Professional Documents
Culture Documents
MEHDI ESHAGH
Elsevier
Radarweg 29, PO Box 211, 1000 AE Amsterdam, Netherlands
The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom
50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States
No part of this publication may be reproduced or transmitted in any form or by any means,
electronic or mechanical, including photocopying, recording, or any information storage
and retrieval system, without permission in writing from the publisher. Details on how to
seek permission, further information about the Publisher’s permissions policies and our
arrangements with organizations such as the Copyright Clearance Center and the Copyright
Licensing Agency, can be found at our website: www.elsevier.com/permissions.
This book and the individual contributions contained in it are protected under copyright by
the Publisher (other than as may be noted herein).
Notices
Knowledge and best practice in this field are constantly changing. As new research and
experience broaden our understanding, changes in research methods, professional
practices, or medical treatment may become necessary.
Practitioners and researchers must always rely on their own experience and knowledge in
evaluating and using any information, methods, compounds, or experiments described
herein. In using such information or methods they should be mindful of their own safety
and the safety of others, including parties for whom they have a professional responsibility.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or
editors, assume any liability for any injury and/or damage to persons or property as a
matter of products liability, negligence or otherwise, or from any use or operation of any
methods, products, instructions, or ideas contained in the material herein.
ISBN: 978-0-12-816936-0
xi
Acknowledgements
xiii
CHAPTER 1
¼ 4pvnm .
(1.7)
Spherical harmonics and potential theory 3
By inserting Eq. (1.8) into Eq. (1.5) and after applying the addition
theorem of spherical harmonics (Eq. 1.9), we obtain another formula for
the Laplace coefficient of the potential:
ZZ
2n þ 1
vn ðq; lÞ ¼ V ðR; q0 ; l0 ÞPn ðcos jÞds. (1.11)
4p
s
n¼0 m¼n
r
s
ZZ (1.12)
0 0
vnm Ynm ðq; lÞYn0 m0 ðq ; l Þds.
s
where Pðr; jÞ is called the kernel of the integral, or the Green function, and
it will have the following spectral form, or Legendre expansion:
XN nþ1
R
Pðr; jÞ ¼ ð2n þ 1Þ Pn ðcos jÞ. (1.18)
n¼0
r
Eq. (1.17) is known as the Poisson integral and Eq. (1.18) as the Poisson
kernel, which is convergent, and a closed-form formula can be derived for
that. Consider the following expression for the Legendre polynomials
(Hobson, 1965):
1 XN pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi R
¼ t n Pn ðe
xÞ where D ¼ 1 þ t 2 2te x; ex ¼ cos j and t ¼ .
D n¼0 r
(1.19)
Spherical harmonics and potential theory 7
when r ¼ R, then t ¼ 1, and the kernel will change to the spherical Diract
delta function, which gives N when j ¼ 0, and for js0, the kernel has
zero value:
X
N
dðjÞ ¼ ð2n þ 1ÞPn ðcos jÞ. (1.21)
n¼0
x 105
10
10 km
50 km
8
Kernel value
0
0 1 2 3 4 5
Geocentric angle (ψ °)
Figure 1.1 The behaviour of the Poisson kernel with respect to different geocentric
angles j.
If we multiply both sides of Eq. (1.23) by Yn0 m0 ðq0 ; l0 Þ and take the
integral of the results over a unit sphere according to the addition theorem
(Eq. 1.9), we can write
ZZ
r r nþ1 1
vnm ¼ Vr ðr; q0 ; l0 ÞYnm ðq0 ; l0 Þds. (1.24)
nþ1 R 4p
s
Eq. (1.24) is an integral for solving vnm from Vr ðr; q; lÞ above the
spherical earth, i.e., r > R. The factor ðr=RÞnþ1 continues Vr ðr; q; lÞ
downward to sea level and r=ðn þ1Þ coverts the spherical harmonics of
Vr ðr; q; lÞ to those of the potential. Since r > R, the ratio ðr =RÞ is larger
than 1, and this means that by the power n þ 1, it will be even larger, and
when n goes to infinity so does this downward continuation factor.
However, r=ðn þ1Þ decreases by increasing the power and when it is
multiplied to ðr=RÞnþ1 , it will reduce the power of this factor. This means
that more frequencies, or vnm with higher degrees and orders, are recov-
erable from Vr ðr; q; lÞ than V ðr; q; lÞ; see Eq. (1.14).
*
X
N
2n þ 1 r nþ1
H ðr; jÞ ¼ Pn ðcos jÞ; (1.27)
n¼0
nþ1 R
with
X
N nþ1
2n þ 1 R
Hðr; jÞ ¼ Pn ðcos jÞ; (1.29)
n¼0
nþ1 r
which are, respectively, known as the extended Hotine integral and func-
tion. Unlike Eq. (1.27), the extended Hotine function is convergent with
the following closed-form formula (Pick et al., 1973):
2 t e
xþD
Hðr; jÞ ¼ ln . (1.30)
D 1e x
1400
10 km
1200
50 km
1000
Kernel value
800
600
400
200
0
0 2 4 6 8 10
Geocentric angle (ψ °)
Figure 1.2
where
nþ1
1X N
R
Pr ðr; jÞ ¼ ð2n þ 1Þðn þ 1Þ Pn ðcos jÞ. (1.33)
r n¼0 r
50
0
10 km
30 km
Kernel value
–50
–100
–150
–200
0 0.5 1 1.5 2 2.5 3
Geocentric angle (ψ °)
Figure 1.3 Behaviour of the first-order derivative of the Poisson kernel, Pr ðr; jÞ.
gravimetry the problem is to find V ðR; q0 ; l0 Þ from Vr ðr; q; lÞ, and Eq.
(1.32) is an integral equation having the unknown inside the integral for-
mula. The behaviour of the kernel of this integral is important to check
how much the potential can be recovered. Similar to those kernels we
discussed before, this kernel has its largest value at j ¼ 0 , but for the case
of 10 km the contribution of near-zone potentials is more significant than
that of those which are far. For the case of 30 km the largest value is closer
to zero than that of 10 km.
The superscript “comb” means combined, and a and b are two con-
stants. The term V comb ðr; q; lÞ is given at the whole surface of the spherical
boundary.
Spherical harmonics and potential theory 13
By integrating the product of both sides of Eq. (1.36) with Ynm ðq0 ; l0 Þ
and simplifying the results based on the orthogonality property of spherical
harmonics given in Eq. (1.6), we have:
ZZ nþ1
0 0 0 0 R nþ1
V comb
ðr; q ; l ÞYnm ðq ; l Þds ¼ 4p ab vnm . (1.37)
r r
s
where
X
N r nþ1 nþ1
Kðr; jÞ ¼ ð2n þ 1Þ ab Pn ðcos jÞ; (1.40)
n¼0
R R
with
XN nþ1 1
* R nþ1
S ðr; jÞ ¼ ð2n þ 1Þ ab Pn ðcos jÞ. (1.42)
n¼0
r R
where according to Eqs. (1.20) and (1.34), the kernel of this integral will be:
v t t3 t 1 3t 2 ðt t 3 Þðt e
xÞ
aþb Pðr; jÞ ¼ a 3 b 3
3 5
. (1.44)
vr D r D D
where the dot ($) means the inner product operator, d stands for the Kro-
2
necker delta and NnðiÞ the squared of the norm of vector spherical
X
harmonics:
2 1 i ¼ 1
NnðiÞ X ¼ : (1.52)
nðn þ 1Þ i ¼ 2 and 3
ðiÞ 0 0
The term vf ;nm is derived by multiplying XðiÞ
nm ðq ; l Þ to both sides of Eq.
(1.47) and taking the integral of the result over the sphere with radius r.
16 Satellite Gravimetry and the Solid Earth
If the orthogonality property (Eq. 1.51) is applied and the result is solved for
ðiÞ
vf ;nm , Eq. (1.53) will be obtained:
ZZ
ðiÞ 1
vf ;nm ¼ vf ðr; q0 ; l0 Þ$XðiÞ 0 0
nm ðq ; l Þds; i ¼ 1; 2 and 3: (1.53)
ðiÞ 2
4p Nn X s
This means that there are three solutions for the SHCs of the vector
field. For the gradient of a scalar field vf ¼ v, see Eq. (1.46) and compare it
with Eqs. (1.48)e(1.50); only the vector spherical harmonics Xð1Þ nm ðq; lÞ and
ð2Þ ð2Þ
Xnm ðq; lÞ are required. For other types of fields, Xnm ðq; lÞ gives the
poloidal components of the field and Xð3Þ nm ðq; lÞ the toroidal ones. This will
be discussed in Chapter 8.
ð1Þ
The solution of Eq. (1.54), which is related to recovery of vnm from
Vr ðr; q; lÞ, is the same as was discussed in Section 1.3.1.1, and the spectral,
spatial, and integral equations have been presented. Therefore, we continue
our discussion only with the horizontal derivatives. By inserting Eq. (1.46)
ð2Þ
into Eq. (1.55) and using the orthogonality property (Eq. 1.51), vnm is
derived:
R r nþ2 1 Z Z vY ðq0 ; l0 Þ
ð2Þ
Vx ðr; q0 ; l0 Þ
nm
vnm ¼
nðn þ 1Þ R 4p vq0
s
vYnm ðq0 ; l0 Þ
þ Vy ðr; q0 ; l0 Þ ds. (1.56)
sinq0 vl0
Spherical harmonics and potential theory 17
Substituting Eqs. (1.61) and (1.62) into Eq. (1.60) and simplification yields:
ZZ
r
V ðR; q; lÞ ¼ Vq ðr; q0 ; l0 ÞGq* ðr; jÞ þ Vl ðr; q0 ; l0 ÞGl* ðr; jÞ ds;
4p
s
(1.63)
where Gq ðr; jÞ and Gl ðr; jÞ are the kernels of the integral (Eq. 1.63):
XN
2n þ 1 r nþ1 vPn ðcos jÞ
Gq* ðr; jÞ ¼ ; (1.64)
n¼1
nðn þ 1Þ R vq
XN
2n þ 1 r nþ1 vPn ðcos jÞ
Gl* ðr; jÞ ¼ . (1.65)
n¼1
nðn þ 1Þ R sin qvl
From the relation between the associated Legendre functions and the
Legendre polynomials (Abramowitz and Stegun, 1964, p. 334):
dm Pn ðcos jÞ
Pnm ðcos jÞ ¼ ð1Þm sinm j m ; (1.66)
dðcos jÞ
and for m ¼ 1, we can write:
vPn ðcos jÞ dPn ðcos jÞ dcos j vj dPn ðcos jÞ
¼ ¼ sin j cos a0
vq dcos j dj vq0 dcos j
¼ Pn1 ðcos jÞcos a0 ; (1.67)
where a0 is called the backward azimuth, meaning the azimuth from the
integration point to the computation point (see Fig. 1.5). Similarly:
vPn ðcos jÞ dPn ðcos jÞ d cos j vj
¼
sin q0 dl0 d cos j dj sin q0 vl0
dPn ðcos jÞ
¼ sin j ðsin a0 Þ
d cos j
After repeating the whole process to get Eq. (1.63), for Eq. (1.68) we
reach:
ZZ
R
V ðr; q; lÞ ¼ ½Vx ðR; q0 ; l0 Þcos a0 Vy ðR; q0 ; l0 Þsin a0 GH ðr; jÞds;
4p
s
(1.72)
where
XN nþ1
2n þ 1 R
GH ðr; jÞ ¼ Pn1 ðcos jÞ; (1.73)
n¼1
nðn þ 1Þ r
Fig. 1.4 shows the behaviour of the kernel (Eq. 1.74) for the cases in
which r is larger than R by 10 and 50 km.
20 Satellite Gravimetry and the Solid Earth
400
350 10 km
50 km
300
250
Kernel value
200
150
100
50
0
0 5 10 15 20
Geocentric angle (ψ °)
Figure 1.4 Behaviour of GH ðr; jÞ.
For r ¼ R, this kernel will have the following spectral and closed-form
expression (see Hwang, 1998):
XN
2n þ 1 dPn ðcos jÞ j
GH ðjÞ ¼ ¼ cot 1: (1.75)
n¼1
nðn þ 1Þ dj 2
Eq. (1.80) is the mathematical relation between the forward and the
backward azimuths. If one of them is given, the other can be simply derived
from this equation.
x 105
0
100 km
150 km
–1
Kernel value
–2
–3
–4
–5
0 2 4 6 8 10
Geocentric angle (ψ °)
Figure 1.6 Behaviour of Pj ðr; jÞ.
to the first-order radial derivative, is the same as Eq. (1.32), but the other
two are:
! ZZ !
Vx ðr; q; lÞ 1 cos a
¼ V ðR; q0 ; l0 ÞPj ðr; jÞ ds; (1.81)
Vy ðr; q; lÞ 4pr sin a
s
where a is the forward azimuth, the azimuth from the computation to the
integration point. The kernel of the integral is derived by taking the deriv-
ative of Eq. (1.20) with respect to j:
pffiffiffiffiffiffiffiffiffiffiffiffi2ffi
3tðt t 3 Þ 1 e x
Pj ðr; jÞ ¼ 5
. (1.82)
D
Fig. 1.6 shows the behaviour of this kernel to j ¼ 10 for two levels of
100 and 150 km above the spherical boundary.
where the colon denotes the double-dot product of the tensors, s is the
unit sphere of integration, ds is the integration elements, d stands for Kro-
h i2
necker’s delta and NnðiÞ is the squared norm of base functions ZðiÞ nm ðq; lÞ:
Z
24 Satellite Gravimetry and the Solid Earth
8
>
>
>
> 1 i¼1
>
>
>
> 1
>
> nðn þ 1Þ i¼2
>
>
>2
>
>
>
2
< 2ðn 1Þnðn þ 1Þðn þ 2Þ i¼3
NnðiÞ ¼ : (1.87)
Z >
> 2n2 ðn þ 1Þ2 i¼4
>
>
>
>1
>
>
>
> nðn þ 1Þ i¼5
>2
>
>
>
>
> 2ðn 1Þnðn þ 1Þðn þ 2Þ i¼6
:
This means that there are six solutions for SHCs of the tensor field.
Spherical harmonics and potential theory 25
r nþ1 1 ZZ
ð2Þ r 2 0 vYnm ðq0 ; l0 Þ 0 0
0 vYnm ðq ; l Þ
vnm ¼ Vxz þ V ds;
nðn þ 1Þðn þ 2Þ R 4p vq0 yz
sin q0 vl0
s
(1.103)
where Vxz 0
¼ Vxz ðr 0 ; q0 ; l0 Þ and Vyz0
¼ Vyz ðr 0 ; q0 ; l0 Þ to keep the formula
shorter, but the prime over the derivatives means that they are at the inte-
gration point. Similarly, we can obtain another solution for the SHCs:
r2 r nþ1
ð3Þ
vnm ¼
ðn 1Þnðn þ 1Þðn þ 2Þ R
Z Z " v2
1 0 v 1 v2
Vxx Vyy
' '
2 cot q
4p vq0 vq0 sin2 q0 vl0 2 (1.104)
s
#
0 0
v 1 vY ðq ; l Þ
Ynm ðq0 ; l0 Þ 2Vxy0 2 0
nm
ds;
vq sin q0 vl0
where Vxx 0
¼ Vxx ðr; q0 ; l0 Þ; Vyy0 ¼ Vyy ðr; q0 ; l0 Þ and Vxy 0
¼ Vxy ðr; q0 ; l0 Þ.
Three independent solutions for the SHCs of the earth’s gravitational po-
tential can be derived. If we look at Eq. (1.103), we will observe that for
n ¼ 0 it is singular. This means that the zero-degree harmonic of the poten-
0 0
tial cannot be solved from the derivatives Vzx and Vzy . In a very similar
manner we can conclude that the zero- and first-degree harmonics cannot
0
be recovered from Vxx , Vyy0 and Vxy
0
.
Again, theoretically, these spectral solutions are equivalent, meaning
ð1Þ ð2Þ ð3Þ
that vnm ¼ vnm ¼ vnm . Therefore, we can write:
1 1
vzz;nm ¼ vVH;nm ¼ vVH;nm for n > 1; (1.105)
n nðn 1Þ
with
ZZ
1
vzz;nm ¼ Vzz ðr; q0 ; l0 ÞYnm ðq0 ; l0 Þds; (1.106)
4p
s
ZZ
1 0 vYnm ðq0 ; l0 Þ 0 0
0 vYnm ðq ; l Þ
vVH;nm ¼ Vxz þ V ds; (1.107)
4p vq0 yz
sin q0 vl0
s
Spherical harmonics and potential theory 27
Z Z " v2
1 v 1 v2
vHH;nm ¼ 0
Vxx cot q 0 2 0 0 2 Ynm ðq0 ; l0 Þ
Vyy0
02
0
4p vq vq sin q vl
s
#
0 v 1 vYnm ðq0 ; l0 Þ
2Vxy 2 0 ds.
vq sin q0 vl0
(1.108)
where
X
N
2n þ 1 r nþ1
Grr* ðr; jÞ ¼ Pn ðcos jÞ (1.110)
n¼0
ðn þ 1Þðn þ 2Þ R
Finally, by inserting Eq. (1.104) into Eq. (1.3) and performing similar
manipulations we obtain:
ZZ "X N r nþ1
r2 1
V ðR; q; lÞ ¼
4p n¼2
ðn 1Þnðn þ 1Þðn þ 2Þ R
s
n 2
X
v v 1 v2
02
cot q 0 2 0 0 2 Ynm ðq0 ; l0 ÞYnm ðq; lÞ
0
m¼n vq
vq sin q vl
!#
X n
v 1 v
0 0 0 0
2Vxy ðr; q ; l Þ 2 0 Ynm ðq ; l ÞYnm ðq; lÞ ds.
m¼n
vq sin q0 vl0
(1.114)
Martinec (2003) presented the following relations for the addition
theorem of the spherical harmonics (see also Eq. 1.66):
Xn 2
v 0 v 1 v2
02
cot q 0 2 0 0 2 Ynm ðq0 ; l0 ÞYnm ðq; lÞ
m¼n vq vq sin q vl
d2 Pn ðcos jÞ
¼ ð2n þ 1Þcos 2a0 sin2 j ; (1.115)
dðcos jÞ2
X n
v 1 vYnm ðq0 ; l0 Þ
2 0 Ynm ðq; lÞ ¼
m¼n
vq sin q0 vl0
(1.116)
d2 Pn ðcos jÞ
0
ð2n þ 1Þsin 2a sin j 2
.
dðcos jÞ2
3 t 2 ðD þ 1Þ x
3e
GVH ðr; jÞ ¼ þ þ 1
2D 2DðD þ 1 te xÞ 2t
1 Dþt 3 D þt e x pffiffiffiffiffiffiffiffiffiffiffiffi2ffi
ln 1e x; (1.126)
1e x DðD þ t e xÞ 2t 1ex
t 3 2 1 e
x2 t 3 e
xðe
x tÞ
GHH ðr; jÞ ¼ þ ext þ Dt þ ð1 DÞ þ þ
2 2 t D þ 1 tex tð1 e
xÞ
e
x2
.
tðD þ t e
xÞ
(1.127)
Fig. 1.7AeC, respectively, shows the behaviour of the kernels Grr ðr; jÞ,
GVH ðr; jÞ and GHH ðr; jÞ. As observed, none of these kernels approaches
the zero value even up to j ¼ 20 . This means that the contributions of
the far-zone data are very significant. Martinec (2003) showed their
(A) (B)
10 2
10 km 10 km
8
50 km 1.5 50 km
Kernel value
Kernel value
6
1
4
0.5
2
0 0
0 5 1015 20 0 5 10 15 20
Geocentric angle ( ψ °) Geocentric angle ( ψ °)
(C) 1
0.8 10 km
20 km
Kernel value
0.6
0.4
0.2
0
0 5 10 15 20
Geocentric angle ( ψ °)
Figure 1.7 Behaviour of the kernels (A) Grr ðr; jÞ, (B) GVH ðr; jÞ and (C) GHH ðr; jÞ.
Spherical harmonics and potential theory 31
significance even up to j ¼ 180 . Therefore, solving a potential outside
the reference sphere from the second-order derivatives of the potential at
the spherical boundary is successful if the values are given with global
coverages.
As seen, all kernel functions are functions of the azimuth between the
computation point and the integration points except Pzz ðr; jÞ. Fig. 1.8
32 Satellite Gravimetry and the Solid Earth
0.014
100 km
0.012
150 km
0.01
Kernel value
0.008
0.006
0.004
0.002
0
0 1 2 3 4 5
Geocentric angle (ψ °)
Figure 1.8 Behaviour of Pzz ðr; jÞ.
represents the behaviour of this kernel for the computation point, which is
located at 100 and 150 km above the reference sphere to j ¼ 5 . The
kernel has large values at the computation point, where j ¼ 0 and
decrease fast to zero.
However, the important issue in integration is the contribution of the
far-zone data, which can be seen on the plot of the isotropic parts of the
kernel, or those parts which are not azimuth dependent. The kernel
functions presented in Eq. (1.130) contain two isotropic parts:
5tð1 e
x2 Þ
Pjj ðr; jÞ þ cot jPj ðr; jÞ ¼ 3t 2 ðt 2 1Þ ; (1.135)
D7
5tð1 ex2 Þ 2e
x
Pjj ðr; jÞ cot jPj ðr; jÞ ¼ 3t ðt 1Þ 2
7
2
5 . (1.136)
D D
(A) x 108
(B) x 10
7
0.5 7
100 km
0 6
150 km
100 km 5
–0.5
Kernel value
Kernel value
150 km
4
–1
3
–1.5
2
–2
1
–2.5 0
0 1 2 3 4 5 0 1 2 3 4 5
Geocentric angle (ψ °) Geocentric angle (ψ °)
Figure 1.9 (A) Pjj ðr; jÞ þ cot jPj ðr; jÞ and (B) Pjj ðr; jÞ cot jPj ðr; jÞ.
Eq. (1.136) and, as we observed, the kernel has its largest minimum at the
computation point and increases up zero.
The kernel Eq. (1.132) is also azimuth dependent, but its isotropic part
has the following formula:
3 pffiffiffiffiffiffiffiffiffiffiffiffi2ffi
1 15t 9t 15ðt t 3 Þðt e
xÞ t 1e x
Pj ðr; jÞ Prj ðr; jÞ ¼ 5
þt 7
;
r D D r
(1.137)
with Pj ðr; jÞ presented in Eq. (1.82). Eq. (1.136) can also be derived from
Eqs. (1.137) and (1.82).
Fig. 1.10 represents the behaviour of this kernel for heights of 100 and
150 km above the surface of the sphere. The kernel is bell shaped and
approaches zero rather fast.
X
N X
k
ð2Þ
V ð2Þ ðR; q; lÞ ¼ vkl Ykl ðq; lÞ. (1.139)
k¼0 l¼k
34 Satellite Gravimetry and the Solid Earth
5
100 km
150 km
0
Kernel value
–5
–10
–15
0 1 2 3 4 5
Geocentric angle (ψ °)
Figure 1.10 1r Pj ðr; jÞ Prj ðr; jÞ
X
N X
n0
ð1;2Þ
¼ vn0 m0 Yn0 m0 ðq; lÞ.
n0 ¼0 m0 ¼n0
(1.140)
As seen, the product of two spherical harmonic expansions can be
ð1;2Þ
written as a new spherical harmonic series with vn0 m0 as the SHCs, which
can be derived from the SHCs of both series by the ClebscheGordan series:
ð1;2Þ
X
N X
i X
N X
k
ð1Þ ð2Þ 0 0
vn0 m0 ¼ vij vkl Qnijklm ; (1.141)
i¼0 j¼i k¼0 l¼k
0 0
where Qnijklm are the Gaunt coefficients, which appear in the product of two
associated Legendre’s functions (Xu, 1996):
X
iþk
0 0
Pij ðcos qÞPkl ðcos qÞ ¼ Qnijklm Pn0 m0 ðcos qÞ. (1.142)
n0 ¼jikj
Spherical harmonics and potential theory 35
The signal spectra are defined as the mean value of the square of each
Laplace coefficient all over the sphere, i.e.:
Z Z nþ1 n0 þ1
1 R 0 0 R
vn ðrÞ ¼ vn ðq ; l Þ vn0 ðq0 ; l0 Þds. (1.148)
4p r r
s
which is the signal spectra of the potential at a sphere with the radius r.
Spherical harmonics and potential theory 37
Inserting each of them into Eq. (1.148), and after performing simplifi-
cations, we obtain:
2 2nþ2 X
nþ1 R n
v r;n ðrÞ ¼ 2
vnm ; (1.154)
r r m¼n
2 2nþ2 X
ðn þ 1Þðn þ 2Þ R n
vrr;n ðrÞ ¼ 2
vnm . (1.155)
r2 r m¼n
If we compare these spectra with the one derived from the potential,
Eq. (1.151), we see that there exist some degree-dependent coefficients
behind the upward continuations factor ðR=rÞ2nþ2 for the spectra of
derivatives of the potential. This means that these coefficients, which are
increasing by degree, reduce the power of this factor and do not allow
the factor to smooth the signal. These degree-dependent coefficients of the
spectra of the second-order radial derivative are stronger and decrease the
power of the upward continuation factor more than the first derivative
does.
the Laplace coefficients should be derived and inserted into Eq. (1.148).
From Eqs. (1.47) and (1.84), it will not be difficult to show that these
coefficients have the following relations to the vector and tensor spherical
harmonics:
nþ1 nþ1 X n
R 1 R
vH;n ðq; lÞ ¼ vnm Xð2Þ
nm ðq; lÞ; (1.156)
r r r m¼n
nþ1 nþ1 X
R nþ2 R n
vVH;n ðq; lÞ ¼ 2 vnm Zð2Þ
nm ðq; lÞ. (1.157)
r r r m¼n
Let us start with Eq. (1.156) and insert into Eq. (1.148). The result will
be:
nþ1 n0 þ1 Z Z X n X n0
1 1 R R ð2Þ ð2Þ
v H;n ðrÞ ¼ v X
nm nm ðq; lÞ$ vn0 m0 Xn0 m0 ðq; lÞds
4p r 2 r r m¼n m0 ¼n0
s
nþ1 n0 þ1 X n X n0 ZZ
1 R R 1 ð2Þ
¼ 2 vnm vn m
0 0 Xð2Þ
nm ðq; lÞ: Xn0 m0 ðq; lÞds.
r r r m¼n m0 ¼n0
4p
s
(1.158)
Now, the orthogonality of the vector spherical harmonics (Eq. 1.51)
will be applied, and after simplification, the spectra will be:
2nþ2 X
nðn þ 1Þ R n
vH;n ðrÞ ¼ 2
vnm . (1.159)
r2 r m¼n
X
n X
n0 ZZ
1 0 0 ð3Þ 0 0
¼ vnm vn0 m0 Zð3Þ
nm ðq ; l Þ: Zn0 m0 ðq ; l Þds;
m¼n m0 ¼n0
4p
s
(1.162)
and finally by applying the orthogonality of the tensor spherical harmonics
(Eq. 1.86), the spectra will be:
2nþ2 X
ðn 1Þnðn þ 1Þðn þ 2Þ R n
v HH;n ðrÞ ¼ 2
vnm . (1.163)
r4 r m¼n
(1.164)
Another random document with
no related content on Scribd:
Elle était des Mac Fergus d’Inverardoch, près du lac Katherine, et
se vantait de connaître, depuis sa jeunesse, des choses profondes
que l’ordinaire des hommes ne connaît pas. Elle m’expliqua que la
lune est un dieu, le plus bienfaisant ou le plus malfaisant des dieux ;
selon qu’on accomplit les rites qu’il exige à son culte, ou qu’on les
méprise. Je me souviens que, ce soir-là, elle portait un « déshabillé »
de Maypoole, alors le couturier le plus à la mode à Londres, et le
plus cher ; des perles à son cou luisaient d’un éclat laiteux, comme
au fond de la mer. Pourtant elle était une petite sauvage, rien qu’une
petite sauvage, n’ignorant rien de ce que savent les sauvages, et
qu’ignorent les civilisés. Elle me dit sérieusement :
— N’oublie jamais !… Il ne faut, sous aucun prétexte, regarder le
nouveau croissant à travers une fenêtre ; ou bien on est en proie à
sa malédiction. Car les maisons des hommes, le croissant ne les
aime pas. Les maisons des hommes sont toujours vieilles ; le
croissant est toujours jeune, il n’aime que ce qui lui ressemble. Si,
par accident, cela t’arrivait, il faut saluer la lune neuf fois, très bas,
comme cela, et lui dire : « Pardon, Beauté ! Je ne l’ai pas fait exprès ;
et maintenant je me détourne… » Malgré tout, méfie-toi, elle peut
demeurer irritée. Si tu es prudent, ne serre de huit jours la main d’un
ennemi, ni même d’un ami, n’entreprends rien, et ne fais pas
l’amour : ta femme accoucherait d’un enfant mort, et ta maîtresse te
trahirait.
« Mais s’il t’arrive, au contraire, de distinguer le croissant dès la
première minute qu’il apparaît, quand tu n’es pas sous un toit, quand
tu es en pleine nature, en une place digne de lui, alors salue encore
neuf fois, aussi respectueusement, mais sans peur, et touche
l’argent que tu as dans ta poche ; tu seras riche tout le mois. Cela
est sûr. »
— Annette, lui dis-je, en l’attirant vers moi, bien qu’elle y fît
obstacle par crainte de l’astre, je n’en doute pas. Dis-moi pourtant
une chose… Comment se fait-il que toutes les femmes qui, comme
toi, possèdent ce secret, ne soient pas riches ?
Elle pencha vers moi ses yeux clairs, devenus dédaigneux de
mon ignorance ou de mon scepticisme :
— Vous ne comprenez pas, dit-elle, vous ne comprenez
réellement pas ?… C’est que, de nos jours, avec les modes
actuelles, les femmes qui pourraient avoir de l’argent n’ont pas de
poches, et celles qui ont encore des poches n’ont pas d’argent…