Applying Wavelet and Fourier Transform Analysis to Large Geophysical Datasets

The recurrence of periodic environmental states is important to many systems of study, and particularly to the life cycles of plants and animals. Periodicity in parameters that are important to life, such as precipitation, are important; to understanding environmental impacts, and changes to their intensity and duration can have far reaching impacts. To keep pace with the rapid expansion of Earth science datasets, efficient data mining techniques are required. This paper presents an automated method for Discrete Fourier transform (DFT) and wavelet analysis capable of rapidly identifying changes in the intensity of seasonal, annual, or interannual events. Spectral analysis is used to diagnose model behavior, and locate land surface cells that show shifting cycle intensity, which could be used as an indicator of climate shift. The strengths and limitations of DFT and wavelet spectral analysis are also explored. Example routines in Octave/Matlab and IDL are provided.