This function calculates a set of statistics for the seismic signal submitted.
signal_stats(data, stats, range_f, res_psd = 1, dt, cut = TRUE)data frame with calculated statsitics
eseis object, data set to be processed.
Character vector, keywords of statistics to be
calculated. If omitted, all statistics will be calculated. Wrongly
spelled keywords will be omitted without warning.
Numerical vector of length two, range of the
frequency spectra used to calculate spectral properties. This is
recommended to account for spurious or unwanted frequency pars, for
example caused by ocean micro seism or high frequency effects.
Numerical value, resolution of the spectrogram
used to calculate statistics, in seconds. Default is 1 sec. The
spectrogram will be calculated with 90
running window of 5 sec.
Numeric value, sampling period. If omitted, dt
is set to 1/200.
Logical value, option to cut output vector to the
required statistics, instead of returning the full length of statistics,
filled with NA values where no statistic was calculated. Default is
TRUE.
Michael Dietze
Available statistics keywords are:
- (1) `"t_duration"` (Duration of the signal)
- (2) `"f_rise"` (Signal rise time, time from start to maximum amplitude)
- (3) `"f_fall"` (Signal fall time, tme from maximum amplitude to end)
- (4) `"t_risefall"` (Ratio of rise to fall time)
- (5) `"a_skewness"` (Skewness of the signal amplitude, see seewave::specprop)
- (6) `"a_kurtosis"` (Kurtosis of the signal amplitude, see seewave::specprop)
- (7) `"a1_kurtosis"` (Kurtosis of the filtered (0.1-1 Hz) signal amplitude, see seewave::specprop)
- (8) `"a2_kurtosis"` (Kurtosis of the filtered (1-3 Hz) signal amplitude, see seewave::specprop)
- (9) `"a3_kurtosis"` (Kurtosis of the filtered (3-10 Hz) signal amplitude, see seewave::specprop)
- (10) `"a4_kurtosis"` (Kurtosis of the filtered (10-20 Hz) signal amplitude, see seewave::specprop)
- (11) `"a5_kurtosis"` (Kurtosis of the filtered (20-50 Hz) signal amplitude, see seewave::specprop)
- (12) `"e_maxmean"` (Ratio of maximum and mean envelope value, see Hibert et al. (2017))
- (13) `"e_maxmedian"` (Ratio of maximum and median envelope value, see Hibert et al. (2017))
- (14) `"e_skewness"` (Skewness of the signal envelope, see seewave::specprop)
- (15) `"e_kurtosis"` (Kurtosis of the signal envelope, see seewave::specprop)
- (16) `"e1_logsum"` (Logarithm of the filtered (0.1-1 Hz) envelope sum, see Hibert et al. (2017))
- (17) `"e2_logsum"` (Logarithm of the filtered (1-3 Hz) envelope sum, see Hibert et al. (2017))
- (18) `"e3_logsum"` (Logarithm of the filtered (3-10 Hz) envelope sum, see Hibert et al. (2017))
- (19) `"e4_logsum"` (Logarithm of the filtered (10-20 Hz) envelope sum, see Hibert et al. (2017))
- (20) `"e5_logsum"` (Logarithm of the filtered (20-50 Hz) envelope sum, see Hibert et al. (2017))
- (21) `"e_rmsdecphaseline"` (RMS of envelope from linear decrease, see Hibert et al. (2017))
- (22) `"c_peaks"` (Number of peaks (excursions above 75
- (23) `"c_energy1"` (Sum of the first third of the signal cross correlation function, see Hibert et al. (2017))
- (24) `"c_energy2"` (Sum of the last two thirds of the signal cross correlation function, see Hibert et al. (2017))
- (25) `"c_energy3"` (Ratio of c_energy1 and c_energy2, see Hibert et al. (2017))
- (26) `"s_peaks"` (Number of peaks (excursions above 75
- (27) `"s_peakpower"` (Mean power of spectral peaks, see Hibert et al. (2017))
- (28) `"s_mean"` (Mean spectral power, see Hibert et al. (2017))
- (29) `"s_median"` (Median spectral power, see Hibert et al. (2017))
- (30) `"s_max"` (Maximum spectral power, see Hibert et al. (2017))
- (31) `"s_var"` (Variance of the spectral power, see Hibert et al. (2017))
- (32) `"s_sd"` (Standard deviation of the spectral power, see seewave::specprop)
- (33) `"s_sem"` (Standard error of the mean of the spectral power, see seewave::specprop)
- (34) `"s_flatness"` (Spectral flatness, see seewave::specprop)
- (35) `"s_entropy"` (Spectral entropy, see seewave::specprop)
- (36) `"s_precision"` (Spectral precision, see seewave::specprop)
- (37) `"s1_energy"` (Energy of the filtered (0.1-1 Hz) spectrum, see Hibert et al. (2017))
- (38) `"s2_energy"` (Energy of the filtered (1-3 Hz) spectrum, see Hibert et al. (2017))
- (39) `"s3_energy"` (Energy of the filtered (3-10 Hz) spectrum, see Hibert et al. (2017))
- (40) `"s4_energy"` (Energy of the filtered (10-20 Hz) spectrum, see Hibert et al. (2017))
- (41) `"s5_energy"` (Energy of the filtered (20-30 Hz) spectrum, see Hibert et al. (2017))
- (42) `"s_gamma1"` (Gamma 1, spectral centroid, see Hibert et al. (2017))
- (43) `"s_gamma2"` (Gamma 2, spectral gyration radius, see Hibert et al. (2017))
- (44) `"s_gamma3"` (Gamma 3, spectral centroid width, see Hibert et al. (2017))
- (45) `"f_modal"` (Modal frequency, see seewave::specprop)
- (46) `"f_mean"` (Mean frequency (aka central frequency), see seewave::specprop)
- (47) `"f_median"` (Median frequency, see seewave::specprop)
- (48) `"f_q05"` (Quantile 0.05 of the spectrum, see seewave::specprop)
- (49) `"f_q25"` (Quantile 0.25 of the spectrum, see seewave::specprop)
- (50) `"f_q75"` (Quantile 0.75 of the spectrum, see seewave::specprop)
- (51) `"f_q95"` (Quantile 0.95 of the spectrum, see seewave::specprop)
- (52) `"f_iqr"` (Inter quartile range of the spectrum, see seewave::specprop)
- (53) `"f_centroid"` (Spectral centroid, see seewave::specprop)
- (54) `"p_kurtosismax"` (Kurtosis of the maximum spectral power over time, see Hibert et al. (2017))
- (55) `"p_kurtosismedian"` (Kurtosis of the median spectral power over time, see Hibert et al. (2017))
- (56) `"p_maxmean"` (Mean of the ratio of max to mean spectral power over time, see Hibert et al. (2017))
- (57) `"p_maxmedian"` (Mean of the ratio of max to median spectral power over time, see Hibert et al. (2017))
- (58) `"p_peaksmean"` (Number of peaks in normalised mean spectral power over time, see Hibert et al. (2017))
- (59) `"p_peaksmedian"` (Number of peaks in normalised median spectral power over time, see Hibert et al. (2017))
- (60) `"p_peaksmax"` (Number of peaks in normalised max spectral power over time, see Hibert et al. (2017))
- (61) `"p_peaksmaxmean"` (Ratio of number of peaks in normalised max and mean spectral power over time, see Hibert et al. (2017))
- (62) `"p_peaksmaxmedian"` (Ratio of number of peaks in normalised max and median spectral power over time, see Hibert et al. (2017))
- (63) `"p_peaksfcentral"` (Number of peaks in spectral power at central frequency over time, see Hibert et al. (2017))
- (64) `"p_diffmaxmean"` (Mean difference between max and mean power, see Hibert et al. (2017))
- (65) `"p_diffmaxmedian"` (Mean difference between max and median power, see Hibert et al. (2017))
- (66) `"p_diffquantile21"` (Mean difference between power quantiles 2 and 1, see Hibert et al. (2017))
- (67) `"p_diffquantile32"` (Mean difference between power quantiles 3 and 2, see Hibert et al. (2017))
- (68) `"p_diffquantile31"` (Mean difference between power quantiles 3 and 1, see Hibert et al. (2017))
References: - Hibert C, Provost F, Malet J-P, Maggi A, Stumpf A, Ferrazzini V. 2017. Automatic identification of rockfalls and volcano-tectonic earthquakes at the Piton de la Fournaise volcano using a Random Forest algorithm. Journal of Volcanology and Geothermal Research 340, 130-142.
## load example data
data(rockfall)
## clip data to event of interest
eq <- signal_clip(data = rockfall_eseis,
limits = as.POSIXct(c("2015-04-06 13:18:50",
"2015-04-06 13:20:10"),
tz = "UTC"))
## calculate full statistics
eq_stats <- signal_stats(data = eq)
## show names of statistics
names(eq_stats)
## calculate and show selected statistics, with truncated frequency range
eq_stats_sub <- signal_stats(data = eq,
stats = c("t_rise",
"c_peaks",
"f_centroid"),
range_f = c(1, 90))
print(eq_stats_sub)
Run the code above in your browser using DataLab