Reconstructing ENSO: the influence of method, proxy data, climate forcing and teleconnections

In this study we compare three newly developed independent NINO3.4 sea surface temperature (SST) reconstructions using data from (1) the central Pacific (corals), (2) the TexMex region of the USA (tree rings) and (3) other regions in the Tropics (corals and an ice core) which are teleconnected with central Pacific SSTs in the 20th century. Although these three reconstructions are strongly calibrated and well verified, inter-proxy comparison shows a significant weakening in inter- proxycoherenceinthe 19thcentury.Thisbreakdownin commonsignalcouldberelatedtoinsufficient data, dating errors in some of the proxy records or a breakdown in El Nino-Southern Oscillation's (ENSO's) influence on other regions. However, spectral analysis indicates that each reconstruction portraysENSO-likespectralproperties.Superposedepochanalysisalsoshowsthateachreconstruction shows a generally consistent 'El Nino-like' response to major volcanic events in the following year, while duringyearsT þ4toT þ7'La Nina-like'conditions prevail.Theseresults suggest thateachofthe series expresses ENSO-like 'behaviour', but this 'behaviour' does not appear to be spatially or temporally consistent. This result may reflect published observations that there appear to be distinct 'types' of ENSO variability depending on location within the tropical Pacific. Future work must address potential dating issues within some proxies (i.e. sampling of multiple coral heads for one location) as well as assessing the time stability of local climate relationships with central Pacific SSTs. More emphasis is needed on sampling new and extending old coral proxy records from the crucial central and eastern tropical Pacific region. Copyright # 2009 John Wiley & Sons, Ltd. and # Crown Copyright 2009.