Geophys. J. Int. (1992) 111, 424-432 A formulation for reliable estimation of active crustal deformation and its application to central Greece Constantinos B. Papazachos and Anastasia A. Kiratzi Geophysical Laboratory, University of Thessaloniki, Mail Box 352-1, GR-540 06 Thessaloniki, Greece Accepted 1992 May 13. Received 1992 March 23; in original form 1991 December 30 SUMMARY A procedure for the analysis of data, concerning seismic moment release and fault-plane solutions, for the estimation of active crustal deformation is proposed. The formulation takes advantage of all the available historical and instrumental data for the calculation of the 'size' of the deformation in a seismic zone and of all the reliable fault-plane solutions which are available for a broader seismic belt for the determination of the 'shape' of the deformation. A detailed analysis of the errors involved in each parameter, based on a Monte Carlo numerical method, is suggested. The proposed procedure is applied to a region of extensional tectonics in central Greece. The resulting seismic strain rates show that the dominant mode of deformation in the area is extension at a rate of the order of 6 mm yr-' and a mean azimuth of N25"W. A vertical contraction rate of about 1 mm yr-' has also been calculated. The results from geodetic measurements performed by Billiris et al. (1991) declare that the calculated seismic deformation represents about 60 per cent of the total strain of this area. Key words: crustal deformation, fault-plane solution, Greece, seismic moment, strain. 1 INTRODUCTION The method of analysis that is usually applied for the estimation of active crustal deformation requires the knowledge of both the fault-plane solution (strike, dip, rake) and the seismic moment for each earthquake of a data sample complete over a certain magnitude threshold (Tselentis & Makropoulos 1986; Jackson & McKenzie 1988a,b; Ekstrom & England 1989). Bearing in mind that reliable fault-plane solutions exist only for recent large earthquakes, this method of analysis is applicable to an observational period of 30yr. It is possible to extend this period back in time, by assuming the fault parameters of the past events, and this has been done by some researchers (Jackson & McKenzie 1988a,b; Ambraseys & Jackson 1990; Papazachos et al. 1990; Kiratzi 1991; Taymaz, Jackson & McKenzie 1991). Anyhow, no matter how well justified these assumptions are, it is very likely that they introduce considerable error and bias. The fact that the known past earthquakes usually have a large magnitude makes the situation more difficult and leads to controversial results. One such example is the application of this method to the Aegean Sea and the surrounding area where values from a few millimetres to some centimetres per year have been determined for the maximum velocity rate by different researchers. To overcome the previously mentioned drawbacks, a method of analysis is suggested which permits: (a) the use of all available complete data, which include information on smaller recent shocks and on strong instrumental and historical earthquakes of a much longer period, for the seismic moment rate calculation in a seismic zone; and (b) the use of fault parameters deduced from reliable fault-plane solutions of recent strong earthquakes and from reliable field observations of past strong earthquakes which occurred in a broader seismic belt. The errors involved in the determination of active crustal deformation is also an important problem (Jackson & McKenzie 1988a,b; Ekstrom & England 1989). For this reason, a detailed analysis of these errors is also performed in the present study. This procedure of data analysis for the estimation of parameters related to active crustal deformation (strain rates, velocities, etc) and the corresponding procedure of 424