NOTES
where
is mantle mean density, g is mean gravity ac-
Um
celeration, D is flexural rigid, there is the relation between
D and lithospheric effective elastic thickness Te ,
ETe3
V is lithospheric Poisson’s ratio, E is
,
D
2
12(1ꢁV )
Young’s modulus. From (5) ü (7), the response function
is the function of crustal density, crustal thickness, mantle
density and lithospheric rigid or lithospheric effective
elastic thickness.
In order to inverse the bathymetry of China seas and
their surroundings from gravity anomalies, the compensa-
tion model of bathymetry and gravity anomalies in China
seas is needed to analyze and investigate. We compute the
correlation coefficients between gravity anomalies on sea
and bathymetry from (4). In computation, we separate
China seas into two areas, one is the continental shelf ar-
eas, the East China Sea and Yellow Sea (hereafter
ECS-YS) (23eü34eN, 120eü130eE), another one is
the South China Sea (hereafter SCS) (0eü 25eN, 105eü
125eE. The computational results are shown in fig. 1.
From this figure, it is easily found that the correlation co-
efficients between gravity anomalies and bathymetry
within 25ü300 km wavelength ranges are the highest.
The correlation coefficients between the basin sea, SCS
and the continental shelf seas, ECS-YS are somewhat dif-
ferent. Within the 25ü300 km wavelengths, the correla-
tion coefficients in two sea areas are almost the same.
Within 25ü50 km and 100ü300 km wavelengths, the
correlation coefficients in SCS are a little higher than ones
in ECS-YS. The response function between gravity
anomalies and ETOPO5 bathymetry is computed by (3),
and shown in fig. 2. If the density contrast of the crust to
the sea water is 1.64 g/cm3, and d = 4 km, 7 km, Tc=20 km,
30 km, Te=0 km, 30 km, respectively, we use (5)ü(7) to
compute the response function to different compensation
models. The computational results are also shown in Fig.
2(a) and (b), respectively. From fig. 2, we can find that the
response function between the gravity anomalies and
bathymetry is nonlinear, but the approximate linear corre-
lation exists within the medium-short wavelengths (25ü
160 km). In different wavelength ranges, the effects of
compensation model parameters (d, Tc, Te) on the response
function are different. In medium-short wavelengths less
than 120 km, the response function mainly depends on the
mean depth of sea water. The effect of crustal thickness
and lithospheric effective elastic thickness on response
function is small in this wavelength range. When we
choose the certain mean depth of sea water, the response
functions of three compensation models (5)ü(7) are al-
most the same. The effects of crustal thickness and litho-
spheric effective elastic thickness on response function are
greater than 120 km in the wavelengths, and the charac-
teristics of the response functions from (6) and (7) are
similar to ones computed by (3) from gravity anomalies
and ETOPO5 bathymetry.
Fig. 1.
Correlation coefficient between altimeter-derived gravity
anomalies and ETOPO5 bathymeter in the South China Sea, the East
China Sea and the Yellow Sea. Bold line denotes the correlation coeffi-
cient of the East China Sea and the Yellow Sea, thin line represents the
correlation coefficient of the South China Sea.
the response functions in China seas and their surround-
ings, the two aspects are needed to consider, while we
inverse the bathymetry from gravity anomalies in fre-
quency domain. On one hand, it had better that response
functions do not depend on more parameters, since we
have poor knowledge about the oceanic crustal thickness
and lithospheric effective elastic thickness; on the other
hand, it had better to use the frequency ranges in which
the correlation between gravity anomalies and bathymetry
is higher and response relation is linear. According to the
above consideration, when we inverse the batymetry from
gravity anomalies, we only compute the bathymetry in
medium-short wavelengths (20ü120 km), and choose the
uncompensation model (5). Thus the gravity anomalies
are divided into long and short wavelength components,
and then we use the Gaussian high-pass filter to remove
the components of gravity anomalies with the wavelength
greater than 120 km. We also use the low-pass filter to
remove the components of ETOPO5 bathymetry with the
wavelength less than 120 km, and remain the components
with long-medium wavelength and higher accuracy. After
the medium-short wavelength components of bathymetry
are obtained from the gravity inversion, we add the
components to the medium-long wavelength component
of ETOPO5 bathymetry. From (6), since the exp (2Sꢂkd)
term exists in response function, it is similar to the gravity
downward continuous, and causes high noise. We multiply
the low-pass filter factor with 20 km truncation wave-
length to the response function for compressing noise.
2
Results and discussion
We separate the China seas into two areas, SCS and
ECS-YS to compute bathymetry, respectively. In compu-
tation, mean ocean depths come form ETOPO5 mean
depths, the density contrast of crust to sea water is 1.64
g/cm3. Gravity anomalies used in this study come from
altimeter-derived 2Ąh2Ągravity anomalies[7], its accu-
racy compared with ship gravity data is 10h10ꢁ5 m/s2[8,9]
.
Based on the correlation analysis between gravity
anomalies and ETOPO5 bathymetry, the characteristics of
We also add some ship gravity data of China seas into the
computation.
1662
Chinese Science Bulletin Vol. 46 No. 19 October 2001