Reconstructing El Nino-southern oscillation

Abstract

El Niño-Southern Oscillation (ENSO) is the most important coupled ocean-atmospheric phenomenon to cause global climate variability on interannual time scales. Efforts to understand recent, apparently anomalous ENSO behaviour are hampered by the lack of long, high-quality climate records. While instrumental data generally covers the past 150 years, record length is insufficient for the assessment of past changes in the frequency, magnitude, and duration of ENSO. Here, multiproxy networks of high-resolution tree-ring, coral, ice and documentary records derived from eastern and western Pacific ENSO ‘centres of action’ are analysed (A.D. 1525-2002). Considerable improvements in ENSO reconstruction are achieved from expanding the use of records from the western Pacific. In particular, ~500 years of a continuous 3,722 year ENSO sensitive tree-ring record from New Zealand is introduced. Although extreme ENSO events are seen throughout a 478-year discrete event analysis, 43% of extreme, 20% of very strong and 28% of all protracted ENSO events occur within the 20th century. Principal component analysis was used to extend instrumental records of the Southern Oscillation Index (SOI) Niño 3.4 Sea Surface Temperature (Niño 3.4 SST) and a newly developed coupled ocean-atmospheric ENSO index (CEI) by 347 years. Significantly, of the three indices reconstructed here, CEI reconstructions were largely found to be the best predictors of ENSO. The results suggest that ENSO may be more effectively characterised using a coupled ocean-atmosphere index, particularly for December-May periods. Compared to the pre-instrumental period, the late 19th and early 20th centuries indicate a clear trend toward increased ENSO variability over the past 150 years. Significantly, spectral analysis of reconstructed indices reveals a marked change in the frequency and intensity of ENSO beginning ~A.D. 1850, coinciding with the end of the Little Ice Age and the boom in global industrialisation. This suggests that ENSO may operate differently under natural (pre-industrial) and anthropogenically influenced background states. This study asserts that recent ENSO variability appears anomalous in the context of the past five centuries. Given the considerable socio-economic impacts of ENSO events, future investigation into the implications an increasingly anthropogenically-warmed world may have on ENSO is vital.