By Brian Makabayi

Overview
‘New’ Geodesy
absolute geocentric
coordinates Classical Geodesy
Floating height datum,
relative coordinates
Geometrical Physical
Geodesy Geodesy
Where am I? What is uphill? Geometrical Physical
Geodesy Geodesy

The level ellipsoid

Stokes’ problem Potential Theory
Convert gravity to potential
Measure gravity on a surface  Find gravity in space

Theory and practice of Laplace’s equation

modern height systems Fourier transforms (Cartesian coordinates)
Spherical harmonics (Spherical polar coordinates)
 Lecture 1
An overview of
what’s new about
Geometrical Geodesy
Geometrical and Physical Geodesy
 There has long been a division of geodesy into its
geometrical and physical branches but the advent of
satellite-based positioning has made geometrical
geodesy relatively more important both in practice and
philosophically.
 Formerly both parts were mutually dependent.
 Now, the object of geometrical geodesy – in essence
the determination of absolute geocentric coordinates
for accessible points on the Earth – is effectively
independent of the results of physical geodesy.
Geometrical and Physical Geodesy
 Physical geodesy is about knowing what is uphill – a
property determined by the direction of the Earth’s gravity
field.
 In the modern world, physical geodesy has to locate level
surfaces in terms of the purely geometrical coordinates
supplied by geometrical geodesy.
 In the post satellite positioning era, the objects of
geometrical geodesy are available without knowing about
or using the irregular form of the Earth’s gravity field.
 Because geometrical geodesy no longer needs a direct
contribution from physical geodesy but stands alone, in
this textbook, we start our introduction to geodesy at a
different point from an introduction to classical geodesy.
Geometrical and Physical Geodesy
 Now, a logically structured account of geometrical
geodesy can take as given an ability to find Cartesian
coordinates {x,y,z} of an isolated survey point on the
surface of the Earth, in a frame that is geocentric and
global.
 This has been a paradigm-shifting event brought about
by satellite positioning.
Satellite orbits and the coordinate
system origin
 Geometrical geodesy is based on the measurement of
distance.
 In the electronic age, the essence of ‘geometrical
geodesy’ has become the measurement of distances
using microwaves or light, either by measuring their
time of travel or counting the number of wavelengths
along their path.
 Except for small engineering applications, measuring
distances using calibrated rods or suspended wires or
tapes is now largely obsolete.
Satellite orbits and the coordinate
system origin
 Formerly, the distance measured would be between two
points on the Earth’s surface but, nowadays the key
measurement is from an artificial satellite to an observer of
the Earth.
 To generate coordinates from practical measurements of
distance, whether from satellite to ground-point or more
directly between two ground-points, the origin and
orientation of the coordinate system must be measurable
and defined.
 Artificial satellites made it possible for the first time to
create a coordinate system whose origin is both globally
absolute and directly observable.
Satellite orbits and the coordinate
system origin
 Newton showed that the gravitational pull of a
spherically symmetrical object is the same as if all its
mass was concentrated into a point at its centre.
 To a first approximation, the Earth is spherically
symmetrical. Newton also showed that the orbit of one
point mass about another is a perfect ellipse whose
shape and orientation does not change.
 Tracking the motion of the satellite from an
observatory on the Earth determines the shape of its
orbit.
Satellite orbits and the coordinate
system origin
 The centre of mass of the Earth lies at the focus of the
ellipse. With satellite-base geodesy, we can generate a
coordinate system whose origin lies at the centre of the
Earth.
 A slow rotation of the orbital plane and very small
differences between a perfect ellipse and the actual
shape of the orbit reflect non-spherically symmetrical
features of the Earth’s mass distribution, but the
perfect elliptical shape remains a very good
approximation.
Satellite orbits and the coordinate
system origin
 Ultimately billions of measurements of distances between
a hundred or so tracking stations and many different
satellites can be processed to give linked solutions for
1) The position of the centre of mass of the Earth, which
becomes the origin of a geocentric coordinate system;
2) The time-averaged direction of the axis about which the
observatories spin daily, which determines the polar z-
axis of the coordinates {x,y,z}
3) The equation of motion of the satellites in geocentric
coordinates;
4) The geocentric coordinates of the tracking stations;
5) Long wavelength, non-spherically symmetric features of
the Earth’s gravity field.
Orientation and the Earth’s
rotation axis
 The simplest system of coordinates uses the
rectangular Cartesian coordinates {x,y,z}.
 The laws of mechanics ensure that axis of the Earth’s
daily rotation must pass through its centre of mass, so
rotation can be used to define the direction of the z-
axis of the coordinate system.
 In detail, the Earth’s spin has a small wobble and the
tectonic movements of the solid parts of the Earth
cause tiny changes in the shape of the Earth, so
defining the orientation of a global coordinate system
needs care and refinement.
 Tracking an The coordinate system origin
artificial satellite
orbit determines m M ~ 1024 kg
its shape
m ~ 104 kg
and that tells you
where the centre m R
of the Earth is M m R

because …

the focus of the

elliptical orbit
of an
artificial satellite
lies at the Before artificial satellites,
centre of mass m was the Moon,
of the Earth- for which m/M ~ 1/80
satellite system and not known very precisely
 The Earth’s spin
defines the z-coordinate axis

(0.01″ ~ 1 foot)
Orientation and the Earth’s
rotation axis
 Careful analysis by the International Earth Rotation
Service (IERS) allows us to adopt a conventional
direction for the z-axis based on the average position
of the corrected rotation axis at a particular epoch.
 The corresponding direction of the x-axes is not based
on Earth dynamics but is a matter of historical
convention, originally related to a reference mark at
the Greenwich Observatory in England.
 Satellite positioning

satellite orbit O
sAB A

sBP

sAP

Triangulation (tri-lateration)
Measuring the distance to two
P
ground point points on the known orbit of
the satellite is like terrestrial
triangulation
Satellite positioning gives geocentric
{x,y,z} of a surface point
z

P(x,y,z)

y
x
 You must define an ellipsoid
in order to transform
rectangular Cartesian coordinates
to
latitude and longitude

How do you choose

the size and shape
of the ellipsoid?
Choosing a reference
Mean sea
ellipsoid level

What is
mean sea level? exact
ellipsoid

Look later at how

to choose a
better datum
 Satellite radar altimetry
gives
geocentric coordinates for a complete grid of surface points

30-90m grid
height errors
 2 - 10 m

plus sparser but much more precise GPS points

 Earth’s surface measured as a polyhedron in

geocentric ellipsoidal coordinates
Topographic height statistics
 Don’t choose an ellipsoid that fits the
topographic surface best

Classical process involves

a fit of the ellipsoid to
‘mean sea level’

– what does ‘mean sea level’ mean?

Now determine the shape from satellite gravity

and the size from a best fit to the coordinates of
satellite tracking stations
 Geodetic latitude
is only
z defined with respect to the parameters of
a particular ellipsoid

b P (x,y,z)
tan   (1  e2 ) tan 
r


The ellipsoid
R  
x
O A
d
Q a
 Transformations
GPS gives {x,y,z} but we often want {h,,l}

 
x    h  cos  cos l
a
 1  e 2 sin 2  
 
  Forward transform –
y   a
 h  cos  sin l
 1  e 2 sin 2   closed formulae
 
 a (1  e 2 ) 
z    h  sin 
 1  e 2 sin 2  
 
y
 tan l
Inverse transform
x
z  he 2 
no closed formulae  tan  1  
(1  e 2 ) ( x 2  y 2 )  (   h) 
– only iterative solution
z2 2  2he2 h 2e 4 2 
x y 
2 2
 (   h) 
1  sin 2
  sin  
(1  e )  (   h) (   h)
2 2 2

 What GPS does and doesn’t do!
Satellite positioning systems provide coordinates that are geocentric
and absolute – distances measured in metres.

We know where we are and the position and shape of the Earth’s
surface in these coordinates, so we have everything we need for
navigation and for aircraft to avoid bumping into mountains!

Geometric coordinates do not tell you what is level and which

direction is uphill – these are determined by the Earth’s gravity field
and are the province of physical geodesy, only necessary (but then
essential) when dealing with the energy needed to transport mass.

‘Energy-based’ heights are needed to plan road and rail inclines,

hydroelectric power plants, ‘gravity-feed’ irrigation, pipelines, etc
 Absolute vertical coordinates
– the crucual innovation

‘New’ geodesy Classical geodesy

levelling plus
GPS gives gravity H
h
this directly estimated inside
the Earth

Gravity inside the Earth geoid

and a worldwide integral N
Ellipsoid
 Overview
‘New’ Geodesy
absolute geocentric
coordinates Classical Geodesy
Floating height datum,
relative coordinates
Geometrical Physical
Geodesy Geodesy
Where am I? What is uphill? Geometrical Physical
Geodesy Geodesy

The level ellipsoid

Stokes’ problem Potential Theory
Convert gravity to potential
Measure gravity on a surface  Find gravity in space

Theory and practice of Laplace’s equation

modern height systems Fourier transforms (Cartesian coordinates)
Spherical harmonics (Spherical polar coordinates)
 Lecture 2
Introduction to Physical Geodesy

1 Gravity, equipotential surfaces and height

Newton’s Law of Universal Gravitation
for a point mass
δF (r2 )   r  r 
 G  (r1 )  V1   1   2 
δg (r2 )  2 
1

m2  r  r   r2  r1 
 2 1 

mass inverse square unit

element term vector

m
r2  r1
r2 2   G  (r1 ) δV1
r2  r1
3
m – r
1 r r1 2
1
The effect of an extended body is found by integrating the point mass formula


r2  r1
g (r2 )   G  (r1 ) dV1
r2  r1
3
V1

The gradient operator  operating of the inverse distance gives

 1  r2  r1
2    
r
 2  r1  r2  r1
3

So we can write gravity as the gradient of a scalar potential V

g (r2 )   2V (r2 )
where


1
V (r2 )  G  (r1 ) dV1
V1
r2  r1
 Level surfaces
V is the differential (ie the slope) of V in the
direction of most rapid change.

This property makes V at right angles to a level surface

‘horizontal’
suface on which the potential V is a constant – equipotential
an equipotential surface

There can therefore be no component of

gravitational force that is tangential to an V = constant
equipotential surface
– an equipotential surface is level
g  V
Provided that there are no other forces acting,
the surface of a fluid at rest will be level
We also use the word horizontal to describe a
level surface
 horizontal is bumpy

go - background gravity of the bulk Earth

plus g - effect of local high density mass

g

g
go
 horizontal is bumpy

go - background gravity of the bulk Earth

plus g - effect of local high density mass
= g – the resultant net gravity is deflected

g

g go
 horizontal is bumpy

‘Vertical’ is the direction of net gravity

 horizontal is bumpy

‘Vertical’ is the direction of net gravity

‘Horizontal’ is a surface at right angles to vertical

 horizontal is bumpy

On the ‘horizontal’, a marble does not roll away

because there is
no component of gravity parallel to the surface
– no where on it is ‘down hill’.
the surface of a fluid at rest will be horizontal

Mean sea level geoid

 We shall see later that the effect of measuring gravity on a rotating Earth is to add
another ‘inertial’ force – the centrifugal force – whose effects are indistinguishable from
‘gravitational’ gravity.
Like gravity, the centrifugal force can be derived as the gradient of a potential, F, so the
net potential W is the sum of both components.
W is called the geopotential
W = V + F
and, what we observe as gravity, is
|g| = | W| 

W

H
 H is an increment of distance along the normal to the equipotential surface
W
so H  
|g| W = W2
H12
Level surfaces are closer together when
gravity is stronger, even though each has a W = W1
constant ‘height’
 Classical heighting uses levelling

Now use  H3
W  H4
|g| 
H H2 P4
 H1 P3

P2
W  | g |H P1

to give


N

WN 0  | g |H  W0
i 0

or


N

cN  W0  WN 0  | g |H
i 0
 Since 1956, the adopted measure of height has been
a direct measure of the gravitational potential energy
Each level surface heighting is labelled by W
and height measured in metres is no longer definitive
(but has remained in widespread use)

c is the geopotential number at a point with coordinates (h,,l)

and Wo is the potential on the zero height datum.

W is the gravitational potential energy

and depends only on where point N is
– it does not depend on the route taken by the levelling

 
N N

cN  | g |H H
i 0 i 0
If is unique, cannot be unique
Geopotential heights without levelling
The essence of heighting for ‘new geodesy’ is determine the geopotential
number directly from a model of the geopotential expressed in terms of
ellipsoidal coordinates
The geopotential number c at a point with coordinates (h,,l) is then
c(h,,l) = Wo – W(h,,l
rather than

(Because the potential decreases with height, this definition ensures that the
geopotential number increases upwards!)

The geopotential number now serves the purpose of the ‘energy-based’

form of height for all cases where the ellipsoidal height is inappropriate


N

cN  W0  WN 0  | g |H
i 0
 How can we find W(h,,l)?
Modelling
Calculate the gravity field (potential and gravity) for an object
with a known shape (perhaps plus density or some other defining property)
This approach is used for example to model the ‘normal’ Earth
and the topographic masses.
We subtract gravity calculated from the models from real gravity
to make residual gravity very small
Potential Theory
Laplace’s equation can transform residual gravity to potential
If a function satisfies Laplace’s equation, measurements over a
coordinate surface determine it in the space above (or below) the surface

 2T   2T  2T 
   2  
2 
z 2  x y 
 Classical geodesy
Pre-satellite geodetic technology could
* measure angles very precisely, at best ~ 0.02 arc second (10-7 radians);
* measure distances (few km to < 0.1mm), referenced to the wavelength
of light, using calibrated invar wires or tapes
* determine positions and large horizontal distances through a network of
triangulation stations, or, latterly, though microwave trilateration;
* determine the absolute direction in space of the local vertical and the
Earth’s spin axis using observations of stars and an adopted star
catalogue (astronomical latitude and longitude);
* compute geodetic latitude from a measured horizontal distance over the
surface of the ellipsoid;
* determine the geopotential number by levelling from mean sea level at
a coastal tide gauge;
* measure gravity at accessible points on land and on ships at sea
Problems
 Regionally different reference ellipsoids
origin of local ellipsoid chosen to minimise local geoid heights
not absolute or geocentric
 Floating height datum
Geopotential number converted to geometric height assuming
gravity inside the Earth; position of geoid unknown
 Stokes’ problem
Need a global integral of gravity anomalies to compute the height
of the geoid above the ellipsoid, in order to find the absolute the ellipsoidal
height of a surface point
 The gravity anomaly
Cannot compare real and normal gravity at the same place
– rotational potential does not cancel when forming the anomalous
potential
 No global gravity model
Stokes’ integral had to deal with the whole of gravity anomaly not a small
residual; no control of far-field effects.
 Note on the meaning of words
original literal meaning came to new word
word of the word mean for original
meaning

Geometry “measuring the Earth” “study of shapes” geodesy

Geodesy “dividing the Earth “measuring the Earth” cadastral

into land parcels survey
for taxation”

Trigonometry “measuring three “relations between Triangulation

angles” angles and sides
of a triangle”
========================================================
Geometrical (measuring the Earth)2 The part of geodesy
Geodesy based on the
measurement of distance
 Bouguer’s 1743 survey of the Peruvian Andes
– first baseline

Bouguer’s first stage was to construct a

baseline. By placing his measuring rods
end-to-end over a 12 225.737 m stretch of
Expansion by triangulation

with repeat base at southern end

 Fitting an Ellipsoid to the Earth’s level surface
Bouguer’s expedition to Peru combined with Clairaut’s to Lapland
Both mMeasured astronomical latitude using known star positions
Both measured the length of an arc measured over the ellipsoid
Modern theory gives the length of an arc from the Equator as
 1 3 5 6 
s  a   1  e 2  e 4  e  
 4 64 256 
3 2  [e2 = f (2 – f)]
 sin 2 
1 15 4
 e  1  e 2  e 
8  4 128 
15 4  
 sin 4 
3
 e  1  e2  
256  4 
35 6   
 e  1    sin 6    
3072   
They had to assume that astronomical latitude = geodetic latitude,
so then the two surveys gave a and f
 More modern fitting of an Ellipsoid to the Earth’s
level surface by varying it size, shape and tilt

By measuring both astronomical latitude

(star-gazing) and geodetic latitude (arc
lengths), the difference – the deflection of the
vertical – maps non-ellipsoidal irregularities
in the geoid.
The mismatch can be minimised the
varying the size and flattening of the
ellipsoid, but this is usually best done on
a continent sized region.
 More modern fitting of an Ellipsoid to the Earth’s
level surface by varying it size, shape and tilt
The next step in minimising the mismatch is to tilt reference ellipsoid
to make it fit the regional geoid better.

Fix a common datum point on the initial ellipsoid and the one to be
tilted. Then swing the line to the centre of the new ellipsoid about this
pivot making its centre moves away from the original centre. This
process aims to minimise the deflection of the vertical at all the
astrogeodetic points on the regional network.

For the 1927 North American Datum (NAD27), setting the deflection of
the vertical to be zero at an astrogeodetic point at Meades Ranch in
Kansas achieved a national best fit.

We can see the effect of the tilting the ellipsoid by looking at the origin
of the local ellipsoid with respect to the Earth’s centre of mass, now
known from satellite positioning. Some early post-satellite
measurements gave
Ethiopia and Sudan DXo = -166 ± 5 m; DYo = -15 ± 5 m; DZo = 204 ±3 m
South Africa DXo = -136 ± 3 m; DYo = -108 ± 6 m; DZo = -292 ±6 m
 Geoid height above the GRS80 ellipsoid
computed from the Earth Gravity Model 1996

Looking at the global

map of a geoid
computed with respect
to the GRS80 ellipsoid
(a geocentric one)
makes it clear why the
centres of the East
African ellipsoids have
a northerly component
of rotation while the
South African ellipsoid
has a southerly
component.

metre
s
Slides illustrating gravity height corrections
and their effect
 11823
gravity
measurement
points
on land
in the
Grampian
Highlands

Data square
200 by 200 km
Measured
gravity

Variations
have a range
of

410 ppm
 Height
Range up to
1500 m
 Height Gravity
Gravity variations correlate with
an up-side-down topographic map
ie, gravity is low where topography is high
 There are two competing effects
that change the gravity measured on a hill
compared with that at sea level

* the free-air effect – just the result of getting

further from the attracting centre of the Earth

* the effect of topographic masses

– on the top of a hill, there is the additional
attraction of the mass between you and
sea level
 GM
go 
R g g R2
o GM
R+h h gh 
( R  h) 2
GM 1

R2  h 2
1  
 (binomial
R  theorem)
But, for small x, (1 x) n  1  n x
 h
Free air Thus g h  g o  g o 1  2   g o
 R
correction 2g 2  9.8

R 6378 000
h
gh  go   2 go  3.0   10  6 m s - 2 m -1
R

 3.0  gu m -1
 Free air anomaly
 2g 
Dg FA  g obs    h
 R 
2g
 g obs    h
R
The free air anomaly is the residual from a model that does not include
the attraction of topographic masses.

Therefore it should correlate positively with local topographic roughness.

Locally, it should be large and rough,

disguising the effects of subsurface density variations

ie geology
 Freeair
anomaly

Make
corrections for
gravity
measured at
an increased
distance from
the centre of
the Earth
(cf Newton’s
inverse square
law).

Variations
have a range
of
160 ppm.
 Height Freeair anomaly
Freeair anomaly correlates with
a ‘right-way-up’ topographic map
ie freeair anomaly is high where topography is high
 The Bouguer anomaly

corrects for both effects of topography

– the free air correction
(increased distance from the centre of
the Earth) and
– the Bouguer correction
(attraction of the extra mass between
the ground surface and sea level)
 Attraction of the topography

To make the Bouguer correction

you need to compute the gravitational attraction
of the topographic masses
– a tough task for rough terrain
– (now use (super-)computers and satellite-derived
digital terrain models)
Bouguer devised a simple and usually good approximation
– represent the terrain by an infinite horizontal slab
whose thickness is equal to the measurement height

g
Then just a very simple formula

 g  2 G  h h

 Calculating the
gravitational attraction of
the topographic masses

 gT
Attraction of
the whole 
density 
topography

 gB
Attraction of
Bouguer 
slab density 

Attraction of
T
topographic 
relief  
 Bouguer anomaly
 2g 
Dg B  g obs    h   gT 
 R 
2g
 g obs    h   gT
R

The Bouguer anomaly is the residual from a model that includes

the attraction of topographic masses.

Therefore it should not correlate with local topographic roughness.

It should be smooth and small,

reflecting only the effects of subsurface density variations
 Bouguer
anomaly
Corrects for
attraction of
topographic
masses as
well as the
inverse
square law.

Locally much
smoother

Variations
have a range
of
97 ppm.
 Height Bouguer anomaly

Bouguer anomaly does not correlate locally with height

Problems of geodesy
The problems
 This section identifies some of the ways in which
classical geodesy dealt with not having readily
observable absolute coordinates or a global gravity
model.
 Both of these benefits only came with the satellite era.
Regionally different reference
ellipsoids
 The parameters and orientation of the reference ellipsoid,
essential for transforming distance measurements to
geodetic latitude, had to be found from a regional fit of the
horizontal surface to mean sea level.
 Because of long-wavelength bumps on the horizontal
surface, different regions of the world found that their
locally best fitting ellipsoid was different in size, shape and
orientation from their neighbours.
 Consequently, latitude and longitude were differently
defined in different regions and there was no precise way to
locate the Earth’s centre of mass and so make it the origin
of the coordinate system.
The floating height datum
 There was no directly observable distance coordinate in the
vertical direction or, alternatively, in the direction normal
to the ellipsoid.
 Levelling rigorously determined a difference in potential,
that is the geopotential number, but this had to be
converted to a geometrical distance, called orthometric
height, by making assumptions about gravity inside the
Earth.
 Orthometric height H was a vertical distance above a
hypothetical equipotential surface inside the Earth called
the geoid, supposed to be fixed by ‘mean sea level’.
The floating height datum
 However, a surveyor did not know where the geoid was
a priori – his height datum was a freely floating
surface:
 even orthometric height constructed from levelling
only told you how far the geoid was below the Earth’s
surface but gave no information about the absolute
location of either the geoid or the Earth’s surface.
Stokes’ problem
 Finding out where the geoid was depended on solving
Laplace’s equation: Stokes derived a surface integral of
gravity over the whole Earth that determined an anomalous
potential T on the ellipsoid.
 Once again, a conventional value of gravity was needed to
convert a difference in potential to a difference in
geometrical height.
 This was N, the height of the geoid above the ellipsoid.
With these assumptions, the sum of the geoid height and
orthometric height gave a vertical coordinate for a point on
the Earth’s surface, in essence defining a geometrical
distance to the centre of the ellipsoid.
Stokes’ problem
 The part involving height of the Earth’s surface above
the ellipsoid was found as

hHN
The gravity anomaly
 There is a practical difficulty in this sequence of
computations that is most clearly seen by returning to
the stage of estimating the geoid height N from a
surface integral of gravity.
 Because it was based ultimately on a solution to
Laplace’s equation, Stokes’ surface integral needed a
gravity residual in a non-rotating, mass-free region of
space.
 Like the modern gravity disturbance, this residual
gravity would be characterised by the difference
between real gravity and normal gravity.
The gravity anomaly
 However, now these quantities could not be evaluated at
the same place.
 The effect of the reference Earth model could only be
calculated at points with known geocentric coordinates,
that is, on the ellipsoid or at a known height H above it.
 In contrast, gravity observations were made on the
topographic surface.
 This was at a height H + N above the ellipsoid. Comparison
of these two quantities at different places produced a more
complicated quantity called the gravity anomaly, Δg, rather
than the gravity disturbance, δg, used by modern
computations.
The gravity anomaly
 where the arguments are heights above the ellipsoid.
Dg  g H  N   l H 
 The principal difficulty was that N remained unknown
until Stokes’ integral had been evaluated so the
process had not only to convert gravity to potential but
also solve for where the input gravity data were.
Lack of a global gravity model
 Molodenskii devised a philosophically more
satisfactory solution that did not need to estimate
gravity inside the Earth but, like Stokes’ approach,
needed a surface integral of gravity.
 The global coverage of good gravity data was very
patchy – very little over most of the oceans or large
tracts of the undeveloped world.
 This meant that there was no reliable spherical
harmonic model of the Earth’s gravity field to subtract
and remove longer wavelength features from local
data.
Lack of a global gravity model
 This had two bad effects: first, the local surface
integrals had to be global and difficult research was
needed to minimise the effect of the poorly known far-
field gravity data;
 secondly, they had to deal with the whole of the gravity
anomaly, rather than just a much smaller residual left
after subtracting a global gravity model.
 With a very much larger integrand, the calculation had
to be performed much more precisely.
Levelling and geopotential height
Geopotential number and orthometric height
 Classical geodesy uses exactly the same approach as
‘New Geodesy’ in its use of levelling to determine the
geopotential number .
 The use of Stokes’ integral theorem showed that the
integral along telescope sight lines and up levelling
staves determined uniquely the potential difference
between the datum point P1 and some other levelling
point, here P4: the result was independent of the path
taken. p4

C p1 p2  W p1  W p4   g.dl   g i DH i
p1 i
Geopotential number and
orthometric height

Figure 4  z3
 z4

 z2 P4
 z1
P3
H

P2

P1
Geopotential number and
orthometric height
 The equation serves to define the orthometric height
at any point P
Cp
Hp 
gp
 The logical difficulty with this approach is that it
needs g p , an average value of gravity inside the Earth.
 Average gravity will depend on how the rock density
varies not only along the descent but more widely
inside the Earth.
Geopotential number and
orthometric height
 In principle, we cannot know this. Helmert devised an
approximate way to estimate g p . Like Stokes’ solution
to estimate the height of the geoid above the ellipsoid,
he started with the Bouguer anomaly.
The Bouguer anomaly
 The Bouguer anomaly ΔgB describes the effect of
unknown density variations below the topographic
surface.
 Although the gravity disturbance, δgB, is not a feature
of classical geodesy, using it instead of the gravity
anomaly simplifies an explanation of Helmert’s
orthometric height.
 The distinction between anomaly and disturbance is
important in other contexts but not here.
gB ( z)  g ( z)   ( z)  DgT ( z)
The Bouguer anomaly
 The ‘model’ part of gravity now includes the attraction of
the topographic masses,
ΔgT (z), in addition to the attraction of the ‘normal Earth’
model γ(z), the contribution from the rotating level
ellipsoid.
 Because the attraction of the topographic masses is rather
well approximated by an infinite horizontal slab whose
thickness is equal to the topographic height at the
observation point,
 it can be convenient to represent the attraction of the
topography by the very simple formula for the infinite slab
plus a correction called the terrain correction, Tc,
accounting for deviations of the topographic surface from
an infinite plain.
The Bouguer anomaly
 The attraction of the Figure 5

whole topography is  gT


density 
DgT  Dg B  Tc
 Attraction of the
Bouguer slab is
 gB

Dg B  2 G  H density 

 Attraction of
topographic relief is the Tc 
terrain correction, Tc 

The Bouguer anomaly
 The terrain correction Tc Figure 5

has to be computed  gT


numerically but is often density 

small enough to be
neglected.
 However, what power the  gB

Bouguer anomaly does 

density 

have comes from

‘geology’, that is,
variations in subsurface Tc 
rock density. 

The Bouguer anomaly
 If there are significant density anomalies within the
topographic masses, they will in principle cause the
downward continuation operator to become large and
eventually unstable.
 Nowadays, the downward continuation problem does
not arise for heighting because the only necessary
output for modern levelling is the geopotential
number on the topographic surface.
ELEMENTS
Introduction
 The law of gravitation allows to calculate a body’s
gravitational potential and attraction if its density
distribution and shape are given.
 In most real life situations the density distribution is
unknown and in physical geodesy even the shape of
the body - the geoid – must be considered unknown.
 The next question is whether the exterior field can be
determined from the function (the potential or its
derivatives ) on the surface. This is a boundary value
problem.
Introduction
 Potential theory is the branch of mathematical physics that
deals with potentials and boundary value problems.
 Its tools are vector calculus, partial differentials equations,
integral equations and several theorems and identities of
gauss (divergence) , stokes (rotation) and green.
 Potential theory describes the behavior of potentials f any
type.
 Thus it finds applications in such diverse disciplines as
electro-magnetics , hydrodynamics and gravitational
theory.
Introduction
 Here we will be concerned with gravitational
potentials and the corresponding boundary value
problems.
Some vector calculus rules
Some vector calculus rules
Divergence - Gauss
 Vector flow – our treatment of the gauss divergence
theorem begins with the concept of vector flow
through a surface.
 Vector flow is loosely speaking the amount of vectors
going through a certain surface - one could think of
water flowing through a section of a river.
 The amount of vectors is quantified by taking the
scalar product of the vector field and the surface
normal.
Divergence - Gauss
Divergence - Gauss
Divergence - Gauss
Divergence - Gauss
Divergence - Gauss
Divergence - Gauss
Divergence - Gauss
Boundary value problem
 Starting of this chapter was the question, whether we
can determine the gravitational field in outer space
without knowing the density structure of the earth,
but the knowledge of the potential on the boundary.
 In our particular case we have two different partial
differential equations used in the boundary value
determination problem.
 The poisson equation leading to an interior BVP, and
 Laplace equation leading to the exterior BVP
 We will only be concerned with the exterior BVP hence
only the Laplace equation
Boundary value problem
Boundary value problem
Solving laplace’s equation
Solving laplace’s equation
Solving laplace’s equation
Solving laplace’s equation
 Solution of dirichlet and Neumann BVPs in x, y and z
 DIRICHLET’S SOLUTION
Solving laplace’s equation
 NEUMANN’S SOLUTION
Solving laplace’s equation
Solving Laplace's equation
Solving Laplace's equation
Solving Laplace's equation
Assignment
1) Solve the Laplace’s equation both in Cartesian and
spherical coordinates?
 This assignment is individual and should be handed
in for assessment to contribute towards your final
mark in the examination.
 It should be hand written and submitted on 22nd Oct
2014 by 12pm.
Gravity Attraction and Potential
Gravity Attraction and Potential
 Gravimetry
 Gravimetry is the measurement of gravity
Gravimetric measurements :
principles:pendulum
Gravimetric measurements :
principles:pendulum
Gravimetric measurements :
principles: spring
Gravimetric measurements :
principles: spring
Astatized spring: LaCoste-
Romberg design
 Lacoste developed the
concept of a zero length
spring as a graduate
student at the university
of Texas in 1932.
Gravimetric measurements :
principles: free fall
Gravimetric measurements :
principles: free fall
Corrections to be made
Gravimetric measurements :
principles: free fall
Gravimetric measurements :
principles: free fall
Gravimetric measurements :
principles: free fall
Gravity Networks
Gravity networks
Gravity networks
Global Gravity Models
Relative gravimetry
 In the pre-satellite era, global
gravity models had to be
constructed from terrestrial
gravity anomalies.
 Most land and marine gravity
data were acquired with portable
gravity meters only capable of
measuring gravity differences
 The disadvantages of this
technique were: first the logistic
one that every point where
Figure 10.1 Field measurement
gravity was needed had to be with a LaCoste & Romberg gravity
visited to make a measurement; meter.
Global Gravity Models
 secondly that this relative survey
had to be made absolute by
being tied in to a recognised
system of reference stations
where the absolute value of
gravity was known.
 In addition, the gravity reading
had to be combined with a
determination of height because
the geodesist wanted a free air
anomaly rather than a raw
gravity value.
Figure 10.1 Field measurement
with a LaCoste & Romberg gravity
meter.
Global Gravity Models
 In the mid 1990s freefall
absolute gravity meters because
available on an almost
commercial basis and now there
is access within most countries
to a gravity site individually
measured with an absolute free-
fall instrument.
 Modern absolute gravity meters
have routine accuracies of about
20-50nm s-2 so the need for an
internationally adopted gravity
reference has disappeared Figure 10.3: Three generations of free-fall
absolute gravity meter at the Bureau of Weights
and Measures BIPM at Sevres, outside Paris.
Satellite sea surface altimetry.
 Before looking at a satellite as a
sensor of the gravity field it
experiences in space, we look at
satellite sea-surface altimetric
satellites.
 They measure from space but
deduce gravity on the Earth’s
surface, so have a foot in both
camps.
 Notable altimetric satellites have
been some early military
satellites called GeoSat, SEASAT,
Earth Resources Satellites ERS1
and ERS2, TOPEX/Poseidon and
Jason 1 and 2.
Satellite sea surface altimetry.
 In the initial phase, the satellite
is tracked from the ground to
establish its orbital parameters
so that its position at any instant
becomes accurately predictable.
 In the data collection phase, the
satellite transmits a narrow-
beam radar pulse vertically
downwards.
 Because of its salinity, sea water
has a high conductivity so acts
like a perfect mirror to
microwaves and reflects the
pulse back to the satellite.
Satellite sea surface altimetry.
 From a measure of the there-and-
back travel time, the satellite
computes the distance to the nadir
point (the point vertically below the
satellite) and thence the ellipsoidal
height of the sea surface.
 There are systematic effects to
correct: tides; the non-sinusoidal
shape of ocean wave makes their
troughs occupy a larger area than
their crests; and the effect of Earth
curvature – more pronounced with
rough seas because normal
reflection from the sloping limbs of
waves can come from much further
away from the nadir point.
Satellite sea surface altimetry.
 After processing, the results give
along-track estimates of the
geoid height, N, plus any
dynamically sustained elevation
of the sea-surface from the
geoid, so-called seasurface
topography, ζ.
 Time-average sea-surface
altimetry data are now available
on 1-2 km grids and have vertical
accuracies of ~ 2 cm. Except for
polar regions with pack ice
cover, oceanic data coverage is
uniform and global.
Satellite sea surface altimetry.
 The best way to separate sea
surface topography ζ from the
true height of the
equipotential N combines the
satellite altimetrty with marin
egravimetry and a model for ζ
based on hydrography.
Subtracting ζ from the
provisional oceanographic
model from the ellipsoidal
height given by altimetry
gives a provisional value for
N;
Satellite gravity measurement
Geodetic satellites
 Geodetic’ satellites tend to be
simple and designed so that
being a freely accelerating mass
is their first priority.
 They are small, have a high
density and a simple geometrical
shape
 All features that minimise or
make calculable the effect of air
drag compared with gravity.
 LAGEOS was a sphere covered Figure 10.9: LAGEOS
with cube-corner reflectors and The Laser Geodynamics Satellite.
tracked by laser pulses from 60 cm diameter
large Earth-based telescopes.
Satellite gravity measurement
Geodetic satellites
 The LAGEOS satellites
are still the best way of
finding the very low
spherical harmonic
degrees of the Earth’s
gravity field (n < ~ 12).

Figure 10.9: LAGEOS

The Laser Geodynamics Satellite.
60 cm diameter
Satellite gravity measurement
High-Low satellite pairs
 A satellite in an orbit with a
large radius will be little
affected by perturbations due
to the higher degree
harmonics.
 If the latter are well
determined by careful
tracking with geodetic
satellites, the high altitude
satellite becomes a known Figure 10.10: CHAMP
datum from which to track
satellites in lower altitude
orbits.
Satellite gravity measurement
 The advantage of this is that
signals from ground based
tracking stations must pass
through the lower
atmosphere, where the effect
of water vapour on the speed
of microwaves introduces
errors.
 An example of this tracking
geometry was the satellite
CHAMP was launched in July
2000 and tracked by GPS Figure 10.10: CHAMP
satellites
Satellite gravity measurement
Low-Low satellite pairs
 GRACE was launched in May 2002
and consists of two identical
satellites, one following about 200
km behind the other in the same
orbit.
 This produces a long baseline
gradiometer when onboard
microwave transmitters monitor
the distance between the two.
 In principle it will not be able to
detect gravity field variation with
wavelengths significantly less than Figure 10.11: GRACE
twice the baseline, say 400 km,
that is, spherical harmonics
beyond degree 100.
Satellite gravity measurement
 Equally, the 200 km baseline will
not be efficient at measuring the
gentle gradients due to very long
wavelength components.
 Although the GRACE mission
succeeded in its aim of getting
snapshots of the gravity field as
a time series in order to
monitors climatological
contributions to the Earth’s
gravity field, the spin off of a
very high precision static field
has made an important Figure 10.11: GRACE
contribution.
Satellite gravity measurement
 A combination of GRACE and
geodetic satellites like
LAGEOS makes the most
important contribution to the
global gravity model for
degrees up to ~80
(wavelengths ~ 500 km) and
continues to make some
contribution to a combined
global model up to degree n ~
130 (wavelengths ~ 300 km).
Figure 10.11: GRACE
Satellite gravity measurement
On-board gradiometer
 The GOCE satellite was
launched in October
2009.
 It carries an on-board
gradiometer that
measures the difference in
acceleration across a
baseline of about 10 cm.
 This gives all components
of the gravity gradient
tensor. Figure 10.12: The GOCE satellite and gradiometer components.
Satellite gravity measurement
On-board gradiometer
 In order to help with this
almost unbelievably
demanding task, GOCE
has a very low altitude
orbit to maximise the
amplitude of gravity
variations but will pay the
penalty of a short life time.

Figure 10.12: The GOCE satellite and gradiometer components.

Satellite gravity measurement
On-board gradiometer
 Air drag will cause the
orbital radius to decrease
until it burns up in the
atmosphere.
 GOCE has already
produced a good gravity
field model for
wavelengths greater than
about 167 km (n > 240)

Figure 10.12: The GOCE satellite and gradiometer components.

END