← Back to Sessions

Visualizing Cyclical Variations in Seismoacoustic Activity Using Circular Spectrograms

Description: 

Environmental and anthropogenic signals are often periodic. For example, wind noise tends to follow diurnal and seasonal patterns whereas human facilities operate on a fixed schedule. These patterns can be difficult to extract and visualize using typical rectilinear time-frequency plots such as Fourier spectrograms. We describe a method to highlight periodic variations in signal power with respect to frequency by constructing a “circular spectrogram.” A data set with suspected cyclic variations is identified, and spectra are calculated over a set of time windows. Spectra corresponding to the same point in the cycle (e. g. a certain time of day) are assembled and the median is taken. The ensemble of spectra is plotted as a circular image, with the angular position of each pixel determined by its point in the time cycle and its radial position determined by frequency. When color coded by power, this produces a visualization of repeating variations in the time span of interest. We show how these circular spectrograms are calculated as well as provide examples of power fluctuations highlighted by this method. Initially developed as a data quality assessment method for the Physics Experiment 1 seismoacoustic network, the circular spectrogram shows promise for broader scientific insights across a variety of environments and time spans.

Session: Geophysics in a Changing World: Monitoring Applications from Seismology and Beyond [Poster]

Type: Poster

Date: 4/15/2025

Presentation Time: 08:00 AM (local time)

Presenting Author: Daniel

Student Presenter: No

Invited Presentation: 

Poster Number: 43

Authors