obspy.signal.spectral_estimation.PPSD

class PPSD(stats, metadata, skip_on_gaps=False, db_bins=(-200, -50, 1.0), ppsd_length=3600.0, overlap=0.5, special_handling=None, period_smoothing_width_octaves=1.0, period_step_octaves=0.125, period_limits=None, **kwargs)[source]

Bases: builtins.object

Class to compile probabilistic power spectral densities for one combination of network/station/location/channel/sampling_rate.

Calculations are based on the routine used by [McNamara2004]. For information on New High/Low Noise Model see [Peterson1993].

Basic Usage

>>> from obspy import read
>>> from obspy.signal import PPSD
>>> st = read()
>>> tr = st.select(channel="EHZ")[0]
>>> paz = {'gain': 60077000.0,
...        'poles': [-0.037004+0.037016j, -0.037004-0.037016j,
...                  -251.33+0j, -131.04-467.29j, -131.04+467.29j],
...        'sensitivity': 2516778400.0,
...        'zeros': [0j, 0j]}
>>> ppsd = PPSD(tr.stats, paz)
>>> print(ppsd.id)
BW.RJOB..EHZ
>>> print(ppsd.times_processed)
[]

Now we could add data to the probabilistic psd (all processing like demeaning, tapering and so on is done internally) and plot it like ...

>>> ppsd.add(st) 
>>> print(ppsd.times) 
>>> ppsd.plot() 

... but the example stream is too short and does not contain enough data.

Note

For a real world example see the ObsPy Tutorial.

Saving and Loading

The PPSD object supports saving to a numpy npz compressed binary file:

>>> ppsd.save_npz("myfile.npz") 

The saved PPSD can then be loaded again using the static method load_npz(), e.g. to plot the results again. If additional data is to be processed (note that another option is to combine multiple npz files using add_npz()), metadata must be provided again, since they are not stored in the numpy npz file:

>>> ppsd = PPSD.load_npz("myfile.npz")  

Note

When using metadata from an Inventory, a Parser instance or from a RESP file, information on metadata will be correctly picked for the respective starttime of the data trace. This means that instrument changes are correctly taken into account during response removal. This is obviously not the case for a static PAZ dictionary!

Attributes

NPZ_STORE_KEYS list() -> new empty list
NPZ_STORE_KEYS_ARRAY_TYPES list() -> new empty list
NPZ_STORE_KEYS_LIST_TYPES list() -> new empty list
NPZ_STORE_KEYS_SIMPLE_TYPES list() -> new empty list
NPZ_STORE_KEYS_VERSION_NUMBERS list() -> new empty list
__dict__
__doc__ str(object=’‘) -> str
__module__ str(object=’‘) -> str
__weakref__ list of weak references to the object (if defined)
channel
current_histogram
current_histogram_count
current_histogram_cumulative
current_times_used
db_bin_centers
db_bin_edges
delta
len Trace length for one psd segment.
location
merge_method
network
nfft
nlap
period_bin_centers Return centers of period bins (geometric mean of left and right edge of
period_bin_left_edges Returns left edges of period bins (same length as number of bins).
period_bin_right_edges Returns right edges of period bins (same length as number of bins).
period_xedges Returns edges of period histogram bins (one element longer than number of bins).
psd_frequencies
psd_periods
psd_values Returns all individual smoothed psd arrays as a list.
station
step Time step between start times of adjacent psd segments in seconds
times_data
times_gaps
times_processed

Public Methods

add Process all traces with compatible information and add their spectral estimates to the histogram containing the probabilistic psd.
add_npz Add previously computed PPSD results to current PPSD instance.
calculate_histogram Calculate and set current 2D histogram stack, optionally with start-
extract_psd_values Extract PSD values for given period in seconds.
get_mean Returns periods and mean psd values (i.e.
get_mode Returns periods and mode psd values (i.e.
get_percentile Returns periods and approximate psd values for given percentile value.
load_npz Load previously computed PPSD results.
plot Plot the 2D histogram of the current PPSD.
plot_coverage Plot the data coverage of the histogram of the current PPSD.
plot_spectrogram Plot the temporal evolution of the PSD in a spectrogram-like plot.
plot_temporal Plot the evolution of PSD value of one (or more) period bins over time.
save_npz Saves the PPSD as a compressed numpy binary (npz format).

Private Methods

Warning

Private methods are mainly for internal/developer use and their API might change without notice.

_PPSD__check_histogram
_PPSD__check_time_present Checks if the given UTCDateTime is already part of the current PPSD
_PPSD__insert_data_times Gets gap information of stream and adds the encountered gaps to the gap
_PPSD__insert_gap_times Gets gap information of stream and adds the encountered gaps to the gap
_PPSD__insert_processed_data Inserts the given UTCDateTime and processed/octave-binned spectrum at
_PPSD__invalidate_histogram
_PPSD__plot_coverage Helper function to plot coverage into given axes.
_PPSD__process Processes a segment of data and save the psd information.
_PPSD__sanity_check Checks if trace is compatible for use in the current PPSD instance.
_add_npz See PPSD.add_npz().
_get_gapless_psd Helper routine to get a list of 2-tuples with gapless portions of
_get_plot_title
_get_response
_get_response_from_inventory
_get_response_from_parser
_get_response_from_paz_dict
_get_response_from_resp
_get_times_all_details
_plot_histogram Reuse a previously created figure returned by plot(show=False)()
_setup_period_binning Set up period binning.
_split_lists
_stack_selection For details on restrictions see calculate_histogram().

Special Methods

__dir__ default dir() implementation
__format__ default object formatter
__init__ Initialize the PPSD object setting all fixed information on the station
__new__ Create and return a new object.
__reduce__ helper for pickle
__reduce_ex__ helper for pickle
__sizeof__ size of object in memory, in bytes
__subclasshook__ Abstract classes can override this to customize issubclass().