEO2002, 2nd Conference of Engineering Surveying.
Kopacik A. and Kyrinovic P. (eds), Bratislava, November: 165–172.