Numerical time-series analysis
Several statistical techniques were employed that are commonly
used with climate data when the time-series are short and noisy, as
well as datasets that are non-evenly sampled or with missing values
(e.g., Ghil et al., 2002). Spectrum Analysis-Multitaper Method (MTM)
(Vautard et al., 1992) was used to perform detailed spectral analysis
on the coral metal and meteorological records. Wavelet analysis was
also employed to determine both the dominant modes of variability
(as detected using singular spectrum analysis) and their variability in
time where the time-series is decomposed into time–frequency space
(Torrence and Compo, 1998).