Processing parameters
Here is the list of processing parameters defined in the mainparam.jl
. You can configure the parameters, which is passed to SeisMonitoring.run_job
key | default value | type | description |
---|---|---|---|
project_name | project | String | project name. |
project_inputdir | ./project_INPUT | String | project input directory which you initiated with init_project(). |
project_outputdir | ./project_OUTPUT | String | project output directory which you initiated with init_project(). |
starttime | 2004-04-01T00:00:00 | DateTime | process start time |
endtime | 2004-04-03T00:00:00 | DateTime | process end time |
sampling_frequency | 20.0 | Float64 | [Hz] Processing sampling frequency. Downsampling is applied when downloaded original data is higher sampling frequency. |
freqency_band | [0.01, 0.1, 0.2, 0.5, 1.0, 2.0] | Float64 | Frequency bands to be analyzed. |
MAX_MEM_USE | 3.0 | Float64 | [GB] Maximum memory use per core in the environment. |
download_time_unit | 86400 | Int64 | [s] Unit time of data request. (e.g. request data for each 10minutes. |
download_margin | 300 | Int64 | [s] Download margin to be clipped to avoid edge effect. |
requeststation_file | ./projectOUTPUT/defaultrequeststations.jld2 | String | Request station dataframe saved in JLD2. See default_requeststations.jld2 |
IsResponseRemove | true | Bool | True if remove instrumental response while downloading data. |
IsLocationBox | false | Bool | True if using lat-lon box for request. |
reg | [35.7, 36.1, -120.7, -120.2] | Float64 | minlat, maxlat, minlon, maxlon |
Istmpfilepreserved | false | Bool | True if you want to preserve temporal files (same size as raw data.) |
IsXMLfilepreserved | false | Bool | True if you want to preserve station xml files. |
numstationperrequest | 1 | Int64 | Advanced: number of station per one HTTP request. |
outputformat | JLD2 | String | JLD2 or ASDF: use JLD2 if you perform the following processes with SeisMonitoring.jl |
RawData_path | default | String | "default" or absolute/relative path to rawdata. "default" links to project output directory. |
IsKurtosisRemoval | true | Bool | Apply Kurtosis removal. |
IsSTALTARemoval | true | Bool | Apply STA/LTA removal. |
IsWhitening | false | Bool | Apply Spectral whitening. |
freqmin_whiten | 0.1 | Float64 | Minimum cutoff frequency for spectral whitening |
freqmax_whiten | 1.0 | Float64 | Maximum cutoff frequency for spectral whitening |
Append_alltraces | false | Bool | Append kurtosis and stalta traces to SeisChannel (this increases data size) |
shorttime_window | 180.0 | Float64 | Short-time window used to compute kurtosis and sta/lta |
longtime_window | 86400.0 | Float64 | Long-time window used to compute sta/lta |
timewindow_overlap | 60.0 | Float64 | Short-time window overlap to compute kurtosis and sta/lta |
kurtosis_threshold | 2.0 | Float64 | Kurtosis removal threshold (The normal distribution of kurtosis is normalized to be zero.) |
stalta_threshold | 1.2 | Float64 | STA/LTA removal threshold (For our purpose, this threshold is smaller than ordinal detection.) |
stalta_absoluteclip | 0.1 | Float64 | [unit-of-data] clip the signal above this value (basically for instrumental error.) |
fixed_tukey_margin | 30.0 | Float64 | [s] Fixed turkey margin; duration of decay outside of zero padding |
IsIsolateComponents | false | Bool | Advanced: isolating comonents at same station |
cc_time_unit | 86400 | Int64 | [s] Unit time of cross-correlation window. e.g. 60*60*24 = 86400 indicates daily-cross correlation. |
cc_len | 3600 | Int64 | [s] short-time window cross-correlation length |
cc_step | 1800 | Int64 | [s] cross-correlation window step |
maxlag | 100.0 | Float64 | [s] Maximum time lag of cross-correlation. |
cc_RawData_path | default | String | "default" or absolute/relative path to rawdata. "default" links to project OUTPUT/EQRemovedData.jld2. |
cc_normalization | deconvolution | String | none, coherence or deconvolution. |
corr_type | CC | String | Type of correlation: CC (standard cross-correlation) or PCC (phase cross-correlation). See also doc in SeisNoise.jl |
pairs_option | SubString{String}["11", "22", "33"] | Vector{String} | "all" or list of component pairs. e.g. XX, YY, ZZ |
chanpair_type | SubString{String}["all"] | Vector{String} | "all" or list of channel pair type. e.g. auto-achan, cross-achan, cross-xchan |
data_contents_fraction | 0.8 | Float64 | Advanced: discard cross-correlation if data fraction within cc_time_unit is less that this value. |
IsOnebit | false | Bool | Apply One-bit normalization. |
smoothing_windowlength | 7 | Int64 | Advanced: number of points for boxcar smoothing window on coherence and deconvolution. |
water_level | 0.0001 | Float64 | Advanced: waterlevel [0.0 if not applied] on spectrum normalization with coherence and deconvolution method. |
cc_bpfilt_method | ButterWorth | String | Frequency decomposition method. "Butterworth" or "Wavelet". |
cc_taper_α0 | 0.1 | Float64 | Advanced: Lowest tapering fraction for frequency adaptive tapering. |
cc_taper_αmax | 0.25 | Float64 | Advanced: Highest tapering fraction for frequency adaptive tapering. |
cc_medianmute_max | 5.0 | Float64 | Advanced: Threshold factor of median mute within cc_time_unit. NCF is removed if maximum(abs.(corr[:,i])) > cc_medianmute_max * median(maximum(abs.(corr)), dims=1) |
cc_medianmute_min | 0.1 | Float64 | Advanced: Threshold factor of median mute within cc_time_unit. NCF is removed if maximum(abs.(corr[:,i])) < cc_medianmute_min * median(maximum(abs.(corr)), dims=1) |
IsPreStack | true | Bool | Advanced: Pre-stacking corrdata within each cc_time_unit when assembling the corrdata for the sake of saving memory use. |
timechunk_increment | 1 | Int64 | Advanced: Number of time chunk increment for parallelization: large number is more efficient, but increase memory use. |
stack_RawData_dir | default | String | "default" or absolute/relative path to cc directory. "default" links to project OUTPUT/cc. |
use_local_tmpdir | true | Bool | True if using local /tmp diretory. Please set true when running in cluster to avoid massive file I/O. |
stack_method | linear | String | stacking method: linear, selective, robust, pws, robustpws are available |
collect_stationpairs | true | Bool | true if correct station pairs. Stacking without this process does not work. |
compute_reference | true | Bool | true if compute reference stack for longterm stack. |
compute_shorttimestack | true | Bool | true if compute shorttime stack for continuous monitoring. |
stack_pairs_option | SubString{String}["11", "22", "33"] | Vector{String} | "all" or list of component pairs. e.g. XX, YY, ZZ |
averagestack_factor | 1 | Int64 | Integer factor of cc_time_unit for stacking duration. e.g. cc_time_unit = 1day and averagestack_factor=30 provides 30days moving window average. |
averagestack_step | 1 | Int64 | Step of averagestack window. |
min_cc_datafraction | 0.5 | Float64 | Advanced: discard cross-correlation if data fraction within stacking period is less that this value. |
reference_starttime | 2004-04-01T00:00:00 | DateTime | reference start time |
reference_endtime | 2004-04-02T00:00:00 | DateTime | reference end time |
dist_threshold | 1.0 | Float64 | Threshold of distance used for selective stacking. |
distance_type | CorrDist | String | Advanced: Distance type used in selective stacking. See https://github.com/JuliaStats/Distances.jl for available types. |
IsZeropadBeforeStack | false | Bool | Zero padding outside of coda window using tukey window before stacking. |
background_vel | 2000.0 | Float64 | [m/s] Approximation of background wave velocity, just used for coda slicing. |
min_ballistic_twin | 1.0 | Float64 | [s] Explicit ballistic time window to remove coherence around zero timelag. This is aimed to remove it mainly for auto-correlation. |
max_coda_length | 60.0 | Float64 | [s] Maximum coda window length. |
mwcc_threshold | 0.5 | Float64 | mwcc slice coda threshold. |
mwcc_len_α | 3.0 | Float64 | moving window size factor (size = (mwcc_len_α/fm)*fs [point]). |
min_codalength_α | 1.0 | Float64 | Threshold of minimum codawindow length: min_codalength = min_codalength_α*mwcc window length. |
coda_init_factor | 2.0 | Float64 | [s] Coda window starts from coda_init_factor*dist/vel. |
coda_minlen_factor | 5.0 | Float64 | [s] Minimumlength is determined by this factor * (1/fm, period of cc) * fs points. |
codaslice_debugplot | false | Bool | If plot debug figures for coda slicing. |
nondim_max_coda_length | 30.0 | Float64 | Deprecated: nondimensional maximum coda window length |
nondim_codamaxlag | 60.0 | Float64 | Deprecated: coda max lag where kinetic energy is evaluated. |
coda_energy_threshold | -1.0 | Float64 | Deprecated: Advanced: Threshold for attenuation decay. |
IsAlternateRefChannel | true | Bool | Advanced: Allow for using alternative station channel for reference. (e.g. BP.LCCB..BP1-BP.MMNB..BP1 is used as reference for BP.LCCB..SP1-BP.MMNB..SP1) |
keep_corrtrace | false | Bool | Advanced: Keep corr trace in CorrData if true. (require the strage to save corrs.) |
measurement_method | mwcs | String | Stretching method for measuring dv/v and dQ^{-1}. "stretching","mwcs","wcc","dtw","dualstretching" |
mwcs_window_length | 6.0 | Float64 | [s] The moving window length |
mwcs_window_step | 3.0 | Float64 | [s] The step to jump for the moving window. |
mwcs_smoothing_half_win | 5 | Int64 | [Points] MWCS smoothing half windown length. |
mwcs_max_dt | 1.0 | Float64 | [s] MWCS threshold on dt. |
stretch_debugplot | false | Bool | If plot debug figures for streching. |
dvv_stretching_range | 0.02 | Float64 | Advanced: dvv stretching trial range for dvv (+- abs(dvv_stretching_range)). |
dvv_stretching_Ntrial | 201 | Int64 | Advanced: dvv stretching trial number for dvv. |
geometricalspreading_α | 0.5 | Float64 | Advanced: geometrical spreading coefficient to compute Qcinv. |
computedqq_smoothing_windowlength | 10.0 | Float64 | [s] smoothing windown length to compute envelope for compute_dvvdqq. |