Empirical Covariance Functions between Seismic, Density and Gravity Data — an Important Constraint in 3D Gravimetric-Seismic Stochastic Inversion

  • G. Strykowski
Part of the Theory and Practice of Applied Geophysics book series (THPAG, volume 5)


In gravity field approximation it is possible to describe the gravity field by a stationary stochastic process. This approach is used in gravity field modeling by Least Squares Collocation (LSC). The purpose of this paper is to include seismic and density information within the framework of the LSC approach. This is done by estimating (and later modeling) covariances between the involved quantities.

In this paper, 41 density and sonic logs were used together with gravity information, to obtain the empirical covariance functions between the involved residual quantities. The area of investigation lies in the North Sea, the horizontal dimensions are 3°x6° and the log data span from 200 m to 5000 m in depth. Regional reference models for density, reciprocal seismic velocity and gravity are modeled and subtracted from the data. A reference model relating the density and the reciprocal seismic velocity is also computed. The correlation coefficient between the density and the reciprocal seismic velocity is -0.66.

The results of this investigation show the characteristic correlation lengths, the variances and the correlation coefficients for the involved residual quantities. The peaks in the variation of the characteristic parameters with depth are related to the geology in the area. The data show a strong correlation between the residual densities both horizontally and vertically. The correlation between the residual reciprocal seismic velocities is not as strong as for the residual density distribution, which indicates that this parameter varies much more than the density distribution. The correlation coefficients between the residual density and the residual reciprocal seismic velocity are large, both positive and negative. There is no correlation between the residual gravity and the residual density and between the residual gravity and the residual reciprocal seismic velocity. This may be due to the constructed reference model for gravity. The residual gravity may still contain strong signal generated from depths which are not covered by the log data.

The covariance functions presented in this paper are probably characteristic only for the area of investigation. They contain however ‘large scale’ information about the distribution of the involved physical parameters within the sedimentary bassin and thus perhaps about the bassin evolution (e. g. compaction, tectonics).


Covariance Function Reference Model Gravity Field Gravity Data Seismic Velocity 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. Hein, G., Sansö, F., Strykowski, G. and Tscherning, C.C.: On the Choice of Norm and Base Functions for the Solution of the Inverse Gravimetric Problem. Presented at 17. Conference on Mathematical Geophysics, Blanes, Cataluna, Spain, June, 1988Google Scholar
  2. Forsberg, R.: Spectral Properties of the Gravity Field in the Nordic Countries. Boll, di Geod. et Sc. Aff., No. 4, 1986, pp. 361–383Google Scholar
  3. Kaula, W.M.: Statistical and Harmonic Analysis of Gravity. Techn. Rep. Army Map Service, Wash.,D.C., 1959Google Scholar
  4. Knudsen, P.: Estimation and Modelling of the Local Empirical Covariance Function using Gravity and Satellite Altimeter Data. Bull. Geod. 61 (1987) pp. 145–160CrossRefGoogle Scholar
  5. Krarup, T.: A Contribution to the Mathematical Foundation of Physical Geodesy, Danish Geodetic Institute, Publication No. 44 Copenhagen, 1969Google Scholar
  6. Lauritzen, S.: The Probabilistic Background of some Statistical Methods in Physical Geodesy. Danish Geodetic Institute, Publication No. 48, Copenhagen, 1973Google Scholar
  7. Moritz, H.: Advanced Physical Geodesy. H.Wichman Verlag. Karlsruhe, 1980Google Scholar
  8. Sanso, F., Barzaghi, R. and Tscherning, C.C.: Choice of Norm for the Density Distribution of the Earth, Geophys. J. R. astr. Soc., 123–141, 1986Google Scholar
  9. Sanso, F.: Statistical Methods in Physical Geodesy. Mathematical and Numerical Techniques in Physical Geodesy. Lecture Notes in Earth Sciences, Vol. 7, pp. 49–155, editor Siinkel H, Springer Verlag, 1986Google Scholar
  10. Strykowski, G.: A Study of Statistical Properties of the Earth’s Density Distribution based on Density Logs from the North Sea. Presented at II Hotine-Marussi Symposium on Mathematical Geodesy, Pisa, June 5–8, 1989.Google Scholar
  11. Tscherning, C.C.: Local Approximation of the Gravity Field by Least Squares Collocation, Proceedings from Beijing International Summer School, Beijing, China, 1984, Edited by K.P. Schwarz, Division of Surveying Engineering, The University of Calgary, Canada.Google Scholar
  12. Tscherning, C.C. and Strykowski, G.: Quasi-harmonic inversion of gravity field data. Proceedings of V Seminar on Model Optimization in Exploration Geophysics, Berlin, 1987Google Scholar
  13. Tscherning, C.C.: Density-Gravity Covariance Functions Produced by Overlapping Rectangular Blocks of Constant Density, Presented at II Hotine-Marussi Symposium on Mathematical Geodesy, Pisa, June 5–8, 1989. Submitted to Geoph. Journ. Intern.Google Scholar
  14. Wei, M.: On the Ergodicity of the Covariance Function in the Case of a Plane, Bull. Geod. 59 (1985) pp. 332–341CrossRefGoogle Scholar

Copyright information

© Springer Fachmedien Wiesbaden 1992

Authors and Affiliations

  • G. Strykowski
    • 1
  1. 1.Kort- og. Matrikelstyrelsen (National Survey and Cadastre)CharlottenlundDenmark

Personalised recommendations