pylot/pylot/core/pick/autopick.py

1588 lines
78 KiB
Python

#!/usr/bin/env python
# -*- coding: utf-8 -*-
"""
Function to run automated picking algorithms using AIC,
HOS and AR prediction. Uses objects CharFuns and Picker and
function conglomerate utils.
:author: MAGS2 EP3 working group / Ludger Kueperkoch
"""
import matplotlib.pyplot as plt
import numpy as np
from pylot.core.pick.charfuns import CharacteristicFunction
from pylot.core.pick.charfuns import HOScf, AICcf, ARZcf, ARHcf, AR3Ccf
from pylot.core.pick.picker import AICPicker, PragPicker
from pylot.core.pick.utils import checksignallength, checkZ4S, earllatepicker, \
getSNR, fmpicker, checkPonsets, wadaticheck, get_pickparams, get_quality_class
from pylot.core.util.utils import getPatternLine, gen_Pool,\
real_Bool, identifyPhaseID
from obspy.taup import TauPyModel
def autopickevent(data, param, iplot=0, fig_dict=None, fig_dict_wadatijack=None, ncores=0, metadata=None, origin=None):
"""
:param data: ObsPy stream object containing waveform data of all stations in the event
:type data: ~obspy.core.stream.Stream
:param param: PylotParameter object containing parameters used for picking
:type param: pylot.core.io.inputs.PylotParameter
:param iplot: logical variable for plotting: 0=none, 1=partial, 2=all
:type iplot: int, Boolean or String
:param fig_dict: dictionary containing Matplotlib figures used for plotting picking results during tuning
:type fig_dict: dict
:param fig_dict_wadatijack: dictionary containing Matplotlib figures used for plotting jackknife-, wadati- and
mediantest results
:type fig_dict_wadatijack: dict
:param ncores: number of cores used for parallel processing. Default (0) uses all available cores
:type ncores: int
:param metadata: tuple containing metadata type string and Parser object read from inventory file
:type metadata: tuple (str, ~obspy.io.xseed.parser.Parser)
:param origin: list containing origin objects representing origins for all events
:type origin: list(~obspy.core.event.origin)
:return: dictionary containing picked stations and pick information
:rtype: dictionary
"""
stations = []
all_onsets = {}
input_tuples = []
try:
iplot = int(iplot)
except ValueError:
if iplot is True or iplot == 'True':
iplot = 2
else:
iplot = 0
# get some parameters for quality control from
# parameter input file (usually pylot.in).
wdttolerance = param.get('wdttolerance')
mdttolerance = param.get('mdttolerance')
jackfactor = param.get('jackfactor')
apverbose = param.get('apverbose')
for n in range(len(data)):
station = data[n].stats.station
if station not in stations:
stations.append(station)
else:
continue
for station in stations:
topick = data.select(station=station)
if iplot is None or iplot == 'None' or iplot == 0:
input_tuples.append((topick, param, apverbose, metadata, origin))
if iplot > 0:
all_onsets[station] = autopickstation(topick, param, verbose=apverbose,
iplot=iplot, fig_dict=fig_dict,
metadata=metadata, origin=origin)
if iplot > 0:
print('iPlot Flag active: NO MULTIPROCESSING possible.')
return all_onsets
# rename str for ncores in case ncores == 0 (use all cores)
ncores_str = ncores if ncores != 0 else 'all available'
print('Autopickstation: Distribute autopicking for {} '
'stations on {} cores.'.format(len(input_tuples), ncores_str))
pool = gen_Pool(ncores)
result = pool.map(call_autopickstation, input_tuples)
pool.close()
for pick in result:
if pick:
station = pick['station']
pick.pop('station')
all_onsets[station] = pick
# quality control
# median check and jackknife on P-onset times
jk_checked_onsets = checkPonsets(all_onsets, mdttolerance, jackfactor, 1, fig_dict_wadatijack)
# check S-P times (Wadati)
wadationsets = wadaticheck(jk_checked_onsets, wdttolerance, 1, fig_dict_wadatijack)
return wadationsets
def call_autopickstation(input_tuple):
"""
helper function used for multiprocessing
:param input_tuple: contains all parameters used for autopicking
:type input_tuple: tuple
:return: dictionary containing P pick, S pick and station name
:rtype: dict
"""
wfstream, pickparam, verbose, metadata, origin = input_tuple
# multiprocessing not possible with interactive plotting
return autopickstation(wfstream, pickparam, verbose, iplot=0, metadata=metadata, origin=origin)
def get_source_coords(parser, station_id):
"""
retrieves station coordinates from metadata
:param parser: Parser object containing metadata read from inventory file
:type parser: ~obspy.io.xseed.parser.Parser
:param station_id: station id of which the coordinates should be retrieved, for example 'BW.RJOB..EHZ'. Only
network and station name is required, channel id (last part) is ignored.
:type station_id: str
:return: dictionary containing 'latitude', 'longitude', 'elevation' and 'local_depth' of station
:rtype: dict
"""
station_coords = None
try:
station_coords = parser.get_coordinates(station_id)
except Exception as e:
print('Could not get source coordinates for station {}: {}'.format(station_id, e))
return station_coords
class PickingParameters(object):
"""
Stores parameters used for picking a single station.
"""
def __init__(self, *args, **kwargs):
"""
Add dictionaries given as positional arguments and the keyword argument dictionary to the instance
as attributes. Positional arguments with types differing from dict are ignored.
"""
# add entries from dictionaries given as positional arguments
for arg in args:
if type(arg) == dict:
self.add_params_from_dict(arg)
# add values given as keyword arguments
self.add_params_from_dict(kwargs)
def add_params_from_dict(self, d):
"""
Add all key-value pairs from dictionary d to the class namespace as attributes.
:param d:
:type d: dict
:rtype: None
"""
for key, value in d.items():
setattr(self, key, value)
class PickingResults(object):
def __init__(self):
# initialize output
self.Pweight = 4 # weight for P onset
self.Sweight = 4 # weight for S onset
self.FM = 'N' # first motion (polarity)
self.SNRP = None # signal-to-noise ratio of P onset
self.SNRPdB = None # signal-to-noise ratio of P onset [dB]
self.SNRS = None # signal-to-noise ratio of S onset
self.SNRSdB = None # signal-to-noise ratio of S onset [dB]
self.mpickP = None # most likely P onset
self.lpickP = None # latest possible P onset
self.epickP = None # earliest possible P onset
self.mpickS = None # most likely S onset
self.lpickS = None # latest possible S onset
self.epickS = None # earliest possible S onset
self.Perror = None # symmetrized picking error P onset
self.Serror = None # symmetrized picking error S onset
self.aicSflag = 0
self.aicPflag = 0
self.Pflag = 0
self.Sflag = 0
self.Pmarker = []
self.Ao = None # Wood-Anderson peak-to-peak amplitude
self.picker = 'auto' # type of picks
class MissingTraceException(ValueError):
"""
Used to indicate missing traces in a obspy.core.stream.Stream object
"""
pass
class AutopickStation(object):
def __init__(self, wfstream, pickparam, verbose, iplot, fig_dict, metadata, origin):
# save given parameters
self.wfstream = wfstream
self.pickparam = pickparam
self.verbose = verbose
self.iplot = iplot
self.fig_dict = fig_dict
self.metadata = metadata
self.origin = origin
# extract additional information
pickparams = self.extract_pickparams(pickparam)
self.p_params, self.s_params, self.first_motion_params, self.signal_length_params = pickparams
self.results = PickingResults()
self.channelorder = {'Z': 3, 'N': 1, 'E': 2} # TODO get this from the pylot preferences
self.station_name = wfstream[0].stats.station
self.network_name = wfstream[0].stats.network
self.station_id = '{}.{}'.format(self.network_name, self.station_name)
# default values used in old autopickstation function #TODO way for user to set those
self.detrend_type = 'demean'
self.filter_type = 'bandpass'
self.zerophase = False
self.taper_max_percentage = 0.05
self.taper_type = 'hann'
def vprint(self, s):
"""Only print statement if verbose picking is set to true."""
if self.verbose:
print(s)
def extract_pickparams(self, pickparam):
"""
Get parameter names out of pickparam dictionary into PickingParameters objects and return them.
:return: PickingParameters objects containing 1. p pick parameters, 2. s pick parameters, 3. first motion determinatiion
parameters, 4. signal length parameters
:rtype: (PickingParameters, PickingParameters, PickingParameters, PickingParameters)
"""
# Define names of all parameters in different groups
p_parameter_names = 'algoP pstart pstop use_taup taup_model tlta tsnrz hosorder bpz1 bpz2 pickwinP aictsmooth tsmoothP ausP nfacP tpred1z tdet1z Parorder addnoise Precalcwin minAICPslope minAICPSNR timeerrorsP'.split(
' ')
s_parameter_names = 'algoS sstart sstop bph1 bph2 tsnrh pickwinS tpred1h tdet1h tpred2h tdet2h Sarorder aictsmoothS tsmoothS ausS minAICSslope minAICSSNR Srecalcwin nfacS timeerrorsS zfac'.split(
' ')
first_motion_names = 'minFMSNR fmpickwin minfmweight'.split(' ')
signal_length_names = 'minsiglength minpercent noisefactor'.split(' ')
# Get list of values from pickparam by name
p_parameter_values = map(pickparam.get, p_parameter_names)
s_parameter_values = map(pickparam.get, s_parameter_names)
fm_parameter_values = map(pickparam.get, first_motion_names)
sl_parameter_values = map(pickparam.get, signal_length_names)
# construct dicts from names and values
p_params = dict(zip(p_parameter_names, p_parameter_values))
s_params = dict(zip(s_parameter_names, s_parameter_values))
first_motion_params = dict(zip(first_motion_names, fm_parameter_values))
signal_length_params = dict(zip(signal_length_names, sl_parameter_values))
p_params['use_taup'] = real_Bool(p_params['use_taup'])
return PickingParameters(p_params), PickingParameters(s_params), PickingParameters(first_motion_params), PickingParameters(signal_length_params)
def get_components_from_waveformstream(self):
"""
Splits waveformstream into multiple components zdat, ndat, edat. For traditional orientation (ZNE) these contain
the vertical, north-south or east-west component. Otherwise they contain components numbered 123 with
orientation diverging from the traditional orientation.
:param waveformstream: Stream containing all three components for one station either by ZNE or 123 channel code
(mixture of both options is handled as well)
:type waveformstream: obspy.core.stream.Stream
:return: Tuple containing (z waveform, n waveform, e waveform) selected by the given channels
:rtype: (obspy.core.stream.Stream, obspy.core.stream.Stream, obspy.core.stream.Stream)
"""
waveform_data = {}
for key in self.channelorder:
waveform_data[key] = self.wfstream.select(component=key) # try ZNE first
if len(waveform_data[key]) == 0:
waveform_data[key] = self.wfstream.select(component=str(self.channelorder[key])) # use 123 as second option
return waveform_data['Z'], waveform_data['N'], waveform_data['E']
def prepare_wfstream(self, wfstream, freqmin=None, freqmax=None):
"""
Prepare a waveformstream for picking by applying detrending, filtering and tapering. Creates a copy of the
waveform the leave the original unchanged.
:param wfstream:
:type wfstream:
:param freqmin:
:type freqmin:
:param freqmax:
:type freqmax:
:return: Tuple containing the changed waveform stream and the changed first trace of the stream
:rtype: (obspy.core.stream.Stream, obspy.core.trace.Trace)
"""
wfstream_copy = wfstream.copy()
trace_copy = wfstream[0].copy()
trace_copy.detrend(type=self.detrend_type)
trace_copy.filter(self.filter_type, freqmin=freqmin, freqmax=freqmax, zerophase=self.zerophase)
trace_copy.taper(max_percentage=self.taper_max_percentage, type=self.taper_type)
wfstream_copy[0].data = trace_copy.data
return wfstream_copy, trace_copy
def modify_starttimes_taupy(self):
"""
Calculate theoretical arrival times for a source at self.origin and a station at self.metadata. Modify
self.pstart and self.pstop so they are based on a time window around these theoretical arrival times.
:rtype: None
"""
def create_arrivals(metadata, origin, station_id, taup_model):
"""
Create List of arrival times for all phases for a given origin and station
:param metadata: tuple containing metadata type string and Parser object read from inventory file
:type metadata: tuple (str, ~obspy.io.xseed.parser.Parser)
:param origin: list containing origin objects representing origins for all events
:type origin: list(~obspy.core.event.origin)
:param station_id: Station id with format NETWORKNAME.STATIONNAME
:type station_id: str
:param taup_model: Model name to use. See obspy.taup.tau.TauPyModel for options
:type taup_model: str
:return: List of Arrival objects
:rtype: obspy.taup.tau.Arrivals
:raises:
AttributeError when no metadata or source origins is given
"""
parser = metadata[1]
station_coords = get_source_coords(parser, station_id)
source_origin = origin[0]
model = TauPyModel(taup_model)
arrivals = model.get_travel_times_geo(source_depth_in_km=source_origin.depth,
source_latitude_in_deg=source_origin.latitude,
source_longitude_in_deg=source_origin.longitude,
receiver_latitude_in_deg=station_coords['latitude'],
receiver_longitude_in_deg=station_coords['longitude'])
return arrivals
def first_PS_onsets(arrivals):
"""
Get first P and S onsets from arrivals list
:param arrivals: List of Arrival objects
:type arrivals: obspy.taup.tau.Arrivals
:return:
:rtype:
"""
phases = {'P': [], 'S': []}
# sort phases in P and S phases
for arr in arrivals:
phases[identifyPhaseID(arr.phase.name)].append(arr)
# get first P and S onsets from arrivals list
arrP, estFirstP = min([(arr, arr.time) for arr in phases['P']], key=lambda t: t[1])
arrS, estFirstS = min([(arr, arr.time) for arr in phases['S']], key=lambda t: t[1])
print('autopick: estimated first arrivals for P: {} s, S:{} s after event'
' origin time using TauPy'.format(estFirstP, estFirstS))
return estFirstP, estFirstS
print('autopickstation: use_taup flag active.')
if not self.metadata[1]:
raise AttributeError('Warning: Could not use TauPy to estimate onsets as there are no metadata given.')
if not self.origin:
raise AttributeError('No source origins given!')
arrivals = create_arrivals(self.metadata, self.origin, self.station_id, self.p_params.taup_model)
estFirstP, estFirstS = first_PS_onsets(arrivals)
# modifiy pstart and pstop relative to estimated first P arrival (relative to station time axis)
self.p_params.pstart += (self.origin[0].time + estFirstP) - self.zstream[0].stats.starttime
self.p_params.pstop += (self.origin[0].time + estFirstP) - self.zstream[0].stats.starttime
print('autopick: CF calculation times respectively:'
' pstart: {} s, pstop: {} s'.format(self.p_params.pstart, self.p_params.pstop))
# make sure pstart and pstop are inside the starttime/endtime of vertical trace
self.p_params.pstart = max(self.p_params.pstart, 0)
self.p_params.pstop = min(self.p_params.pstop, len(self.zstream[0]) * self.zstream[0].stats.starttime)
def autopickstation(self):
self.zstream, self.nstream, self.estream = self.get_components_from_waveformstream()
self.ztrace, self.ntrace, self.etrace = self.zstream[0], self.nstream[0], self.estream[0]
try:
self.pick_p_phase()
#TODO handle exceptions correctly
# requires an overlook of what should be returned in case picking fails at various stages
except MissingTraceException as mte:
print(mte)
return
def pick_p_phase(self):
"""
Pick p phase, return results
:return: P pick results
:rtype: PickingResults
:raises:
MissingTraceException: If vertical trace is missing.
"""
if not self.zstream or self.zstream is None:
raise MissingTraceException('No z-component found for station {}'.format(self.station_name))
msg = '##################################################\nautopickstation:' \
' Working on P onset of station {station}\nFiltering vertical ' \
'trace ...\n{data}'.format(station=self.station_name, data=str(self.zstream))
self.vprint(msg)
z_copy, tr_filt = self.prepare_wfstream(self.zstream, self.p_params.bpz1[0], self.p_params.bpz2[0])
if self.p_params.use_taup is True and self.origin is not None:
Lc = np.inf # what is Lc? DA
try:
self.modify_starttimes_taupy()
except AttributeError as ae:
print(ae)
else:
Lc = self.p_params.pstop - self.p_params.pstart
Lwf = self.ztrace.stats.endtime - self.ztrace.stats.starttime
if Lwf < 0:
print('autopickstation: empty trace! Return!')
return
Ldiff = Lwf - abs(Lc)
if Ldiff < 0 or self.p_params.pstop <= self.p_params.pstart:
msg = 'autopickstation: Cutting times are too large for actual waveform!\nUsing entire waveform instead!'
self.vprint(msg)
self.p_params.pstart = 0
self.p_params.pstop = len(self.ztrace.data) * self.ztrace.stats.delta
cuttimes = [self.p_params.pstart, self.p_params.pstop]
if self.p_params.algoP == 'HOS':
cf1 = HOScf(z_copy, cuttimes, self.p_params.tlta, self.p_params.hosorder)
elif self.p_params.algoP == 'ARZ':
cf1 = ARZcf(z_copy, cuttimes, self.p_params.tpred1z, self.p_params.Parorder, self.p_params.tdet1z,
self.p_params.addnoise)
else:
cf1 = None
assert isinstance(cf1, CharacteristicFunction), 'cf2 is not set ' \
'correctly: maybe the algorithm name ({algoP}) is ' \
'corrupted'.format(algoP=self.p_params.algoP)
# AICcf needs stream object -> build it
tr_aic = tr_filt.copy()
tr_aic.data = cf1.getCF()
z_copy[0].data = tr_aic.data
aiccf = AICcf(z_copy, cuttimes)
# get preliminary onset time from AIC-CF
fig, linecolor = get_fig_from_figdict(self.fig_dict, 'aicFig')
aicpick = AICPicker(aiccf, self.p_params.tsnrz, self.p_params.pickwinP, self.iplot,
Tsmooth=self.p_params.aictsmooth, fig=fig, linecolor=linecolor)
# add pstart and pstop to aic plot
if fig:
for ax in fig.axes:
ax.vlines(self.p_params.pstart, ax.get_ylim()[0], ax.get_ylim()[1], color='c', linestyles='dashed', label='P start')
ax.vlines(self.p_params.pstop, ax.get_ylim()[0], ax.get_ylim()[1], color='c', linestyles='dashed', label='P stop')
ax.legend(loc=1)
fig, linecolor = get_fig_from_figdict(self.figdict, 'slength')
if aicpick.getpick() is not None:
z_copy[0].data = tr_filt.data
zne = z_copy
if len(self.nstream) == 0 or len(self.estream) == 0:
msg = 'One or more horizontal component(s) missing!\n' \
'Signal length only checked on vertical component!\n' \
'Decreasing minsiglengh from {0} to {1}' \
.format(self.signal_length_params.minsiglength, self.signal_length_params.minsiglength/2)
self.vprint(msg)
minsiglength = self.signal_length_params.minsiglength/2
else:
trH1_filt, _ = self.prepare_wfstream(self.estream, freqmin=self.s_params.bph1[0], freqmax=self.s_params.bph1[1])
trH2_filt, _ = self.prepare_wfstream(self.nstream, freqmin=self.s_params.bph1[0], freqmax=self.s_params.bph1[1])
zne += trH1_filt
zne += trH2_filt
minsiglength = self.signal_length_params.minsiglength
Pflag = checksignallength(zne, aicpick.getpick(), self.p_params.tsnrz,
minsiglength,
self.signal_length_params.noisefactor,
self.signal_length_params.minpercent, self.iplot, fig, linecolor)
if Pflag == 0:
Pmarker = 'shortsignallength'
Pweight = 9
if Pflag == 1:
if len(self.nstream) == 0 or len(self.estream) == 0:
msg = 'One or more horizontal components missing!\n' \
'Skipping control function checkZ4S.'
self.vprint(msg)
else:
if self.iplot > 1:
fig, linecolor = get_fig_from_figdict(self.fig_dict, 'checkZ4s')
Pflag = checkZ4S(zne, aicpick.getpick(), self.s_params.zfac, self.p_params.tsnrz[2], self.iplot, fig, linecolor)
if Pflag == 0:
Pmarker = 'SinsteadP'
Pweight = 9
# go on with processing if AIC onset passes quality control
slope = aicpick.getSlope()
if not slope: slope = 0
if slope >= self.p_params.minAICPslope and aicpick.getSNR() >= self.p_params.minAICPSNR and Pflag == 1:
aicPflag = 1
msg = 'AIC P-pick passes quality control: Slope: {0} counts/s, ' \
'SNR: {1}\nGo on with refined picking ...\n' \
'autopickstation: re-filtering vertical trace ' \
'...'.format(aicpick.getSlope(), aicpick.getSNR())
self.vprint(msg)
# refilter waveform with larger bandpass
z_copy, tr_filt = self.prepare_wfstream(self.zstream, freqmin=self.p_params.bpz2[0])
cuttimes2 = [round(max([aicpick.getpick() - self.p_params.Precalcwin, 0])),
round(min([len(self.ztrace.data) * self.ztrace.stats.delta,
aicpick.getpick() + self.p_params.Precalcwin]))]
if self.p_params.algoP == 'HOS':
cf2 = HOScf(z_copy, cuttimes2, self.p_params.tlta, self.p_params.hosorder)
elif self.p_params.algoP == 'ARZ':
cf2 = ARZcf(z_copy, cuttimes2, self.p_params.tpred1z, self.p_params.Parorder, self.p_params.tdet1z, self.p_params.addnoise)
else:
cf2 = None
# get refined onset time from CF2
assert isinstance(cf2, CharacteristicFunction), 'cf2 is not set ' \
'correctly: maybe the algorithm name ({algoP}) is ' \
'corrupted'.format(algoP=self.p_params.algoP)
fig, linecolor = get_fig_from_figdict(self.fig_dict, 'refPpick')
refPpick = PragPicker(cf2, self.p_params.tsnrz, self.p_params.pickwinP, self.iplot, self.p_params.ausP,
self.p_params.tsmoothP, aicpick.getpick(), fig, linecolor)
mpickP = refPpick.getpick()
if mpickP is not None:
# quality assessment, get earliest/latest pick and symmetrized uncertainty
fig, linecolor = get_fig_from_figdict(self.fig_dict, 'el_Ppick')
epickP, lpickP, Perror = earllatepicker(z_copy, self.p_params.nfacP, self.p_params.tsnrz, mpickP,
self.iplot, fig=fig, linecolor=linecolor)
SNRP, SNRPdB, Pnoiselevel = getSNR(z_copy, self.p_params.tsnrz, mpickP)
# weight P-onset using symmetric error
#todo shorter expression for this
if Perror is None:
Pweight = 4
else:
if Perror <= self.p_params.timeerrorsP[0]:
Pweight = 0
elif self.p_params.timeerrorsP[0] < Perror <= self.p_params.timeerrorsP[1]:
Pweight = 1
elif self.p_params.timeerrorsP[1] < Perror <= self.p_params.timeerrorsP[2]:
Pweight = 2
elif self.p_params.timeerrorsP[2] < Perror <= self.p_params.timeerrorsP[3]:
Pweight = 3
elif Perror > self.p_params.timeerrorsP[3]:
Pweight = 4
if Pweight <= self.first_motion_params.minfmweight and SNRP >= self.first_motion_params.minFMSSNR:
fig, linecolor = get_fig_from_figdict(self.fig_dict, 'fm_picker')
FM = fmpicker(self.zstream, z_copy, self.first_motion_params.fmpickwin, mpickP, self.iplot,
fig, linecolor)
else:
FM = 'N'
msg = "autopickstation: P-weight: {0}, " \
"SNR: {1}, SNR[dB]: {2}, Polarity: {3}".format(Pweight, SNRP, SNRPdB, FM)
print(msg)
msg = 'autopickstation: Refined P-Pick: {} s | P-Error: {} s'.format(mpickP, Perror)
print(msg)
Sflag = 1
else:
msg = 'Bad initial (AIC) P-pick, skipping this onset!\n' \
'AIC-SNR={0}, AIC-Slope={1}counts/s\n' \
'(min. AIC-SNR={2}, ' \
'min. AIC-Slope={3}counts/s)'.format(aicpick.getSNR(), aicpick.getSlope(),
self.p_params.minAICPSNR, self.p_params.minAICPslope)
self.vprint(msg)
Sflag = 0
def get_fig_from_figdict(figdict, figkey):
"""
:param figdict:
:type figdict: dict
:param figkey:
:type figkey: str
:return:
:rtype:
"""
fig = figdict.get(figkey, None)
linecolor = figdict.get('plot_style', 'k')
linecolor = linecolor['linecolor']['rgba_mpl']
return fig, linecolor
def autopickstation(wfstream, pickparam, verbose=False,
iplot=0, fig_dict=None, metadata=None, origin=None):
"""
picks a single station
:param wfstream: stream object containing waveform of all traces
:type wfstream: ~obspy.core.stream.Stream
:param pickparam: container of picking parameters from input file, usually pylot.in
:type pickparam: pylot.core.io.inputs.PylotParameter
:param verbose: used to control output to log during picking. True = more information printed
:type verbose: bool
:param iplot: logical variable for plotting: 0=none, 1=partial, 2=all
:type iplot: int, (Boolean or String)
:param fig_dict: dictionary containing Matplotlib figures used for plotting picking results during tuning
:type fig_dict: dict
:param metadata: tuple containing metadata type string and Parser object read from inventory file
:type metadata: tuple (str, ~obspy.io.xseed.parser.Parser)
:param origin: list containing origin objects representing origins for all events
:type origin: list(~obspy.core.event.origin)
:return: dictionary containing P pick, S pick and station name
:rtype: dict
"""
# declaring pickparam variables (only for convenience)
# read your pylot.in for details!
plt_flag = 0
# get picking parameter dictionaries
p_params, s_params, first_motion_params, signal_length_params = get_pickparams(pickparam)
# initialize output
Pweight = 4 # weight for P onset
Sweight = 4 # weight for S onset
FM = 'N' # first motion (polarity)
SNRP = None # signal-to-noise ratio of P onset
SNRPdB = None # signal-to-noise ratio of P onset [dB]
SNRS = None # signal-to-noise ratio of S onset
SNRSdB = None # signal-to-noise ratio of S onset [dB]
mpickP = None # most likely P onset
lpickP = None # latest possible P onset
epickP = None # earliest possible P onset
mpickS = None # most likely S onset
lpickS = None # latest possible S onset
epickS = None # earliest possible S onset
Perror = None # symmetrized picking error P onset
Serror = None # symmetrized picking error S onset
aicSflag = 0
aicPflag = 0
Pflag = 0
Sflag = 0
Pmarker = []
Ao = None # Wood-Anderson peak-to-peak amplitude
picker = 'auto' # type of picks
def get_components_from_waveformstream(waveformstream):
"""
Splits waveformstream into multiple components zdat, ndat, edat. For traditional orientation (ZNE) these contain
the vertical, north-south or east-west component. Otherwise they contain components numbered 123 with
orientation diverging from the traditional orientation.
:param waveformstream: Stream containing all three components for one station either by ZNE or 123 channel code
(mixture of both options is handled as well)
:type waveformstream: obspy.core.stream.Stream
:return: Tuple containing (z waveform, n waveform, e waveform) selected by the given channels
:rtype: (obspy.core.stream.Stream, obspy.core.stream.Stream, obspy.core.stream.Stream)
"""
#TODO: get this order from the pylot preferences
channelorder_default = {'Z': 3, 'N': 1, 'E': 2}
waveform_data = {}
for key in channelorder_default:
waveform_data[key] = waveformstream.select(component=key) # try ZNE first
if len(waveform_data[key]) == 0:
waveform_data[key] = waveformstream.select(component=str(channelorder_default[key])) # use 123 as second option
return waveform_data['Z'], waveform_data['N'], waveform_data['E']
def prepare_wfstream_component(wfstream, detrend_type='demean', filter_type='bandpass', freqmin=None, freqmax=None, zerophase=False, taper_max_percentage=0.05, taper_type='hann'):
"""
Prepare a waveformstream for picking by applying detrending, filtering and tapering. Creates a copy of the
waveform the leave the original unchanged.
:param wfstream:
:type wfstream:
:param detrend_type:
:type detrend_type:
:param filter_type:
:type filter_type:
:param freqmin:
:type freqmin:
:param freqmax:
:type freqmax:
:param zerophase:
:type zerophase:
:param taper_max_percentage:
:type taper_max_percentage:
:param taper_type:
:type taper_type:
:return: Tuple containing the changed waveform stream and the first trace of the stream
:rtype: (obspy.core.stream.Stream, obspy.core.trace.Trace)
"""
wfstream_copy = wfstream.copy()
trace_copy = wfstream[0].copy()
trace_copy.detrend(type=detrend_type)
trace_copy.filter(filter_type, freqmin=freqmin, freqmax=freqmax, zerophase=zerophase)
trace_copy.taper(max_percentage=taper_max_percentage, type=taper_type)
wfstream_copy[0].data = trace_copy.data
return wfstream_copy, trace_copy
# split components
zdat, ndat, edat = get_components_from_waveformstream(wfstream)
if not zdat:
print('No z-component found for station {}. STOP'.format(wfstream[0].stats.station))
return
if p_params['algoP'] == 'HOS' or p_params['algoP'] == 'ARZ' and zdat is not None:
msg = '##################################################\nautopickstation:' \
' Working on P onset of station {station}\nFiltering vertical ' \
'trace ...\n{data}'.format(station=wfstream[0].stats.station, data=str(zdat))
if verbose: print(msg)
z_copy, tr_filt = prepare_wfstream_component(zdat, freqmin=p_params['bpz1'][0], freqmax=p_params['bpz1'][1])
##############################################################
# check length of waveform and compare with cut times
# for global seismology: use tau-p method for estimating travel times (needs source and station coords.)
# if not given: sets Lc to infinity to use full stream
if p_params['use_taup'] is True:
Lc = np.inf
print('autopickstation: use_taup flag active.')
if not metadata[1]:
print('Warning: Could not use TauPy to estimate onsets as there are no metadata given.')
else:
station_id = wfstream[0].get_id()
parser = metadata[1]
station_coords = get_source_coords(parser, station_id)
if station_coords and origin:
source_origin = origin[0]
model = TauPyModel(p_params['taup_model'])
arrivals = model.get_travel_times_geo(
source_origin.depth,
source_origin.latitude,
source_origin.longitude,
station_coords['latitude'],
station_coords['longitude']
)
phases = {'P': [],
'S': []}
for arr in arrivals:
phases[identifyPhaseID(arr.phase.name)].append(arr)
# get first P and S onsets from arrivals list
arrP, estFirstP = min([(arr, arr.time) for arr in phases['P']], key=lambda t: t[1])
arrS, estFirstS = min([(arr, arr.time) for arr in phases['S']], key=lambda t: t[1])
print('autopick: estimated first arrivals for P: {} s, S:{} s after event'
' origin time using TauPy'.format(estFirstP, estFirstS))
# modifiy pstart and pstop relative to estimated first P arrival (relative to station time axis)
p_params['pstart'] += (source_origin.time + estFirstP) - zdat[0].stats.starttime
p_params['pstop']+= (source_origin.time + estFirstP) - zdat[0].stats.starttime
print('autopick: CF calculation times respectively:'
' pstart: {} s, pstop: {} s'.format(p_params['pstart'], p_params['pstop']))
elif not origin:
print('No source origins given!')
# make sure pstart and pstop are inside zdat[0]
pstart = max(p_params['pstart'], 0)
pstop = min(p_params['pstop'], len(zdat[0])*zdat[0].stats.delta)
if p_params['use_taup'] is False or origin:
Lc = p_params['pstop'] - p_params['pstart']
Lwf = zdat[0].stats.endtime - zdat[0].stats.starttime
if not Lwf > 0:
print('autopickstation: empty trace! Return!')
return
Ldiff = Lwf - abs(Lc)
if Ldiff < 0 or pstop <= pstart:
msg = 'autopickstation: Cutting times are too large for actual ' \
'waveform!\nUsing entire waveform instead!'
if verbose: print(msg)
pstart = 0
pstop = len(zdat[0].data) * zdat[0].stats.delta
cuttimes = [pstart, pstop]
cf1 = None
if p_params['algoP'] == 'HOS':
# calculate HOS-CF using subclass HOScf of class
# CharacteristicFunction
cf1 = HOScf(z_copy, cuttimes, p_params['tlta'], p_params['hosorder']) # instance of HOScf
elif p_params['algoP'] == 'ARZ':
# calculate ARZ-CF using subclass ARZcf of class
# CharcteristicFunction
cf1 = ARZcf(z_copy, cuttimes, p_params['tpred1z'], p_params['Parorder'], p_params['tdet1z'],
p_params['addnoise']) # instance of ARZcf
##############################################################
# calculate AIC-HOS-CF using subclass AICcf of class
# CharacteristicFunction
# class needs stream object => build it
assert isinstance(cf1, CharacteristicFunction), 'cf2 is not set ' \
'correctly: maybe the algorithm name ({algoP}) is ' \
'corrupted'.format(algoP=p_params['algoP'])
tr_aic = tr_filt.copy()
tr_aic.data = cf1.getCF()
z_copy[0].data = tr_aic.data
aiccf = AICcf(z_copy, cuttimes) # instance of AICcf
##############################################################
# get preliminary onset time from AIC-HOS-CF using subclass AICPicker
# of class AutoPicking
key = 'aicFig'
if fig_dict:
fig = fig_dict[key]
linecolor = fig_dict['plot_style']['linecolor']['rgba_mpl']
else:
fig = None
linecolor = 'k'
aicpick = AICPicker(aiccf, p_params['tsnrz'], p_params['pickwinP'], iplot, Tsmooth=p_params['aictsmooth'],
fig=fig, linecolor=linecolor)
# add pstart and pstop to aic plot
if fig:
for ax in fig.axes:
ax.vlines(pstart, ax.get_ylim()[0], ax.get_ylim()[1], color='c', linestyles='dashed', label='P start')
ax.vlines(pstop, ax.get_ylim()[0], ax.get_ylim()[1], color='c', linestyles='dashed', label='P stop')
ax.legend(loc=1)
##############################################################
if aicpick.getpick() is not None:
# check signal length to detect spuriously picked noise peaks
# use all available components to avoid skipping correct picks
# on vertical traces with weak P coda
z_copy[0].data = tr_filt.data
zne = z_copy
if len(ndat) == 0 or len(edat) == 0:
msg = 'One or more horizontal component(s) missing!\n' \
'Signal length only checked on vertical component!\n' \
'Decreasing minsiglengh from {0} to {1}' \
.format(signal_length_params['minsiglength'], signal_length_params['minsiglength'] / 2)
if verbose: print(msg)
key = 'slength'
if fig_dict:
fig = fig_dict[key]
linecolor = fig_dict['plot_style']['linecolor']['rgba_mpl']
else:
fig = None
linecolor = 'k'
Pflag = checksignallength(zne, aicpick.getpick(), p_params['tsnrz'],
signal_length_params['minsiglength'] / 2,
signal_length_params['noisefactor'], signal_length_params['minpercent'], iplot,
fig, linecolor)
else:
trH1_filt, _ = prepare_wfstream_component(edat, freqmin=s_params['bph1'][0], freqmax=s_params['bph1'][1])
trH2_filt, _ = prepare_wfstream_component(ndat, freqmin=s_params['bph1'][0], freqmax=s_params['bph1'][1])
zne += trH1_filt
zne += trH2_filt
if fig_dict:
fig = fig_dict['slength']
linecolor = fig_dict['plot_style']['linecolor']['rgba_mpl']
else:
fig = None
linecolor = 'k'
Pflag = checksignallength(zne, aicpick.getpick(), p_params['tsnrz'],
signal_length_params['minsiglength'],
signal_length_params['noisefactor'], signal_length_params['minpercent'], iplot,
fig, linecolor)
if Pflag == 1:
# check for spuriously picked S onset
# both horizontal traces needed
if len(ndat) == 0 or len(edat) == 0:
msg = 'One or more horizontal components missing!\n' \
'Skipping control function checkZ4S.'
if verbose: print(msg)
else:
if iplot > 1:
if fig_dict:
fig = fig_dict['checkZ4s']
linecolor = fig_dict['plot_style']['linecolor']['rgba_mpl']
else:
fig = None
linecolor = 'k'
Pflag = checkZ4S(zne, aicpick.getpick(), s_params['zfac'],
p_params['tsnrz'][2], iplot, fig, linecolor)
if Pflag == 0:
Pmarker = 'SinsteadP'
Pweight = 9
else:
Pmarker = 'shortsignallength'
Pweight = 9
##############################################################
# go on with processing if AIC onset passes quality control
slope = aicpick.getSlope()
if not slope:
slope = 0
if slope >= p_params['minAICPslope'] and aicpick.getSNR() >= p_params['minAICPSNR'] and Pflag == 1:
aicPflag = 1
msg = 'AIC P-pick passes quality control: Slope: {0} counts/s, ' \
'SNR: {1}\nGo on with refined picking ...\n' \
'autopickstation: re-filtering vertical trace ' \
'...'.format(aicpick.getSlope(), aicpick.getSNR())
if verbose: print(msg)
# re-filter waveform with larger bandpass
z_copy, tr_filt = prepare_wfstream_component(zdat, freqmin=p_params['bpz2'][0], freqmax=p_params['bpz2'][1])
#############################################################
# re-calculate CF from re-filtered trace in vicinity of initial
# onset
cuttimes2 = [round(max([aicpick.getpick() - p_params['Precalcwin'], 0])),
round(min([len(zdat[0].data) * zdat[0].stats.delta,
aicpick.getpick() + p_params['Precalcwin']]))]
cf2 = None
if p_params['algoP'] == 'HOS':
# calculate HOS-CF using subclass HOScf of class
# CharacteristicFunction
cf2 = HOScf(z_copy, cuttimes2, p_params['tlta'],
p_params['hosorder']) # instance of HOScf
elif p_params['algoP'] == 'ARZ':
# calculate ARZ-CF using subclass ARZcf of class
# CharcteristicFunction
cf2 = ARZcf(z_copy, cuttimes2, p_params['tpred1z'], p_params['Parorder'], p_params['tdet1z'],
p_params['addnoise']) # instance of ARZcf
##############################################################
# get refined onset time from CF2 using class Picker
assert isinstance(cf2, CharacteristicFunction), 'cf2 is not set ' \
'correctly: maybe the algorithm name ({algoP}) is ' \
'corrupted'.format(algoP=p_params['algoP'])
if fig_dict:
fig = fig_dict['refPpick']
linecolor = fig_dict['plot_style']['linecolor']['rgba_mpl']
else:
fig = None
linecolor = 'k'
refPpick = PragPicker(cf2, p_params['tsnrz'], p_params['pickwinP'], iplot, p_params['ausP'],
p_params['tsmoothP'], aicpick.getpick(), fig, linecolor)
mpickP = refPpick.getpick()
#############################################################
if mpickP is not None:
# quality assessment
# get earliest/latest possible pick and symmetrized uncertainty
if iplot:
if fig_dict:
fig = fig_dict['el_Ppick']
linecolor = fig_dict['plot_style']['linecolor']['rgba_mpl']
else:
fig = None
linecolor = 'k'
epickP, lpickP, Perror = earllatepicker(z_copy, p_params['nfacP'], p_params['tsnrz'],
mpickP, iplot, fig=fig,
linecolor=linecolor)
# get SNR
SNRP, SNRPdB, Pnoiselevel = getSNR(z_copy, p_params['tsnrz'], mpickP)
# weight P-onset using symmetric error
Pweight = get_quality_class(Perror, p_params['timeerrorsP'])
##############################################################
# get first motion of P onset
# certain quality required
if Pweight <= first_motion_params['minfmweight'] and SNRP >= first_motion_params['minFMSNR']:
if iplot:
if fig_dict:
fig = fig_dict['fm_picker']
linecolor = fig_dict['plot_style']['linecolor']['rgba_mpl']
else:
fig = None
FM = fmpicker(zdat, z_copy, first_motion_params['fmpickwin'], mpickP, iplot, fig, linecolor)
else:
FM = fmpicker(zdat, z_copy, first_motion_params['fmpickwin'], mpickP, iplot)
else:
FM = 'N'
msg = "autopickstation: P-weight: {0}, " \
"SNR: {1}, SNR[dB]: {2}, Polarity: {3}".format(Pweight, SNRP, SNRPdB, FM)
print(msg)
msg = 'autopickstation: Refined P-Pick: {} s | P-Error: {} s'.format(mpickP, Perror)
print(msg)
Sflag = 1
else:
msg = 'Bad initial (AIC) P-pick, skipping this onset!\n' \
'AIC-SNR={0}, AIC-Slope={1}counts/s\n' \
'(min. AIC-SNR={2}, ' \
'min. AIC-Slope={3}counts/s)'.format(aicpick.getSNR(),
aicpick.getSlope(),
p_params['minAICPSNR'],
p_params['minAICPslope'])
if verbose: print(msg)
Sflag = 0
else:
print('autopickstation: No vertical component data available!, '
'Skipping station!')
if ((len(edat) > 0 and len(ndat) == 0) or (len(ndat) > 0 and len(edat) == 0)) and Pweight < 4:
msg = 'Go on picking S onset ...\n' \
'##################################################\n' \
'Only one horizontal component available!\n' \
'ARH prediction requires at least 2 components!\n' \
'Copying existing horizontal component ...'
if verbose: print(msg)
# check which component is missing
if len(edat) == 0:
edat = ndat
else:
ndat = edat
pickSonset = (edat is not None and ndat is not None and len(edat) > 0 and len(
ndat) > 0 and Pweight < 4)
if pickSonset:
# determine time window for calculating CF after P onset
cuttimesh = [
round(max([mpickP + s_params['sstart'], 0])), # MP MP relative time axis
round(min([
mpickP + s_params['sstop'],
edat[0].stats.endtime-edat[0].stats.starttime,
ndat[0].stats.endtime-ndat[0].stats.starttime
]))
]
if not cuttimesh[1] >= cuttimesh[0]:
print('Cut window for horizontal phases too small! Will not pick S onsets.')
pickSonset = False
if pickSonset:
msg = 'Go on picking S onset ...\n' \
'##################################################\n' \
'Working on S onset of station {0}\nFiltering horizontal ' \
'traces ...'.format(edat[0].stats.station)
if verbose: print(msg)
if s_params['algoS'] == 'ARH':
# re-create stream object including both horizontal components
hdat = edat.copy()
hdat += ndat
h_copy = hdat.copy()
# filter and taper data
trH1_filt = hdat[0].copy()
trH2_filt = hdat[1].copy()
trH1_filt.detrend(type='demean')
trH2_filt.detrend(type='demean')
trH1_filt.filter('bandpass', freqmin=s_params['bph1'][0], freqmax=s_params['bph1'][1],
zerophase=False)
trH2_filt.filter('bandpass', freqmin=s_params['bph1'][0], freqmax=s_params['bph1'][1],
zerophase=False)
trH1_filt.taper(max_percentage=0.05, type='hann')
trH2_filt.taper(max_percentage=0.05, type='hann')
h_copy[0].data = trH1_filt.data
h_copy[1].data = trH2_filt.data
elif s_params['algoS'] == 'AR3':
# re-create stream object including all components
hdat = zdat.copy()
hdat += edat
hdat += ndat
h_copy = hdat.copy()
# filter and taper data
trH1_filt = hdat[0].copy()
trH2_filt = hdat[1].copy()
trH3_filt = hdat[2].copy()
trH1_filt.detrend(type='demean')
trH2_filt.detrend(type='demean')
trH3_filt.detrend(type='demean')
trH1_filt.filter('bandpass', freqmin=s_params['bph1'][0], freqmax=s_params['bph1'][1],
zerophase=False)
trH2_filt.filter('bandpass', freqmin=s_params['bph1'][0], freqmax=s_params['bph1'][1],
zerophase=False)
trH3_filt.filter('bandpass', freqmin=s_params['bph1'][0], freqmax=s_params['bph1'][1],
zerophase=False)
trH1_filt.taper(max_percentage=0.05, type='hann')
trH2_filt.taper(max_percentage=0.05, type='hann')
trH3_filt.taper(max_percentage=0.05, type='hann')
h_copy[0].data = trH1_filt.data
h_copy[1].data = trH2_filt.data
h_copy[2].data = trH3_filt.data
##############################################################
if s_params['algoS'] == 'ARH':
# calculate ARH-CF using subclass ARHcf of class
# CharcteristicFunction
arhcf1 = ARHcf(h_copy, cuttimesh, s_params['tpred1h'], s_params['Sarorder'], s_params['tdet1h'],
p_params['addnoise']) # instance of ARHcf
elif s_params['algoS'] == 'AR3':
# calculate ARH-CF using subclass AR3cf of class
# CharcteristicFunction
arhcf1 = AR3Ccf(h_copy, cuttimesh, s_params['tpred1h'], s_params['Sarorder'], s_params['tdet1h'],
p_params['addnoise']) # instance of ARHcf
##############################################################
# calculate AIC-ARH-CF using subclass AICcf of class
# CharacteristicFunction
# class needs stream object => build it
tr_arhaic = trH1_filt.copy()
tr_arhaic.data = arhcf1.getCF()
h_copy[0].data = tr_arhaic.data
# calculate ARH-AIC-CF
haiccf = AICcf(h_copy, cuttimesh) # instance of AICcf
##############################################################
# get prelimenary onset time from AIC-HOS-CF using subclass AICPicker
# of class AutoPicking
if fig_dict:
fig = fig_dict['aicARHfig']
linecolor = fig_dict['plot_style']['linecolor']['rgba_mpl']
else:
fig = None
linecolor = 'k'
aicarhpick = AICPicker(haiccf, s_params['tsnrh'], s_params['pickwinS'], iplot, None,
s_params['aictsmoothS'], fig=fig, linecolor=linecolor)
###############################################################
# go on with processing if AIC onset passes quality control
slope = aicarhpick.getSlope()
if not slope:
slope = 0
if (slope >= s_params['minAICSslope'] and
aicarhpick.getSNR() >= s_params['minAICSSNR'] and aicarhpick.getpick() is not None):
aicSflag = 1
msg = 'AIC S-pick passes quality control: Slope: {0} counts/s, ' \
'SNR: {1}\nGo on with refined picking ...\n' \
'autopickstation: re-filtering horizontal traces ' \
'...'.format(aicarhpick.getSlope(), aicarhpick.getSNR())
if verbose: print(msg)
# re-calculate CF from re-filtered trace in vicinity of initial
# onset
cuttimesh2 = [round(aicarhpick.getpick() - s_params['Srecalcwin']),
round(aicarhpick.getpick() + s_params['Srecalcwin'])]
# re-filter waveform with larger bandpass
h_copy = hdat.copy()
# filter and taper data
if s_params['algoS']== 'ARH':
trH1_filt = hdat[0].copy()
trH2_filt = hdat[1].copy()
trH1_filt.detrend(type='demean')
trH2_filt.detrend(type='demean')
trH1_filt.filter('bandpass', freqmin=s_params['bph2'][0], freqmax=s_params['bph2'][1],
zerophase=False)
trH2_filt.filter('bandpass', freqmin=s_params['bph2'][0], freqmax=s_params['bph2'][1],
zerophase=False)
trH1_filt.taper(max_percentage=0.05, type='hann')
trH2_filt.taper(max_percentage=0.05, type='hann')
h_copy[0].data = trH1_filt.data
h_copy[1].data = trH2_filt.data
#############################################################
arhcf2 = ARHcf(h_copy, cuttimesh2, s_params['tpred2h'], s_params['Sarorder'], s_params['tdet2h'],
p_params['addnoise']) # instance of ARHcf
elif s_params['algoS'] == 'AR3':
trH1_filt = hdat[0].copy()
trH2_filt = hdat[1].copy()
trH3_filt = hdat[2].copy()
trH1_filt.detrend(type='demean')
trH2_filt.detrend(type='demean')
trH3_filt.detrend(type='demean')
trH1_filt.filter('bandpass', freqmin=s_params['bph2'][0], freqmax=s_params['bph2'][1],
zerophase=False)
trH2_filt.filter('bandpass', freqmin=s_params['bph2'][0], freqmax=s_params['bph2'][1],
zerophase=False)
trH3_filt.filter('bandpass', freqmin=s_params['bph2'][0], freqmax=s_params['bph2'][1],
zerophase=False)
trH1_filt.taper(max_percentage=0.05, type='hann')
trH2_filt.taper(max_percentage=0.05, type='hann')
trH3_filt.taper(max_percentage=0.05, type='hann')
h_copy[0].data = trH1_filt.data
h_copy[1].data = trH2_filt.data
h_copy[2].data = trH3_filt.data
#############################################################
arhcf2 = AR3Ccf(h_copy, cuttimesh2, s_params['tpred2h'], s_params['Sarorder'], s_params['tdet2h'],
p_params['addnoise']) # instance of ARHcf
# get refined onset time from CF2 using class Picker
if fig_dict:
fig = fig_dict['refSpick']
linecolor = fig_dict['plot_style']['linecolor']['rgba_mpl']
else:
fig = None
linecolor = 'k'
refSpick = PragPicker(arhcf2, s_params['tsnrh'], s_params['pickwinS'], iplot, s_params['ausS'],
s_params['tsmoothS'], aicarhpick.getpick(), fig, linecolor)
mpickS = refSpick.getpick()
#############################################################
if mpickS is not None:
# quality assessment
# get earliest/latest possible pick and symmetrized uncertainty
h_copy[0].data = trH1_filt.data
if iplot:
if fig_dict:
fig = fig_dict['el_S1pick']
linecolor = fig_dict['plot_style']['linecolor']['rgba_mpl']
else:
fig = None
linecolor = 'k'
epickS1, lpickS1, Serror1 = earllatepicker(h_copy, s_params['nfacS'], s_params['tsnrh'], mpickS,
iplot, fig=fig, linecolor=linecolor)
else:
epickS1, lpickS1, Serror1 = earllatepicker(h_copy, s_params['nfacS'], s_params['tsnrh'], mpickS, iplot)
h_copy[0].data = trH2_filt.data
if iplot:
if fig_dict:
fig = fig_dict['el_S2pick']
linecolor = fig_dict['plot_style']['linecolor']['rgba_mpl']
else:
fig = None
linecolor = ''
epickS2, lpickS2, Serror2 = earllatepicker(h_copy, s_params['nfacS'], s_params['tsnrh'], mpickS,
iplot, fig=fig, linecolor=linecolor)
else:
epickS2, lpickS2, Serror2 = earllatepicker(h_copy, s_params['nfacS'], s_params['tsnrh'], mpickS, iplot)
if epickS1 is not None and epickS2 is not None:
if s_params['algoS'] == 'ARH':
# get earliest pick of both earliest possible picks
epick = [epickS1, epickS2]
lpick = [lpickS1, lpickS2]
pickerr = [Serror1, Serror2]
if epickS1 is None and epickS2 is not None:
ipick = 1
elif epickS1 is not None and epickS2 is None:
ipick = 0
elif epickS1 is not None and epickS2 is not None:
ipick = np.argmin([epickS1, epickS2])
elif s_params['algoS'] == 'AR3':
[epickS3, lpickS3, Serror3] = earllatepicker(h_copy, s_params['nfacS'], s_params['tsnrh'],
mpickS, iplot)
# get earliest pick of all three picks
epick = [epickS1, epickS2, epickS3]
lpick = [lpickS1, lpickS2, lpickS3]
pickerr = [Serror1, Serror2, Serror3]
if epickS1 is None and epickS2 is not None \
and epickS3 is not None:
ipick = np.argmin([epickS2, epickS3])
elif epickS1 is not None and epickS2 is None \
and epickS3 is not None:
ipick = np.argmin([epickS2, epickS3])
elif epickS1 is not None and epickS2 is not None \
and epickS3 is None:
ipick = np.argmin([epickS1, epickS2])
elif epickS1 is not None and epickS2 is not None \
and epickS3 is not None:
ipick = np.argmin([epickS1, epickS2, epickS3])
epickS = epick[ipick]
lpickS = lpick[ipick]
Serror = pickerr[ipick]
msg = 'autopickstation: Refined S-Pick: {} s | S-Error: {} s'.format(mpickS, Serror)
print(msg)
# get SNR
[SNRS, SNRSdB, Snoiselevel] = getSNR(h_copy, s_params['tsnrh'], mpickS)
# weight S-onset using symmetric error
if Serror <= s_params['timeerrorsS'][0]:
Sweight = 0
elif s_params['timeerrorsS'][0] < Serror <= s_params['timeerrorsS'][1]:
Sweight = 1
elif Perror > s_params['timeerrorsS'][1] and Serror <= s_params['timeerrorsS'][2]:
Sweight = 2
elif s_params['timeerrorsS'][2] < Serror <= s_params['timeerrorsS'][3]:
Sweight = 3
elif Serror > s_params['timeerrorsS'][3]:
Sweight = 4
print('autopickstation: S-weight: {0}, SNR: {1}, '
'SNR[dB]: {2}\n'
'##################################################'
''.format(Sweight, SNRS, SNRSdB))
################################################################
# get Wood-Anderson peak-to-peak amplitude
# initialize Data object
# re-create stream object including both horizontal components
hdat = edat.copy()
hdat += ndat
else:
msg = 'Bad initial (AIC) S-pick, skipping this onset!\n' \
'AIC-SNR={0}, AIC-Slope={1}counts/s\n' \
'(min. AIC-SNR={2}, ' \
'min. AIC-Slope={3}counts/s)\n' \
'##################################################' \
''.format(aicarhpick.getSNR(), aicarhpick.getSlope(), s_params['minAICSSNR'], s_params['minAICSslope'])
if verbose: print(msg)
############################################################
# get Wood-Anderson peak-to-peak amplitude
# initialize Data object
# re-create stream object including both horizontal components
hdat = edat.copy()
hdat += ndat
else:
print('autopickstation: No horizontal component data available or '
'bad P onset, skipping S picking!')
##############################################################
try:
iplot = int(iplot)
except ValueError:
if iplot is True or iplot == 'True':
iplot = 2
else:
iplot = 0
if iplot > 0:
# plot vertical trace
if fig_dict is None or fig_dict == 'None':
fig = plt.figure()
plt_flag = 1
linecolor = 'k'
else:
fig = fig_dict['mainFig']
linecolor = fig_dict['plot_style']['linecolor']['rgba_mpl']
fig._tight = True
ax1 = fig.add_subplot(311)
tdata = np.arange(0, zdat[0].stats.npts / tr_filt.stats.sampling_rate,
tr_filt.stats.delta)
# check equal length of arrays, sometimes they are different!?
wfldiff = len(tr_filt.data) - len(tdata)
if wfldiff < 0:
tdata = tdata[0:len(tdata) - abs(wfldiff)]
ax1.plot(tdata, tr_filt.data / max(tr_filt.data), color=linecolor, linewidth=0.7, label='Data')
if Pweight < 4:
ax1.plot(cf1.getTimeArray(), cf1.getCF() / max(cf1.getCF()),
'b', label='CF1')
if aicPflag == 1:
ax1.plot(cf2.getTimeArray(),
cf2.getCF() / max(cf2.getCF()), 'm', label='CF2')
ax1.plot([aicpick.getpick(), aicpick.getpick()], [-1, 1],
'r', label='Initial P Onset')
ax1.plot([aicpick.getpick() - 0.5, aicpick.getpick() + 0.5],
[1, 1], 'r')
ax1.plot([aicpick.getpick() - 0.5, aicpick.getpick() + 0.5],
[-1, -1], 'r')
ax1.plot([refPpick.getpick(), refPpick.getpick()],
[-1.3, 1.3], 'r', linewidth=2, label='Final P Pick')
ax1.plot([refPpick.getpick() - 0.5, refPpick.getpick() + 0.5],
[1.3, 1.3], 'r', linewidth=2)
ax1.plot([refPpick.getpick() - 0.5, refPpick.getpick() + 0.5],
[-1.3, -1.3], 'r', linewidth=2)
ax1.plot([lpickP, lpickP], [-1.1, 1.1], 'r--', label='lpp')
ax1.plot([epickP, epickP], [-1.1, 1.1], 'r--', label='epp')
ax1.set_title('%s, %s, P Weight=%d, SNR=%7.2f, SNR[dB]=%7.2f '
'Polarity: %s' % (tr_filt.stats.station,
tr_filt.stats.channel,
Pweight,
SNRP,
SNRPdB,
FM))
else:
ax1.set_title('%s, P Weight=%d, SNR=None, '
'SNRdB=None' % (tr_filt.stats.channel, Pweight))
else:
ax1.set_title('%s, %s, P Weight=%d' % (tr_filt.stats.station,
tr_filt.stats.channel,
Pweight))
ax1.legend(loc=1)
ax1.set_yticks([])
ax1.set_ylim([-1.5, 1.5])
ax1.set_ylabel('Normalized Counts')
# fig.suptitle(tr_filt.stats.starttime)
try:
len(edat[0])
except:
edat = ndat
try:
len(ndat[0])
except:
ndat = edat
if len(edat[0]) > 1 and len(ndat[0]) > 1 and Sflag == 1:
# plot horizontal traces
ax2 = fig.add_subplot(3, 1, 2, sharex=ax1)
th1data = np.arange(0,
trH1_filt.stats.npts /
trH1_filt.stats.sampling_rate,
trH1_filt.stats.delta)
# check equal length of arrays, sometimes they are different!?
wfldiff = len(trH1_filt.data) - len(th1data)
if wfldiff < 0:
th1data = th1data[0:len(th1data) - abs(wfldiff)]
ax2.plot(th1data, trH1_filt.data / max(trH1_filt.data), color=linecolor, linewidth=0.7, label='Data')
if Pweight < 4:
ax2.plot(arhcf1.getTimeArray(),
arhcf1.getCF() / max(arhcf1.getCF()), 'b', label='CF1')
if aicSflag == 1 and Sweight < 4:
ax2.plot(arhcf2.getTimeArray(),
arhcf2.getCF() / max(arhcf2.getCF()), 'm', label='CF2')
ax2.plot(
[aicarhpick.getpick(), aicarhpick.getpick()],
[-1, 1], 'g', label='Initial S Onset')
ax2.plot(
[aicarhpick.getpick() - 0.5,
aicarhpick.getpick() + 0.5],
[1, 1], 'g')
ax2.plot(
[aicarhpick.getpick() - 0.5,
aicarhpick.getpick() + 0.5],
[-1, -1], 'g')
ax2.plot([refSpick.getpick(), refSpick.getpick()],
[-1.3, 1.3], 'g', linewidth=2, label='Final S Pick')
ax2.plot(
[refSpick.getpick() - 0.5, refSpick.getpick() + 0.5],
[1.3, 1.3], 'g', linewidth=2)
ax2.plot(
[refSpick.getpick() - 0.5, refSpick.getpick() + 0.5],
[-1.3, -1.3], 'g', linewidth=2)
ax2.plot([lpickS, lpickS], [-1.1, 1.1], 'g--', label='lpp')
ax2.plot([epickS, epickS], [-1.1, 1.1], 'g--', label='epp')
ax2.set_title('%s, S Weight=%d, SNR=%7.2f, SNR[dB]=%7.2f' % (
trH1_filt.stats.channel,
Sweight, SNRS, SNRSdB))
else:
ax2.set_title('%s, S Weight=%d, SNR=None, SNRdB=None' % (
trH1_filt.stats.channel, Sweight))
ax2.legend(loc=1)
ax2.set_yticks([])
ax2.set_ylim([-1.5, 1.5])
ax2.set_ylabel('Normalized Counts')
# fig.suptitle(trH1_filt.stats.starttime)
ax3 = fig.add_subplot(3, 1, 3, sharex=ax1)
th2data = np.arange(0,
trH2_filt.stats.npts /
trH2_filt.stats.sampling_rate,
trH2_filt.stats.delta)
# check equal length of arrays, sometimes they are different!?
wfldiff = len(trH2_filt.data) - len(th2data)
if wfldiff < 0:
th2data = th2data[0:len(th2data) - abs(wfldiff)]
ax3.plot(th2data, trH2_filt.data / max(trH2_filt.data), color=linecolor, linewidth=0.7, label='Data')
if Pweight < 4:
p22, = ax3.plot(arhcf1.getTimeArray(),
arhcf1.getCF() / max(arhcf1.getCF()), 'b', label='CF1')
if aicSflag == 1:
ax3.plot(arhcf2.getTimeArray(),
arhcf2.getCF() / max(arhcf2.getCF()), 'm', label='CF2')
ax3.plot(
[aicarhpick.getpick(), aicarhpick.getpick()],
[-1, 1], 'g', label='Initial S Onset')
ax3.plot(
[aicarhpick.getpick() - 0.5,
aicarhpick.getpick() + 0.5],
[1, 1], 'g')
ax3.plot(
[aicarhpick.getpick() - 0.5,
aicarhpick.getpick() + 0.5],
[-1, -1], 'g')
ax3.plot([refSpick.getpick(), refSpick.getpick()],
[-1.3, 1.3], 'g', linewidth=2, label='Final S Pick')
ax3.plot(
[refSpick.getpick() - 0.5, refSpick.getpick() + 0.5],
[1.3, 1.3], 'g', linewidth=2)
ax3.plot(
[refSpick.getpick() - 0.5, refSpick.getpick() + 0.5],
[-1.3, -1.3], 'g', linewidth=2)
ax3.plot([lpickS, lpickS], [-1.1, 1.1], 'g--', label='lpp')
ax3.plot([epickS, epickS], [-1.1, 1.1], 'g--', label='epp')
ax3.legend(loc=1)
ax3.set_yticks([])
ax3.set_ylim([-1.5, 1.5])
ax3.set_xlabel('Time [s] after %s' % tr_filt.stats.starttime)
ax3.set_ylabel('Normalized Counts')
ax3.set_title(trH2_filt.stats.channel)
if plt_flag == 1:
fig.show()
try:
input()
except SyntaxError:
pass
plt.close(fig)
##########################################################################
# calculate "real" onset times
if lpickP is not None and lpickP == mpickP:
lpickP += zdat[0].stats.delta
if epickP is not None and epickP == mpickP:
epickP -= zdat[0].stats.delta
if mpickP is not None and epickP is not None and lpickP is not None:
lpickP = zdat[0].stats.starttime + lpickP
epickP = zdat[0].stats.starttime + epickP
mpickP = zdat[0].stats.starttime + mpickP
else:
# dummy values (start of seismic trace) in order to derive
# theoretical onset times for iteratve picking
lpickP = zdat[0].stats.starttime + p_params['timeerrorsP'][3]
epickP = zdat[0].stats.starttime - p_params['timeerrorsP'][3]
mpickP = zdat[0].stats.starttime
if edat:
hdat = edat[0]
elif ndat:
hdat = ndat[0]
else:
return
if lpickS is not None and lpickS == mpickS:
lpickS += hdat.stats.delta
if epickS is not None and epickS == mpickS:
epickS -= hdat.stats.delta
if mpickS is not None and epickS is not None and lpickS is not None:
lpickS = hdat.stats.starttime + lpickS
epickS = hdat.stats.starttime + epickS
mpickS = hdat.stats.starttime + mpickS
else:
# dummy values (start of seismic trace) in order to derive
# theoretical onset times for iteratve picking
lpickS = hdat.stats.starttime + s_params['timeerrorsS'][3]
epickS = hdat.stats.starttime - s_params['timeerrorsS'][3]
mpickS = hdat.stats.starttime
# create dictionary
# for P phase
ccode = zdat[0].stats.channel
ncode = zdat[0].stats.network
ppick = dict(channel=ccode, network=ncode, lpp=lpickP, epp=epickP, mpp=mpickP, spe=Perror, snr=SNRP,
snrdb=SNRPdB, weight=Pweight, fm=FM, w0=None, fc=None, Mo=None,
Mw=None, picker=picker, marked=Pmarker)
# add S phase
ccode = hdat.stats.channel
ncode = hdat.stats.network
spick = dict(channel=ccode, network=ncode, lpp=lpickS, epp=epickS, mpp=mpickS, spe=Serror, snr=SNRS,
snrdb=SNRSdB, weight=Sweight, fm=None, picker=picker, Ao=Ao)
# merge picks into returning dictionary
picks = dict(P=ppick, S=spick, station=zdat[0].stats.station)
return picks
def iteratepicker(wf, NLLocfile, picks, badpicks, pickparameter, fig_dict=None):
"""
Repicking of bad onsets. Uses theoretical onset times from NLLoc-location file.
:param wf: waveform, obspy stream object
:type wf: ~obspy.core.stream.Stream
:param NLLocfile: path/name of NLLoc-location file
:type NLLocfile: str
:param picks: dictionary of available onset times
:type picks: dict
:param badpicks: picks to be repicked
:type badpicks:
:param pickparameter: picking parameters from autoPyLoT-input file
:type pickparameter: pylot.core.io.inputs.PylotParameter
:param fig_dict: dictionary containing Matplotlib figures used for plotting results
:type fig_dict: dict
:return: dictionary containing iterative picks
:rtype: dict
"""
msg = '##################################################\n' \
'autoPyLoT: Found {0} bad onsets at station(s) {1}, ' \
'starting re-picking them ...'.format(len(badpicks), badpicks)
print(msg)
newpicks = {}
for i in range(0, len(badpicks)):
if len(badpicks[i][0]) > 4:
Ppattern = '%s ? ? ? P' % badpicks[i][0]
elif len(badpicks[i][0]) == 4:
Ppattern = '%s ? ? ? P' % badpicks[i][0]
elif len(badpicks[i][0]) < 4:
Ppattern = '%s ? ? ? P' % badpicks[i][0]
nllocline = getPatternLine(NLLocfile, Ppattern)
res = nllocline.split(None)[16]
# get theoretical P-onset time from residuum
badpicks[i][1] = picks[badpicks[i][0]]['P']['mpp'] - float(res)
# get corresponding waveform stream
msg = '##################################################\n' \
'iteratepicker: Re-picking station {0}'.format(badpicks[i][0])
print(msg)
wf2pick = wf.select(station=badpicks[i][0])
# modify some picking parameters
pstart_old = pickparameter.get('pstart')
pstop_old = pickparameter.get('pstop')
sstop_old = pickparameter.get('sstop')
pickwinP_old = pickparameter.get('pickwinP')
Precalcwin_old = pickparameter.get('Precalcwin')
noisefactor_old = pickparameter.get('noisefactor')
zfac_old = pickparameter.get('zfac')
twindows = pickparameter.get('tsnrz')
tsafety = twindows[1]
pstart = max([0, badpicks[i][1] - wf2pick[0].stats.starttime - pickparameter.get('tlta')])
if abs(float(res)) <= tsafety / 2 or pstart == 0:
print("iteratepicker: Small residuum, leave parameters unchanged for this phase!")
else:
pickparameter.setParam(pstart=pstart)
pickparameter.setParam(pstop=pickparameter.get('pstart') + (pickparameter.get('Precalcwin')))
pickparameter.setParam(sstop=pickparameter.get('sstop') / 2)
pickparameter.setParam(pickwinP=pickparameter.get('pickwinP') / 2)
pickparameter.setParam(Precalcwin=pickparameter.get('Precalcwin') / 2)
pickparameter.setParam(noisefactor=1.0)
pickparameter.setParam(zfac=1.0)
print(
"iteratepicker: The following picking parameters have been modified for iterative picking:")
print(
"pstart: %fs => %fs" % (pstart_old, pickparameter.get('pstart')))
print(
"pstop: %fs => %fs" % (pstop_old, pickparameter.get('pstop')))
print(
"sstop: %fs => %fs" % (sstop_old, pickparameter.get('sstop')))
print("pickwinP: %fs => %fs" % (
pickwinP_old, pickparameter.get('pickwinP')))
print("Precalcwin: %fs => %fs" % (
Precalcwin_old, pickparameter.get('Precalcwin')))
print("noisefactor: %f => %f" % (
noisefactor_old, pickparameter.get('noisefactor')))
print("zfac: %f => %f" % (zfac_old, pickparameter.get('zfac')))
# repick station
newpicks = autopickstation(wf2pick, pickparameter, fig_dict=fig_dict)
# replace old dictionary with new one
picks[badpicks[i][0]] = newpicks
# reset temporary change of picking parameters
print("iteratepicker: Resetting picking parameters ...")
pickparameter.setParam(pstart=pstart_old)
pickparameter.setParam(pstop=pstop_old)
pickparameter.setParam(sstop=sstop_old)
pickparameter.setParam(pickwinP=pickwinP_old)
pickparameter.setParam(Precalcwin=Precalcwin_old)
pickparameter.setParam(noisefactor=noisefactor_old)
pickparameter.setParam(zfac=zfac_old)
return picks