COORDINATE SYSTEMS

IN GEODESY

E. J. KRAKIWSKY
D. E. WELLS

May 1971

TECHNICAL REPORT
LECTURE NOTES
NO.
NO. 217
16
COORDINATE SYSTElVIS IN GEODESY

E.J. Krakiwsky
D.E. \Vells

Department of Geodesy and Geomatics Engineering

University of New Brunswick
P.O. Box 4400
Fredericton, N .B.
Canada
E3B 5A3

May 1971
Latest Reprinting January 1998
 PREFACE

In order to make our extensive series of lecture notes more readily available, we have
scanned the old master copies and produced electronic versions in Portable Document
Format. The quality of the images varies depending on the quality of the originals. The
images have not been converted to searchable text.
 TABLE OF CONTENTS

page

LIST OF ILLUSTRATIONS iv

LIST OF TABLES ... vi

l. INTRODUCTION l

1.1 Poles~ Planes and -~es 4

1.2 Universal and Sidereal Time 6
1.3 Coordinate Systems in Geodesy ... 7

2. TERRESTRIAL COORDINATE SYSTEMS 9

2.1 Terrestrial Geocentric Systems • ... 9

2.1.1 Polar Motion and Irregular Rotation of the
Earth • . • • . . • • • • .. 10
2.1.2 Average and Instantaneous Terrestrial Systems • 12
2.1. 3 Geodetic Systems • • • • • • • • • • . 1 17
2.2 Relationship between Cartesian and Curvilinear
Coordinates • • • • • • • . • • 19

2.2.1 Cartesian and Curvilinear Coordinates of a

Point on the Reference Ellipsoid • • • • • 19
2.2.2 The Position Vector in Terms of the Geodetic
Latitude • • • • • • • • • • • • • • • • • • • 22
2.2.3 Th~ Position Vector in Terms of the Geocentric
and Reduced Latitudes . . • • • • • • • • • • • 27
2.2.4 Relationships between Geodetic, Geocentric and
Reduced Latitudes • . • • • • • • • • • • 28
2.2.5 The Position Vector of a Point Above the
Reference Ellipsoid . • • . • • • • • • . ..
.• 28
2.2.6 Transformation from Average Terrestrial
Cartesian to Geodetic Coordinates • 31

2.3 Geodetic Datums 33

2.3.1 Datum Position Parameters . 36

2.3.2 Establishment of a Dat'lllli • . 4o
=2. 3. 3 The North· American "Datum .• : . • · • • • 42
2.3.4 Datum Transformations • • • • 45

2.4 Terrestrial Topocentric Systems 48

2.4.1 Local Astronomic System . .. 50
2.4.2 Local Geodetic System • 54
2.5 Summary of Terrestrial Systems 57
ii
 page

3. CELESTIAL COORDINATE SYSTEMS • • 61

3.1 The Ecliptic System ••. • • • • • • .... 63

3.2 The Right Ascension System 65
3.3 The Hour Angle System • • • 67
3.4 The Horizon System • • • • • • • • • 71
3.5 Variations of the Right Ascension System 73

3.5.1 Precession and Nutation 74

3.5.2 Mean Celestial Systems • 76
3.5.3 The True Celestial System 81
3.5.4
3.5.5
The Apparent Place System
The Observed Place System
.. ·~
83
84

3.6 Transformation between Apparent Celestial and

Average Terrestrial Coordinate Systems • • • • • • 85
3.7 Summary of Celestial Systems • • • • • • • 87

4. THE ORBITAL COORDINATE SYSTEM 92

•
4.1 The Orbital Ellipse and Orbital. Anomalies • • • • • 92
4.2 The Orbital Coordinate System •. • • • • • · 96
4.3 Transformation from Orbital to Average Terrestrial
System. . . • • • -. • • . . • • . . • . • • . 98
4.4 Variations in the Orbital Elements 99
4.5 The Satellite ·subpoint • • • • • • • .•.• 99
4.6 Topocentric Coordinates of Satellite • • 101
•
5. SUMMARY OF COORDINATE SYSTEMS 104

5.1 Terrestrial Systems • • • • • • • • • 104

5.2 Ce1estial Systems • • • • 106
5.3 Duality Paradox in the Apparent and Observed
Celestial Systems • • • • • • • • • • • • • • • 1"07
5.4 The Connections between Terrestrial, Celestial and
Orbital Systems • • • • • • • • •. • • • • • • • • • 108

REFERENCES . 109
r
APPErlDIX A: Swmnary of Reflection and Rotation ll.atrices 111

iii
 LIST OF ILLUSTRATIONS

Figure tio. Title Page

l-1 Terrestrial, Celestial, and OrbitaL

Coordinate Systems 2

l-2 Types of Coordinate Systems 3

2-l Polar Motion 11

2-2 Position of Point Moving Uniformly Along

Equator ~linus Position on Actual Equator 13

2-3 Terrestrial and Geodetic Coordinate Systems 14

2-4 Transformations from Instantaneous to Average
Terrestrial System 16
2-5 Reference Ellipsoid 20

2-6 Various Latitudes 23

2-7 Tangent Line to the Meridian Ellipse 24

2-8 Point Above Reference Ellipsoid 29
2-9 Meridian Section of the Earth 3liA

2-10 Orientation of Ellipsoid to Geoid 37A

.& I • a

2-12 Geodetic and Local Geodetic Coordinate

Systems 51
2-13 Equations Relating Terrestrial Systems 59

3-1 Ecliptic System 64

3-2 Right Ascension System 66

3-3 Hour Angle System 68

3-4 Time, Longitude, and Right Ascension 70

iv
·Figure No. Title .:. Page

3-5 Horizon Sys.tem 72

3-6 Variations of the Celestial Right Ascension
..
System 75
3-7 Motion of Celestial Pole 77
3-8 The Effect of Precession and Nutation 78
3-9 Mean Celestial Coordinate Systems 80
3-10 True.and Mean Celestial Coordinate Systems 82
3-ll Celestial Coordinate Systems 91

4-l Orbit Ellipse 93

4-2 Keplerian Orbital Elements 97

4-3 Satellite Subpoint 100

4-4 Topocentric Coordinates of Satellite 102

5-l Coordinate Systems 105

v
 LIST OF TABLES

Table No. Title Page

2-1 Parameters Defining the 1927 North American 44

Datum

2-2 Translation Components 44

2-3 Example of Datum Transformations 49

2-4 Reference Poles~ Planes and Axes Defining
Terrestrial Coordinate Systems 58

2-5 Transformations Among Terrestrial Coordinate

Systems 60

3-1 Reference Poles, Planes and Axes Defining

Celestial Coordinate Systems 89
3-2 Transformations Among Celestial Coordinate
Systems 90

vi
 l. INTRODUCTION

These notes discuss the precise definitions of, and transformations

between, the coordinate systems to which coordinates of stations on or

above the surface of the earth are referred. To define a coordinate

system we must specify:

a) the location of the origin,

b) the orientation of the three axes,

c) the parameters {Cartesian, curvilinear) which define the

position of a point referred to the coordinate system.

The earth has two different periodic motions .in space. It rotates

about its axis, and it revolves about. the sun (see Figure 1-1). There

is also one natural satellite (the moon) and many artificial satellites

which have a third periodic motion in space: orbital motion about the

earth. These periodic motions are fundamental to the definition of

systems of coordinates and systems of time.

Terrestrial coordinate systems are earth fixed and rotate with the

earth. They are used to define the coordinates of points on the surface

of the earth. There are two kinds of terrestrial systems called

geocentric systems and topocentric systems (see Figure 1-2).

Celestial coordinate systems do not revolve but ~ rotate with the

 *STAR
""'·
-- -------- OBSERVER ON EARTH

1\)

SUN
EARTH•s
ROTATtON
ORBIT

/
EARTH'S OAIIT

TIER RESTRtAb , CELI!STt'Al Att:n ""'1JTAL COOROIN·ATE SYSTEMS.

•. Fi.gur7 1-l~~t
 3

Figure l-2. Types of Coordinate Systems.

Terrestrial

I
I l

Geocentric ~ Topocentric
' ~

I
i Celestial
I
I I I l

Ecliptic '
f-+ Right Ascension
I
Hour Angle Horizon

L...- J Orbital
l~....---___.
earth. They are used to define the coordinates of celestial bodies

such as stars~ There are four different celestial systems~ called the

ecliptic, right ascension~ hour angle~ and horizon systems.

The orbital system does not rotate with the earth~ but revolves

with it. It is used to define the coordinates of satellites orbiting

around the earth.

l.l POLES , PLANES AND AXES .

The orientation of axes of coordinate systems can be described in

terms of primary and secondary poles~ primary and secondary planes, and

primary~ secondary and tertiary axes.

The primary pole is the axis of symmetry of the coordinate system,

for example the rotation axis of the earth. The primary plane is the

plane perpendicular to the primary pole, for example the earth's

equatorial plane. The secondary plane is perpendicular to the

primary plane and contains the primary pole. It sometimes must be

chosen arbitrarily,.for example the Greenwich meridian plane, and

sometimes arises naturally, for example the equinoctial plane. The

secondary pole is the intersection of the primary and secondary planes.

The primary axis is the secondary pole. The tertiary axis is the primary

pole. The secondary axis is perpendicular to the other two axes,

chosen in the direction which makes the coordinate system either right-

handed or left-handed as specified.

We will use either the primary plane or the primary pole, and the

primary axis to'specify the orientation of each of the coordinate

systems named above.

 5

For terrestrial geocentric systems:

a) the origin is near the centre of the earth,

b) the primary pole is aligned to the earth's axis of

rotation, and the primary plane perpendicular to this pole is called

the equatorial plane,

c) the primary axis is the intersection between the

equatorial plane and the plane containing the Greenwich meridian,

d) the systems are right-handed.

For terrestrial topocentric systems:

a) the origin is at a point near the surface of the earth,

b) the primary plane is the plane tangential to the surface

of the earth at that point,

c) the primary axis is the north point (the intersection

between the tangential plane and the plane containing the earth's north

rotational pole),

d) the systems are left-handed.

For the celestial ecliptic system:

a) the origin is near the centre of the sun,

b) the primary plane is the plane of the earth's orbit,

called the ecliptic plane,

c) the primary axis is the intersection between the ecliptic

plane and the equatorial plane, and is called the vernal equinox,

d) the system is right-handed.

For the celestial right ascension system:

a) the origin is near the centre of the sun,

b) the primary plane is the equatorial plane,

c) the primary axis is the vernal equinox,

 6

d) the system is right-handed.

For the celestial hour angle system:

a) the origin is near the centre of the sun,

b) the primary plane is the equatorial plane,

c) the secondary plane is the celestial meridian (the plane

containing the observer and the earth's rotation axis),

d) the system is left..-handed.

For the celestial horizon system:

a) the origin is near the centre of the sun,

b) the primary plane is paralle1 to the tangential plane at

the observer (the horizon plane),

c) the primary axis is parallel to the observer's north point,

d) the system is left-handed.

For the orbital system:

a) the origin is the centre of gravity of the earth,

b) the primary plane is the plane of the satellite orbit

around the earth,

c) the primary axis is in the orbital plane and is oriented

towards the point of perigee (the point at which the satellite most

closely approaches the earth) and is called the line of apsides,

d) the system is right-handed.

1.2 UNIVERSAL AND SIDEREAL TIME

Also intimately involved with the earth's periodic rotation and

revolution are two systems of time called universal (solar)· time (UT)

and sidereal time (ST). A time system is defined by specifying an

 7

interval ~,d an ~= The solar day is the interval between successive

passages of the sU."l over the sa-ne terrestrial meridia."l. The sidereal

day is the interval between two successive passages of the vernal

equinox over the same terrestrial meridian. The sidereal epoch is the

angle between the vernal equinox and some terrestrial meridian: if

this is the Greenwich meridian then the epoch is Greenwich Sidereal

Time (GST). The solar epoch is rigorously related to the sidereal

epoch by a mathematical formula. Sidereal time is the parameter

relating terrestrial and celestial systems.

1.3 COORDINATE SYST&~S IN GEODESY

Geodesy is the study of the size and shape of the earth and the

determination of coordinates of points on or above the earth's surface.

Coordinates of one station are determined with respect to

coordinates of other stations by making one or more of the following

four categories of measurements: directions, distances, distance

differences, and heights. Horizontal and vertical angular measurements

between two stations on the earth (as are measured by theodolite for

example) are terrestrial directions. Angular measurements between a

station on the earth and a satellite position (as are measured by

photographing the satellite in the star background for example) are

satellite directions. Angular measurements between a station on the

earth and a star (as are measured by direct theodolite paintings on the

star for example) are astronomic directions. Distances between two

stations on the earth (as are measured by electromagnetic distance

 8

measuring instruments for example) are terrestrial distances. Distances

between a station on the earth and a satellite position (as are measured

by laser ranging for example) are satellite distances~ Measurements

of the difference in distance between one station on the earth and two

other stations (as are measured by hyperbolic positioning systems for

example) are terrestrial distance differences. Measurements of the

difference in distance between one station on the earth and two satellite

positions (as are measured by integrated Doppler shift systems for

example) are satellite distance differences. All these measurements

determine the geometrical relationship between stations, and are the

subject of geometric geodesy [e.g. Bamford 1962].

Spirit level height differences and enroute gravity values are

measurements related to potential differences in the earth's gravity

field, and are the subject of physical geodesy [e.g. Heiskanen and

Moritz 1967].

The functional relationship between these measurements and the

coordinates of the stations to and from which they are made is

incorporated into a mathematical model. A unique solution for the

unknown coordinates can be obtained by applying the least squares

estimation process [Wells and Krakiwsky 1971) to the measurements

and mathematical model.

Details on coordinate systems as employed for terrestrial and

satellite geodesy can be found in Veis [1960] and Kaula [1966], and

for geodetic astronomy in Mueller [1969].

 9

2. TERRESTRIAL COORDINATE SYSTEMS

In this chapter we will discuss terrestrial geocentric and

terrestrial topocentric coordinate systems.

We first discuss terrestrial geocentric systems using only

Cartesian coordinates, and considering in detail what is meant by "the

earth's axis of rotation" and "the Greenwich meridian". Then the

relationship between Cartesian and curvilinear coordinates is described.

Geodetic datums are discussed. Finally terrestrial topocentric systems

are considered, with .attention paid to what is meant by "the surface

of the earth".

2.1 TERRESTRIAL GEOCENTRIC SYSTEMS

In the introduction it was stated that for terrestrial geocentric

systems:

a) the origin is near the centre of the earth,

b) the primary pole is aligned to the earth's axis of

rotation,

c) the primary axis is the intersection between the primary

plane and the plane containing the Greenwich meridian,

d) the systems are right-handed.

 10

The last specification is unambiguous. As we shall see the other

three are not. We will first discuss problems in defining the earth's

axis of rotation and the Greenwich meridian. The~ we will discuss

translations of the origin from the centre of the earth.

2.1.1 Polar Motion and Irreg~lar Rotation of the Earth

We think of the earth as rotating about a fixed axis at a uniform

rate. In fact, the axis is not fixed and the rate is not uniform.

Over 70 years ago, it was discovered that the direction of the

earth's rotation axis moves with respect to the earth's surface. This

polar motion is principally due to the fact that the earth's axes of

rotation and maximum inertia do not coincide. The resultant motion is

irregular but more or less circular and counterclockwise (when viewed

from North), with an amplitude of about 5 meters and a main period of

430 days (called the Chandler period}.

Two international organizations, the International Polar Motion

Service (IPMS) and the Bureau International de l'Heure (BIH) routinely

measure this motion through astronomic observations; the IPMS from five

stations at the same latitude, and the BIH from about ~0 stations

scattered worldwide. The results are published as the coordinates of

the true rotation axis with respect to a reference point called the

Conventional International Origin (CIO) which is the average position

of the rotation axis during the years 1900-1905( IUGG (1967) Bull Geed

86, 379 (1967) Resolution 19). Figure 2-1 shows the polar motion during 1969
as determined by IPMS and BIH.

Over 30 years ago irregularities in the rotation of the earth were

 JAN. i969
-5
DEC.I969

/
I
I
Y--~-----r--------~~--~,o--------------~5~--~----~c~~~r--
, 15 METERS
fTOWAROS 90° ) {
'WEST LONGITUDE \

\
',>---/ IPMS
/
/
5

X
Figure 2-l. . POLAR MOTION ( TO.....a6 Gftftft'WI'Qt)
 12-

discovered (other than polar motion). There are three types of

i~regularities: seasonal variations probably due to meteorological

changes or earth tides; secular decrease due to tidal friction; and

irregular fluctuations [Mueller 1969]. The seasonal variation is the

only one of these presently being taken into account, and it is more or

less reproducible from year to year, and produces a displacement along

the equator of up to 15 meters with respect to a point rotating uniformly

throughout the year (see Figure 2-2).

Because of this seasonal variation, the Greenwich m"eridian (the

plane containing the earth's rotation axis and the center of the transit

instrument at Greenwich Observatory) does not rotate uniformly.. A

ficticious zero meridian which does rotate uniformly (so far as the

effects of polar motion and seasonal variations are concerned) is

called the Mean Observatory o~ Greenwich mean astronomic meridian.

Its location is defined by the BIH.

2.1.2 Average and Instantaneous Terrestrial Systems

The average terrestrial (A. 'I'.) system is the ideal world geodetic

coordinate system (see Figure 2-3):

a) Its origin is at the centre of gravity of the earth.

b) Its primary pole is directed towards the CIO (the

average north pole of 1900-1905), and its primary plane is the plane

perpendicular to the primary pole and containing the earth's center of

gravity (the average equatorial plane).

c) Its secondary plane is the plane containing the primary

 l3

+10 .

METERS

-10

JAN.1967 APR. JULY OCT. JAN. 1968

POSITION OF POINT MOVING UNIFORMLY ALONG

Figure 2-2 •., EQUATOR MINUS POSITION OF POtNT ON ACTUAL
EQUATOR.
 ROTATION AXIS
OF EARTH

AVERAGE

GREENWICH
MEAN
MERIDIAN

CENTRE OF
GRAVITY

X
Figure ..:_-~ ..

TERRESTRIAL AND GEODETIC COORDINATE SYSTEMS

 15

pole and the Mean Observatory. The intersection of these two planes is

the secondary pole, or primary axis.

d) It is a right-handed system.

We can then define the position vector Ri of a terrain point i in

terms of its Cartesian coordinates x, y, z as

y 2-1

z A.T.
The instantaneous terrestrial {I.T.) system is specified as follows:

a) Its origin is at the centre of gravity of the earth.

b) Its primary pole is directed towards the true (instan-

taneous)rotation axis of the earth.

c) Its primary axis is the intersection of the primary plane

and the plane containing the true rotation axis and the Mean Observatory.

d) It is a right-handed system.

The main characteristic of these two systems is that they are

geocentric systems having their origins at the centre of gravity of the

earth and the rotation axis of the earth as the primary pole.

By means of rotation matrices [Thompson 1969; Goldstein 1950;

Wells 1971] the coordinates of a point referred to the instantaneous

terrestrial system are transformed into the average system by the

following equation (see Figure 2-4):

r:lz
X

= R2 ( -xp )' R ( -y ) y 2-2

1 p
z
'
A.T. I.T.
 16

y - . . . . . - - - , - - - - - - - - • CIO y -~-----------•,.. CtO

I
I
I
"P I
I
I
I
~~---~p-------
INSTANTANEOUS
TERRESTRIAL
POLE

Figure 2-4

TRANSFORMATION FROM INSTANTANEOUS TO AVERAGE

TERRESTRIAL SYSTEM.
 17

where (xp , y p } are expressed in arcseconds, and the rotation matrices

are

1 0 0

Rl (-yp) = 0 cos{-y )
p
sin(-y )
p
0 -sin(-y ) cos(-y )
p p

a clockwise (negative) rotation about the x axis, and

cos(-x ) 0 -sin(-x }
p p
R2 (-xp) = 0 1 0

sin(-x } 0 cos(-x }
p p

a clockwise (negative) rotation about they-axis. The inverse is

X X

z z
I.T. A.T.

and because of the orthogonal characteristic of rotation matrices, that is

-1
R (a) = RT (a) = R (-e)

X X

y y • 2-3

z z A.T.
I.T.

2.1.3 Geodetic Systems

In terms of Cartesian coordinates, the geodetic (G) coordinate

system is that system which is introduced into the earth such that its

three axes are coincident with or parallel to the corresponding three

 18

axes of the average terrestrial system (see Figure 2-3). The first

situation defines a geocentric geodetic system while the latter non-

geocentric system is commonly referred to as a relative geodetic s~stem,

whose relationship to the average terrestrial systeo is given by the

three datum translation components

X
0

r
0
=
z
0

and in vector equation form, the relationship is

Ri = ro + ri

where the position vector ri is referred to the geodetic system, that is

z
G

and·

X X X
0

y = 2-4
z z z
A.T. 0 G

A more detailed account of how a relative geodetic system is

established within the earth is in order {Section 2.3),but before this

can be done, it would be useful to review the relationship between

Cartesian and curvilinear coordinates.

 19

2. 2 RELATIONSHIP BETWEEN CARTESIAN AliD CURVILINEAR COORDIIIATES

In this section we first describe the Cartesian (x, y, z) and

curvilinear (latitude, longitude, height) coordinates for a point ·on the

reference ellipsoid. We then develop expressions for its position

vector in terms of various latitudes. Finally the transfOrmation tram

geodetic coordinates (41, A:, h) to (x, y, .z) and its inverse are

discussed.

2.2.1 Cartesian and Curvilinear Coordinates of a Point

on the Reference Ellipsoid

. The specific ellipsoid used in geodesy as a reference surface is

a rotational ellipsoid formed from the rotation of an ·e~ipse about its

semi-minor axis b (Figure 2-5). The semi-ma.lor axis a and -the

flattening

f = .;;;;a~-_b;;. 2-5
a

are the defining parameters of the reference ellipsoid.

Other useful parameters associated with this particular ellipsoid

are the first eccentricity

2
e = '

and the second eccentricity

2-7

A Cartesian coordinate system is superimposed on the reference

ellipsoid (see Figure 2-5) so that:

 20

i!

Figure 2-5. REFERENCE ELLIPSOID

 21

a) The origin of the Cartesian system is the centre of the

ellipsoid.

b) The primary pole (z-axis) of the Cartesian system is the

semi-minor axis of the ellipsoid. The primary plane is perpendicular

to the primary axis and is called the equatorial plane.

· c) Any plane containing the semi-minor axis and cutting

the surface of the ellipsoid is called a meridian plane. The particular

meridian plane chosen as the secondary plane is called the Greenwich

meridian plane. The secondary pole (x-axis) is the intersection of

the equatorial plane and the Greenwich meridian plane.

d) The y-axis is chosen to form a right-handed system, and

lies in the equatorial plane, 90° counterclockwise from the x-axis.

The equation of this ellipsoid, in terms of Cartesian coordinates

is

1? sE x = 1 '
2-8
where
-T
X = (x y z]
'

l/a2 0 0

SE = 0 l/a2 0 2-9
'
0 0 l/b 2

or
x2 + Y2 2
z
2
+
b2
=1 • 2-10
a

The latitude of a point is the acute angular distance between the

equatorial plane and the ellipsoid normal through the point measured in

the meridian plane of the point. The line perpendicular to the ellipsoid

at a point is called
 22

the ellipsoid normal at the point. Ellipsoid normals only pass through

the geometric center of the ellipsoid in the equatorial plane or along

the semi-minor axis. Therefore there are two different kinds of

latitude. The angle between the ellipsoid normal at the point and the

equatorial plane is called the geodetic latitude ~. The angle between

the line joining the point to the centre of the ellipse, and the

equatorial plane is called the geocentric latitude ~- There is also a

third latitude, used mostly as a mathematical convenience, called the

reduced latitude 6 (see Figure 2-6).

The longitude A of a meridian plane is the counterclockwise angular

distance between the Greenwich meridian plane and the meridian plane of

the point, measured in the equatorial plane (see Figure 2-5).

The ellipsoid height h of a point is its linear distance above the

ellipsoid, measured along the ellipsoidal normal at the point (see Figure 2-8}.

2.2.2 The Position Vector in Terms of the Geodetic Latitude

Consider a point P on the surface of the ellipsoid. The coordinates

of P referred to a system with the primary axis (denoted x*) in the meridian

plane of P are

2-11
r = 0

Zj
The plane perpendicular to the ellipsoid normal at P, and passing

through P is called the tangent plane at P. From Figure 2-7 the slope

of the t.angent plane is

dz cosp
dx* sinljl 2-12
 23

V GEOCENTRIC LATITUDE
+ GEODETIC LATIYI"uVE
.. . .
/J REDUCED LATITOOE
P ACTUAL POINT
Q,R PROJECTED POINTS

Figure 2-6.
. VARIOUS LATITUDES .
 24

p•

Figure 2-T. TANGENT LINE TO THE MERIDIAN ELLIPSE

 25

The slope can also be computed from the equation of the meridian ellipse

as follows:

2-13

or

b 2 ( x *)2 +a 2 z 2 =a2 b2 2-14

2-15

dz 2
b x*
dx* =- 2 2-16
a z

It follows from·~he above two equations for the slope, that

2 .
b x* _ cos;
2 - 2-17
a z sin~
or

b1 (x*) sin ~ = a1 z cos ~ 2-18

and after squaring the above

4 2 . 2
b (x*) s~n ~ = a4 z
2 2
cos ~ 2-19

Expressing equations 2-14 and 2-19 in matrix form

. 2-20

The inverse of the coefficient matrix is

a2
1 [
2
ao 2(a2 cos 4>·+ b 2 sin 2~ )
2 -b2

a
2
cos
2~
~
1
+
b2 . 2~
s~n ~
[a
4
4 2
b sin $
cos 2~]
and finding the square root

[x*J =
z (a2 cos 2 ~ 1+ b2 s~n
. 2~)1/2
"'.
2-21
~·rom Figure 2-6

cos ¢." = x*
N

but from equation 2-21

2
a cos$
x* = --~~~~~----------~-
2 2 2 n , jn
• • .::; ) .i.{ t::.
(
a cos ~ + o sin ~

therefore

2
N = ------~a~-------------
( a 2cos 2~i' + b2 s1.n
. 2~)1/2
'I' 2-22

r =
x*

0
= r :os$ N
2-23

z LN b 2 /a2 sin$

N is the radius of curvature of the ellipsoid surface in the plane

perpendicular to the meridian plane (called the Prime vertical plane).

We now refer the position vector P to a system with the primary axis

in the Greenwich meridian plane, that is we rotate the coordinate system

about the z-axis clockwise (negative rotation) through the longitude A.

X x*

r = y = R3 (->.) 0

z z

cos(->.) sin(->..) 0 N cos4>

= -sin(->.) cos(-/c) 0 0
2
0 0 1 N b 2/a sin4>
 27

or

X cos' COSA

r = y = N cos' sinA 2-24

z b2/a2 sin'

2.2.3 The Position'Vector in Terms of the Geocentric and Reduced Latitudes

From Figure 2-6 the position vector of the point P in terms of the

geocentric latitude ~ is

x* ·cos~

r = 0 = lrl 0

z sin~

where lrl is the magnitude of r.

Rotating the coordinate system to introduce longitude as before,

X x* cos~ cosA

r= y 0 = lrl cos~ sinA 2-25

z z sin~

From Figure 2-6 the reduced latitude 8 of the point P is the

geocentric latitude of both the points Q and R, where Q is the

projection of P parallel to the semi-minor axis to intersect a circle

with radius equal to the semi-major axis, and R is the projection of the

point P parallel to the semi-major axis to intersect a circle with

radius equal to the semi-minor axis.

The position vector of P in terms of the reduced latitude 8 is

x* a cos8

r= 0 = 0

z b sinS·
 28

Rotating the coordinate system to introduce longitunes

X x* a case COSA

r = y = R3(-A.) 0 = a case sin>.. 2-26

z z b sine

2.2.4 Relationships between Geodetic, Geocentric and Reduced Latitudes

From equations 2-24, 2-25, and 2-26

-Xz = tan$ cosA. =tan~

b
cosA. =-tanS cosA..
a

Cancelling the cos A. term,

b
tanS =-r tan$
a ' 2-27
a
tanS = tamjl 2-28
b '
b2
tantjl = - tancjl
2
.
a

2.2.5 The_Position Vector of a Point Above the Reference Ellipsoid

Let us consider a terrain point i, as depicted in Figure 2-8,

whose coordinates are the geodetic latitude $ and longitude >.., and

the ellipsoid height h. The projection of i onto the surface of the

ellipsoid is along the ellipsoidal normal defined by the unit vector u •

z
The position vector of i is then the sum of two vectors, namely

r.
1.
=rp + h uz s 2-30

where r is defined by equation 2-24 and u is the unit vector defined

p z
by equation 2-68c, that is
 29

TERRAIN POINT

Figure 2-8. POINT ABOVE REFERENCE ELLIPSOtO

 30

COS$ COSA

U
z
= COS$ SinA

sin$

Thus

X cos$ COSA cos$ COSA

ri = y =N cos$ sinA. + h cos$ sinA.

z b2/a2 sin$ sin$

or
~

(N+h) cos$ COSA -

= (N+h) cos$ sinA. 2-31

z sin$
I. - -
Now the position vector -
r. in equation 2-31 refers to a coordinate
1

system whose origin is at the geometrical centre of the ellipsoid. If

this ellipsoid defines a·relative geodetic system, then its centre will

not in general coincide with the centre of gravity of. the earth. The

expression for the position vector in the average terrestrial system is,

from equation 2-4

X
0

(ri)A.T. = + ("I\)
'l. G
z
0

or

X X (N+h) cos$ COSA

0
• 2-32
cr.>
1
= y = + (N+h) cos$ sinA.
A.T.
z z (Nb 2 /a 2+h) sin$
0

This expression gives the general transformation from relative

geodetic coordinates ($, A., h) to average terrestrial coordinates

 31

(x, y, z), given the size of the ellipsoid (a, b) and the translation

components (x y z )
o' o' o •

2.2.6 Transformation from Average Terrestrial Cartesian to

Geodetic Coordinates

A very useful transformation is the inverse of equation 2-32.

Given the average terrestrial coordinates (x, y, z), the translation

components (x , y , z ), and the size of the ellipsoid (a, b), compute

0 0 0

the relative geodetic coordinates (,, A, h).

First we translate the origin from the centre of gravity to the

centre of the ellipsoid. From equation 2-32

X X X
0

y = y 2-33

z z z
G A.T. 0

The longitude A is computed directly from

X = tan-l (~) ·I 2-34

The latitude ' and ellipsoid height h are more difficult to compute

sine~ N is a function of~. from equation 2-22

2-35

and h is not known. We begin by computing

e2 =1 - b 2 /a2 2-36

p = (X 2 + y
2)1/2 2-37
 32
From equation 2-31
2
p =
p = (N+h) cos<P

J
or p

Also from 2-31 I h

z
=

= (N
coscj> - N

b 2 /a2 + h)
2-38

a2-b2
= (N - 2 N + h) sin$
a
= (N + h - e 2 N) sincfl .

Therefore
-pz = (N·+ h-e~) sinp _ t
- ancf!
(l e~)
-- 2-39
(N + h) coslj> N+h
This equation can be developed in two wayss to produce either a direct

solution for 4> which is quite involved, or an iterative solution which is

simpler. We consider the iterative solution first. We have

0 = tan-1 [<~) (1 - ;~) 1

- ] •

The iterative procedure is initiated by setting

N0 =a

Each iteration then consists of evaluating in order

a
N.
J.
= 2 . 2q, . )1/2
cos <fl.J.- 1 + b2/a2 s1.n 1.-1

= E ~ N.
hi 1.
cosq,i-l

cj>i = tan-1 [t· 2

eN.
p) (1- N.+h.
1.
)-1] .
J. J.
 33

The iterations are repeated until

(h.l. - h.l.-1) < 8£

and

..:.10
for some appropriately chosen value of E (for example E= 10 for double

precision Fortran on the IBM 360 computer).

Returning to equation 2-39, we eliminate h using equation 2-38

to obtain
.!. = tan ~(l _ e~ cosp )
p p
or
p tan~ - z = e~ sin~ •
In this equation the only unknown is $. We will now modify this

equation to obtain an equation which can be solved for tan $. Substituting

the expression for·N from equation 2-35 we have

a e 2 sinp

Dividing the numerator and denominator of the right hand side by cos
2
p tan ~ - z = .;;;a.....;..e_t.-a.n~-4>~-~~~~-­
(1 + b2/a2 tan2~ )l/2

or
2
ae tan ~

Squaring this equation to eliminate the square root

(p2 tan 2 ~ - ~ p z tan~ + z2 ) (1 + (l - e 2 ) tan 2 ~)

4
= a2 e tan
2
<f;
or
2
2 4 3 2 2 + ---'-z_ _ =0
p tan ~ -2 p z tan ~ + (3 +z ) tan ~
(l - e 2 )
where
3 -
 34

This is a quartic (biquadratic) equation in tan' , in which the

values of all coefficients are knowh. Standard procedures for solving

quartic equations exist (see for example Kern and Kern, 1968), and have

been applied to this equation by Paul (1973), to prod~ce a computer

program which is about 25% faster than iterative programs. Once a solution

for tan~ is obtained, N and h are computed from equations 2-35 and 2-38

respectively.

2.3 GEODETIC DATUMS

There are two natural figures of the earth (see Figure 2-9);

the topographic or physical surface of the earth including the surface of

the oceans (the terrain), and the equipotential surface of the earth's

gravity field which coincides with an idealized surface of the oceans (the

geoid}.

Control measurements (e.g. distances, angles, spirit levelling)

are made between points on the terrain which we call control points. These

measurements are used to determine the geometrical relationship between

the control points in a computation called network adjustment. Other points

are then related to the network of control points through further measurements

and computations called densification. The classical approach is to treat

the vertical measurements, networks and computations aeparately from the

horizontal measurements, networks and computations. However the unified

three dimensional approach is currently gaining favour. {Hotine, 1969].

In the classical approach vertical measurements·and networks are

referred to a coordinate surface or (vertical) datum which is th~ geoid.

Rather than using the geoid as the coqrdinate surface or datum for the

horizontal measurements and networks as well, a third, unnatural figure of

 34A

--~= -~ DEFL~CTI~N OF
J ~ THE VERT-ICAL

AXIS AXIS OF
OF EARTH ELLIPSOID

CENTRE OF
ELUPSOID.,_-~.._-¥-_ _ _ _ _ _.__.......;._ _....,._

LATITUDE~

CENTRE OF
GRAVITY

Figure 2-9. _ MERIOlAN SECTION -OF TH£ E·ARTH

 35

the earth is introduced - the ellipsoid of rotation discussed earlier.

The reason a mathematical figure like the ellipsoid is used as the horizontal

~ is to simplify the computations required both for network adjustment

and densification.

Correction terms are necessary in these computation to account

for the fact that the datum is not the geoid. An ellipsoid can be chosen
to approximate the geoid closely enough that.these correction terms can

be assumed linear and for some applications even ignored. For a well-chosen

ellipsoid (see Figure 2-9), the geoid-ellipsoid.separation (geoid height)

is always less than 100 metres, and the difference between the geoid and

ellipsoid normals at any point (deflection of the vertical) is usually le~s

than 5 arc seconds, very rarely exceeding 1 arcminute.

Even simpler surfaces than the ellipsoid (such· as the sphere

or the plane} can be sufficient approximationsoto the.geoid if the area

under consideration is sufficiently small, and/or the control application

permits lower orders of accuracy.

s
The introduction of a new surface (the ellipsoid) has a price.

The horizontal control network (that is . the coordinates of the points of the

network) is to be referred t? the eilipsoid. Therefore before network

computations can begin, the control measurements must first be reduced so

that they too "refer" to the ellipsoid.

It is important to distinguish between the datum (the coordinate

surface or ellipsoid surface) and the coordinates of the points of the net-

work referred to the datum. It is a common but confusi_ng practice (part-

icularly in North.America) to use the term "datum" for the set of coordinates.
 36

2.3.1 Datum Position Parameters

In order to establish an· ellipsoid as the reference surface for

a system of control we must specify its size and shape (usually by assigning

values to the semi-major axis and flattening) and we must specify its pos-

ition with respect to the ea1;th. A well~positioned ellipsoid will closely

approximate the geoid over the area covered by the network for which it is

in datum. The parameters to vhich we assign values in order to specify the

ellipsoid position we call the datum position parame;;ers.;·

In three-dimensional space~ any figure {and particularly our

ellipsoid) has six degrees of freedom~ that is six ways in vhich its posi-

tion with respect to a fixed figure (in our case the earth) can be changed.

Thus there are -six datum position parameters.

Another vay of looking at this is to consider tvo three-dimensional

Cartesian coordinate systems, one fixed to the ellipsoid and one fixed to

the earth. In general the origins of the two systems will not coincide,

and the axes vill not be parallel. Therefore, to define the transformation

from one system to the other we must specify the location of one origin

with respect to the other system, and the orientation of one set of axes

with respect to the other system, that is three coordinates, and three

rotation angles. These six parameters provide a description of the six

degrees of freedom and _assigning values to them positions the ellipsoid

with respect to the earth. They are our six datum position parameters. A

datum then is completely specified by assigning values to eight parameters

the ellipsoid size and shape, and the six datum position parameters.
 37

There are in fact two kinds of datum position parameters in

use. One kind is obtained by considering the ellipsoid-fixed and earth-

. >v.·

fixed coordinate systems to have their origins in the neighbourhood of the

geocentre. The other kind is obtained by considering the ellipsoid-fixed

and earth-fixed coordinate systems to have their origins near the surface

of the earth at a point we call the initial point of the datum.

In the first (geocentric) case the earth-fixed system is the

Average Terrestrial system of section 2.1.2, and the ellipsoid-fixed system

is the geodetic system of Equation 2-31 (except that here we assume the

geodetic and average terrestrial axes are not in general parallel). In

this case the datum position parameters are tne Average Terrestrial coor-

dinat.es ·of the ellipsoid origin (x , y , z of Equation 2-32} and three

0 0 0

rotation angles (say w1 , w2 , w3 ) required to define the misalignment bet-

ween the axes. It is of course highly desirable that the ellipsoid be

positioned so that these angles are as small as possible, particularly

that the two axes of symmetry (the ellipsoid minor axis and earth's average

rotation axis or Average Terrestrial z-.axis) be parallel.

In the second (topocentric) case the earth-fixed system is a

local astronomic system at the initial point, and the ellipsoid-fixed

system is a local geodetic system at the same point (local astronomic and

geodetic systems are discussed in section 2.4).

Before proceeding further let us consider the geometry in the

neighbourhood of a point on the earth's surface. Figure 2-10 is an exag-

gerated view of the geodetic meridian plane at such a point, showing the

sectioned ellipsoid, geoid, several equipotential surfaces related to the

geoid, and the terrain. A particular ellipsoid normal intersects the

 37A

EQUIPOTENTIAL.
SURFACES

SUR.F,\c:E
GaRA\I•TY
VE(TICAt..

FIGURE 2-10

ORIENTATION OF ELLIPSOID TO GEOID

 38

ellipsoid, geoid, and terrain at Q,P and T respectb·ely. There are three

"natural" normals corresponding to this ellipsoid normal; the surface

gravity vertical (perpendicular to the equipotential surface at T, passing

through T) , the geoid gravity vertical (perpendicular to the geoid pass_ing

through P), and the plumbline (perpendicular to all equipotential surfaces

between terrain and· geoid, passing through T}. In general, the plumbline

is curved while the others are straight lines, and none of these three

actually lie in the geodetic meridian plane - they are shown here as pro-

jections onto this plane. If the curvature of the plumbline is ignored

the two gravity verticals are parallel.

The astronomic meridian plane is the plane containing one of

the gravity verticals and a parallel to the Average Terrestrial z-axis.

The angle between the gravity vertical and the parallel to the A.T •. z-axis

i.s the astronomic co-latitude <; - t). The angle between the astronomic

meridian plane and a reference meridian plane (Greenwich) is the astronomic

longitude A. The angle between the ellipsoid normal and the gravity

vertical is the deflection of the vertical, which can be resolved into a

component ~ in the geodetic meridian plane and a component n in the

geodetic prime vertical plane (the plane perpendicular to the geodetic mer-

idian plane which contains the ellipsoid normal). Thus corresponding to

the two gravity verticals, there are two sets of values for the astrqnomic

latitude and longitude and deflection components, and if the curvature of

the plumbline is ignored, these two sets are equal.

If the ellipsoid is positioned so that itE geocentric axes are

paralled to the Average Terrestrial axes (that is w1 = w2 = w3 = 0) then

2-40

n = (A -A) cos ~ 2-41

 39

where ($, A) are the common geodetic coordinates of Q~ P. and T.

The distance between the ellipsoid and geoid, measured along

the ellipsoid normal (QP) is the geoid height N*. The distance between the

ellipsoid and terrain, measured along the ellipsoid normal (QT.} is the

ellipsoid height h. The distance between the geoid and terrain, measured

along the plumbline (P'T) is the orthometric height H. If the curvature

of the plumbline is ignored

h = N* + H.

Given a point some distance from T, the angle between the geodetic

meridian plane and the plane containing this point and the ellipsoid ndriaa.J.

QPT is the geodetic azimuth a ·of that point with respect to Q, P or T.

(Actually this is the azimuth of the normal section, and is related to the

geodetic azimuth by small corrections ( Bamford, 1971)}. The angle betweenthe

astronomic meridian plane and the plane containing this point and the cor-

responding gravity vertical is the astronomic azimuth A of that point with

respect to either P or T depending on which gravity vertical is used.

Because the deflection of the vertical is small, then for all such points

the difference

oa =A -a 2-43

is nearly constant, and is the angle between the geodetic and astronomic

meridian planes.

Returning to the topocentric datum position parameters, it is

natural to specify that our local geodetic system at the initial point have

its origin on the datum surface, that is on the ellipsoid. In the classical

(non-three-dimensional) approach the orthometric height H enters into hori-

zontal networks only in the reduction of surface quantities to the geoid,

 40

therefore it is natural to take our local astronomic system at the initial

point to have its origin on the g~oid. Denoting quantities at the initial

point by a zero subscript, we. then see that the six datum position para-

meters are in this case the geodetic coordinates of the local astronomic

origin (q, , :A ·, N*) and the rotation angles required to define the trans-
0. 0 0

formation between the local geodetic and local astronomic . systems (~ 0 , n0 ,

oa0 ) .

2.3.2 Establishment of a Datum

We have seen that a datum is defined by assigning values either

to the eight parameters (a, b, x 0 , y0 , z0 , w1 , w2 , w3 ) or to the eight

parameters (a, b, q, 0 , A0 , N* 0 , ~0 , n0 , oa0 ). However, an arbitrary set

of values will not in general result in a satisfactory datum~ We recall

that it is important that a datum closely approximate the geoid over the

area of the network for which it is a datum, and that the geocentric axes

of the geodetic .coordinate system be closely parallel

to the Average Terrestrial axes, particularly that the axes of symmetry be

parallel. The process of assigning values to the eight datum parameters in

such a way that these characteristics are obtained is called establishment

of a datum.

To begin with, in establishing a datum values are always assigned

to the topocentric set (a, b, 4>

0
,A , N* ,
0 0
~
0
, n ,
0
oa0 ) rather than the geo~

centric set (a, b, x 0 , y0 , z0 , w1 , w2 , w3 ) because it is the set which is

related to the geodetic and astronomic measurements which we must use in

establishing the datum. We see that we must somehow choose values for

(a, b, q,, A, N*, ~, n0

0 0 0 0
, oa 0 ) so that the values of (N*,~,n) elsewhere
in the network are not excessive (the datum approximates the geoid), and

so that w1 = w2= w3= 0 (the axes are parallel). Additionally ror networks

of global extent we require that x

0
= y0 = z0 = 0, in which case the datum
.

is termed a geocentric datum. Otherwise the datum is a local datum.

The problem of approximating the geoid can be ignored, in which

case the values

N* 0 = ~ 0 = n0 = o
are assigned, which rorces the ellipsoid to intersect and be tangent to the

geoid at the initial point.

The geoid can be approximated in two ways, by choosing values

of (a, b, N* 0 , ~0 , n0 } such that either values of (~, n) or values or N*

throughout the network are minimized (Vanicek, 1972). Note that values

of (N*, ~. n) are available throughout the network only ir some adjusted

network already exists, which points up the iterative nature or datum

establishment - a "best ritting" datum can be established only as an improve

ment on an already existing datum.

The classical method of' "ensuring" that the axes of' symmetry

are parallel is to enforce the Laplace azimuth condition at the initial

point, that is to assign a value to a according to

0

oa0 = A0 - a0 = n0 tan ~
0
2-44
where A is an observed astronomic azimuth.
0
This condition forces the

geodetic and astronomic meridians to be parallel at the initial point, and

thus forces both axes of' symmetry to ·,lie in this corJmon plane. However,

the axes of' symmetry can still be misaligned within the meridian plane.

The solution to this dilemma has been to apply the Laplace condition at

several geodetic. meridians parallel to their corresponding astron6mic

 42

meridians. In essence this constrains the adjusted network to compensate

for misalignment of the datum, rather than ensuring that the datum minor

axis is parallel to the earth's rotation axis. Note that enforcing the

Laplace condition throughout the network presumes th~ existance of an

adjusted network, which again _points up the iterative nature of datum

establishment.

2.3.3 The North American Datum

The iterative nature of datum establishreent is illustrated by

the history of the North American Datum.

Towards the close of the last century geodetic networks existed

in several parts of North America, each defined on its c~ datum. The

largest of these was the New England Datum established in 1879 with an

initial point at Principia, Maryland. The New England Datum used the Clarke

1866 ellipsoid, still used by the North American Datum today.

By 1899 the U.S. Transcontinental Network linking the Atlantic

and Pacific coasts was complete. When an attempt was made to join the neWer

networks to those of the New England Datum large discrepancies occurred.

Therefore in 1901 the United States Standard Datum was established. The

Clarke 1866 ellipsoid was retained from the New England Datum, but the

initial point was moved from Principio to the approximate geographical centre

of the U.S. at Meades Ranch, Kansas. The coordinates ruid azimuth at Meades

Ranch were selected so as to cause minimum change in existing coordinates

and publications (mainly in New England) while providiP.g a better fit to

the geoid for the rest of the continent.

 43

Meanwhile additional networks were being established in the

United States, Canada and Mexico. In 1913 Canada an~. Mexico agreed to

accept Meades Ranch as the initial point tor all North American networks,

and the datum was renamed the North American Datum.

This eventually led to the readjustment, between 1927 and 1932,

of all the North American networks then in existance. The 1901 coordinates

of Meades Ranch and the Clarke 1866 ellipsoid remained unchanged, however

the value of the geodetic azimuth was changed by abo~t 5 arcseconds {Mitchell,

1948). Thus the new datum was called the 1927 North American Datum.
~ definition of the North American Datum was not yet complete. It

was only in 1948 that astronomic coordinates were observed at Meades Ranch,

allowing specification of values tor ~ , 11 • The final datum parameter

0 0

was defined in 1967 when the U.S. Army Map Service chose a value of N'
0
=0
at Meades Ranch tor their astrogeodetic geoid [Fischer, et al 1967]. Table

2-1 lists the values assigned to the datum parameters for the North American

Datum, and the date at which they were determined.

Since the 1927 readjustment many new networks have been added

to what was then available. However, these new networks have been adjusted

by "tacking them on~' to· previously adjusted networks, ,.the latter being held

fixed in the process. Until the recent advent of large fast digital computers

it was impractical to consider readjusting all the networks on the continent

again, consequently distortions have crept in to the networks, a notorious

case being the 10 metres discrepancy which has been "drowned" in Lake Superior

by international agreement. The day is fast approaching when a massive new

readjustment and perhaps redefinition of the North American Da.tum ·will occur

[Smith, 1971]. One landmark on this path is the International Symposium on

Problems Related to the Redefinition of North American Geodetic Networks,

 44

Table 2-1

PARAMETERS DEFINING THE

1927 NORTH AMERICAN DATUM

Date Ado;eted
Clarke 1866 Ellipsoid semi-major axis a = 6378206.4 metr.es.

Clarke 1866 Ellipsoid semi-minor axis b = 6356583.8 metr~s

} 1879

Initial Point Latitude of Meade's Ranch 41 0 = 39° 13' 26".686 N

1
-+<
1901
Initial Point Longitude of Meade's Ranch A
0
= 98° 32' 30".506 W
Initial Point Azimuth (to Waldo) a
0~
.= 75° 28' 9".64
{clockwlse from south} 1927
Initial Point Meridian Deflection
= -1.02"
}
Component ~
0

Initial Point Prime Vertical

1948
Deflection Component n0 = -1. 79"
N.~t
Initial Point.Geoid Height
0
=0 1967

Table 2-2

TRANSLATION COMPONENTS
X yo z:0 axo· a .a zo
0 yo
Merry & Vanicek -28.7 150.5 179-9 1.7 1.0 1.2
Krakiwsky et al. -35 164 186 2 3 3
May 1974 at the University or New Brunswick.

The North American Datum is a local datums that is its geo-

metrical centre does not coincide with the origin of the Average Terres-

trial system. Because of the distortions in the networks just mentioneds

determinations of x , y , z vary depending on the locations at which

0 0 0

they are measured. Two recent sets of values obtained by different methods

are listed in Table 2-2. Merry and Vanicek [1973] used data within 1000

km of Heades Ranch. Krakiwsky et al [1973] used data from New Brunswick

and Nova Scotia. The discrepancies of order 10 metres likely reflect the

distortions which exist in the present North American networks.

2.3.4 Datum Transformations

If the curvilinear coordinates of an observing station referring

to one particular datum are given, then a problem which often occurs is to

obtain the curvilinear coordinates for the station referred to another datum.

In transforming coordinates from one datum to another it is

necessary to account for two items:

a) the location of the geometric centres of each reference

ellipsoid with respect to the centre of gravity of the earths or with

respect to each other,

b) the difference in size and shape between the ellipsoids.

It is usually assumed that the axes of both datums are parallel to the axes

of the average terrestrial system.

Consider the ellipsoids with sizes and shapes defined by (a1 ,b1 )

and (a2 ,b 2 ) (or alternatively (a1 , f 1 ) .and (a1 ,f2 ), where f=(a-b)/a)and with

locations of the geometric centres with respect to the centre of gravity

defined by
 46

( r- 'i~
0 j_
= r~:J,
Lz0 -
.J.

and

X
0

cr: )2 0
= Yo
z r.
0 c.

Let us define the coordinates of a point referred to the first

ellipsoid as ( ~l, >.. 1 , h 1 ). We want to find the coordinates of the

same point, referred to the second ellipsoid {~ 2 , >..

2 , h 2 ).
The average terrestrial coordLnates of the point are given by

equation 2-32:

X X
0
(Nl + hl) cos..P 1 CO SAl
y = Yo + (N + h1 ) cos$ 1 sinA. 1 • 2-45
l 2 2
z (N1 b/a1 + hl) sin..p 1
l
But the average terrestrial coordinates are not affected by a

X X
0
(N2 + h2) cos$ 2 COSA2
y = yo + (N2 + h2) cos..p 2 sinA. 2 2-46

z
2, 2
z (N2 b2;a2 + h ) sin$ 2
A.T. 0
2
z
There are two methods for obtainiP~ (..p 2 , >..
2 , h 2 ). The first

method, called the iterative method is to find the average terrestrial

coordinates directly from equation 2-45, and then to invert equation

2-46 to find (..p 2 , >.. 2 , h 2 ), using the iterative method described in

section 2.2.6.
 47

The second method, called the differential method can be applied

when the parameter differences (oa, of, ox ' oy ' oz ) between the two datums
0 0 0

is small enough that we can use the Taylor's series linear approximation.

Taking the total differential of equation 2032, keeping the average terres-

trial coordinates invariant, and setting the differential quantities equal

to differences between the datums we have

ox

oy
0

0
+ J + B [::] = 0 2-47
ox oh
0

where

- (M+h) sin' COSA , - (N+h) cos~ sinA. cos~ COSA

'
J = (M+h) sin~ sinA., (N+h) cos~ cos>. cos~ sin>. 2-48
(M+h) cos~ , 0 ) sin ~

li coscp cos>./a . 2~
M s~n cos~ cosA./(1-f)
'
B = N cos' sin>./a . 2~
M s~n cos~ sinA/(1-f) 2-49
'
N (l-f) 2 sincp/a > (M sin2 <J> - 2N) sincp (1-f)

2 2 2 . 2<jl)3/2
f4 = a(l-f) /(cos <P + (1-f) s~n 2-50

Solving for the coordinate differences

o.x
o<J>
[::]
I
0
-1
a>. = -J oy
0
+ B 2-51
oh oz 0
where

cos>../ (M+h) , - sin~ sinA./(M+h) ·~ cos,/(M+h)

[- sinf
-1 - sin>./ (N+h) cos<jl cos>./ (N+h) cos~ ~ 0 2-52
J = '
cos~ COSA
'
coscp sin). , sin <I>
 48

Note that the matrices can be evaluated in either of the two coordinate

systems, since the differences in quantities has been assumed small. Further

it is reasonable to simplify the evaluation of the matrices by using the

spherical approximation (f = 0, H = M = N+h = M+h = a) in which case ve

obtain the transformation equations of Heiskanen and Moritz (1967, equation

5-55).

Table 2-3 shows an example of datum transformation computations.

This particular example transforms the coordinates of a station in Dartmouth,

Nova Scotia froi!l the 1927 North American Datum ("Old"Datum) to the 1950

European Datum ("New" Datum). The datum translation components used are

those given by Lambeck [1971). Both the iterative method of equations

2-45 and 2-46, and the differential method of equation 2-51 were used. The

discrepancies between the two results are about 0.4 meters in latitude,

0.3 meters in longitude, and 0.2 meters in height.

2.4 TERRESTRIAL TOPOCENTRIC SYSTEMS

In the introduction it was stated that terrestrial topocentric

systems are defined as follows:

a) the origin is at a point near the surface of the earth,

b) the primary plane is the plane tangential to the earth's

surface at the point,

c) the primary axis is the north point,

d) the systems are left-handed.

The last two specifications present no problem!:. However, "the

surface of the earth" can be interpreted in three ways to mean the

earth's physical surface, the earth's equipotential surface, or the

 Table 2-3.

EXAMPLE OF DATUM TRANSFORMATIONS

Parameter "Old" Datum "New" Datum "New'~-"Old"

Given:
semi-major axis a 6378206.4 meters 6378388.0 181.6
flattening f 1/294-98 1/297-0 -2-30,Tno-5

{::
-25.8 -64.5 -38.7
o ffset from geocentre 168.1 -154.8 -322.9
(from Lambeck [1971])
167.3 -46.2 -213.5

observer's coordinates cl Al
hl
44.683°N
63.612°W
37.46 meters
?
?

.
?

Solution b:t: Iterative Method:

2018917.91 2018917.91
observer's coordinates
in average Terrestrial
System (Equation2-45)
{: -4069107.35
4462360.64
-40691·07. 35
4462360.64

44.6847lQ0 N

{
lj>l
observer's coordinates A_t 63.609752°W
(Equation 2-46)
h2. -259-U meters

Solution b:t: differential Method:

change in semi-major axis cSa 181.6
change in flattening of -2.3057xlo- 5
change in offsets from
geocenter

observer's coordinates
.
r·
cSy:
oz 0
q,2 44.6841Q.6°N
-38.7
-322.9
-213.5

(Equation 2-50).
{ ).2
h2
63.6097~ 0 W
-259. ,2g, meters
 50

surface of a reference ellipsoid. It is not practical to define a

coordinate system in terms of a planetangential to the earth's physical

surface. Two kinds of terrestrial topocentric coordinate systems can

be defined, however~ The system in which the primery pole is the

normal to the equipotential surface at the observation station is

called a local astronomic system; The system in which the primary

pole is the ellipsoid normal passing through the observation station

is called a locai geodetic system [Krakiwsky 1968].

2.4.1 Local Astronomic System

A local astronomic (L.A.) system is speciried:

a} The origin is at the observation station.

b) The primary pole (z-axis) is the normal to the equipo-

tential surface (the gravity vertical) at the observation station. The

primary plane is the plane containing the origin and perpendicular to

the gravity vertical.

c) The primary axis (x-axis) is the intersection of the

primary plane and the plane containing the average terrestrial pole and

the observation station, and is called the astronomic north.

d) The y-axis is directed east to form a left-handed system.

The position vector of an observed station 1, expressed in the

local astronomic system of the observation station k, is given by

cos vtl cos

X
~1
(rkl)L.A. = y = rkl cos vtl sin
~ 2-54

z . v
L.A. s~n kl
 51

z
z

ELLIPSOID
PARALLEL TO NORMAL
THE AVERAGE
ROTATION
AXIS x NORTH

y
EAST

PARALLEL TO MERIDIAN
THE GREENWICH
MEAN MERIDIAN

·.,

X
Figilre 2-12
GEODETIC AND LOCAL GEODETIC COORDINATE
SYSTEMS
 52

where rkl is the terrestrial spatial distance, vkl the vertical angle,

and ~1 the astronomic azimuth.

Note that the relationship of the local astronomic system to the

average terrestrial system is given by the astronomic latitude ~k and

longitude Ak only after the observed quantities ~k' Ak, ~l have been

corrected for polar motion •. Thus the position vector rkl of 2-54

expressed in the average terrestrial system is:

X X

y = y 2-55

z A.T. z
L.A.
vhere the reflection matrix

1 0 0

0 -1 0 2-56

0 0 1

accomplishes the transformation from a left-handed system into a right-

handed system, while the rotation matrices

cos (90':.~k) 0 -sin (90-~k)

R2 = 0 1 0 2-57

sin ( 90.._\) 0 cos (9rf-~k)

and

cos(l~O:_Ak) sin(l8o:.Ak) 0

R3 = -sin(lBo-Ak) cos ( 18o'1-Ak) 0 2-58

0 0 1

. bring the three axes of the local astronomic systeiD parallel to the

corresponding axes in the average terrestrial system.

The inverse transformation is

 53

" 2-60

Note that so far no translation!; have taken place. We have merely

rotated the position vector (rk1 ) of station 1 with respect to station

k into the average terrestrial system. If the position vector of

station k with respect to the centre of gravity in the average

terrestrial system is{Rk)A.T~, then the total position vector R1 of

the observed station 1 with respect to the centre of gravity in the

average terrestrial system is given by

The unit vectors uX , uy , uZ directed along the axes of the local

astronomic system have the following components in the average

terrestrial system:

1
A

ux = R3 (180°-A) R2 (90°-~) P2 0
0

s:Ln9 cosA

2-62
cos<P

uy = R3 (180°-A) R2 (90°-~) P 2 1
0

-sin A

=
A

u cos A 2-63
y
0
 54

u
z
= [:::: :::] .
sin~
2-64

The local astronomic coordinate system is unique for every obser-

vat ion point. Because of this fact~ this system is tile basis for

treating terrestrial three-dimensional measurements at several sta-

tions together in one solution.

2.4.2. Local Geodetic System

A local geodetic (L.G.) system is specified (see Figure 2-12):

a) The origin lies along the ellipsoidal normal passing

through the observation station. Note that in principle the origin may

lie anywhere along the ellipsoidal normal. In practice it is chosen to

be at the observation station~ at the ellipsoid, or at the intersection

of the ellipsoidal normal with the geoid.

b) The primary pole (z-axis) is the elli-psoidal normal. The

primar:; plane is the plane containing the origin and perpendicular to

the primary pole.

c) The primary axis (x-axis) is the intersection of the

primary plane and the plane containing the semi-minor axis of the

ellipsoid and the origin, and is called t~e.geodetic north.

d) The y-axis is directed east to form a left-handed system.

 55

Trar1sformations between local geodetic and local astronomic

systems sharing a common origin can be expressed in terms of the angle

between the ellipsoid normal and gravity vertical (the deflection of the

vertical) and the angle between the geodetic north and astronomic north.

Given the meridian and prime vertical deflection components ~. n respectively,

and the geodetic and astronomic azimuths a, A to a particular point, then

a vector in the local astronomic system is transformed into a vector in the

local geodetic system by

2.65

Note that the order in which the rotations are performed in this case is

not important, since the ~ngles r,, n (A- a) are sma~l enough that their

rotation matrices can be assumed to commute. Note also that if the Laplace

condition is enforced at the origin of these local systems, we have

A- a= (A- A) sin$ = n tan$

If the origin is not at the observation station, the position vector

Rk in 2-61 would refer to the origin, not the observation station. That

is, for a point on the geoid, the computation of (~, yk, zk) is made

from (¢k' Ak' Nk) (geoid undulation), while on the ellipsoid ($k' \k, 0)

are used. Note that when a small region of the earth is taken as a

plane it is a local geodetic system that is implied.

Similar to equations 2-54 and 2-55 the position vector from

observing station k to observed station 1 is given by

X cos cosakl
~1
<r:k1) T ,., = y = rkl cos
~1 sinakl 2-66
L.u.
z
sin
L.G. ~1

and
 56

X X

y 2-67
z z
G L.G.
where (a~ a~ r) are the geodetic altitude, azimuth and range, and

(~, A.} are the geodetic latitude and longitude. Note that the geodetic

system (G) and the average terrestrial system (A.T.) are related by

equation 2-4

X X X
0

y = + y

z A.T. z
0
z
G

where (x , y , z ) are the translation components of the origin of

0 0 0

the geodetic system in the average terrestrial system.

The unit vectors corresponding to the three Cartesian axes in the

local geodetic system are

,_
-sin4l COSA -
" =
u.
X
-sin4l sinA , 2-68a
--
cos4l
I.
-
-sinA.
-
u" = COSA ~ 2-68b
y

I.
0

COSA-
,..
cos4l

=
A.

u cos4l sinA. • 2-68c

z
sin4>
-
 57

2. 5 SUMMARY OF TERRESTRIAL SYSTEMS

In this chapter we have precisely defined five specific terrestrial

coordinate systems:

a) Average Terrestrial .(A.T.),

b) Instantaneous Terrestrial (I.T.),

c) Geodetic (G),

d) Local Astronomic (L.A.),

e) Local Geodetic (L.G.),

of which the first three are geocentric and the last two topocentric.

Table 2-4 summarizes the planes, poles and axes defining these systems.

We have also precisely defined four kinds of coordir..ates:

a) Cartesian (x, y, z) - used by all systems,

b) Curvilinear (', !., h) - used by Geodetic system,

c) Curvilinear (v, A, r) - used by Local Astronomic system,

d) Curvilinear (a, a, r) - used by Local Geodetic system.

Finally we have defined the principal transformations between these

coordinates and coordinate systems. Figure 2-13 lists the equation

numbers which define these transformations, which are tabulated in

Table 2-5.
 Table 2-4.
REFERENCE POLES, PLANES AND AXES DEFINING TERRESTRIAL COORDINATE SYSTE~lli

Reference Poles i Reference Planes

System Primary· : Secondary Primary I Secondary Handedness
( z-axis) i (x-axis) (1 to Primary Pole)
I I I
i
H
Average
Terrestrial
Average Terrestriali
Pole (CIO) ~
Cb
11
ltAverage Terrestrial
equator containing
Greenwich mean
meridian
right

Cll 1 centre of gravity.

Cb
n
C1
1-'•
0
::I
Instantaneous Instantaneous 0 Instantaneous Greenwich mean right V1
111 co
Terrestrial Terrestrial Terrestrial meridian
1-0
Pole 11 equator.
~
-~
, Parallel To
Geodetic Semi-minor axis
(parallel to
[ ! Average
i
Terrestrial
equator
Parallel to
Greenwich mean
right
Cll I

-~--r--==+--
terrestrial pole) (I) meridian
n I
0

g left
Local Gravity Vertical Local Horizon Astronomic
Astronomic at Station ;:g Meridian
§
Cll
of station.

Local Ellipsoidal Normal Tangent Plane Coincident with I left

Geodetic at Station. Geodetic Meridian
of station.
 TERRESTRIAL

Geocentric 1 Tol>ocentric

Geodetic Average Instantaneous Local Astronomic Local Geodetic

(G) Terrestrial Terrestrial (L.A.) · (L.G.)
(A.T.) (I.T.)

Curvilinear Cartesian Cartesian Cartesian Cartesian Curvilinear Cartesian Curvilinear

(«jj,A,h~ (v,A,r) (a,a,r) V1
\0

Figure 2-13.

EQUATIONS RELATING TERRESTRIAL SYSTEMS

 .·. .,. .... . . ·•.. . ·. ·• .
~ ..... ' . .•.•... ,. ... .. t.,:.·

':'a'cle 2-5.
TRANSFORMATIONS AMONG TERRESTRIAL COORDINATE SYSTEMS.

Original System
Average Instantaneous Geodetic Local Local
Terrestrial Terrestrial Astronomic Geodetic
··•...
'
Average
T errestrial !i via
R2 (-xp)R1 (-yp) R (1809.-A)R (90~~)P Geodetic

..
[:JAS. + [::]
3 2 2
.

Instantaneoun via I via

Terrestrial Average Average via
r:L.T.
R1 (yp R2 (xp) 0'\
Terrestrial Terrestrial Geodetic ' 0

"'.l

-[~:]
'

I
r:L
f-'•
:::1
via via
il'
!-'
Geodetic Average Local R 3 (~80°-A)R 2 (90°-~)P 2
r:n Terrestrial Geodetic
~
(/1
-+
~

Local .via via

Astronomic P 2R 2 (~-90°)R 3 (A-180°) Average Local R 2 (+~) R1 (+n)
Terrestrial Geodetic
r:L.A.
Local via via
Geodetic Geodetic Average P 2R 2 ~~-90°)R3(~-180) Ri (-n )R 2 t~)
·-·
Terrestrial
..
r:L.G.
Note: ~, A have been corrected fnr polar motion. ...,.
.:,;.. v ..... ,.,., '-...
...., .... . ..,., ../
 61

3. CELESTIAL COOP~INATE SYSTEMS

Celestial coordinate systems are used to define the coordinates

of celestial bodies such as stars. The distance from the earth to

~

the nearest star is more than 10~ earth radii, therefore, the

~imensions of the earth (indeed of the solar syste12) are almost

negligible compared to the distances to the stars. A second

consequence of these great distances is that, although the stars

themselves are believed to be moving at velocities near the velocity

of light, to an observer on the earth this motion is perceiv~d to be

very small, very rarely exceeding one arcsecond per year. Therefore~

the relationship between the earth and stars can be closely approx-

imated by considering the stars all to be equidistant from the earth,

on the surface of the celestial sphere, the dimension of which is so

large that the earth (and indeed the solar system) can be considered

as a dimensionless point at the centre. Although this point may be

dimensionless, relationships between directions on the earth and in

the solar system can be extended to the celestial aphere.

The earth's rotation axis is extended outward to intersect the

celestial sphere at the north celestial pole (NCP) and south celestial

pole (SCP). The earth's equatorial plane extended outward intersects

 62

the celestial sphere at the celestial eguator. The gravity vertical

at a station on the earth is extended upwards to intersect the celestial

sphere at the zenith (Z), and downwards to intersect at the nadir (N).

The plane of the earth's orbit around the sun {the ecliptic plane) is

extended outward to intersect the celestial sphere at the ecliptic.

The line of intersection between the earth's equatorial plane and the

ecliptic plane is extended outwards to intersect the celestial sphere

at the vernal equinox or first point of Aries, and the autumnal

equinox. The vernal equinox is denoted by the symbol ~, and is the

point at which the sun crosses the celestial equator from south to

north.

There are two fundamental differences between celestial systems

and terrestrial or orbital systems. First, only directions and not

distances are considered in celestial coordinate systems. In effect

this means that the celestial sphere can be considered the unit

sphere, and all vectors dealt with are unit vectors. The second

difference is related to the first, in that the celestial geometry is

spherical rather than ellipsoidal as in terrestrial and orbital

systems, which simplifies the mathematical relationships involved.

As discussed in the introduction, there are four main celestial

coordinate systems, called the ecliptic, right ascension, hour angle,

and horizon. Sometimes the right ascension and hour angle systems are

referred to collectively as equatorial systems. We will begin this

chapter by discussing each of these systems in turn.

We noted above that the celestial sphere is only an approximation

of the true relationship between the stars and an observer on the earth.
 63

Therefore, like all approximations, there are a number of corrections

which must be made to precisely represent the true relationship. These

corrections represent the facts that the stars are not stationary

points on the celestial sphere but are really moving (proper motion);

the earth's rotation axis is not stationary with·respect to the stars

(precession and nutation); the earth is displaced ~rom the centre of

the celestial sphere, which is at the sun (parallax); the earth is in

motion around the centre of the celestial sphere (aberration); and

directions measured through the earth's atmosphere are bent by refraction.

All these effects will be discussed in section 3.5 in terms of variations

in the right ascension system.

3.1 THE ECLIPTIC SYSTEM

The ecliptic (E) system is specified as follows (see Figure 3-1}!

a) The origin is heliocentric (at the centre of the sun).

b) The primary plane is the ecliptic plane (the plane of

the earth's orbit) and the primary pole (z-axis) is the north ecliptic

pole (NEP).

c) The primary axis (x-axis) is the vernal equinox.

d) The y-axis is chosen to make the system right-handed.

The ecliptic system is the celestial system which is closest to

being inertial, that is motionless with respect to the stars. However,

due to the effect of the planets on the sun-earth~ystem, the ecliptic

plane is slowly rotating (at 0~5 per year) about a slowly moving axis

of rotation.
 64

NCP

[x]
y
z &
[enC..···~
•c.s ~J
...
E

NEP STAR

ECLIPTIC IIERtOtMI

ECLIPTIC
MERIDIAN
PLANE OF
C(.)
y

Figure 3-l. ECLIPTIC SYSTEM

 65

The ecliptic meridian is the great circle which contains the

ecliptic poles and the celestial body in question while the ecliptic

meridian of 1r contains the vernal equinox. Ecliptic latitude B is

the angle from the ecliptic and in the ecliptic meridian to the line

connecting the origin to the body. Ecliptic longitude A is the angle

of the ecliptic meridian of the body measured eastwards in the ecliptic

plane from the vernal equinox. The unit vector to a celestial Qody in

the celestial system is

X cosB cosA

y = cosB sin>.. 3-1

sinS

and the angles are related to the Cartesian components by

. -1
B = Sl.n z
' 3-2

3-3

3.2 THE RIGHT ASCENSION SYSTEM

The right ascension (RA) system is specified as follows (see

Figure 3-2).

a) The origin is heliocentric~

b) The primary plane is the equatorial plane, and the

primary pole (z-axis) the north celestial pole (NCP).

c) The primary axis (x-axis) is the vernal equinox.

d) The y-axis is chosen to make the system right-handed.

The right ascension system is the most important celestial system. It

 66

• t ...
.. a
cCH 8 c. •]·.
ltft e.
:

NCP ...........

EQUINOCTIAL
COLURE PLANE
.__CEL&STtM.
STM81011Y

CELESTIAL
X EWIITOftiAL
PLAt~£

Figure 3-2. RIGHT ASCENSION SYSTE·M

 67

is in this system that star and satellite coordinates are published,

and-it serves as the connection between terrestrial,celestial and

orbital systems.

The secondary plane contains the north celestial pole and the

vernal equinox and is called the eguinoctialcolure plane. The hour

circle is the great circle_containing the celestial poles and the

body in question. The declination o of a body is then the angle

between the celestial equator and a line joining the origin to the

body. The right ascension a is the angle measured in the equatorial

plane eastwards from the vernal equinox to the hour circle passing

through the body in question. The unit vector describing the direction

of a body in the right-ascension system is

X coso cosa

y = coso sina • 3-4

z sino
R.A.
The right ascension system is related to the ecliptic system

by the acute angle between the ecliptic and celestial equator,

called the obliquity of the ecliptic, and denoted~. Therefore

X X

y = y 3-5
z z
R.A. E

3.3 THE HOUR ANGLE SYSTEM

The hour angle (HA) system is specified as follows (see Figure 3-3):

a) The origin is heliocentric.

b) The primary plane is the equatorial plane.

 68

CELESTIAL
MERIDIAN _ _-""
PLANE

STAR

Figure 3-3. HOUR ANGLE SYSTEM

 c) The secondary plane is the celestial meridian plane of

the observer. The primary axis (x-axis) is the intersection between

the equatorial and observer's celestial meridian planes.

d) The y-axis is chosen so that the system is left-handed.

The hour angle system rotates with the observer.

·T~e hour angle h is the angle measured westward in the equatorial

plane from the observer's celestial meridian to th~ hour circle of

the body in question. The angle measured up from the equatorial plane

to the line directed from the origin towards the body is the declin-

ation o.
The unit vector describing the direction of a celestial body in

the hour angle system is

X cos 0 cos h

·y = cos o sin h • 3-6

z H.A. sin o
We have so far defined four meridians on the celestial sphere:

that containing the vernal equinox (the equinoctial colure); the

Greenwich meridian; that containing the observer (the celestial

meridian); and that containing the star (the hour circle). Figure 3-4
shows the relationships between these meridians.

From the vernal equinox counterclockwise to

a) the Greenwich meridian is called Greenwich Sidereal

Time (GST),

b) the celestial meridian is called l-ocal Sidereal Time

(LST),

c) the hour circle is called the right ascension (a).

 7n
i"""

TOP VIE:W

OF CEl.ESTIAL
CELESTIAL EQUATOf't SPMI:.W:

NCP

X-AXIS OF
RIGHT ASCENStON
SYSTEM
X-AXtS OF
110Uft AMGLE
GREENWICH HOUR CtRCLE STSTft
MERtOIAH T~STAft

X-AXIS OF
TERRESTRIAL
~YS'f'£'M

·-~ ....... ___ __

......

Figure 3-4.
TIME, LONGITUDE, AND RteHT ASOENS"'"
 71

From the Greenwich meridian counterclockwise to the celestial meridian

is called the astronomic longitude {A). From the celestial meridian

clockwise to the hour circle is called the hour angle (h). Therefore,

LST = GST + A 3-7

LST =h + a 3-8
and

h = GST + A- a 3-9
The hour angle system is related to the right ascension system

by the local sidereal time (LST). That is

X X

y y 3-10

z
HA

3.4 THE HORIZON SYSTEM

The horizon (H) system is specified as follows {see Figure 3-5).

a) The origin is heliocentric.

b) The primary pole (z-axis) is the observer's zenith

(gravity vertical}. The primary plane is the observer's horizon.

c) The primary axis (x-axis) is the north point.

d) The y-axis is chosen so that the system is left-handed.

The horizon system is used to describe the position of a

celestial body in a system peculiar to a topocentrically located

observer, similar to the local astronomic system described in the

chapter on terres~rial systems. The main difference is that the

origin of the horizon system is heliocentric instead of topocentric.

 NORTH

ZENITH

\
'

CELESTIAL y
HORIZON
Pl.ANE

SOUTH
Figure 3-5.
HORIZON SYSTEM
 73

The great circle containing the primary pole and the celestial

body being observed is called the vertical circle. The location of

this great circle is given by the astronomic azimuth A, the angle

measured clockwise in the horizon plane from north to the vertical

circle. The altitude a of the body is the angle between the horizon

plane and the line directed .from the origin of the system toward the

body. The unit vector to a celestial body in the horizon system is

X cos a cos A

y = cos a sin A 3-11

z sin a

The horizon system is related to the hour angle system by the

astronomic latitude ~. That is

X X

y y 3-12

z z
H HA

3.5 VARIATIONS OF THE RIGHT ASCENSION SYSTEM

As mentioned at the beginning of this chapter, the celestial

sphere approximation requires corrections for precise work. These

corrections are for proper motion, precession, nutation, aberration,

parallax and refraction, and are applied in four stages between the

system in which observations are actually made (which we will call

the "observed place system at epoch T") and the most absolute right

ascension system (which we will call the "mean celestial system at

standard epoch T "). We will consider these systems in the reverse

0

order, that is:

 74

a) mean celestial system at standard epoch T0 ,

b) mean celestial system at epoch T,

c) true celestial system at epoch T,

d) apparent place system at eppch.T,

e) observed place system at epoch T.

The connections between these five systems are shown in Figure 3-6.

The first three systems are related by motions of the coordinate

system, while the last two are related by physical effects which

cause the position of the celestial body to vary.

3.5.1 Precession and Nutation

The earth is not perfectly spherical, but has an equatorial bulge

which is attracted by the sun, moon, and planets in a non-~etrical

way. This causes the earth's axis of rotation (th~ north celestial

pole) to move around the north ecliptic pole with a period of about

25,800 years and an amplitude equal to the obliquity of the ecliptic

(23~5). This motion is called precession and is similar to the

precession of an ordinary gyroscopic top about the gravity vector

[Mueller 1969, pages 59-62].

Precession is itself not a regular motion sin~e the earth's

orbit is not circular and the moon's orbit does not lie in the ecliptic

plane, and is not circular. Therefore, the added effects of the sun

and moon are constantly changing as their configuration changes.

Irregularities in precession are called nutation, and for the celestial

pole have a period of about 18.6 years and a maximum amplitude of

about 9". The added irregularity due to the chang) ng configuration of

 1.1

Mean @ T
0

Proper Motion
Precession

Nutation

Annual Aberation
Annual Parallax

Diurnal Aberation
Geocentric Parallax
Refraction

Observed @T

Figure 3-6.
VARIATIONS OF THE CELESTIAL RIGHT ASCENSION SYSTEM
 76

the planets is called planetary precession, and causes the very slow

motion of the ecliptic plane mentioned in section 3.1. In principle,

precession exists only because the earth has an equatorial bulge, and

nutation and planetary precession exist only because precession

exists.

Precession and nutation are shown schematically in Figure 3-7.

The celestial equator is defined as being perpendicular to the

celestial pole, so that it too follows the precession and nutation of

the pole. The vernal equinox is defined as being at the intersection

of the celestial equator and ecliptic, so that it ~ill follow both

the precession and nutation of the celestial equator, and the motion

of the ecliptic due to planetary precession. The effects are shown

in Figure 3-8.

3.5.2 Mean Celestial Systems

A mean celestial (M.C.) system is specified as follows.

a) The origin is at the centre of the sun.

b) The primary pole (z-axis) is a precessing (but not

nutating) pole which follows the precession of the north celestial

pole, and is called the mean celestial pole.

c) The primary axis (x-axis) is a prece!:sing (but not

nutating) axis which follows the motion of the vernal equinox due both

to precession of the celestial equator and rotation of the ecliptic,

and is called the mean vernal eguinox.

d) The y-axis is chosen so the system is right-handed.

 77

GENERAL
PRECESSION

ECLIPTIC ~TH OF TRUE

POLE
CELESTtAL 'POLE

P=PftECESStON
NT :: NUT.ION
~T

PATH OF MEAN
CELESTIAL POLE

Fig,ure :-1. MOTION OF CELESTIAL POLE

 TRUE CELESTIAL AXIS
ECLIPTIC AXIS
TRUE EQUATOR OF EARTH
AT T 1 •

. :::::::_-1\ INTERSECTION OF
("\ ECLIPTIC WITH EARTH.

TRUE CELESTIAL AXIS

-·~
0:.)
(EXTENSION OF ROTA-
T tON AXtS OF EARTH ) .

/
ECLIPTIC PLANE TRUE EQUATOR OF EARTH AT
T0 ( PLAME 1 TRUE C!LESTtAL
AXtS).

Gfc>

Figure 3-8. ·THE EFFECT OF PRfCI!S9tON M«> NUTAltON

 79

Because the mean celestial system is moving, the coordinates

(right ascension a, and declination 6) of celestial bodies vary with

time. Therefore, for each epoch of time~' a different mean celestial

system is defined. Certain epochs T0 have been chosen as standard

epochs, to which tabulated mean celestial coordinates of celestial

bodies refer. The relationship between mean celestial systems of times·

T0 and T is usually defined in terms of the precessional elements

(' 0 , e, z) as shown in Figure 3-9. Expressions for these elements as

a function of time were derived over 70 years ago by Simon Newcomb

[Mueller 1969, p. 63]. The angles (90°-t ) and (90°+z) are the right
0

ascensions of the ascending node of the equator at T measured

respectively in the systems at T

0
and at T. The angle e is the

inclination of the equator at T with respect to the equator at T0 •

The transformation from a mean celestial system at T to one·at Tis

0

given by

X X

y y 3-13

z z _ M.C.T
M.C.T. 0

Independent of the motion of the mean celestial coordinate

system due to precession, each star is changing in position due

to proper motion. Because this proper motion is uniform, it is

most appropriate to account for it in the most uniform right

ascension system, that is the mean celestial system. The proper

motion components for each star of interest (usuall~· tabulated

 NCP(T0 )

80

NCP(T)

MI•N EOtJaTOR
. ·····'t)\• ..

~(T)

PftECESSION

Gl.· .
. MC.To

Figure· 3-9. MEAN CELESTIAL COOftOINATE SYST£MS

 81

as rates of changes in right ascension and declination) must therefore

be included in the conversion of mean place at T to mean place at T.

o·

3.5.3 The True Celestial System

A true celestial system (T.C.) is specified as follows.

a) The origin is at the centre of the sun.

b) The primary pole (z-axis) is a precessing and nutating

pole which follows the precession and nutation of the north celestial

pole, and is called the true celestial pole.

c) The primary axis (x-axis) is a precessing and nutating

axis which follows the motion of the vernal equinox due to precession

and nutation of the celestial equator, arid to rotation of the ecliptic,

and is called the true vernal eguinox.

d) The y-axis is chosen so the system is right-handed.

As in the case of mean celestial systems, a different true

celestial system is defined for each epoch of time T. The true

celestial sy"stem at epoch T differs from the mean celestial system

at epoch T only by the effect of nutation, and the relationship is

usually defined in terms of the nutation in longitude 8~ and

nutation in obliquity f:l.e. shown in Figure 3-10. Expressions for these

two elements as a function of time, and other parameters were derived

by Woolard [~~eller 1969, p. 69]. The transformation from a mean

celestial system at T to a true celestial system at T is given by

r-
X
- :- -
X

y 3-14

z ·z M.C.T •
__
._ - T.C.T.
 82

NCP (TRUE)
NCP(MEAN);;

t«-'

. TRUE EOUATOMAL
PLANE AT T

O'fO (TRUE)

NUTATION

TRUE AND MEAN CELESTIAL COORDINATE SYSTEMS.

Fi~~~-3=1o~.
 83

3.5.4 The Apparent Place System

An apparent place (A.P.) system is specified as follows.

a) The origin is at the centre of the earth.

b) The primary pole is parallel to the true celestial pole.

c) The primary axis is parallel to the true vernal equinox.

d) The system is right-handed.

Therefore, an apparent place system is a true celestial system with

the origin shifted from the centre of the sun to the centre of the

earth. This means the origin is no longer at the centre of the true

celestial sphere which causes annual parallax, and the origin is

revolving around the centre of the true celestial sphere which causes

annual aberration.
If the. earth's orbit is regarded as circular, the earth has a

constant of aberration.
n
K = vc cosec 1" = 20.4958 3-15

where v is the earth's velocity and c the velocity of light; and the

radius of the earth's orbit will subtend a different angle H at each

star, called the stellar parallax for that star. The nearest star

has a stellar parallax of 0':16.

The right ascension a and declination o of a star expressed in

the apparent place system is then [Mueller 1969, pages 93 and 61].

+ 3-16

where
 Ah

[::p] p
=n
[coso cos£ seco - sina cosA 5 seco

coso sin£ sinA 5 - cosa sino cos).. 5 - sina sino COSE sinAS
}-17
and

[AoA1 = -K
[coso casAs cos£ seco + sina sinAs secO J 3-18
MA COS As cos£(tanc coso - sina sino) + ccsa sino sinAs

and As is the longitude of the sun, £ the obliquity of the ecliptic,

and (a, o) in 3-17 and 3-18 expressed in the true celestial system.

The fact that the earth's orbit is not circular introduces

errors of about 1% in equation 3-17 and up to 0'!343 in equation 3-18.

3.5.5 The Observed Place System

An observed place (O.P.) system is specified as follows.

a) The origin is at the observing station.

b) The primary pole is parallel to the true celestial pole.

c) The primary axis is parallel to the true vernal equinox.

d) The system is right-handed.

Therefore, an observed place system is an apparent place system

with the origin shifted from the centre of the earth to the observing

station. This means the origin is no longer at the centre of the

earth, which causes geocentric parallax, and the·origin is rotating

around the centre of the earth, which causes diurnal aberration. In

fact, the effect of geocentric parallax is always negligible when

observing stars. The diurnal constant of aberration is

k = vc cosec 1" = O'! 320 p coscp 3-19

~here v is the earth's surface rotational velocity, c is the velocity

of light, p is the radial distance from geocenter to observer in units

..
of earth radius, and~ is the geodetic latitude of the observer.

There is a third effect due to the fact that the earth is

blanketed with an atmosphere of varyi,ng optical density. This

·~auses a complex change in the direction of the light ray from a star

which depends on the incident angle. Mueller [1969, pages 103-109]

discusses this atmospheric refraction in detail.

The right ascension and declination of a star in the observed

place system is then

3-20

where

k [cos h seco J 3-21

sinh sino

where h is the hour angle of the star, and (6aR, 6oR} are the

corrections due to refraction.

3.6 TRANSFORHATION BETWEEN APPARENT CELESTIAL AND AVERAGE

TERRESTRIAL COORDINATE SYSTEHS

The apparent celestial and average terrestrial coordinate systems

both have

a) their origins at the centre of gravity of the earth,

b) their primary poles as the era pole, that is the

average terrestrial pole is parallel to the true celestial pole,
 86

c) both are right-handed.

'l'he only difference between the two systems is that the primary

axis of the apparent celestial system is parallel to the true vernal

equinox, ami the primary axis of the average terrestrial system lies

in the Greenwich mean astronomic meridian. The angle between these

two axes varies with the rotation of the earth, and is called the

Greenwich Apparent Sidereal Time (GAST). Therefore, the transformation

from apparent celestial to average terrestrial is (see Figure 3-4).

X X

y = R 3 (GAST) y 3-22
z z
A.T. A.P.
'l'o use this equation, we require some means of computing GAST

from the Universal (Solar) Time used for broadcasts of standard time.

'.-le will describe two methods.

First, if GAST is known for some epoch T of Universal Time,

0

then it may be computed for some other epoch T from the relation

GAST (T) = GAST (T ) + w (T-T )~

o e o 3-23
where it is assumed that sidereal and universal time are related by a

uniform rotation rate of the earth

we = 360.98565 degrees/UT day

= 4.3752695 x 10- 3 radians/minute. 3-24

This is not precisely true, but a difference with respect to

ti1e more accurate method presented below of less than 10-7 radians

(equivalent to about 0.02 arcseconds, 1 millisecond, or 1/2 meter

along the earth's equator) is introduced if (T-T ) is less than a day.

0
 1-\ more a<:curate :relation is given by Veis [1966, p. 19}:

r--·
I JAS1' =

1!~392 X 10- 3 sin (12'?1128 - 0'?052954 T)

I + 0'?053 X 10- 3 sin 2 ( 12'?1128 - 0'?052954 T) 3-25

0'?325 X l0- 3 sin 2(280~0812 + 0'?9256473 T)

:.'~1ere T is the number of Julian Days since the epoc~ CL 5 January 1950

(that 1s midnight of December ~1, 1949). For 1971

T = 7669 + D + (~ + S/60)/1440 3-26

the Ja:v number during 1971,

= the minute of UT time,

S = the second of UT time,

·md 7669 is the nu1nber of days beb,-een January 1, 1950 and December 31,

'970. This expression is accurate to 0.2 arcseconds, 10 milliseconds,

•Jr 5 !rleters along the equator for any value (T - T 0 ). r·1ore accuracy

c:an be obtained by adding more terms (:-Jautical Almanac Office 1961].

3. '( SUN!·JI..F:Y C.F C:::LESTIAL SYST2·~S

r~ this chapter we have d~fi~ed four celestial coordinate systems:

a) Ecliptic (E),

b) Rig~t Ascension (5.A. ),

c) Hour Angle (H.A.),

d) Horizon (H.).

~able 3-l SlW~arizes the reference poles, planes and axes defining

the s:.rstems. 'l'ab2_e 3-2 SUElffiari=es ·~he transformations between these

 88

systems.

\~P hav.:• a1~~n precisely d0t'i.ned four variations to the right

a:.~c!c:'Tl!~ iun Bystem;

a) ~lean Celestial (i.L C. ) ,

b) 'i'ruc~ t~elesti aJ ('J' .C.),

c) Apparent Place (A.P.),

d) Observed Place (O.P.),

all of which vary with time, so that the epoch T to which they refer

must be specified. Figure 3-ll shows the parameters which connect all

of these celestial coordinate systems.

 'l'able 3-1

REFERENCE POLES, PLANES AND AXES DEFINING CELESTIAL COORDINATE SYSTEMS

-
Reference Poles Reference Planes
r--------1
-j
System Primary pole Secondary Pole Primary Secondary Handedness !

(z-axis) (x-axis) (y-axis)

Ecliptic North Vernal eq_uinox Ecliptic Ecliptic meridian of right

ecliptic. of the eq_uinox (half
pole containing vernal co
I \0
eq_uinox)

Right ascensio North Vernal eq_uinox Celestial Eq_uinoctial colure right

celestial eq_uator (half containing
pole vernal equinox)

Hour angle North Celestial Hour Circle of left

celestial eq_uator observer's zenith
pole (half conta:ining
zenith)

Horizon Zenith North point Celestial Celestial meridian left

horizon (half containing
north pole)
 Table 3-2
TRANSFORMATIONS ANONG CELESTIAL COORDINATE SYSTEMS.
'
Original System

Ecliptic Right Ascension Hour Angle Horizon

Ecliptic
r:t Rl (£)
'

....::s
'">j

Right R3 (-LST) P2 \D
Rl (-£) 0
I»
1-' Ascension
en [:] R.A.
«
(ll
ct
(1)
s
Hour
Angle
P2 R3 (tLST)
[:L.A. R 2 (~-90°) R3 (180°)

Horizon R3 (180°) R 2 (90°-~)

r:L.
 Figure 3-11. Celestial Coordinate Systems.

Ecliptic Right Ascension Hour Angle Horizon

Apparent Observed 1.0

Mean Celestial Mean Celestial True Celestial 1-'
@T @T @T @T @T
0
 92

4. THE ORBITAL COORDINATE SYSTn4

In this chapter we discuss the orbital system, which is used to

define the coordinates of a satellite orbiting around the earth. We

first discuss the orbital ellipse, and the coordinate system in the

orbit plane. Then we transform this system into the apparent celestial

and average terrestrial systems, and discuss variations in the orbital

elements. Finally expressions for the coordinates of the satellite

subpoint,and the topocentric coordinates of the satellite are derived.

4.1 THE ORBITAL ELLIPSE AJID ORBITAL ANOMALIES

The trajectory of a body moving in a central force field describes

an ellipse, with the attracting force centred at one of the foci of the

ellipse.

In the case of a satellite orbiting around the earth, this is

called the orbital ellipse, and the centre of gravity of the earth is

at one of the foci (see Figure 4-1). The point of closest approach of

the satellite to the earth is called the perigee, and the farthest

point is called the apogee. Both perigee and apogee lie on the semi-

major axis of the ellipse, called the line of apsides. The size and

shape of the orbital ellipse are usually defined using the semi-major

axis, a and the eccentricity e, where

2
e = 3-1

and b is the semi-minor axis of the ellipse.

 93

f~ TftUE MIOMALY
a
E- ECGdl'fMt AW-ALY

APOGEE X
ae
EARTH (focus) -·
PStllt
·6EOME;fRI C
CENTRE
l ....
a
)

FIGURE 4-l ORBIT ELLIPSE

 94

Consider the satellite to be at a point m on the orbital ellipse.

The angular distance between perigee and m is called the satellite

anomaly. There are three anomalies. The true anomaly f is the angle

between the line of apsides and the line joining the focus to the

satellite.

Consider the projection of the satellite position m along a line

parallel to the semi-minor axis to intersect a circle with radius

equal to the semi-major axis at a point m'. The eccentrLc anomaly E

is the angle between the line of apsides and the line joining the

geometric centre of the ellipse tom'.

The mean anomaly M is the true anomaly corresponding to the motion

of an imaginary satellite of uniform angular velocity, that is M= 0

at the perigee and then increases uniformly at a rate of 360° per

revolution. When this is expressed as a rate per unit time, then it is called
the mean anomalistic motion n.
The relationship between the true anomaly f and the eccentric

anomaly E is from Figure 4-1

[J =
cos E

sin E
- ae] = [a(cos
a(l - sin
4-2

or

tan f
cos E - e
4-3

The relationship between the eccentric anomaly E and the mean

anomaly Mis Kepler's e~uation [Kaula 1966, p. 23]

M= E - e sin E. 4-4
where Mand E are in radians.
 95

We are usually given the mean anomaly M, and want to find the

eceentrlc anomaly E from equation 4-4. We will present three ways.

If the eccentricity is very small (say e ~ 0.002), then thee sinE

term will be small, and M~ E. Therefore we can write

I E =M+ e sin E = M + e sin M ·I 4-5

For an eccentricity of e = 0.002 this approximation introduces an error

of about 10
-6 radians.

If greater precision is required or the eccentricity is not so

small, we can solve 4-4 iteratively. Taking the total differential of

4-4
6!-1 = (l - e cos E) 6E

or

6E
oM
= --'----- 4-6
1 - e cos E

Given M, the iterative solution of 4-4 begins by making an initial

approximation from 4-5 as

E0 = r.1 + e sin M

The following equations are then iteratively evaluated in order

M.
1
= E.-
1
e sin E.
1

t1M = M.1 H

tlE =l - e cos E.
1

until the difference frM is less than some chosen £.

 96

A third method of evaluating E is to use a power series in e, for

example [Brouwer and Clemence 1961, p. 76]:

1 3 1 5 1 7
E = l'-1 + (e - 8 e + 192 e - 9216 e ) sin M +

+ <21 e
2
- 61 e 4 1 6
+ 98 e ) sin 2M +

27 5 243 7
128 e + 5120 e ) sin 3M + 4-7

+ (31 e 4 - 154 e 6 ) sin 4M + (125

384 e
5

+ 27 6 sin ~~ + 16807 e7 sin 7M

Bo e ... Ol'l 46o8o

4.2 THE ORBITAL COORDINATE SYSTEM

The orbital (ORB) coordinate system is specified as follows (see

Figure 4-2) :

a) The origin is at the centre of gravity of the earth.

b) The primary plane is the plane of the orbital ellipse,

and the primary pole (z - axis) is perpendicular to this

plane (see Figure 4-1) .

c) The primary axis (x- axis) is the line of apsides.

d) The y-axis is chosen so that the system is right-handed.

The position vector of the satellite in its orbit is given by

X cos f a (cos E - e)

r = y =r 4-8

z ORB 0 0

Note z =0 because the satellite is assumed not to be out of the orbit

plane.
 9T

GREENWICH

y TRUE VERNAL
EQUINOX

FIGURE 4-2 KEPL£ft1Aft ORBITAL ELEMENTS

0 - SEMI MAJOR } SIZE AND SHAPE OF OMIT

e- ECCENTRICITY
w- ARGUMENT OF PERIGEE- OfRECTION OF LlN£ Of' APseES}
.0... -RIG. HT ASCENSION OF lfSCENOING NOD£ Ns!TIGN OF ORBtT
i -l.tiQ.tMAT ION
~ f- TRUE :ANOMALY AT TiME T ·
 98

4.3 TRANSFORMATION FROM ORBITAL TO AVERAGE TERRESTRIAL SYSTEM

The orbital plane does not rotate with the earth, but· remains fixed

in the celestial system. The orbital system and the apparent celestial

both have their origins at the centre of gravity of the earth.

From Figure 4-2 we see that when the orbital plane is extended to

meet the celestial sphere, it intersects the celestial equator at the

ascending node (where the satellite croses the equator from south to

north), and the descending node. The angle between the celestial

equator and the orbital plane is the inclination i. The angle between

the ascending node and the line of apsides, measured in the orbital

plane is the argument of pergiee w. The angle between the vernal

equinox and the ascending node, measured in the celestial equatorial

plane is the right ascension of the ascending node n.

The transformation from the orbital system to the apparent

celestial system is

X X

y y 4-9
z z ORB

The transformation from apparent celestial to average terrestrial

is given by equation 3-22, so that

X X

y y 4-10

z A.T.
z
ORB
 99

4. 4 VARIATIONS IN THE ORBITAL ELEMENTS

So far we have assumed that the satellite orbit does not vary with

time, and is completely specified by the six Keplerian orbital elements

a, e, f, w, i, n. The earth 1 s gravitational force field is not

spherically symmetric, as evidenced by geoid undulations and the

equatorial bulge. Also the atmosphere exerts a fluctuating drag force

on the satellite. Because of these and other smaller effects, the

satellite trajectory cannot be assumed to be a fixed ellipse. However,

for each epoch of time T there will be a different orbital ellipse

tangent to the satellite trajectory at that epoch. Each of these

different ellipses will have a set of Keplerian orbital elements, and

if the variation in these orbital elements with time is known, they

are said to describe an osculating orbital ellipse which describes the

satellite trajectory accurately.

The variation with time of the inclination angle is equivalent to

introducing a time-varying out-of-plane component, and this is often

done. For nearly circular orbits, the eccentricity is small to start

with, so variations are usually neglected.

4.5 THE SATELLITE SUBPOINT

The subpoint of a satellite is simply the trace of the path of

the satellite on the ellipsoid (see Figure 4-3}. The coordinates of

the subpoint are given by the geodetic latitude ~ and longitude A of

the ellipsoi.dal normal passing through the satellite.

The average terrestrial Cartesian coordinates of the satellite

are Roown from equation 4-10. From equation 2-32

 10.0 ..

ROTATIONt--•L
AXIS SAT£ lUTE

Figure 4-3 SATELLITE SU8POfNT

 101

X (N + h} cos' COSA

y = Yo + (N + h) cos~ sinA

z
A.T. Zo rnb 2 /a 2 +h) sin'

and if the reference ellipsoid parameters (a, b, x 0 , y0 , z0 ) are known

this equation can be iteratively inverted to solve for (,, A, h) using

the method of section 2.2.6.

4.6 TOPOCENTRIC COORDINATES OF SATELLITE

If we are observing a satellite at position j from a station i

on the earth (see Figure 4-4), then we will require an expression for

the coordinates of the satellite in the local geodetic system of station

i.

If the coordinates (<Pi, A.i_, IJ_) of station i are known with respect

to a reference ellipsoid (a, b, x 0 , y0 , ~0 ) then the geodetic Cartesian

coordinates of i
x.
~

r.~ = yi 4-11

z. G
~

can be computed from equation 2-31.

If the average terrestrial Cartesian coordinates of j have been

computed by the methods outlined in this chapter, then the geodetic

coordinates of j are

xj Xo

rj = yj Yo 4-12

z. Zo
J A.T.
 102

LOCAL GEOOETIC
SYSTEM

\
\
\
\
I

'
I
I
I
I
I
I
I
I
I
I

Figure 4-4 TOPOCEHTPtlC CO()ft{)fNATES (W MTE'LLIT£

 103

The range vector from l to j is

.... ...,
Ilv - xil
I::l~zJI
~.

I J
.....
~
= r. r.
= • 4-13
lJ J l. =iv j v yi

Lzj - z.
l.J G. "
\.1

The coordinates of the range vector can be expressed in the local

geodetic system by using equation 2-67

r~x

l
~y 4-14
~z
L.G.

But from equation 2-66

cos a cos a
r
t:,.y = 6r cos a sin a 4-15
~ G•
fJ.zJ L. lsin a

where the range

4-16

and the altitude a and azimuth a are given by

a = sin -1 4-17

a = -<-an - 1 (!&._) 4-18

" t-,x
 104

5. SUMMARY OF COORDINATE SYSTEMS

In this chapter we will summarize the relationships between the

coordinate systems dealt with in these notes. We will also explain a

duality paradox which has arisen earlier in the notes. This chapter

is, in effect, an explanation of the symbols and abbreviations used in

Figure 5-l.

5.1 TERRESTRIAL SYSTEMS

We discussed five terrestrial systems:

a) I.T. = Instantaneous Terrestrial Coordinate System,

b) A.T. = Average Terrestrial Coordinate System,

c) G. = Geodetic Coordinate System,

d) L.G. = Local Geodetic Coordinate System,

e) L.A. = Local Astronomic Coordinate System,

which are related to each other by the four sets of parameters

a) Polar Motion (xp, yp) -relates I.T. and A.T .. ,

b) Translation of the origin (x 0 , y0 , z0 ) -relates A,T, and G.,

c) Geodetic Latitude and Longitude(~, X}- relates G. and L.G.,

d) Astronomic Latitude and Longitude (~, A} - relates A.T •.

and L.A.
 Figure 5-l

COORDINATE SYSTEMS.

I~ERRESTRIAL ""§ys~EMS]

.....
0
\11

Diurnal

HELIOCENTRIC GEOCENTRIC TOPOCEN'rRIC

 106

5. 2 CELESTIAL SYSTEMS

We discussed four main celestial systems

a) E. = Celestial Ecliptic Coordinate System,

b) R.A. = Celestial Right Ascension Coordinate System,

'
c) H.A. = Celestial Hour Angle Coordinate System,

d) H. = Celestial Horizon Coordinate System,

which are related to each other by the three parameters

a) Obliquity of the ecliptic (e:) - relates E. and R.A.,,

b) Local Sidereal Time (LST)- relates R.A. and H.A.,

c) Astronomic Latitude (~} - relates H.A. and H.

The R.A. system has four variations

a} M.C. =Mean Celestial Coordinate System,

b) T.C. = True Celestial Coordinate System,

c) A.P. = Apparent Place Coordinate System,

d) O.P. = Observed Place Coordinate System.

which all vary with time and thus are defined only when the epoch T

to which they refer is specified.

The parameters relating these systems are

a) Precession and Proper Motion - relates M.C. at standard

epoch T and M.C. at epoch T,

0

b) Nutation - relates M.C. and T.C., both at epoch T,

c) Annual Aberration and Parallax - relates T.C. ru1d A.P.

both at epoch T,

d) Diurnal Aberration, Geocentric Parallax and Refraction -

relates A.P. and O.P. at epoch T.

 107

5.3 DUALITY PARADOX IN THE APPARENT AND

OBSERVED CELESTIAL SYST~fS

The reason there are apparent and observed systems is because

the observer is not at the centre of the celestial sphere (the centre

of the sun) and this must so~ehow be accounted for. There are two

ways of making this correction, and this difference has not been made

explicit earlier in these notes.

First we can retain the true celestial system (with heliocentric

origin) as our coordinate system, and apply corrections to the positions

of the stars. This is the approach described in sections 3.5.4 and

3.5.5 where the aberration and parallax corrections are applied to the
right ascension and declination, and do not change the coordinate system.

Therefore, we then say that the stars have "apparent places" or

"observed places" in the true celestial system.

The second approach is to actually move the origin of the true

celestial system from the centre of, the sun to the centre of the earth

(for the apparent system) and to the observer's position (for the

observed system}. This is what we have done when we related the average

terrestrial to the celestial system in section 3.6 and the orbital

system to the celestial system in section 4.3. In ~his case we called

the shifted true celestial system the "apparent celestial system". In

other words we have adopted the convention that

a) "true" means heliocentric,

b) "apparent" means either geocentric or corrected :for the

heliocentric-geocentric shift,

c} "observed" means either topocentric or corrected for the

heliocentric-topocentric shift.
 108

Two connecting parameters, Proper Motion .. and Refraction, do not

fit in to this second scheme. Proper Motion is the changes in the

positions of stars, and is different for each star. Therefore a

different coordinate system would have to be defined for each star,

which is nonsense. The magnitude of the refraction correction

depends on the incident angle and the ambient conditions. Therefore

specifying that the coordinate system follow refraction would mean a

different coordinate system for each incident angle, all of which would

be jumping around with the temperature and wind.

5 •4 THE CONNECTIONS BETWEEN TERRESTRIAL,

CELESTIAL AND ORBITAL SYSTEMS

The average terrestrial and apparent celestial systems are connected

by GAST (Greenwich Apparent Sidereal Time). Note that the use of

"apparent" in GAST is consistent with the convention we adopted in the

previous section. That is, "apparent" means "geocentricR.

The orbital and apparent celestial systems are connected by the

Euler angles

a} w = argument of perigee,

b) i = orbital inclination,

c) Q =right ascension of the ascending node,

To summarize the differences between terrestrial, celestial and

orbital systems:

a) terrestrial systems rotate and revolve with the earth,

b) celestial systems do not revolve with the earth,

c) orbital systems do not rotate with the earth.

 109

REFERENCES

Bamford, G. (1971). Geodesy, Third Edition, Oxford University Press,

London.

Brouwer, D. and G.M. Clemence (1961). Methods of Celestial Mechanics.

Academic Press, New York.

Goldstein, H. (1950). Classical Mechanics. Addison-Wesley Publishing

Company, Reading, Massachusetts.

Heiskanen, W.A. and H. Moritz (1967). Physical Geodesy. W.H. Freeman

and·co., San Francisco.

Kaula, William M. (1966). Theory of Satellite Geodesy. Blaisdell

Publishing Company, Toronto.

Krakiwsky, E.J. (1968). "Sequential Least Squares Adjustment of

Satellite Triangulation and Trilateration in Combination
with Terrestrial Data." Reports of the Department of
Geodetic Science No. 114, The Ohio State University.

Lambeck, K. (1971). "The Relation of Some Geodetic Datums to a Global

Geocentric Reference System", Bulletin Geodesigue, No. 99,
page 37, March.

Mueller, Ivan I. (1969). Spherical and Practical Astronomy as Applied

to Geodesy. Frederick Ungar Publishing Co., New York.

Nautical Almanac Offices of the United Kingdom and the United States of
America {1961). Explanatory Supplement to the Astronomical
Ephemeris and Nautical Almanac. H.M. Stationary Office,
London.

Schmid, H.H. (1970). "A World Survey Control System and Its Implications
for National Control Networks." In Papers of the 63rd. Annual
Meeting of The Canadian Institute of Surveying, Halifax, 1970.

Thompson, E.H. (1969). An Introduction to the Algebra of the Matrices

With Some Applications. The University of IDoronto Press,
Toronto.

Veis, G. (1960). "Geodetic Uses of Artificial Satellites." Smithsonian

Contributions to Astrophysics, l , 9. Smithsonian Institution,
Washington, D.C.

Veis G. (1966). "Geodetic Interpretation" in Geodetic Parameters for

a 1966 Smithsonian Institution Standard Earth, Volume 1,
Edited by C.A. Lundquist and G. Veis, Smithsonian Astrophysical
Observatory Special Report 2D_Q.
 110

Wells, David E. ( 1971). "Matrices." Department of Surveying Engineering

Lecture Notes No. 15, The University of New Brunswick,
Fredericton.

Wells, D.E. and E.J. Krakiwsky (1971). "The Method of Least Squares."
Department of Surveying Engineering Lecture Notes No. 18,
Univercity of New Brunswick, Fredericton.

Korn, G. A. and T. M. Korn (1968}. "Mathematical Handbook for Scientists

and Engineers" Second Edition, McGraw-Hill, Toronto.

Paul, M. K. ( 1973). "A Note on the Computation of Geodetic (Cartesian)

Coordinates". Bulletin Geodesique, No. 108, page 135, June.

Hotine, M. (1969). "Mathematical Geodesy", ESSA Monograph 2, Washington,

D.C.

Vanicek,,P. (1972). "Physical Geodesy II", Department of Surveying Engineering

Lecture Notes No. 24, University of New Brunswick, Fredericton

Mitchell, H.C. (1948). "Definitions of terms used in geodetic and other

Surveys". U.S. Coast and Geodetic Survey Special Publication 242.

Fischer, I., M. Slutsky, R. Shirley, P. Wyatt (1967). "Geoid Charts of

North and Central America". U.S. Army Map Service Tech. Rep.
No. 62.

Smith, P.A. (1971). "North American Datum" National Academy of Sciences-

National Academy of Engineering, ESSA E53(N), Washington, D.C.

Merry, C.M. and P. Vanicek (1973). "The Geoid and the Positioning of
Geodetic Datums". The Canadian Surveyor, 27, No. 5 (December)

Krakiwsky, E.J., D.E. Wells and D.B. Thomson (1973). "Geodetic Control
From Doppler Satellite Observations for lines under 200 km"
The Canadian Surveyor, 27, 141-148.
 111

APPENDIX A

SUMMARY OF REFLECTION AND ROTATION MATRICES

A.l Orthogonal Transformations

The matrix equation

Y =A X

where A is a matrix and X andY are column vectors, can be regarded as a

linear transformation, in which case the matrix A is called the

transformation matrix. If the two vectors X and Y have the same length,

then both the transformation and the matrix are said to be orthogonal.

Orthogonal matrices have the property that the product of the matrix and

its transpose (or vice versa) is the identity matrix, that is

AT A = A AT = I.
From this property it follows that the determinant of an orthogonal matrix

is either +1 or -1. There are two kinds of orthogonal transformations

called reflections and rotations. The determinant of reflection matrices

is -1, and the determinant of rotation matrices is +1.

There are two interpretations of the linear transformation above. The

first is that the transformation describes the relationship between two

coordinate systems, in which case X and Y are the same vector, but their

elements refer to the two different systems. The second is that the

transformation describes the relationship between different vectors X and Y

in the same coordinate system. In these notes, we are interested only in

the first interpretation.

 112

A.2 Right and Left Handed Cartesian Coordinate Systems

A three dimensional Cartesian coordinate system can be orthogonally

transformed in only six different ways. It can be rotated about each of

its axes. Each of its axes can be reflected. In such a coordinate system,

the vectors X and Y will have only three elements. Let us define the axis

to which the first, second, and third elements of X and Y are referred as

the 1-axis, 2-axis, and 3-axis respectively (we could equally well label

them the x 1 , x 2 , x 3 axes or x, y, z axes).

These three axes may define either a right-handed or a left-handed

coordinate system. Right handed systems follow the right hand rule:

if the fingers of the right hand are curled around any axis so that the

thumb points in the positive direction, then the fingers will point from a

second axis to the third axis, numbered in cyclic fashion. Grasping the

1-axis, the fingers point from the 2-axis to the 3-axis. Grasping the

2-axis, the fingers point from the 3-axis to the 1-axis. Grasping the

3-axis, the fingers point from the 1-axis to the 2-axis. Left-handed

coordinate systems follow the left hand rule, which differs from the above

only in that the left hand is used.

A.3 Reflections

If we denote a reflection of the kth axis by Pk, then the following

expressions define the three reflection matrices:

l~ ~]
0
pl = 1
0

~]
1 0
p2 0 -1
=[ 0 0

-~J
1 0
0 1
p3 =[ 0 0
 113

Note that reflection_matrices commute (e.g. P2P 3 = P3P2 ), so that it

makes no difference in what order a sequence of reflections are performed.

Note also that an odd number of reflections changes the handedness of the

coordinate system.

A.4 Rotations
If we denote a rotation of angle e about the kth axis by Rk(e),
then the following expressions define the three rotation matrices:

~
0 0
R1 (e)
·U cos e
-sin e
sin
cos

[cos e 0 -sin
R2 (e) = ~
s1.n e
1
0
0
cos J
cos e sin e
{ -sin e cos e
0 0

Note that rotation matrices do not commute. The product of several

rotations is performed from right to left, for examrle in

the rotations are performed about the 3-axis of the original system, the

2-axis of the transformed system, and the 1-axis of the doubly transformed

system, to yield the final triply transformed system.

If the rotation angles are all so small that their cosines can be

assumed to be unity, then the rotation matrices become commutative. This

is the case for differential rotations, for example.

The above expressions define positive rotations, which are right-

hand rotations for right-handed coordinate systems and left-hand rotations

 114

for left-handed coordinate systems. A right-hand rotation is related to

the right hand rule given above: if the fingers of the right hand are

curled around the rotation axis so that the thumb points in the positive

direction» then the fingers curl in the direction of a right hand rotation.

A similar statement for left hand rotations is obvious.

A.5 Inverse Transformations

~

The inverse of a transformation A (denoted A-~) is the transformation

which returns conditions to their original state, that is

Re£1ections are self-inverse, that is

Common sense tells us that the inverse of a positive rotation is a

negative rotation, that is

R~1 (e) = Rk(-e)

and this conclusion is verified by taking the orthogonal property

AT A =I
from which it is evident that for orthogonal matrices

and for each of the above expressions for rotation matrices it can be shown that

Applying the rule for the inverse of products

[A B)-l = B-l A-l
we have
 115

A product transformation consisting of one rotation and one

reflection commutes only if the rotation and reflection refer to the same

axis, that is

if j =k
otherwise

if j 1 k.