Modelling Soil Radon Concentration for Earthquake Prediction

We use regression/model trees to build predictive models for radon concentration in soil gas on the basis of environmental data, i.e., barometric pressure, soil temperature, air temperature and rainfall. We build model trees (one per station) for three stations in the Krsko basin, Slovenia. The trees predict radon concentration with a (cross-validated) correlation of 0.8, provided radon is influenced only by environmental parameters (and not seismic activity). In periods with seismic activity, however, this correlation is much lower. The increase in prediction error appears a week before earthquakes with local magnitude 0.8 to 3.3.

[1]  Saso Dzeroski,et al.  Discovering dynamics: From inductive logic programming to machine discovery , 1993, Journal of Intelligent Information Systems.

[2]  B. Zmazek,et al.  RADON MONITORING FOR EARTHQUAKE PREDICTION IN SLOVENIA , 2000 .

[3]  Ta-Liang Teng,et al.  Some recent studies on groundwater radon content as an earthquake precursor , 1980 .

[4]  Ian H. Witten,et al.  Data mining: practical machine learning tools and techniques with Java implementations , 2002, SGMD.

[5]  Ian H. Witten,et al.  Data mining: practical machine learning tools and techniques, 3rd Edition , 1999 .

[6]  S. Tasaka,et al.  Ground-Water Radon Anomaly Before the Kobe Earthquake in Japan , 1995, Science.

[7]  Peter Fajfar,et al.  Seismic hazard reassessment of an existing NPP in Slovenia , 1997 .

[8]  Pier Francesco Biagi,et al.  Difficulties with interpreting changes in groundwater gas content as earthquake precursors in Kamchatka, Russia , 2001 .

[9]  Yuji Nakamura,et al.  Short-term and intermediate-term geochemical precursors , 1988 .

[10]  M. Bidovec,et al.  Soil radon monitoring in the Krsko Basin, Slovenia. , 2002, Applied radiation and isotopes : including data, instrumentation and methods for use in agriculture, industry and medicine.

[11]  Andrej Gosar,et al.  Seismic reflection surveys of the Krško basin structure: implications for earthquake hazard at the Krško nuclear power plant, southeast Slovenia , 1998 .

[12]  I. P. Dobrovolsky,et al.  Estimation of the size of earthquake preparation zones , 1979 .

[13]  H. S. Virk,et al.  Investigation of radon-222 in soil-gas as an earthquake precursor , 1994 .

[14]  Vivek Walia,et al.  Helium/radon precursory anomalies of Chamoli earthquake, Garhwal Himalaya, India , 2001 .

[15]  David W. Aha,et al.  Instance-Based Learning Algorithms , 1991, Machine Learning.

[16]  Jan M. Zytkow,et al.  Handbook of Data Mining and Knowledge Discovery , 2002 .

[17]  M. Kumar,et al.  Radon in ground water related to seismic events , 1999 .

[18]  Sago Deroski,et al.  Discovering Dynamics: From Inductive Logic Programming To Machine Discovery , 2002 .

[19]  S. Setayeshi,et al.  Layered neural networks based analysis of radon concentration and environmental parameters in earthquake prediction. , 2002, Journal of environmental radioactivity.

[20]  James H. Dieterich,et al.  Two models for earthquake forerunners , 1975 .

[21]  Maria Macchiato,et al.  Robust Statistical Methods to Discriminate Extreme Events in Geoelectrical Precursory Signals: Implications with Earthquake Prediction , 2000 .

[22]  F. Italiano,et al.  Geochemical monitoring of thermal waters in Slovenia: relationships to seismic activity. , 2002, Applied radiation and isotopes : including data, instrumentation and methods for use in agriculture, industry and medicine.

[23]  C H Scholz,et al.  Earthquake prediction: a physical basis. , 1973, Science.

[24]  Y. Yasuoka,et al.  Anomaly in atmospheric radon concentration: a possible precursor of the 1995 Kobe, Japan, earthquake. , 1997, Health physics.

[25]  Leo Breiman,et al.  Classification and Regression Trees , 1984 .