Eos, Transactions American Geophysical Union

VAN method of short-term earthquake prediction shows promise


  • Seiya Uyeda

    1. Institute for Physical and Chemical Research (RIKEN) International Program on Earthquake Research, Earthquake Prediction Research Center, Tokai University, Shimizu, 424-8610, Japan
    Search for more papers by this author


Although optimism prevailed in the 1970s, the present consensus on earthquake prediction appears to be quite pessimistic. However, short-term prediction based on geoelectric potential monitoring has stood the test of time in Greece for more than a decade [VarotsosandKulhanek, 1993] Lighthill, 1996]. The method used is called the VAN method.

The geoelectric potential changes constantly due to causes such as magnetotelluric effects, lightning, rainfall, leakage from manmade sources, and electrochemical instabilities of electrodes. All of this noise must be eliminated before preseismic signals are identified, if they exist at all. The VAN group apparently accomplished this task for the first time. They installed multiple short (100–200m) dipoles with different lengths in both north-south and east-west directions and long (1–10 km) dipoles in appropriate orientations at their stations (one of their mega-stations, Ioannina, for example, now has 137 dipoles in operation) and found that practically all of the noise could be eliminated by applying a set of criteria to the data.