bvmed {GRTo}

Repeated medians regression for the determination of the Gutenberg-Richter b-value

Description

This function determines the Gutenberg-Richter b-value from a set of magnitude values, using the repeated medians regression method.

Usage

bvmed(file, lis, hd = FALSE, colid = 1, nrep = 200, tm = NULL,
findtm = TRUE, title = "bvmed")

Arguments

file

file to be loaded

lis

list to be loaded

hd

whether file contains header or not

colid

field number for the magnitude values in file

nrep

number of replicates for the bootstrap (calculation of the standard-error for the b-value)

tm

threshold magnitude value

findtm

logical. Whether an automatic procedure is engaged to determine the threshold magnitude value tm or not

title

character. The title of the plot. The name for the PNG file that includes the plot begins with title

Details

This function reads magnitude values in the field which number is indicated by colid in file. Magnitude values are read in the list lis if the file does not exist. This function produces a plot showing the FMD and the linear model line. The plot is stored into a file with name file_bvmed.png (png format file). It includes the mblm function from the version 0.11 (2007) of the mblm library by Lukasz Komsta (known e-mail address : [e-mail]).

Value

A list containing the following components:

quantm

the 5%, 50% and 95% quantiles of the bootstrap replicates for the threshold magnitude value

mmed

the median of the bootstrap replicates for the threshold magnitude value

quantb

the 5%, 50% and 95% quantiles of the bootstrap replicates for the b-value

valid

the number of valid replicates

brm

the b-value

bse

the bootstrap standard-error value for the b-value

bme

the bootstrap margin of errors value for the b-value

Note

Thanks to Paul Friberg for telling us about the bug in the graphical part of bvmed.R. Thanks to Scott Kostyshak for telling us about the extra bootstrap package dependency. GRTo version 1.2 fixes these issues. Thanks to Andrew Barbour (USGS) for detecting and fixing minor problems of bvmed.R of version 1.2. In version 1.3, an option coded by Andrew Barbour allows the processing of data from a list. Willy Aspinall (University of Bristol) highlights bugs in the graphical parts of bvmed: An inappropriate magnitude shift in the FMD plot is definitely fixed in version 1.3. Many thanks to Willy!

Author(s)

Daniel Amorese <[e-mail]>, Paul A. Rydelek <[e-mail]> and Jean-Robert Grasso <[e-mail]>

Maintainer: Daniel Amorese <[e-mail]>

References

D. Amorese, J.-R. Grasso and P. A. Rydelek, "On varying b-values with depth: results from computer-intensive tests for Southern California", Geophys. J. Int. (2010) 180, 347-360

Siegel, A.F., "Robust regression using repeated medians", Biometrika (1982) 69, 242-244.

Examples

bvmed(file=system.file("extdata","IDYLLdeep.data.txt",package="GRTo"), colid = 15, nrep = 150, 
tm = NULL, findtm = TRUE, title = "IDYLLWILD")