Tree-growth cycles of sub-fossil tree ring chronologies from southern South America: could trees reveal ENSO dynamics at former times?

FIDEL ROIG

Akumal, Mexico

Workshop; Annual Science Meeting IAI CRN03; 2005

Tree rings repeatedly have offered many opportunities to deduce histories of late Holocene climates but too few tree-ring records exists to bring detailed insights into climate variability during earlier periods. This void is being partially filled by the recovery of large amounts of still in situ old stumps in environments of the Lake District and Chiloé Island in southern Chile. Three accelerator mass spectrometry dating revealed wood with a radiocarbon minimum ages of ca. 50 kyrs BP, indicating that these forest stands probably became established during the interglacial corresponding to the oxygen isotope stage 3. Interdrift well-developed organic horizons suggest intervals with swampy habitats that provided suitable conditions for the development of conifer forests at all the studied sites. These forests revealed a mix composition including Pilgerodendron uviferum, Fitzroya cupressoides and Nothofagus sp. (probably dombeyi type) trees(1,2). This wood material permitted the construction of well-replicated tree-ring chronologies allowing the study of growth details in segments of several hundred-years of record. Thus, the spectral analysis reveals growth oscillations at periods of around 150-250, 82-87, 45-46, 22-24, 11-12, and 9-3 years. Comparisons between the power spectra of sub-fossil chronologies and living trees show strong resemblances, suggesting that similar growth forcing factors operated during this glacial interstadial phase as in the current interglacial conditions(3). These results provide evidence for the interesting debate about particularly around the structure of the high frequency band of the spectra. The climate in areas where Fitzroya presently grows appears to be teleconnected with ENSO dynamics. Above-normal temperatures during summer follow a mature phase of a warm ENSO event(4,5). Thus, ENSO-induced changes in temperature leave a measurable signal in the Fitzroya tree-rings(6,7). At the wavelength range of the ENSO variability domain (2-7 yr), the modern Fitzroya tree-ring based temperature reconstructions show a high percentage of total variance contained in the high-frequency bands, a fact that could be attributed to the modulation of climate by ENSO activity. Similarly, our sub-fossil chronologies express significant peaks within these frequency bands, accounting for a variable amount of the total variance. Could these periodicities be a reflection of a paleo-ENSO system? It is known that ENSO has operated for most of the Holocene; but for earlier times, data from marine cores imply that during low glacial sea-levels, the ocean/atmosphere conditions in the inter-tropical Pacific could have prevented or dramatically reduced the functioning of the ENSO system (at least El Niño-like states as deduced by models(8)). Furthermore, there are increasing indications that the ENSO system probably operated during middle-late Pleistocene interglacials and warm interstadial periods, although overall climate and boundary conditions at these stages may have been too different to permit the characteristic present ENSO climate variability in terms of frequency, intensity and recurrence. New data provided by fossil corals and planktonic foraminifera revived the discussion(9,10). For example, the Papua New Guinea fossil coral records from selected time periods that correspond to the last interglacial- glacial- Holocene and modern periods all display some concentrations of variance in the typical ENSO frequency band(9). Although the ENSO strength has changed through time, the coral proxies suggest that ENSO has been a persistent component of the climate system over the past 130 kyrs. This evidence is supported by our tree-ring sub-fossil chronologies and other proxy biological records. This opens an interesting field of research because we can now use different biological sources to identify ancient evidence of high frequency oscillations at timescales in the frequency domain of the ENSO dynamics. We will discuss our findings in the frame of all the available evidences.