I can perform a wavelet cross coherence using Matlab but I do not know how to tell what parts of the resulting transform are statistically significant. Torrence and Compo do this by using a montecarlo method against red noise background. I have failed to manage to change their code to test against white noise and I don't see any obvious way to test the normal wavelet cross coherence results. Can anyone help please? Or if someone has figured out how to alter Torrence's code please please let me know! Thanks!Kim
MATLAB: Wavelet cross coherence statistics
wavelet cross coherence statistics wavelets white noise
Related Solutions
From my understanding of the problem, it sounds like you would like to determine the correlation between a given wavelet and the mother wavelet. Have you tried to use "wavefun" to generate the mother wavelet as a separate signal and then pass that to "wcoher"? One of the output arguments of "wcoher", called "xval", can be used to evaluate your signal such that the sizes are consistent when passed to "wcoher". "wcoher" ends up using "wavefun" to evaluate the wave you specify in 'wname' so this might give the desired result. For more information, please see:
Please let me know if I misunderstood your question.
In wcoherence, smoothing is done by calculating weighted running average (or convolution) in both the time and scale directions. Taking reference from Torrence and Webster, smoothing operator in wcoherence function is given by 
Best Answer