Premium
Application of Hurst Resecaling to Geophysical Serial Data
Author(s) -
Outcalt Samuel I.,
Hinkel Kenneth M.,
Meyer Erika,
Brazel Anthony J.
Publication year - 1997
Publication title -
geographical analysis
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.773
H-Index - 65
eISSN - 1538-4632
pISSN - 0016-7363
DOI - 10.1111/j.1538-4632.1997.tb00947.x
Subject(s) - hurst exponent , autocorrelation , detrended fluctuation analysis , series (stratigraphy) , cluster analysis , statistics , sign (mathematics) , standard deviation , precipitation , sunspot , statistical physics , mathematics , meteorology , geology , geography , physics , paleontology , mathematical analysis , geometry , quantum mechanics , scaling , magnetic field
The empirical investigation of several geophysical time series indicates that they are composed of segments representing different natural regimes, or periods when events are strongly autocorrelated. Using a data transformation method developed by Hurst, these regimes are differentiated by rescaling the time series and examining the resulting transformed trace for inflections. As regime signals are not completely mixed and have rather long run lengths, Hurst rescaling produces a clustering of extremes of the same sign and elevates the Hurst exponent to values greater than 0.5. These regimes have a characteristic distribution, as defined by the mean and standard deviation, which differ from the statistical characteristics of the complete record. Analysis of sunspot numbers, monthly precipitation records from the midwestern United States, and annual watershed discharge yield regimes that correlate to well‐documented periods of extreme activity. Rescaling appears to be useful in screening time series of unknown characteristics to differentiate periods dominated by regime‐specific physical processes.