Reconstructing the Tropical Pacific ENSO Sea Surface Temperature Field

We present new climate field reconstructions (CFR) of tropical Pacific ENSO sea surface temperatures (SST) based on a circum-Pacific tree-ring network from regions known to be sensitive to El Niño rainfall. The target field is based on 1x1° HadISST data for the boreal winter (December-February) season. We used two different CFR methods: Point-by-Point Regression (PPR) and reduced-space Orthogonal Spatial Regression (OSR). The two show very similar results. Therefore, OSR is preferred because it is far more efficient. Both methods have high levels of validation skill as far back as 1100 CE over the Niño4, Niño3.4, Niño3 regions. In contrast, we were unable to reconstruct the Niño1+2 coastal region. Consistent with this, the leading EOF of the SST field (EOF1) can be reconstructed with a very high level of skill, but only EOF1; the EOF2 reconstruction did not validate with either method. The success of EOF1 is a consequence its importance for ENSO impacts; the failure with EOF2 reflects the lack of impacts sensitive to it, at least for available tree-ring chronologies. Though many more recent ENSO indices (e.g. flavors, ELI) are thought to require several tropical Pacific SST measures, we will show that they can be reconstructed from EOF1 alone.