No change in ENSO hydroclimate variability after the industrial revolution as recorded in ?18O of Tectona grandis L.F. from Southeast Sulawesi, Indonesia

Abstract

El Niño-Southern Oscillation (ENSO) is a quasi-periodic interannual oscillation of the ocean-atmosphere system in the tropical Pacific which greatly influences global climate variability. However, the long-term response to greenhouse gas forcing is still controversial. In this study, we measured the oxygen isotopic composition of ?-cellulose samples at intraannual resolution from independently crossdated teak cores (Tectona grandis L. f.) collected at Muna, Indonesia (5.3ºS, 123ºE, elev. 10m). The site and observation has been previously shown to provide an indirect measure of ENSO activity via local precipitation amount variations associated with ENSO. We created an ensembled composite of the interannual variability for the period 1680-2005 (316 years) using empirical high pass filtering and random sampling of intra-annual resolution measurements. In processing this time series composite, we used Singular Spectrum Analysis (SSA) to high pass filter the data for the interannual variability associated with ENSO. The annually-resolved composite time series of ?18O that we constructed has a higher resolution than other studies that have been conducted to reconstruct ENSO-hydroclimate activities in the western tropical Pacific region over this period. Using this ?18O composite, we compared the distribution of events in the period before and after the industrial revolution using the two-sample Kolmogorov-Smirnov(KS) test. We found no statistically significant change in the distribution of ?18O anomalies. The same statistical test was applied to the Niño 3.4 reconstruction from the Last Millennium Reanalysis (LMR). The results of this study suggest that if there is indeed a forced response of ENSO, it is as yet indetectable. This may be because the forcing is not yet large enough or the forced response is small relative to the unforced variability. Additional factors that might explain this result in the ?18O composite include its observational and interpretational uncertainty, and in the LMR reconstruction, the scarcity of tropical observational constraints and systematic error in the representation of ENSO in climate simulations.