I was talking the other day about wavelets and I found this, from @johncarlosbaez 's blog "Azimuth". Here, John and a colleague used the Gabor transform, very related to wavelets, to analyze the Milankovich cyles, an important concept in climate science.
It's really interesting. The Gabor transform, as the wavelets, show information in both frequency and temporal domain, contrary to a Fourier transform that takes you completely to the frequency world or the temporal one.
For a signal whose components may have changed through time, the Gabor can help more than a Fourier analysis. This in fact is what happened with Milankovich, where "recently", in the last million years aprox. the eccentricity seems to be the stronger signal.
