I have extracted different cycles from temperature and precipitation at kyanging, langtang river basin, Nepal using EMD. Temperature data shows cycles of 1 year 3.13 years, 6.25 years, 8.33 years and 12.5 years. Similarly, precipitation data is embedded with half year, 1 year, 2 years, 2.27 years and 8.33 years cycle. Now i want to know the causes for this oscillation. How can i do this?
Many of your "cycles" are multiples of each other. You need to make sure that this is not purely an artefact first. The annual cycle (1 year) is expected in climate data, if the annual cycle is not first removed.
As far as I know, EMD should be robust against harmonics, as the actual shape of the signal is (in contrast to Fourier transformation) not fixed. Actually, the intrinisc mode functions should give you "local" frequencies that are not necessarily the same over all of the temporal domain.
Regarding the evaluation / cause of the time scales: Check SSTs in the Indic and Pacific Oceans, and probably climate index series such as ENSO (though off the top of my head this has a different "frequency").
