A power spectrum describes how the power of a signal, or the variance of a time series, is distributed over frequency.12
Marchal calls the spectral density function the power spectrum, or more simply the spectrum, and describes spectral analysis as the estimation of that function from a given time series.1 Aster and Borchers apply the power spectral density to signals that are unbounded in time or continuous, and they observe that a power spectral density computed from finite data is always an estimate.2 The simplest estimate is the periodogram, which truncates the data with a rectangular window.2
SWEET declares the class equivalent to its general Spectrum class, so the ontology does not separate the power form from the wider notion of a spectrum.3 The retained sources are general time-series and geophysics notes, and they do not establish any remote-sensing procedure for the power spectrum.12
Background noise at seismic stations is the one application the retained sources describe.2 They give no numerical example that this page relies on.
Footnotes
-
Marchal (WHOI), “Notes of Time Series Analysis”. https://www.whoi.edu/cms/files/notes_on_time_series_analysis_268904.pdf ↩ ↩2 ↩3
-
Aster and Borchers, “Data Processing and Analysis”. http://www.ce.memphis.edu/7137/PDFs/signal%20Processing/power.pdf ↩ ↩2 ↩3 ↩4 ↩5
-
ESIP Federation (SWEET), “reprDataServiceAnalysis.ttl”. https://raw.githubusercontent.com/ESIPFed/sweet/master/src/reprDataServiceAnalysis.ttl ↩