Utility Functions (seislib.utils
)
This module provides support for several useful functions for processing and analysing seismic data. These mainly build on ObsPy, and include:
Vectorized calculation of great-circle distances, azimuths, and back azimuths
Slicing / resampling of pairs of seismograms so as to adapt them to a common time window / sampling rate
Pre-processing operations such as zero-padding and bandpass filtering
Rotation of horizontal-component recordings to an arbitrary amount of degrees. SeisLib’s implementation also takes into account not uncommon ValueErrors raised by ObsPy due to the presence of slightly (sub-sample) differences in the time spanned by the two traces
Interpolation of scattered data on equal-area grids
Calculation of the Pearson correlation coefficient between physical parameters defined on equal-area grids
- seislib.utils._utils.adapt_sampling_rate(st1, st2)
If the input streams (or traces) have different sampling rates, the one characterized by the largest sampling rate is downsampled to the sampling rate of the other stream (or trace).
The downsampling is carried out via the
resample()
, which modifies the input streams in place.- Parameters
- st1, st2obspy.Stream, obspy.Trace
- Returns
- st1, st2obspy.Stream, obspy.Trace
Obspy stream or trace depending on the input. The input is permanently modified
- seislib.utils._utils.adapt_timespan(st1, st2)
Slices all traces from the two input streams to the overlapping timerange. Then returns a copy of the sliced streams.
- Parameters
- st1, st2obspy.Stream, obspy.Trace
- Returns
- st1, st2obspy.Stream, obspy.Trace
Obspy stream or trace depending on the input. The original input is not permanently modified (a copy is returned)
- Raises
- TimeSpanException
If no overlap is available
Notes
The maximum precision achieved by this function is governed by the samling rate. If sub-sample precision is required, consider using
adapt_timespan_interpolate()
- seislib.utils._utils.adapt_timespan_interpolate(st1, st2, min_overlap=0)
Slices all traces from the two input streams to the overlapping timerange. Then returns a copy of the sliced streams. If the starttime of the sliced traces do not fit exactly (because of sub-sample time shifts), the traces are interpolated to remove the time shift.
- Parameters
- st1, st2obspy.Stream, obspy.Trace
- Returns
- st1_out, st2_outobspy.Stream, obspy.Trace
Obspy stream or trace depending on the input. The original input is not permanently modified (a copy is returned)
- Raises
- TimeSpanException
If no overlap is available
Notes
Interpolation can require a relatively long time depending on the size of the sliced stream. If speed is preferred to (sub-sample) precision, consider using
adapt_timespan()
- seislib.utils._utils.azimuth_backazimuth(lat1, lon1, lat2, lon2)
Calculates the azimuth and backazimuth (in degrees) between coordinate points (in degrees). This function calls directly the obspy gps2dist_azimuth, it only extends its functionality through the numpy.vectorize decorator
- Parameters
- lat1, lon1, lat2, lon2float or array-like of shape (n,)
Coordinates of the points on the Earth’s surface, in degrees.
- Returns
- tuple of shape (2,) or ndarray of shape (n, 2)
The columns consist of azimuth azimuth and backazimuth
- seislib.utils._utils.bandpass_gaussian(data, dt, period, alpha)
Gaussian filter of real-valued data carried out in the frequency domain
The bandpass filter is carried out with a Gaussian filter centered at period, whose width is controlled by alpha:
exp(-alpha * ((f-f0)/f0)**2)
where f is frequency and f0 = 1 / period.
- Parameters
- datandarray of shape (n,)
Real-valued data to be filtered
- dtfloat
Time sampling interval of the data
- periodfloat
Central period, around which the (tight) bandapass filter is carried out
- alphafloat
Parameter that controls the width of the Gaussian filter
- Returns
- numpy.ndarray of shape (n,)
Filtered data
- seislib.utils._utils.gc_distance(lat1, lon1, lat2, lon2)
Calculates the great circle distance (in m) between coordinate points (in degrees). This function calls directly the obspy gps2dist_azimuth, it only extends its functionality through the numpy.vectorize decorator.
- Parameters
- lat1, lon1, lat2, lon2float or array-like of shape (n,)
Coordinates of the points on the Earth’s surface, in degrees.
- Returns
- Great-circle distance (in m)
If the input is an array (or list) of coordinates, an array of distances is returned
- seislib.utils._utils.load_pickle(path)
Loads a .pickle file
- Parameters
- pathstr
Absolute path to the file
- Returns
- Object contained in the .pickle file
- seislib.utils._utils.next_power_of_2(x)
Closest power of two larger or equal to x
- Parameters
- xint
- Returns
- int
- seislib.utils._utils.pearson_corrcoef(v1, v2)
Pearson coerrelation coefficient between two vectors
- Parameters
- v1, v2lists or ndarrays (n,)
- Returns
- corrcoefffloat
Pearson correlation coefficient
- pvaluefloat
- Raises
- ValueError
If v1 and v2 have different shapes
Notes
This function calls scipy.stats.pearsonr, bu handles the presence of nan values.
- seislib.utils._utils.remove_file(file)
Removes a file from disk, handling eventual exceptions
- Parameters
- filestr
Absolute path to the file to be removed
- seislib.utils._utils.resample(x, fs)
If the input streams (or traces) have different sampling rates, the one characterized by the largest sampling rate is downsampled to the sampling rate of the other stream (or trace).
- Parameters
- st1, st2obspy.Stream, obspy.Trace
- Returns
- st1, st2obspy.Stream, obspy.Trace
Obspy stream or trace depending on the input. The input is permanently modified
- seislib.utils._utils.rotate(r, t, dphi)
Rotation of radial and transverse component of the seismogram by a specified angle, following the obspy signs convention.
- Parameters
- r, tnumpy.ndarray
Radial (r) and transverse (t) components
- dphifloat
Angle in degrees
- Returns
- rnew, tnewnumpy.ndarray
Rotated components
- seislib.utils._utils.rotate_stream(st, **kwargs)
The method calls the obspy.Stream.rotate method, which sometimes runs into errors if differences are present among the starttimes and/or endtimes of the traces constituting the stream. These are prevented by slicing the stream to a common time window and (if necessary) interpolating it so as to avoid sub-sample differences.
- Parameters
- stobspy.Stream
- **kwargs
Optional arguments passed to obspy.Stream.rotate
- Returns
- stobspy.Stream
Rotated copy of the input Stream
- seislib.utils._utils.running_mean(x, N)
Moving average
- Parameters
- xndarray of shape (m,)
Data vector
- Nint
Controls the extent of the smoothing (larger values correspond to larger smoothing)
- Returns
- runmeanndarray of shape (m,)
Smoothed input
Notes
This is a simple implementation of a moving average. More sofisticated functions can be found, e.g., in scipy.signal.savgol_filter or scipy.ndimage.filters.uniform_filter1d
- seislib.utils._utils.save_pickle(file, obj)
Saves an object to a .pickle file
- Parameters
- filestr
Absolute path to the resulting file
- objpython object
Object to be saved (see documentation on the pickle module to know more on which Python objects can be stored into .pickle files)
- seislib.utils._utils.scatter_to_mesh(lats, lons, c, mesh, method='linear')
Translates scattered data into a seislib mesh
- Parameters
- lats, lonsndarray (n,)
Coordinates of the scattered data
- cndarray (n,)
Values of the scattered data
- meshndarray (m, 4)
mesh in SeisLib’s format, where the four columns correspond to the boundaries of each pixel, i.e., lat1, lat2, lon1, lon2
- method{‘linear’, ‘nearest’, ‘cubic’}
Interpolation method. Supported: ‘linear’ (default), ‘nearest’, and ‘cubic’. The three methods call LinearNDInterpolator, NearestNDInterpolator, and CloughTocher2DInterpolator of the scipy.interpolate module, respectively.
- Returns
- 1-D ndarray containing the c values interpolated on mesh
- Raises
- NotImplementedError
If method is not supported
- seislib.utils._utils.zeropad(tr, starttime, endtime)
Zeropads an obspy.Trace so as to cover the time window specified by starttime’and endtime
- Parameters
- trobspy.Trace
- starttime, endtimeobspy.UTCDateTime
- Returns
- traceobspy.Trace
Zeropadded copy of the input trace.