Proxies generally record frequency-dependent climate signals. I would think that this frequency aspect is perhaps handled by careful selection of proxies or by e.g. choosing proper indexing methods if tree-rings are concerned. Do any of the approaches called climate field reconstruction (CFR methodologies) have clear advantages in this respect? Opinions, views and relevant references welcome!
By “careful” I simply mean paying attention to known properties of the proxies, which may be frequency related.
Different indexing (or standardization) methods used in tree-ring analysis likely handle those frequency-related properties in different ways.
Dear Markus,
by frequency-related properties do you mean frequency signals embedded into the reconstruction records you study?
In that sense, the most critical parameter is time resolution, which forces you to observe a precise frequency range. After the identification of precise signals, each related to a frequency, you can isolate and link them to some property. May be I'm not catching your question properly, so sorry in advance.
Let me know,
Kind regards,
Gianna
Agree with Dr Gianna VIvaldo. Time resolution is one of the most important parameters. Under the time constrain, both the observation of climate-senstive sediments and proxies of geochemistry are significant on frequent cylces.
Thank you Gianna, Xianghui and Kenneth for thoughtful comments! I realize my question was poorly formulated.
Keeping the above short discussion in mind, do you have any recommendations or clues regarding the method to build a transfer function?
Dear Prof. Lindholm,
did you mean, given a dataset, how to manage to find cycles/trends?
Sorry again if I need more info, hope to help you soon!
Kind regards.
Dear Prof. Kenneth M Towe,
I cannot understand your question, from the abstract I cannot see time resolution, neither the cycles you are looking for. I just see that you found a growth period: is it a trend or a large period cycle?
Do you have more detail about the analysis you performed?
Yes Gianna,
Given a dataset, how would you proceed to reconstruct cycles? I was actually thinking known oceanic oscillations with decadal and multidecadal and even lower oscillations.
Markus
Thanks Kenneth:
I was not aware of such an early work on the subject. I will certainly study this carefully!
Dear Prof. Lindholm,
there are several approaches to cycles detection in a given time series. You can start from "classical/older" ones, such as Fast Fourier Transform, continuous wavelets transform, maximum entropy methods; or you can apply some more recent techniques (multi-taper methods, singular-spectrum analysis). If you need an help, feel free to contact me directly, since the field is a very huge one (digital signal processing and spectral analysis).
Have a nice day,
Gianna
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Graphs