Reply II to “VAN: Candidacy and validation with the latest laws of the game,” by F. Mulargia and P. Gasperini and “Precursor candidacy and validation: The VAN Case so far,” by F. Mulargia and P. Gasperini

Abstract

A simple inspection of the Tables of Mulargia and Gasperini [1992, 1996a] that refer to the period 1987–1989 reveals the following two facts: (i) by recalling that predictions are issued only when the expected magnitude is Mpred ≥ 5.0, the true VAN success rate is 79% (for Δr ≤ 120km) while the (incorrect) success rate of 45% emerges when restricting only correlations that include earthquakes with MEQ≥5.0 are considered as successes, and (ii) the VAN alarm rate increases significantly for larger earthquake magnitudes, i.e., for MEQ≥5.0, 5.3, 5.5 and 5.8, the alarm rates are 22%, 36%, 50% and 60% respectively. These facts are inconsistent with their claim that “VAN results can be ascribed to chance.” This claim also contradicts the conclusions obtained from the calculations by Aceves et al. [1996], and by Honkura and Tanaka [1996].In a separate Section we give answers, point by point, to various statements made in the two papers by Mulargia and Gasperini [1996a,b]. Most of these statements are obviously untrue. For example, they claim that “VAN predictions have been documented primarily by circulating telegrams among the members of the VAN‐group” although: (i) the Greek Authorities (as well as 29 International Institutes) receive the VAN predictions well before the earthquake occurrence and (ii) it is extensively known that the two most destructive activities in Greece during the period 1987–1994 were publicly announced by VAN well in advance. Furthermore, Mulargia and Gasperini [1996a] claim that, for the period 1990–1992, VAN correctly predicted only 2 out of 16 EQs with Ms(=mb+0.3) ≥ 5.0; however, a check of the VAN data of this period (see Varotsos et al., 1993c) shows that the alarm rate is 8/20, for EQs with mb > 4.7, and increases to 3/4 for EQs with mb > 5.0.