|
Titel |
Climate wavelet spectrum estimation under chronology uncertainties |
VerfasserIn |
G. Lenoir, M. Crucifix |
Konferenz |
EGU General Assembly 2012
|
Medientyp |
Artikel
|
Sprache |
Englisch
|
Digitales Dokument |
PDF |
Erschienen |
In: GRA - Volume 14 (2012) |
Datensatznummer |
250062509
|
|
|
|
Zusammenfassung |
Several approaches to estimate the chronology of palaeoclimate records  exist in the
literature: simple interpolation between the tie points, orbital tuning, alignment on other
data... These techniques generate a single estimate of the chronology. More recently,
statistical generators of chronologies have appeared (e.g. OXCAL, BCHRON) allowing the
construction of thousands of chronologies given the tie points and their uncertainties.
These techniques are based on advanced statistical methods. They allow one to take
into account the uncertainty of the timing of each climatic event recorded into the
core.
On the other hand, when interpreting the data, scientists often rely on time series analysis,
and especially on spectral analysis. Given that paleo-data are composed of a large spectrum
of frequencies, are non-stationary and are highly noisy, the continuous wavelet
transform turns out to be a suitable tool to analyse them. The wavelet periodogram,
in particular, is helpful to interpret visually the time-frequency behaviour of the
data.
Here, we combine statistical methods to generate chronologies with the power of continuous
wavelet transform. Some interesting applications then come up: comparison of
time-frequency patterns between two proxies (extracted from different cores), Â between a
proxy and a statistical dynamical model, and statistical estimation of phase-lag between two
filtered signals. All these applications consider explicitly the uncertainty in the chronology.
 The poster presents mathematical developments on the wavelet spectrum estimation under
chronology uncertainties as well as some applications to Quaternary data based on marine
and ice cores. Â |
|
|
|
|
|