Документ взят из кэша поисковой машины. Адрес оригинального документа : http://geophys.geol.msu.ru/berd/IPSE719.pdf Дата изменения: Tue Oct 28 14:59:59 2003 Дата индексирования: Mon Oct 1 19:55:43 2012 Кодировка:

MAGNETOVARIATIONAL SOUNDING: NEW POSSIBILITIES CR 0
1 2 345 6

719 DP , m
1230 869 614

WV
78 9 10

WC
11 12

HC
13 14 15

­10

434 307 217 154 109 76.8 54.3

­20

­30 z, km

38.4 27.2 19.2 13.6 9.60 6.79 4.80

­40

­50

3.39 2.40 1.70 1.20 0.85 0.60 0.42

­60

­70 ­50 0 50 100 150 200 y, km

0.30

Fig. 18. The 2-D inversion of Re Wzy, Im Wzy, and || performed with the use of the smoothing REBOCC program: CR, Coast Range; WV, Willamette Valley; WC, Western Cascades; HC, High Cascades; DP, Deschutes Plateau.

The EMSLAB-I model is open for criticism. A cold continental mantle contradicts current geodynamic ideas of the Cascadian subduction zone (compare the EMSLAB-I model with the predictive CASCADIA model shown in Fig. 16). Analysis of the EMSLAB-I model has shown that the TM-mode is weakly sensitive to variations in the electrical conductivity of the mantle and that the bimodal inversion alone, using both the TE and TM modes, can be effective in studying the asthenosphere [Berdichevsky et al., 1992; Vanyan et al., 2002]. Experiments on the bimodal interpretation of MTS and MVS data obtained in the Cascadian subduction zone resulted in the construction of the 2-D EMSLAB-II model (Fig. 17b). It was constructed with the use of the automatic inversion program INV2D-FG, optimizing resistivities on 20 blocks of a fixed geometry [Varentsov et al., 1996]. An algorithm of parallel weighted inversion was applied to , Re Wzy, and Im Wzy (maximum weight), || and (normal weight), and || (minimum weight). The EMSLAB-II model has much in common with EMSLAB-I. Both have the same
IZVESTIYA, PHYSICS OF THE SOLID EARTH Vol. 39

oceanic asthenosphere, subducting slab, and crustal conducting layer. However, the EMSLAB-II plate joins the crustal conductor, and a conducting asthenosphere is present in the EMSLAB-II continental mantle. Thus, the geoelectric data have revealed partial melting in the continental mantle. The main drawback of the EMSLAB-II model is its sketchiness due to the limited possibilities of the INV2D-FG program. Presently, the INV2D-FG program has given way to more efficient software tools designed for the automatic 2-D inversion of MVS and MTS data. These are the smoothing program REBOCC [Siripunvaraporn and Egbert, 2000] and the programs IGF-MT2D [Nowozynski and Pushkarev, 2001] and II2DC [Varentsov, 2002]. These programs enable the optimization of models containing 512 and more blocks of a fixed geometry and provide new possibilities for interpreting the EMSLAB experimental data [Vanyan et al., 2002]. Three-dimensional model estimates obtained for the Pacific coast of North America and analysis of experimental field characteristics, induction arrows, and polar diagrams show that the regional structure along the LinNo. 9 2003


720 Initial START model

BERDICHEVSKY et al. Inversion of Re Wzy and Im Wzy

TP model

TE model

Inversion of ||

Starting TP model

Starting TE model

Inversion of and

TM model

EMSLAB-III model

Synthesis of the TP, TE, and TM models

Fig. 19. Algorithm of interpretation of EMSLAB experimental data in the class of block models.

coln line is favorable for the 2-D interpretation of MVD and MTS data. The interpretation consisted of three stages. At the first stage, the 1-D inversion of short-period magnetotelluric (MT) curves (T = 0.01­100 s) was performed and an approximate geoelectric section of the continental volcanic-sedimentary cover was constructed to a depth of 3.5 km. This section agrees with the near-surface portion of the EMSLAB-I model [Wannamaker et al., 1989b]. At the second stage, the REBOCC program was applied in experiments with 2-D smoothed inversion. With the complicated conditions of the Cascadian subduction zone, the parallel inversion of the TE- and TMmodes yielded whimsical alternation of low- and highresistivity spots with a poor minimization of the misfit. It is difficult to recognize real structures of the subduction zone in these spots. The most interesting result was obtained from the partial inversion of Re Wzy, Im Wzy , and || (Fig. 18). The western and eastern conducting zones identified here are separated by a T-shaped region of higher resistivity that can be associated with the subducting slab. An oceanic asthenosphere whose top can be fixed at a depth of about 30 km is recognizable in the western conducting zone. The eastern conducting zone coincides with the crust-mantle zone of wet melting in the predictive CASCADIA model shown in Fig. 16. It is noteworthy that the upper boundary of the eastern conductor closely resembles the topography of the

crustal conducting zone in the EMSLAB-I and EMSLAB-II models shown in Fig. 17. At the third, final stage, the method of partial inversions was applied and a new 2-D geoelectric model of the Cascadian subduction zone was constructed [Vanyan et al., 2002]; this model was called EMSLAB-III. It was constructed with the use of the II2DC program [Varentsov, 2002], minimizing the model misfit in the class of media with a fixed geometry of blocks. The algorithm of interpretation consisting of partial inversions is illustrated in Fig. 19. The interpretation was conducted in the regime of testing hypotheses. We consider three hypotheses on the structure of the Cascadian subduction zone: (1) CASCADIA predictive model, (2) EMSLAB-I model, and (3) EMSLAB-II model. The interpretation model is shown in Fig. 20a. The ocean floor topography and thicknesses of the seafloor, accretionary prism, and shelf sediments were specified from bathymetric and sedimentary thickness maps [Connard et al., 1984a, 1984b]. The resistivities of the water, sediments, and oceanic crust were set at 0.3, 2, and 10 000 m, respectively. The depth to the oceanic mantle and its resistivities were taken in accordance with the CASCADIA, EMSLAB-I, and EMSLAB-II models. The slab surface was constructed from seismic [Trehu et al., 1994] and seismic tomography [Weaver and Michaelson, 1985; Rasmussen and Humphries, 1988] data. The structure of the volcanic-sedimentary cover was determined from the results of a 1-D inversion of short-period MT curves. The crust and mantle of
Vol. 39 No. 9 2003

IZVESTIYA, PHYSICS OF THE SOLID EARTH


MAGNETOVARIATIONAL SOUNDING: NEW POSSIBILITIES (a) 0
0.3 2 35 100 35 100
4

721

CR
1 2 3 45 678
4 4 90 130 130 14 14

WV
9
40 40 40 9 30 7 10 000

WC HC
10 11 12 13 14 15

DP

11 30 60 20 140 6 18 8 10 000
1000 1000

3 3.5 10 37.5

0 0.6 1.5 4 10 25 35 45 70

2 10 000

2 20
1000

1000 1000 1000 1000 1000 1000 1000 1000 1000 1000

10 000

1000

1000

1000 1000

1000 1000 30 1000 1000
1000 1000

1000 1000 1000

1000

1000

1000

1000

110
40

20

1000 1000
1000

110
1000 1000

15 10 5

170
1000 1000

1000

280 z, km ­ 200 ­ 100 (b) 0
0.3 5.7 49 78 56 118
5

280 z, km 150
WC HC

0
CR
1 2 3

50

100
WV

200
DP

km

45

678

9

10 11 12 13 14 15
9 33 71 19 127 5.9 23 6.9

81 120 158 21 17 4.7 4.1 37 34 34 13 26 11 9990

3 3.5 10 37.5

0 0.6 1.5
8310

1.7 10 010

3.3 20 22

452 47
2060

183 1760 72 38 23 89 53 42

58

10 080

27

25 35 45 70

234 143 234

15 490 71

53

147 337
1860 1940

16 1360 19

20

71

20

69

110
35

58

5450 2080
1310

110
17 38

8.4 7.5 0.79

170
22 35

1330

280 z, km ­ 200 ­ 100 0 50

280 z, km 150 200 km

100

Fig. 20. Models obtained by partial inversions of the EMSLAB experimental data (resistivity values in units of m are shown within blocks): (a) interpretation model (resistivity values of the initial START model are shown); (b) TP-2 model (inversion of Re Wzy and Im Wzy); (c) TE model (inversion of ||); (d) TM model (inversion of and ). CR, Coast Range; WV, Willamette Valley; WC, Western Cascades; HC, High Cascades; DP, Deschutes Plateau.

the continent were divided into homogeneous blocks. The division density and block geometry were chosen with regard for the configuration of the eastern conducting zone, delineated with the help of the REBOCC program, and admitted a free choice of crust and mantle structures within the framework of the three hypotheses considered. A hypothesis best fitting the observed data is chosen automatically in the process of optimization of resistivities and minimization of misfits. The crust and mantle of the continent have a resistivity of 1000 m
IZVESTIYA, PHYSICS OF THE SOLID EARTH Vol. 39

in the starting START model constructed on the basis of the interpretation model. Below, we consider the successive partial inversions. Inversion of Re Wzy and Im Wzy. The START model was taken as the starting one. The TP model resulting from the inversion is shown in Fig. 20b. The tipper misfit (the rms deviation of model tippers from observed values) in this model is 5­10 times smaller than the tipper amplitude (the difference between the maximum and minimum tipper values), which is evidence of good
No. 9 2003