Docstoc

20320130406015-2-3-4

Document Sample
20320130406015-2-3-4 Powered By Docstoc
					International Journal of Civil Engineering and Technology (IJCIET), ISSN 0976 – 6308
    INTERNATIONAL JOURNAL OF CIVIL ENGINEERING AND
(Print), ISSN 0976 – 6316(Online) Volume 4, Issue 6, November – December (2013), © IAEME
                                TECHNOLOGY (IJCIET)

ISSN 0976 – 6308 (Print)
ISSN 0976 – 6316(Online)                                                        IJCIET
Volume 4, Issue 6, November – December, pp. 145-159
© IAEME: www.iaeme.com/ijciet.asp
Journal Impact Factor (2013): 5.3277 (Calculated by GISI)
www.jifactor.com
                                                                             ©IAEME


     ELLIPSOIDAL APPROXIMATION FOR TOPOGRAPHIC-ISOSTATIC
       MASSES EFFECTS ON AIRBORNE AND SATELLITE GRAVITY
                         GRADIOMETRY

                                         A.A. MAKHLOOF
               Civil Engineering Department, Faculty of Engineering-Minia University



ABSTRACT

         The topographic-isostatic masses represent an important source of gravity field information
especially in the high-frequency band of the gravity field spectrum, even if the detailed density func-
tion inside the topographic masses is unknown. If this information is used within a remove-restore
procedure, then the instability problems related to the downward continuation of gradiometer from
airplane or satellite altitude can be reduced. In this paper, integral formulae are derived for the de-
termination of gravitational effects of topographic-isostatic masses of the second order derivatives of
the gravitational potential for various topographic-isostatic models. The application of these formu-
lae is useful especially for airborne gradiometry and satellite gravity gradiometry. The computation
formulae are presented in ellipsoidal approximation by separating the three-dimensional integration
in an analytical integration in ellipsoidal element direction and integration over the unit area. There-
fore, in the numerical evaluation procedure the ellipsoidal volume elements can be considered as
being approximated by mass-lines, located in the centre of the discretization compartments (the mass
of this element is condensated mathematically along its ellipsoidal normal axis). The formulae are
applied to various scenarios of satellite gradiometry measurement campaigns. The gravitational ten-
sor in the ellipsoidal normal direction component at a satellite altitude of 230 km for ESA’s gravity
satellite mission GOCE (Gravity Field and Steady-State Ocean-Circulation Explorer) has been com-
puted. The numerical computations are based on digital elevation models with five arc-minute reso-
lutions for gravity gradiometry effects at satellite altitude.

Keywords: Topographic-Isostatic Models, Ellipsoidal Approximation, Mass-Lines, Satellite Gravity
Gradiometry, Downward Continuation, Regularization, GOCE.




                                                  145
International Journal of Civil Engineering and Technology (IJCIET), ISSN 0976 – 6308
(Print), ISSN 0976 – 6316(Online) Volume 4, Issue 6, November – December (2013), © IAEME

1.     INTRODUCTION

         The determination of the gravity field from observations at aircraft or satellite altitudes is an
improperly posed problem in the sense that small changes in the observations at flight level produce
large effects in the gravity field parameters on the Earth’s surface or geoid level. This holds espe-
cially for the high-frequency constituents of the observation spectrum. To prevent the results from
unrealistic oscillations in the parameters, regularization techniques are usually applied in very poorly
conditioned cases (e.g. Ilk 1993). Most of the regularization methods represent a filtering procedure
and the filtering property can be controlled by a regularization parameter (e.g. Ilk 1998). This is
critical in those cases where the signal shows similar spectral characteristics as the observation noise.
The topographic-isostatic masses represent a gravity field information especially in the high fre-
quency band of the gravity field spectrum which can be superposed with the measurement noise in
aircraft or satellite altitude. Therefore, it is helpful if those signal parts are reduced before the down-
ward continuation and restored afterwards. In this case, it can be assumed that the high-frequency
part in the observations is mainly caused by the observation noise, which can be filtered without
loosing gravity field information. This procedure is only a first step to process airborne or satellite
gravity field information in an integrated computation environment by involving all available Earth
system information as sketched in Ilk (2000).
         In global applications as envisaged here, the frequently applied planar approximations of the
topographic-isostatic models cannot be used anymore (see Novák el al. 2001). Therefore, the very
efficient fast Fourier transformation (FFT) techniques cannot be applied for the present computations
as demonstrated by Schwarz et al. (1990) in case of airborne gravimetry applications. Also, spherical
approximation can not give the to be applied, especially for global or large-scale regional applica-
tions. There are two principal possibilities for calculating the effects of the topographic-isostatic
masses on gravitational functionals in ellipsoidal approximation: the representation of the topog-
raphic masses by any ellipsoidal discretisation in form of ellipsoidal compartments (e.g. defined by
geographical coordinate lines) and a subsequent integration (Abd-Elmotaal 1995b, Smith et al. 2001;
Tenzer et al. 2003; Heck 2003) or the representation of the Newton’s integral by a spherical har-
monic expansion (e.g. Sünkel 1985; Rummel et al. 1988; Tsoulis 1999, 2001). Sjöberg (1998) im-
plemented the formulae for the exterior Airy-Heiskanen topographic/isostatic gravitational potential
and the corresponding gravity anomalies. Geoid determinations with density hypotheses from
isostatic models based on geological information have been studied by Kuhn (2003).
         The investigations performed thus far are limited to the determination of the second deriva-
tives of the gravitational potential of the topographic-isostatic masses, necessary for Satellite Gravity
Gradiometry (SGG) are not treated for the general case. Only the topographic-isostatic effects on the
vertical component of the gravitational tensor have been studied by Wild and Heck (2004a,b) and
Heck and Wild (2005). The effects of topographic-isostatic masses on satellite-to-satellite tracking
(SST) data and SGG functionals based on spherical harmonic series are investigated by Makhloof
and Ilk (2004). This procedure is very efficient but limited to an upper spherical harmonic degree of
about 2700 which corresponds to a 4 arc-minute resolution. Beyond this degree numerical computa-
tion problems concerning the stability of the recursive computation of Legendre’s polynomials occur
(see e.g. Holmes and Featherstone 2002).
         In this paper integral formulae in ellipsoidal approximation and based on mass-lines, ap-
proximating the ellipsoidal volume elements of the topographic-isostatic masses are presented which
can be evaluated numerically based on global digital elevation/bathymetry models over land and
oceans. The formulae are derived for the gravitational potential of topographic-isostatic masses itself,
as well as the first and second derivatives. Different topographic-isostatic models have been investi-
gated such as the Airy-Heiskanen, Pratt-Hayford the formulae for Helmert’s first and second conden-
sation method are derived as well. The computation formulae are applied to the gravitational ele-

                                                   146
International Journal of Civil Engineering and Technology (IJCIET), ISSN 0976 – 6308
(Print), ISSN 0976 – 6316(Online) Volume 4, Issue 6, November – December (2013), © IAEME

ments at satellite altitude envisaged for European Space Agency’s (ESA’s) gravity satellite mission
GOCE (Gravity Field and Steady-State Ocean-Circulation Explorer). The effect of the topographic
and isostatic masses

2.      THE EFFECT OF THE TOPOGRAPHIC AND COMPENSATING MASSES

        In the following, the geoid used as reference surface for the heights given by the DEMs is
approximated by a geocentric reference ellipsoid of major radius (a=6378 km). The geocentric
heights of the computation and the integration points are given from DEM heights, interpreted here
as orthometric heights. Therefore, the Cartesian coordinates of the points can be determined from the
ellipsoidal coordinates as follows (Fig. 1):

       x = [ N (ϕ ) + h (ϕ , λ ) ] cos ϕ cos λ
       y = [ N (ϕ ) + h (ϕ , λ ) ] cos ϕ sin λ                                                      (1)
       z =  N (ϕ )(1 − E 2 ) + h (ϕ , λ )  cos ϕ ,
                                          




                                                 hP         l1   l

                                                       P`

                                                                         Surface of the earth
                                                       hP`
       Reference ellipsoid                  hQ
                                                                     ξ




              Fig. 1: Geometry of the topographical masses in ellipsoidal approximation


        where h(ϕ , λ ) is the ellipsoidal height, refers the topographical surface to the surface of the
geocentric biaxial ellipsoid used in geodesy as a reference body for geometric and the ellipsoidal
prime vertical radius of curvature
                 a
N (ϕ ) =                                                                                           (2)
               2
         (1 − e sin 2 ϕ )1 2

where e is the first eccentricity of the reference ellipsoid.

                                                       147
International Journal of Civil Engineering and Technology (IJCIET), ISSN 0976 – 6308
(Print), ISSN 0976 – 6316(Online) Volume 4, Issue 6, November – December (2013), © IAEME

        The potential of the topographical masses can be computed from Newton integral in ellip-
soidal coordinates as follows (Novák and Grafarend 2005, Fig. 1):

                                                      2π   π2   hQ
                             ρ(ϕ,λ,ξ)                                ρ(ϕ, λ,ξ)
  V ( p) = G∫∫∫
       T
                                           dv = G ∫          ∫ ∫                         (N(ϕQ) +ξ)(M(ϕQ) +ξ)dξdσ ,    (3)
                     v
                                  l                        ϕ π ξ
                                                  λ=0 =− 2 =0                l

with dσ = cos ϕQ dϕQ d λQ , n is the geoid undulation and the ellipsoidal meridian of curvature is given
by:
                      a
M (ϕQ ) =                                                                                                              (4)
                         2
               (1− E sin2 ϕQ )3 2

and the distance between the computation and the integrated point is

                                  12
l = ξ 2 + 2uξ + l12 
                                                                                                                    (5)

with

u =xQh=0 −xPcosϕξ cosλ +yQh=0 −yPcosϕξ sinλ +zQh=0 −zPsinϕξ
                      ξ                    ξ                                                                    (6)

and
                                                                     2 12

     {
l1 = xξ =0 − xP  +  yξ =0 − yP  + zξ =0 − zP 
                  
                         2

                                   
                                              2

                                                                    }               .                                (7)

Eq. (3) can be written as the sum of spherical effect and ellipsoidal correction to the spherical ap-
proximation (in case of constant density):
V T ( p) = V Ts + V Te

Where

                                  hQ (ϕ,λ)
VTs = Gρcr ∫∫ a2k1 + 2ak2 + k3 
                                         dσ                            and                                          (8)
                                                  0
                 σ


                                                                         hQ (ϕ,λ )
  VTe = Gρcr E2 ∫∫ a2k1 + 2k2 (2sin2 ϕ −1)
                                                                                   dσ                               (9)
                                                                         0
                              σ
with

k1 = ln u + ξ + l ,
k 2 = l − uk1 ,                                                                                                       (10)
k 3 =  (ξ − 3u )l + (3u − l ) k1  .
      
           1
           2                      
                                       2
                                             1
                                              2




                                                                             148
International Journal of Civil Engineering and Technology (IJCIET), ISSN 0976 – 6308
(Print), ISSN 0976 – 6316(Online) Volume 4, Issue 6, November – December (2013), © IAEME

       The direct topographical effect on gravity (the first derivatives of the potential of the topog-
raphical masses with respect to the ellipsoidal surface normal) can be given by:

                   hQ (ϕ,λ)
∂VT                            ∂ (N(ϕQ ) +ξ )(M(ϕQ ) +ξ ) 
    = Gρcr ∫∫         ∫        ∂h                         dξdσ .                                   (11)
 ∂h        σ         ξ =n                    l            

        Also, the first derivative of potential can be separated into spherical and ellipsoidal correction
to the spherical term as follow:

∂V T ∂V T   ∂V T
    =     +                                                                                         (12)
 ∂h   ∂h s ∂h e

Where

∂V T                                      hQ (ϕ , λ )
      = G ρ cr ∫∫  a 2 P1 + 2 aP2 + P3 
                                                    dσ                                             (13)
 ∂h s          σ                          0



∂V T                                             hQ (ϕ , λ )
      = G ρcr ∫∫  a 2 P + 2 P2 (2sin 2 ϕ − 1) 
                  1                           0            dσ                                      (14)
 ∂h e         σ


With

       Ru − Sl12      R − Su
P=
1         2   2
                  +ξ 2           ,                                                                   (15)
       (l1 − u )l   (l1 − u 2 )l


       Rl12 − Sl12u    Sl 2 + Ru − 2 Su 2
P2 =                 −ξ 1 2               + Sk1 ,                                                   (16)
       (l12 − u 2 )l        (l1 − u 2 )l



   2Sl14 + Rl12u − 3Sl12u2    2Ru2 + 5Sl12u − 6Su3 − Rl12 2 S
P=
3                          +ξ                            + ξ + (R − 3Su)k1 ,                        (17)
         (l12 − u2 )l                 (l12 − u2 )l           l

and

R = xQh=0 − xP  cosϕP cos λP +  yQh=0 − yP  cosϕP sin λP + zQh=0 − zP  sinϕP ,
                                                                                             (18)


S =cosϕP cosλP cosϕQ cosλQ +cosϕP sinλP cosϕQ sinλQ +sinϕP sinϕQ .                                 (19)



                                                           149
International Journal of Civil Engineering and Technology (IJCIET), ISSN 0976 – 6308
(Print), ISSN 0976 – 6316(Online) Volume 4, Issue 6, November – December (2013), © IAEME

        Then, the effect of the topographic masses at airborne or satellite altitude (the second deriva-
tives of the potential of the topographic masses with respect to ellipsoidal height) can be given by:

                                   hQ (ϕ , λ )
∂ 2V T                                             ∂ 2 ( N (ϕ ) + ξ )( M (ϕ ) + λ ) 
       = G ρ cr ∫∫                         ∫       ∂h 2                             d ξ dσ                                             (20)
 ∂h 2           σ                          0                         l              

This equation can be transformed to the following equation:

                              hQ (ϕ ,λ )
∂ 2V T                                          ∂ 2 ( N (ϕ ) + ξ )( M (ϕ ) + λ ) 
       = G ρcr ∫∫                  ∫            2                                d ξ dσ
 ∂h 2          σ                   0            ∂h                l              
                               hQ (ϕ ,λ )
                                           ∂2 1
                                                                          hQ (ϕ ,λ )
                                                                                     ∂2 ξ
                                                                                                hQ (ϕ , λ )
                                                                                                            ∂2 ξ 2     
     = G ρ cr ∫∫  N (ϕ ) M (ϕ ) ∫                d ξ + ( N (ϕ ) + M (ϕ )) ∫                dξ + ∫                 d ξ dσ .
              σ                   0
                                           ∂h 2 l                             0
                                                                                     ∂h 2 l         0
                                                                                                            ∂h 2 l     
                                                                                                                       
                                                                                                                         (21)
Three integral of Eq. (21) can be estimated separately; the first term

                      hQ (ϕ ,λ )                                     hQ (ϕ , λ )
                                   ∂2 1                                     1      ( R 2 + 2 RSξ + S 2ξ 2 ) 
N (ϕ ) M (ϕ )            ∫0
                                   ∂h 2 l
                                          d ξ = N (ϕ ) M (ϕ )          ∫  l
                                                                       0   
                                                                             − 3 +3
                                                                                               l5
                                                                                                              dξ ,
                                                                                                             
                                                                     −(ξ + u )                            
                                                    = N (ϕ ) M (ϕ )  2       2
                                                                                + 3R 2 A + 6 RSB + 3S 2C                                (22)
                                                                     (l1 − u )                            
                                                    := N (ϕ ) M (ϕ ) w1
Where

     hQ (ϕ ,λ )
                  1         (ξ + u )     2(ξ + u ) 
A=       ∫
         0
                  l 5
                      dξ =  2       2 3
                                         + 2     2 
                            3(l1 − u )l 3(l1 − u l 
                                                                                                                                         (23)


     hQ (ϕ ,λ )
                  ξ        −1    
B=      ∫0
                  5
                  l
                    d ξ =  3 + uA
                           3l    
                                                                                                                                         (24)


     hQ (ϕ ,λ )
                  ξ2        −ξ u    l2 
C=       ∫
         0
                  l5
                     d ξ =  3 − B + 1 A .
                            2l 2    2 
                                                                                                                                         (25)


The second integral is given by:

                               hQ (ϕ , λ )                                          hQ (ϕ , λ )
                                                ∂2 ξ                                      1       ( R 2ξ + 2 RS ξ 2 + S 2ξ 3 ) 
[ N (ϕ ) + M (ϕ )] ∫                                   d ξ = [ N (ϕ ) + M (ϕ ) ]     ∫  l − 3 +3                                dξ ,
                                       0        ∂h 2 l                               0                        l5               
                                                                                  (l + uξ )
                                                                                      2
                                                                                                                           
                                                           = [ N (ϕ ) + M (ϕ ) ]  1  2    2
                                                                                              + 3 R 2 B + 6 RSC + 3S 2 D  ,
                                                                                  (l1 − u )l                              
                                                           := [ N (ϕ ) + M (ϕ ) ] w2

                                                                                                                                           (26)
                                                                                   150
International Journal of Civil Engineering and Technology (IJCIET), ISSN 0976 – 6308
(Print), ISSN 0976 – 6316(Online) Volume 4, Issue 6, November – December (2013), © IAEME

where

         hQ (ϕ , λ )
                       ξ3        ξ 3              
D=            ∫
              o
                       l 5
                           d ξ =  3 + uC + 2l12 B  .
                                 l                
                                                                                                    (27)


The third integral of Eq. ( 21 ) is given by:

hQ (ϕ , λ )                         hQ ( ϕ , λ )
              ∂2 ξ 2                                ξ2     ( R 2ξ 2 + 2 RS ξ 3 + S 2ξ 4 ) 
    ∫
    0
              ∂h 2 l
                     dξ =               ∫n
                                                    − 3 +3
                                                    l                    l5
                                                                                            dξ
                                                                                           
                                  ξ (l12 − 2u 2 )                                  
                               = 2          2
                                                   − k1 + 3 R 2 C + 6 RSD + 3 S 2 F  ,             (28)
                                  (l1 − u )l                                       
                              := w3

where

        hQ (ϕ ,λ )
                       ξ4       ξ3               ξ (l 2 − 2u2 ) 
F=            ∫
              0
                       l5
                          dξ =  3 − uD + 2l12 B − 12 2 + k1  .
                                −3l                (l1 − u )    
                                                                                                    (29)


Using the binomial expansion

                  1
              −              1    3
(1 − x)           2
                      = 1+     x + x 2 + ..........                                                 (30)
                             2    8

that can be successfully be truncated for x p 1 , both radii of curvature can be written in the follow-
ing form:

            1               
N (ϕ ) = a  1 + E 2 sin 2 ϕ                                                                     (31)
            2               

            3                    
M (ϕ ) = a 1 + E 2 sin 2 ϕ − E 2                                                                (32)
              2                  

Then, Eq. (21) can be written in the following formula:


∂ 2V T  2      (1 − E 2 )             2 − E 2 (1 + sin 2 ϕ )                  
    2
      =  a ∫∫     2     2
                           w1dσ + a ∫∫         2    2     32
                                                              w2 dσ + ∫∫ w3 dσ  .                (33)
 ∂h      σ (1 − E sin ϕ )          σ (1 − E sin ϕ )                  σ        


        The effect of the topographic masses can be written also in two terms: one of the spherical
effect and the other is the correction to spherical term.
        Eq. (33) are used to compute the effect of the topographic masses for the case of Bouguer
model, Airy-Heiskanen model. For calculating the effect of condensation masses in case of Helmert
                                                                             151
International Journal of Civil Engineering and Technology (IJCIET), ISSN 0976 – 6308
(Print), ISSN 0976 – 6316(Online) Volume 4, Issue 6, November – December (2013), © IAEME

second method of condensation and generalized Helmert method of condensation (Heck 2003), the
effect of the condensation masses can be given as follow:

                   k
V CM ( p) = G ∫∫      ( N − D1 )( M − D1 )dσ                                                       (34)
              σ    l2

        where k is he surface layer density and given by (Novak and Grafarend (2005), D1 is the
Helmert condensation depth and it equals zero in Helmert second method of condensation and 32 km
in case of generalized Helmert of condensation , l2 here denotes the distance between the computa-
tion point and the integrated point at the condensation surface. In case of constant mass density of the
topographical masses, the surface mass density is given by (Novák and Grafarend 2005)

                    hQ NQ + M Q hQ3
                                       1                 
kQ (ϕ , λ ) = ρ hQ 1 +          +                                                               (35)
                       2 NQ * MQ 3 NQ * M Q             
                                                        

        Eq. (34) can be transformed also in two terms: one for the spherical approximation and the
other is the ellipsoidal correction to the spherical approximation. The first and second derivative of
the condensated topography can be determined as follows:

∂V CM ( p)         ∂ 1
           = G ∫∫ k   ( N − D1 )( M − D1 )dσ
   ∂h          σ   ∂h  l2 
                                                                                                  (36)
∂ 2V CM ( p)         ∂2  1 
             = G ∫∫ k 2  ( N − D1 )( M − D1 )dσ
    ∂h2          σ   ∂h  l2 

        In case of the Airy-Heiskanen model the topographic masses of constant density ρ float on a
mantle of constant but larger density ρ m . An elevation column of height h is compensated by a cor-
responding root of thickness t. The higher the topographic features are, the deeper they sink. Thus
the thickness of a root column under a mountain column with height h can be determined by the for-
mula
   hQ                                          −T

∫∫ ξ∫ ρ ( N +ξ )( M +ξ ) dξdσ = ∫∫ ξ ∫
σ   =0                          σ          =−T −tQ
                                                     ∆ρ ( N + ξ )( M + ξ ) dξ dσ .                 (37)


       After Performing the numerical integration for finding the root for height larger than 10 km
and comparing the result with results from spherical formula and the error was lesser than 0.01%.
Then the spherical formula for computing the roots is applied in the present investigation. The root is
given by (Khun 2000)

                         ( R + h)3 − R 3  ρ 
t = ( R − T ) 1 − 3 1 −                    ,                                                  (38)
                             ( R − T ) 3 ∆ρ 
                                              
and the anti-roots of thickness t ′ in case of oceanic water columns of height h′ by,



                                                             152
International Journal of Civil Engineering and Technology (IJCIET), ISSN 0976 – 6308
(Print), ISSN 0976 – 6316(Online) Volume 4, Issue 6, November – December (2013), © IAEME

                      R 3 − ( R − h′)3  ( ρ − ρ ) 
 t′ = (R − T )  3 1+                           w
                                                    − 1 ,                                       (39)
                          ( R − T )3          ∆ρ      
                                                      

with the water density ρ w and the mean radius of the Earth R (R=6371 km). The density difference
 ∆ρ = ρ m − ρ in the roots and anti-roots produces a buoyancy so that hydrostatic equilibrium is
reached.
       The above formulae can be simplified if the heights of the Earth’s surface are expressed by
the rock-equivalent topographic heights heq (Rummel et al. 1988). It can be expressed in planar ap-
proximation as follows:

            land: h eq = h    
                              
heq   =               ρ − ρw  ,                                                                 (40)
        ocean: h eq =       h′
                        ρ     

        The potential of the isostatic masses at the computation point P can be determined analogous-
ly as the potential of the topographic masses by Newton’s integral,

                                         2π   π2    −T
              ∆ρ(ϕ, λ,ξ)                          ∆ρ(ϕ, λ,ξ)
V ( p) = G∫∫∫
  I
                        dv = G ∫ ∫ ∫                         (N(ϕQ ) +ξ)(M(ϕQ ) +ξ)dξdσ .        (41)
           v
                  l           λ=0 ϕ=−π 2 ξ =−T −t     l

This integration can be written as the sum

      V I ( p ) = V Is + V Ie                                                                   (42)

where

                                         −T
V Is = G ρcr ∫∫  a 2 k1 + 2ak2 + k3 
                                     −T −tQ dσ                                                (43)
                σ


                                                   −T
V Ie = G ρcr E 2 ∫∫  a 2 k1 + 2k2 (2sin 2 ϕ − 1) 
                                                  −T −tQ dσ                                    (44)
                    σ


        The combined effects of the topographic-isostatic masses on the different gravity functionals
are the differences between the effect of the topographic masses and the effect of the isostatic com-
pensation masses. It reads e.g.,


VTI h = V T                −V C                                                                 (45)
       P            h=hP          h=hP

       In case of the Pratt-Hayford model a certain adjustment surface is defined, in case of a ellip-
soidal approximation an ellipse in a constant depths D2 ( D2 = 100 km ). At this ellipsoidal surface
hydrostatic equilibrium is anticipated, i.e. the pressure of any topography column is identical at this

                                                            153
International Journal of Civil Engineering and Technology (IJCIET), ISSN 0976 – 6308
(Print), ISSN 0976 – 6316(Online) Volume 4, Issue 6, November – December (2013), © IAEME

ellipse which requires constant mass but different density depending on the elevation of the surface
of the Earth. The effect of the topographic masses according to the Pratt-Hayford model is calculated
for two cases: one for the land areas and the other for the ocean areas. The same has to be done for
the effects of the isostatic compensation masses.

Then, the effect of the topographic masses is given by

V P −T ( p ) = VTland + VTOcean                                                                              (46)

where
                                     2π   π2        hQ
                          ρL                             ρL (ϕ, λ,ξ)
VT
  land
         ( p) = G∫∫∫           dv = G ∫    ∫ ∫                         (N(ϕQ ) +ξ )(M(ϕQ ) +ξ )dξdσ ,        (47)
                  v
                          l          λ=0 ϕ=−π 2 ξ =0           l


                                               2π        π 2   0
                          (ρ − ρw )                         (ρ − ρw )
VTOcean ( p) = G∫∫∫                dv = G ∫ ∫ ∫                       ( N (ϕQ ) + ξ )(M (ϕQ ) + ξ )dξ dσ     (48)
                      v
                              l          λ =0 ϕ =−π 2 ξ =h′     l

The effect of the compensation masses for land areas is given by


                                               2π        π2    hQ
                     (ρ − ρL )                      (ρ − ρL )
Vland
 C       ( p) = G∫∫∫          dv = G ∫ ∫ ∫                    (N(ϕQ ) +ξ)(M(ϕQ ) +ξ)dξdσ ,                  (49)
                  v
                        l           λ=0 ϕ=−π 2 ξ =0    l

and for ocean areas by:

                                               2π        π2    −hQ
                      (ρ − ρO)                      (ρ − ρO)
VOcean
 C        ( p) = G∫∫∫         dv = G ∫ ∫ ∫                  (N(ϕQ) +ξ)(M(ϕQ) +ξ)dξdσ                        (50)
                   v
                         l          λ=0 ϕ=−π 2 ξ=−D    l

where density ρ L in land areas and density under ocean ρO can be determined from Kuhn (2000).

            R 3 − ( R − D )3       
ρL =          3           3
                              ρ,                                                                           (51)
      ( R + h) − ( R − D ) 


      R 3 − ( R − D )3  ρ − ρ w  R 3 − ( R − h′)3 
ρO =                                              ,                                                    (52)
                   ( R − h′) − ( R − D ) 
                   
                           3             3
                                          




                                                                       154
International Journal of Civil Engineering and Technology (IJCIET), ISSN 0976 – 6308
(Print), ISSN 0976 – 6316(Online) Volume 4, Issue 6, November – December (2013), © IAEME

3.     TEST REGIONS

        The numerical tests are based on the ETOPO5 with five arc-minute resolution. One typical
test regions have been selected: This test area is covering the Himalaya region (Fig. 2), shall be used
to visualize the damping of the topographic-isostatic effects in the component of the gravity gradi-
ometry in the ellipsoidal normal direction at 230km above sea surface. In this case resolution for the
elevation compartments has been used. The accuracy of this DEM is sufficient to demonstrate the
topographic-isostatic effects at satellite altitude.




                            Fig. 2: Topography of the test area (ETOPO5)


       The densities of crust and topography are considered to be constant and equal to 2670 kg/m3.
As density of sea water a value of 1030 kg/m3 has been taken and as density of the mantle the fre-
quently used value of 3270 kg/cm3 has been assumed. The Airy-Heiskanen depth of compensations
is considered to be 30 km, The depth of the condensation surface of Helmert’s first condensation
method is 21 km and 32 km for the generalized Helmert’s condensation model.


4.     NUMERICAL ANALYSIS

        To give an impression of the size of the topographic-isostatic effects on the gravity
gradiometry at satellite altitude, the topographic-isostatic effects on the of the gravitational tensor at
a satellite altitude of 230 km are computed. It can be found that, the structure of the topographic-
isostatic effects at a satellite altitude of 230 km shows still steep changes with a gradient of 0.80
Eötvös per 100km in north-south direction and of 0.60 Eötvös per 100 km in east-west direction. If it
is possible to remove this sort of roughness from the observations with a noise level of approximate-
ly 5 mE for GOCE the downward continuation can be considerably simplified. The GOCE mission is
designed to derive the static part of the gravity field with an extremely high precision. Therefore, it is
very important to filter the observations by the topographic-isostatic gravity field effects to ease the
requirements for the downward continuation. An additional regularization might be avoided in that
case; but this depends on the envisaged resolution of the gravity field model and on the error level of
the observations. Fig. 3 gives an impression of the size of these effects for the hh-component of the
gravitational tensor.



                                                   155
International Journal of Civil Engineering and Technology (IJCIET), ISSN 0976 – 6308
(Print), ISSN 0976 – 6316(Online) Volume 4, Issue 6, November – December (2013), © IAEME

5.     CONCLUSIONS

         In this paper the formulae for the calculation of the first and second derivatives of the gravita-
tional potential of the topographic-isostatic masses are derived for various frequently applied topo-
graphic-isostatic models in ellipsoidal approximation by approximating the ellipsoidal volume ele-
ments by mass-lines located in the middle of the compartments. Only the formula for the determina-
tion of the first derivatives in the ellipsoidal approximation has been derived by one author. The situ-
ation is different for the second derivatives of the potential; the second derivatives of the potential in
the ellipsoidal normal direction have not been studied till now.
         The formulae can be used to determine the topographic-isostatic effects at aircraft altitudes
for airborne applications or for satellite altitudes to reduce the observation functionals of airborne or
satellite gravity gradiometry missions. The integral formulae presented here allow to use DEMs with
an - in principle - arbitrary high resolution depending on the numerical integration method. Obvious-
ly such a high resolution is necessary for airborne altitudes. In these cases it might not be to apply
formulae which are based on the expansion of Newton’s integral in spherical harmonics at least for
global applications (see e.g., Makhloof 2007). As the higher resolution of the DEMs is indispensable
and the computation of the corresponding spherical harmonic coefficients can be critical because of
numerical problems (Holmes and Featherstone 2002). Therefore only a maximum degree of 2700
corresponding to a compartment size of 4 arc-minutes resolution can be selected. Due to this limita-
tion a spherical harmonic expansion of the topographic-isostatic masses cannot be used for exact
determinations of the geoid (see Kuhn and Seitz 2005).




       a) Bouguer method (topography)                          b) Generalized Helmert (D=32km)




      b) Helmert first method (D=21 km)                           b) Airy-Heiskanen method


                                                   156
International Journal of Civil Engineering and Technology (IJCIET), ISSN 0976 – 6308
(Print), ISSN 0976 – 6316(Online) Volume 4, Issue 6, November – December (2013), © IAEME




                                        f) Helmert first method

 Fig. 1: Effects of the topographic-isostatic masses on the tensor component Vzz at an altitude of 230
                         km for different topographic-isostatic models (Eötvös).

         Now the question arises which model should preferably be used for the filtering of satellite-
borne observations such as gravity gradients as preparation for the subsequent downward continua-
tion procedure? While Helmert’s second condensation method might be useful for geoid computa-
tions, its usefulness for the processing of satellite in-situ observations cannot be answered in such a
simple generally valid way. Indeed, the task of filtering topographic-isostatic effects in the satellite
observations is helpful only if these quantities have a significant magnitude larger than the observa-
tion noise. Obviously, this fact may depend on the validity of the topographic-isostatic hypothesis in
specific geographic regions. This can be decided only after a careful analysis of the specific gravity
field features within the various geographical regions of the Earth to find out which model describes
the reality in these regions in the best possible way. It is well-known that the Earth is isostatically
compensated by an amount of approximately 90% (Heiskanen and Moritz 1967), but it is difficult to
decide which model fits best. Although seismic measurement results indicate the validity of an Airy-
Heiskanen type of topographic-isostatic compensation, but in some parts of the Earth the isostatic
compensation seems to follow anther model (Heiskanen and Moritz 1967). The change of the con-
densation level by using a sort of a generalized Helmert model (Heck 2003) could be used to fit the
topographic-isostatic model to the reality. Very promising seems to be to introduce geophysical
models in coincidence with modern models of plate tectonics.
         If the specific topographic-isostatic model holds more or less uniformly for a larger region
then this model can be used to filter the satellite observations before the application of the regional
gravity field recovery procedure. The situation is more complicated in case of regionally varying
deviations of the reality from a specific model; further investigations are necessary to consider these
cases. Because of the varying effects of the topographic-isostatic models depending on the type of
observables such as gravity vectors or tensor components the frequently expressed argument that a
high resolution gravity field model might be sufficient to reduce observations at aircraft or satellite
altitude is not valid. Therefore, additional investigations are necessary to demonstrate the benefit of a
remove-restore procedure taking into account individually selected topographic-isostatic models for
the processing of airborne measurements and SGG observables. Finally, the results computed here
are computed with the results computed using spherical approximation and it is found that the ellip-
soidal approximation gives exact results.



                                                  157
International Journal of Civil Engineering and Technology (IJCIET), ISSN 0976 – 6308
(Print), ISSN 0976 – 6316(Online) Volume 4, Issue 6, November – December (2013), © IAEME

REFERENCES

 1.    Abd-Elmotaal H (1995b), "Attraction of the topographic masses", Bull. Geod (69): 304-307
 2.    Heck B (2003), "On Helmert’s methods of condensation". J Geod 77 (3-4): 155-170
 3.    Heck B and Wild F (2005), Topographic-isostatic reductions in satellite gravity gradiometry
       based on a generalized condensation model. In: Sansó, F. (ed): A Window on the Future of
       Geodesy, Springer Verlag, IAG Symposia 128: 294-299
 4.    Heiskanen WA, Moritz H (1967) Physical Geodesy. WH Freeman, San Francisco
 5.    Holmes SA and Featherstone WE (2002), "A unified approach to the Clenshaw summation
       and the recursive computation of very high degree and order normalized associated Legendre
       functions", J Geod 76(5): 279-299
 6.    Ilk KH (1993), "Regularization for high resolution gravity field recovery by future satellite
       techniques", in: G Anger el al. (eds.): Inv. Probl. Principles and Applications in Geophysics,
       Technology and Medicine, Mathematical Research, Vol. 74, Akademie Verlag GmbH, Berlin
 7.    Ilk KH (1998), "A proposal for the determination of optimal regularization parameters in Tik-
       honov-type regularization methods", Proc. of the International Seminar on Model Optimiza-
       tion in Exploration Geopyhsics 2, Friedr. Vieweg und Söhne, Braunschweig-Wiesbaden
 8.    Ilk KH (2000), "Envisaging a new area of gravity field research, in: Towards an Integrated
       Global Geodetic Observing System (IGGOS)", Rummel R, Drewes H, Bosch W, Hornik H
       (eds.): IAG Symposium 120, pp. 53-62, Springer-Berlin Heidelberg New York
 9.    Kuhn M (2000) Geoidbestimmung unter Verwendung verschiedener Dichtehypothesen.
       Deutsche Geodätische Kommission, Rh. C,Heft Nr 250, Munich
 10.   Kuhn M (2003), "Geoid determination with density hypotheses from isostatic models and
       geological information", J Geod 77 (1-2): 50-65
 11.   Kuhn M, Seitz K (2005), "Comparison of Newton’s integral in the space and frequency do-
       main, In: Sansó", F. (ed): A Window on the Future of Geodesy, Springer Verlag, IAG Sympo-
       sia 128: 386-391
 12.   Makhloof A. A, Ilk KH (2004), "The use of topographic-isostatic gravity field information in
       satellite-to-satellite tracking and satellite gravity gradiometry", Poster presented at the IAG In-
       ternational Symposium Gravity, Geoid and Space Missions, GGSM2004, Porto
 13.   Makhloof, A. A. (2007)," The use of Topographic- isostatic mass information in geodetic ap-
       plications", Ph. D. dissertation, Institute of geodesy and geoinformation, Bonn University,
       Bonn, Germany
 14.   NOAA (1988), "Data Announcement 88-MGG-02, Digital relief of the surface of the Earth,
       National Geophysical Data Centre", Boulder
 15.   Novák P, Vaniček P, Martinec Z, Véronneau M (2001), "Effect of the spherical terrain on
       gravity and the geoid", J Geod 75 (9-10): 491-504
 16.   Novák P, Grafarend EW (2005), "The ellipsoidal representation of the topographical potential
       and its vertical gradient", J Geod 78 (11-12): 691-706
 17.   Rummel R, Rapp R H, Sünkel H, Tscherning CC (1988), "Comparisons of global topograph-
       ic-isostatic models to the Earth’s observed gravity field". Rep 388, Department of Geodetic
       Science and Surveying, The Ohio State University, Columbus
 18.   Schwarz KP, Sideris MG, Forsberg R (1990), "The use of FFT technique in physical geode-
       sy", Geophysical Journal International (100): 485-514
 19.   Smith DA, Robertson DS, Milbert DG (2001), "Gravitational attraction of local crustal masses
       in spherical coordinates". J Geod 74 (11-12): 783-795
 20.   Sjöberg LE (1998), "The exterior Airy/Heiskanen topographic-isostatic gravity potential,
       anomaly and the effect of analytical continuation in Stokes’ formula", J Geod (72): 654-662


                                                   158
International Journal of Civil Engineering and Technology (IJCIET), ISSN 0976 – 6308
(Print), ISSN 0976 – 6316(Online) Volume 4, Issue 6, November – December (2013), © IAEME

 21. Sünkel H (1985), "An isostatic Earth model", Report. 367, Department of Geodetic Science
     and Surveying, The Ohio State University, Colubus
 22. Tenzer T, Vaníček P, Novák P (2003), "Far-zone contributions to topographical effects in the
     Stokes-Helmert method of the geoid determination", Studia Geophysica et geodaetica 47 (3):
     467-480
 23. Tsoulis D (1999), "Spherical harmonic computations with topographic/isostatic coefficients",
     Reports in the series IAPG / FESG (ISSN 1437-8280), Rep. No. 3 (ISBN 3-934205-02-X), In-
     stitute of Astronomical and Physical Geodesy, Technical University of Munich
 24. Tsoulis D (2001), "A Comparison between the Airy-Heiskanen and the Pratt-Hayford isostatic
     models for the computation of potential harmonic coefficients", J Geod 74 (9): 637-643
 25. Wild F, Heck B (2004a), "Comparison of different isostatic models applied to satellite gravity
     gradiometry". Jekeli, C. Bastos, L. Fernandes, J. (eds.): Gravity, Geoid and Space Missions,
     GGSM 2004, pp 230-235, Springer-Berlin Heidelberg New York
 26. Wild F, Heck B (2004b), "Effects of topographic and isostatic masses in satellite gravity
     gradiometry". Proc. 2nd Int. GOCE User Workshop, Frascati Mach 2004
 27. A.N.Satyanarayana, Dr Y.Venkatarami Reddy and B.C.S.Rao, “Remote Sensing Satellite Data
     Demodulation and Bit Synchronization”, International Journal of Advanced Research in
     Engineering & Technology (IJARET), Volume 4, Issue 3, 2013, pp. 1 - 12, ISSN Print:
     0976-6480, ISSN Online: 0976-6499.




                                               159

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:1
posted:12/31/2013
language:Latin
pages:15