6.6.5 Regional Variability in Quantities Other than Temperature

6.6.5.1 Changes in the El Niño-Southern Oscillation System

Considerable interest in the ENSO system has encouraged numerous attempts at its palaeoclimatic reconstruction. These include a boreal winter (December–February) reconstruction of the Southern Oscillation Index (SOI) based on ENSO-sensitive tree ring indicators (Stahle et al., 1998), two multi-proxy reconstructions of annual and October to March Niño 3 index (average SST anomalies over 5°N to 5°S, 150°W to 90°W; Mann et al., 2005a,b), and a tropical coral-based Niño 3.4 SST reconstruction (Evans et al., 2002). Fossil coral records from Palmyra Island in the tropical Pacific also provide 30- to 150-year windows of ENSO variability within the last 1.1 kyr (Cobb et al., 2003). Finally, a new 600-year reconstruction of December to February Niño-3 SST has recently been developed (D’Arrigo et al., 2005), which is considerably longer than previous series. Although not totally independent (i.e., the reconstructions share a number of common predictors), these palaeorecords display significant common variance (typically more than 30% during their respective cross-validation periods), suggesting a relatively consistent history of El Niño in past centuries (Jones and Mann, 2004). In most coral records from the western Pacific and the Indian Ocean, late 20th-century warmth is unprecedented over the past 100 to 300 years (Bradley et al., 2003b). However, reliable and consistent interpretation of geochemical records from corals is still problematic (Lough, 2004). Reconstructions of extratropical temperatures and atmospheric circulation features (e.g., the North Pacific Index) correlate significantly with tropical estimates, supporting evidence for tropical/high-latitude Pacific links during the past three to four centuries (Evans et al., 2002; Linsley et al., 2004; D’Arrigo et al., 2006).

The El Niño-Southern Oscillation may have responded to radiative forcing induced by solar and volcanic variations over the past millennium (Adams et al., 2003; Mann et al., 2005a). Model simulations support a statistically significant response of ENSO to radiative changes such that during higher radiative inputs, La Niña-like conditions result from an intensified zonal SST gradient that drives stronger trade winds, and vice versa (Mann et al., 2005a). Comparing data and model results over the past millennium suggests that warmer background conditions are associated with higher variability (Cane, 2005). Numerical experiments suggest that the dynamics of ENSO may have played an important role in the climatic response to past changes in radiative forcing (Mann et al., 2005b). Indeed, the low-frequency changes in both amplitude of variability and mean state indicated by ENSO reconstructions from Palmyra corals (Cobb et al., 2003) were found to correspond well with the model responses to changes in tropical volcanic radiative forcing over the past 1 kyr, with solar forcing playing a secondary role.

Proxy records suggest that ENSO’s global climate imprint evolves over time, complicating predictions. Comparisons of ENSO and drought indices clearly show changes in the linkage between ENSO and moisture balance in the USA over the past 150 years. Significant ENSO-drought correlations occur consistently in the southwest USA, but the strength of moisture penetration into the continent varies substantially over time (Cole and Cook, 1998; Cook et al., 2000). Comparing reconstructed Niño 3 SST with global temperature patterns suggests that some features are robust through time, such as the warming in the eastern tropical Pacific and western coasts of North and South America, whereas teleconnections into North America, the Atlantic and Eurasia are variable (Mann et al., 2000). The spatial correlation pattern for the period 1801 to 1850 provides striking evidence of non-stationarity in ENSO teleconnections, showing a distinct absence of the typical pattern of tropical Pacific warming (Mann et al., 2000).