Configuration Parameter Reference

All MSNoise configuration parameters, auto-generated from msnoise/config/config_*.csv at documentation build time. Grouped by workflow category in pipeline order.

Global

Configuration set category: global

Parameter	Default	Type	Description
response_path	inventory	str	Instrument correction file(s) location (path relative to db.ini), defaults to ‘./inventory’, i.e. a subfolder in the current project folder. All files in that folder will be parsed.
startdate	1970-01-01	str	Start Date to process: [1970-01-01]=’since beginning of the archive’
enddate	2100-01-01	str	End Date to process: [2100-01-01]=’No end’
crondays	1	str	Number of days to monitor with scan_archive, typically used in cron (should be a float representing a number of days, or a string designating weeks, days, and/or hours using the format ‘Xw Xd Xh’)
plugins		str	Comma separated list of plugin names. Plugins names should be importable Python modules.
hpc	N	bool	Is MSNoise going to run on an HPC? [Y/N]
output_folder	OUTPUT	str	Folder in which all the results will be output (base of the folder structure)
fdsn_retries	3	int	Number of retry attempts for transient FDSN/EIDA network errors (timeout etc)
fdsn_min_coverage	0.5	float	Minimum fraction of the requested time window that must be returned by FDSN/EIDA for the job to succeed (0.0-1.0)
fdsn_keep_raw	N	bool	Y: cache raw fetched waveforms to _output/raw/<date>/ for re-use by other preprocess configsets without re-fetching [Y/N]

Preprocessing

Configuration set category: preprocess

Parameter	Default	Type	Description
cc_sampling_rate	20.0	float	Sampling Rate for the CrossCorrelation (in Hz)
resampling_method	Decimate	str	Resampling method [Lanczos/Decimate]
preprocess_lowpass	8.0	float	Preprocessing Low-pass value (in Hz)
preprocess_highpass	0.01	float	Preprocessing High-pass value (in Hz)
preprocess_max_gap	10.0	float	Preprocessing maximum gap length that will be filled by interpolation (in seconds)
preprocess_taper_length	20.0	float	Duration of the taper applied at the beginning and end of trace during the preprocessing, to allow highpass filtering (in seconds)
remove_response	N	bool	Remove instrument response [Y/N]
response_prefilt	(0.005, 0.006, 30.0, 35.0)	eval	Remove instrument correction pre-filter (in Hz, see <a href=’https://docs.obspy.org/packages/autogen/obspy.core.trace.Trace.remove_response.html’ target=”_blank”>ObsPy Doc</a> for what it means.)
preprocess_components	E,N,Z	str	Components to preprocess: Z only is enough for ZZ Cross-Correlation (CC) jobs, while E,N,Z is needed if e.g. SC jobs are desired.

Cross-Correlation

Configuration set category: cc

Parameter	Default	Type	Description
cc_sampling_rate	20.0	float	Sampling Rate for the CrossCorrelation (in Hz)
cc_normalisation	NO	str	Normalisation method for individual corr_duration) CCFs, default is not normalized (NO), but can be normalized based on the power of the two traces (POW), or by the maximum (MAX) or absolute maximum (ABSMAX) of the CCF [NO/POW/MAX/ABSMAX]
maxlag	120.0	float	Maximum lag (in seconds)
corr_duration	1800.0	float	Data windows to correlate (in seconds)
overlap	0.0	float	Amount of overlap between data windows [0:1]
winsorizing	3.0	float	Winsorizing at N time RMS , 0 disables winsorizing, -1 enables 1-bit normalization
whitening	A	str	Whiten Traces before cross-correlation: [A]ll (except for autocorr), [N]one, or only if [C]omponents are different [A/N/C]
whitening_type	B	str	Type of spectral whitening function to use: [B]rutal (amplitude to 1.0), divide spectrum by its [PSD] or by band-passing the white spectrum with a hanning-window [HANN]. WARNING: only works for compute_cc, not compute_cc_rot, where it will always be B [B/PSD/HANN]
clip_after_whiten	N	bool	Do the clipping (winsorizing) after whitening? [Y/N]
stack_method	linear	str	Stack Method: Linear Mean or Phase Weighted Stack [linear/pws]
pws_timegate	10.0	float	If stack_method =’pws’, width of the smoothing (in seconds)
pws_power	2.0	float	If stack_method =”pws”, Power of the Weighting
components_to_compute	ZZ	str	List (comma separated) of components to compute between two different stations
cc_type	CC	str	Cross-Correlation type: CC for classical FFT-based correlation ; PCC for Phase Crosse Correlation (FastPCC’s PCC2, Ventosa et al) [CC/PCC]
components_to_compute_single_station		str	List (comma separated) of components within a single station. ZZ would be the autocorrelation of Z component, while ZE or ZN are the cross-components. Defaults to [], no single-station computations are done.
cc_type_single_station_AC	CC	str	Auto-Correlation type (same as for cc_type) [CC/PCC]
cc_type_single_station_SC	CC	str	Cross-Correlation type for Cross-Components (same as for cc_type: CC or PCC) [CC/PCC]
keep_all	Y	bool	Keep all cross-corr (length: corr_duration) [Y/N]
keep_days	Y	bool	Keep all daily cross-corr [Y/N]
cc_taper_fraction	0.04	float	Fraction of each trace to taper (cosine) before cross-correlation. Applied symmetrically to both ends. Standard seismological convention is 0.04 (4%). Range: 0.0 (no taper) to 0.5 (full Hann window).

Filters

Configuration set category: filter

Parameter	Default	Type	Description
freqmin	0.1	float	Lower frequency bound (Hz)
freqmax	1.0	float	Upper frequency bound (Hz)
CC	Y	bool	Compute cross-correlation between different stations [Y/N]
SC	N	bool	Compute cross-correlation between different components of same station [Y/N]
AC	N	bool	Compute auto-correlation from single station components [Y/N]
used	Y	bool	Whether this filter is active [Y/N]

Moving Stack

Configuration set category: stack

Parameter	Default	Type	Description
stack_method	linear	str	Stack Method: Linear Mean or Phase Weighted Stack [linear/pws]
pws_timegate	10.0	float	If stack_method =’pws’, width of the smoothing (in seconds)
pws_power	2.0	float	If stack_method =”pws”, Power of the Weighting
mov_stack	((‘1D’,’1D’))	eval	A list of two parameters: the time to “roll” over (default 1 day) and the granularity (step) of the resulting stacked CCFs (default 1 day) to stack for the Moving-window stacks. This can be a list of tuples, e.g. ((‘1d’,’1d’),(‘2d’,’1d’)) corresponds to the MSNoise 1.6 “1,2” before. Time deltas can be anything pandas can interpret (“d”, “min”, “sec”, etc).
wienerfilt	N	bool	Apply wiener filter before stacking? Y[N] [Y/N]
wiener_mlen	24h	str	Smoothing along date axis (time delta)
wiener_nlen	0.5s	str	Smoothing along lag time axis (time delta)

Reference Stack

Configuration set category: refstack

Parameter	Default	Type	Description
ref_begin	1970-01-01	str	Start of REF period. Absolute date (YYYY-MM-DD) OR negative integer for rolling-index mode (e.g. -5 means 5 windows before current)
ref_end	2100-01-01	str	End of REF period. Absolute date (YYYY-MM-DD) OR negative integer (e.g. -1 means exclude self). Must be > ref_begin when both are negative.
stack_method	linear	str	Stack method for REF computation (Mode A only) [linear/pws]
pws_timegate	10.0	float	If stack_method=pws: smoothing width in seconds (Mode A only)
pws_power	2.0	float	If stack_method=pws: power of the weighting (Mode A only)
wienerfilt	N	bool	Apply Wiener filter to daily CCFs before REF stacking (Mode A only). Y/N [Y/N]
wiener_mlen	24h	str	Wiener smoothing along date axis (time delta, e.g. 24h)
wiener_nlen	0.5s	str	Wiener smoothing along lag-time axis (time delta, e.g. 0.5s)

MWCS

Configuration set category: mwcs

Parameter	Default	Type	Description
freqmin	0.1	float	The lower frequency bound to apply MWCS
freqmax	1.0	float	The upper frequency bound to apply MWCS
mwcs_wlen	10.0	float	Window length to perform MWCS
mwcs_step	2.0	float	Step of the windowing procedure in MWCS

MWCS dt/t

Configuration set category: mwcs_dtt

Parameter	Default	Type	Description
dtt_minlag	5.0	float	Minimum lag time
dtt_width	30.0	float	Width of the time lag window
dtt_lag	static	str	How is the lag window defined for MWCS [static/dynamic]
dtt_v	1.0	float	What velocity to use to avoid ballistic waves
dtt_sides	both	str	Which sides to use [both/left/right]
dtt_mincoh	0.5	float	Minimum coherence on dt measurement
dtt_maxerr	1.0	float	Maximum error on dt measurement
dtt_maxdtt	1.0	float	Maximum dt/t value

MWCS dv/v Aggregate

Configuration set category: mwcs_dtt_dvv

Parameter	Default	Type	Description
dvv_split_pair_type	Y	bool	Split dv/v by pair type (CC/SC/AC) separately. CC=inter-station; SC=same station different location; AC=autocorrelation. [Y/N]
dvv_split_components	Y	bool	Split dv/v by component pair (ZZ/RR/TT/etc) separately. [Y/N]
dvv_cc_value	m	str	dv/v column to use for CC (inter-station) pairs. m=free-intercept slope; m0=forced-through-zero slope. [m/m0]
dvv_cc_error	em	str	Error column for CC pairs. Must match dvv_cc_value: em for m; em0 for m0. [em/em0]
dvv_sc_value	m0	str	dv/v column for SC (single-station different location) pairs. m0 recommended: no clock error expected. [m/m0]
dvv_sc_error	em0	str	Error column for SC pairs. [em/em0]
dvv_ac_value	m0	str	dv/v column for AC (autocorrelation) pairs. m0 recommended: no clock error expected. [m/m0]
dvv_ac_error	em0	str	Error column for AC pairs. [em/em0]
dvv_quality_min	0.0	float	Minimum mean coherence (mcoh) to include a pair at a given time step. Set to 0 to disable.
dvv_weighted_mean	Y	bool	Compute 1/error^2 weighted mean across pairs. [Y/N]
dvv_trimmed_mean	Y	bool	Compute trimmed mean: reject pairs whose dv/v deviates more than dvv_trim_limit sigma from the mean. [Y/N]
dvv_trim_limit	3.0	float	Sigma cutoff for trimmed mean.
dvv_percentiles	5,25,75,95	str	Comma-separated list of percentiles to save (e.g. 5,25,75,95).
dvv_output_percent	Y	bool	Output dv/v in percent (multiply slope by 100). If N output as fraction. [Y/N]

Stretching

Configuration set category: stretching

Parameter	Default	Type	Description
stretching_minlag	5.0	float	Minimum lag time
stretching_width	30.0	float	Width of the time lag window
stretching_lag	static	str	How the lag window is defined for stretching [static/dynamic]
stretching_v	1.0	float	What velocity to use to avoid ballistic waves
stretching_sides	both	str	Which sides to use [both/left/right]
stretching_max	0.02	float	Maximum stretching coefficient
stretching_nsteps	1000	int	Number of stretching steps

Stretching dv/v Aggregate

Configuration set category: stretching_dvv

Parameter	Default	Type	Description
dvv_split_pair_type	Y	bool	Split dv/v by pair type (CC/SC/AC) separately. CC=inter-station; SC=same station different location; AC=autocorrelation. [Y/N]
dvv_split_components	Y	bool	Split dv/v by component pair (ZZ/RR/TT/etc) separately. [Y/N]
dvv_quality_min	0.65	float	Minimum cross-correlation coefficient (Coeff) between stretched reference and data. Pairs below this threshold at a given time step are excluded.
dvv_weighted_mean	Y	bool	Compute 1/Error^2 weighted mean across pairs. Error is the half-width at full maximum of the Gaussian fit. [Y/N]
dvv_trimmed_mean	Y	bool	Compute trimmed mean: reject pairs whose dv/v deviates more than dvv_trim_limit sigma from the mean. [Y/N]
dvv_trim_limit	3.0	float	Sigma cutoff for trimmed mean.
dvv_percentiles	5,25,75,95	str	Comma-separated list of percentiles to save (e.g. 5,25,75,95).
dvv_output_percent	Y	bool	Output dv/v in percent: dv/v(%) = (Delta-1)*100. If N output as raw Delta-1 fraction. [Y/N]

Wavelet Coherence

Configuration set category: wavelet

Parameter	Default	Type	Description
wct_freqmin	0.1	float	The lower frequency bound to apply WCT
wct_freqmax	1.0	float	The upper frequency bound to apply WCT
wct_ns	5	float	Smoothing parameter in frequency
wct_nt	5	float	Smoothing parameter in time
wct_vpo	12	float	Voices per octave: controls the scale resolution of the CWT. Values of 10-12 are standard in ambient-noise WCT literature and give adequate time-frequency resolution for dv/v measurements. 20 (old default) is very fine and rarely changes results while tripling computation time. Default reduced from 20 to 12.
wct_nptsfreq	100	int	Number of frequency points sampled linearly between wct_freqmin and wct_freqmax. Because adjacent wavelet scales are correlated over ~1/vpo octaves, having more points than ~vpo*(log2(freqmax/freqmin)) adds no independent information. For a 0.1-1.0 Hz band at vpo=12 that is ~40 independent scales; 100 points is a safe over-sample. Default reduced from 300 (overkill).
wct_norm	True	bool	If the REF and CCF are normalized before computing wavelet [True/False]
wavelet_type	(‘Morlet’,6.)	str	Type of wavelet function used
wct_compute_dtt	True	bool	When True (default), compute dt/t inline during the wavelet step using ALL downstream wavelet_dtt config sets, and write only the compact DTT/ERR/COH output — no intermediate WCT files are stored on disk. This eliminates ~14000x of intermediate storage (a WCT file for a single pair/component/year is ~2.5 GB compressed vs ~175 KB for DTT). Set to False to store the full WCT arrays (WXamp, Wcoh, WXdt) for inspection or to re-run wavelet_dtt separately with different parameters without recomputing the CWT. [True/False]

Wavelet dt/t

Configuration set category: wavelet_dtt

Parameter	Default	Type	Description
wct_dtt_freqmin	0.1	float	The lower frequency bound to compute the dv/v from the WCT
wct_dtt_freqmax	1.0	float	The upper frequency bound to compute the dv/v from the WCT
wct_minlag	5.0	float	Minimum lag time
wct_width	30.0	float	Width of the time lag window
wct_lag	static	str	How is the lag window defined for WCT [static/dynamic]
wct_v	1.0	float	Velocity parameter used for dynamic lag calculation
wct_sides	both	str	Which sides to use [both/left/right]
wct_mincoh	0.0	float	Minimum coherence on dt measurement
wct_maxdt	1.0	float	Maximum dt values
wct_codacycles	5	int	Number of cycles of period
wct_min_nonzero	0.1	float	Percentage of data points with non-zero weighting

Wavelet dv/v Aggregate

Configuration set category: wavelet_dtt_dvv

Parameter	Default	Type	Description
dvv_freqmin	0.1	float	Lower frequency bound (Hz) for dv/v extraction from the WCT-DTT output. Can be a subset of the parent wct_dtt computation band — enables multi-band analysis from one broadband WCT run.
dvv_freqmax	1.0	float	Upper frequency bound (Hz) for dv/v extraction. Multiple wavelet_dtt_dvv configsets with different freqmin/freqmax can be linked to the same wct_dtt step.
dvv_freq_agg	mean	str	How to collapse the frequency dimension to a scalar dv/v time series: mean or median over the selected band. [mean/median]
dvv_quality_min	0.5	float	Minimum wavelet coherence (coh) to include a time-frequency cell before frequency averaging. Cells below this threshold are masked.
dvv_split_pair_type	Y	bool	Split dv/v by pair type (CC/SC/AC) separately. CC=inter-station; SC=same station different location; AC=autocorrelation. [Y/N]
dvv_split_components	Y	bool	Split dv/v by component pair (ZZ/RR/TT/etc) separately. [Y/N]
dvv_weighted_mean	Y	bool	Compute 1/err^2 weighted mean across pairs after frequency averaging. [Y/N]
dvv_trimmed_mean	Y	bool	Compute trimmed mean: reject pairs whose dv/v deviates more than dvv_trim_limit sigma from the mean. [Y/N]
dvv_trim_limit	3.0	float	Sigma cutoff for trimmed mean.
dvv_percentiles	5,25,75,95	str	Comma-separated list of percentiles to save (e.g. 5,25,75,95).
dvv_output_percent	Y	bool	Output dv/v in percent. [Y/N]

PSD

Configuration set category: psd

Parameter	Default	Type	Description
psd_components	E,N,Z	str	Components to process for PSD, defaults to [E,N,Z], but can be any comma separated list (e.g. ‘Z,E,N’)
psd_ppsd_length	3600.0	float	Length of data segments passed to psd in seconds. In the paper by McNamara et al (2004) a value of 3600 (1 hour) was chosen. Longer segments increase the upper limit of analyzed periods but decrease the number of analyzed segments.
psd_ppsd_overlap	0.0	float	Overlap of segments passed to psd. Overlap may take values between 0 and 1 and is given as fraction of the length of one segment, e.g. psd_ppsd_length=3600 and psd_ppsd_overlap=0.5 result in an overlap of 1800s of the segments.
psd_ppsd_period_smoothing_width_octaves	0.125	float	Determines over what period/frequency range the psd is smoothed around every central period/frequency. Given in fractions of octaves (default of 1 means the psd is averaged over a full octave at each central frequency).
psd_ppsd_period_step_octaves	0.0125	float	Step length on frequency axis in fraction of octaves (default of 0.125 means one smoothed psd value on the frequency axis is measured every 1/8 of an octave).
psd_ppsd_period_limits	(0.01,100)	eval	Set custom lower and upper end of period range (e.g. (0.01, 100) seconds). The specified lower end of period range will be set as the central period of the first bin (geometric mean of left/right edges of smoothing interval). At the upper end of the specified period range, no more additional bins will be added after the bin whose center frequency exceeds the given upper end for the first time.
psd_ppsd_db_bins	(-200, -50, 1.)	eval	Specify the lower and upper boundary and the width of the db bins. The bin width might get adjusted to fit a number of equally spaced bins in between the given boundaries.

PSD RMS

Configuration set category: psd_rms

Parameter	Default	Type	Description
psd_rms_frequency_ranges	[(1.0, 20.0), (4.0, 14.0), (4.0, 40.0), (4.0, 9.0)]	eval	Specify the frequency bounds (in Hz) to compute the RMS from PSDs
psd_rms_type	DISP	str	What units do you want for the exported RMS [DISP/VEL/ACC]