Do our reconstructions of ENSO have too much low-frequency variability?
Abstract
Reconstructing the spectrum of Pacific SST variability has proven to be difficult both because of complications with proxy systems such as tree rings and the relatively small number of records from the tropical Pacific. We show that the small number of long coral δ18O and Sr/Ca records has caused a bias towards having too much low-frequency variability in PCR, CPS, and RegEM reconstructions of Pacific variability. This occurs because the individual coral records used in the reconstructions have redder spectra than the shared signal (e.g. ENSO). This causes some of the unshared, low-frequency signal from local climate, salinity and possibly coral biology to bleed into the reconstruction. With enough chronologies in a reconstruction, this unshared noise cancels out but the problem is exacerbated in our longest reconstructions where fewer records are available. Coral proxies tend to have more low-frequency variability than SST observations so this problem is smaller but can still be seen in pseudoproxy experiments using observations and reanalysis data. The identification of this low-frequency bias in coral reconstructions helps bring the spectra of ENSO reconstructions back into line with both models and observations. Although our analysis is mostly constrained to the 20th century due to lack of sufficient data, we expect that as more long chronologies are developed, the low-frequency signal in ENSO reconstructions will be greatly reduced.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2017
- Bibcode:
- 2017AGUFMPP31A1245L
- Keywords:
-
- 3305 Climate change and variability;
- ATMOSPHERIC PROCESSES;
- 3344 Paleoclimatology;
- ATMOSPHERIC PROCESSES;
- 1620 Climate dynamics;
- GLOBAL CHANGE;
- 4928 Global climate models;
- PALEOCEANOGRAPHY