Spectral analysis is the process of calculating and interpreting a spectrum, that is a representation of a time or space series in the frequency domain.12

SWEET makes frequency domain analysis an equivalent class and lists wavelet analysis among the classes beneath it.2 A 1980 NASA report on geophysical applications adds that some calculations are easier in the spectral domain than in the time or space domain.1 That is a reason to transform a series, not a claim that the spectrum replaces the original data.

Wavelet analysis is one family within the topic. Torrence and Compo describe it as useful for time series with many different timescales or changes in variance, and note that significance testing needs a background spectrum, usually white noise or red noise.3

All retained sources concern frequency or wavenumber. They do not cover spectral analysis of radiance across wavelength, and they do not tie the term to a NASA processing level or to a particular sensor.

Footnotes

  1. NASA Technical Reports Server, “Spectral analysis and applications (1980 report)”. https://ntrs.nasa.gov/api/citations/19800013337/downloads/19800013337.pdf ↩ ↩2

  2. ESIP SWEET, “reprDataServiceAnalysis.ttl”. https://raw.githubusercontent.com/ESIPFed/sweet/master/src/reprDataServiceAnalysis.ttl ↩ ↩2

  3. NOAA Physical Sciences Laboratory (Torrence and Compo), “A Practical Guide to Wavelet Analysis”. https://psl.noaa.gov/people/gilbert.p.compo/Torrence_compo1998.pdf ↩