Verification of wavelet analysis interpretation of µ-XRF paleoclimate data from lake sediments using PCA, cluster analysis and empirical climate data
No Thumbnail Available
Date
2019-07-30
Journal Title
Journal ISSN
Volume Title
Publisher
International Union for Quaternary Research (INQUA)
Abstract
Reliable interpretation of annual resolution climate proxies for wind, precipitation, and detrital influx are a necessity for understanding the onset and periodicities of climatic events. In particular, verification of individual climate proxies in a µ-XRF time series data set is essential for the evaluation of inter-annual, decadal, and centennial periodicities driven by shifting positions of the South Westerly Winds (SWW) and subsequent storm belts associated with the Southern Annular Mode (SAM) and El Niño Southern Oscillation (ENSO) using wavelet analysis. Chronology for the Late Holocene section of lake sediment cores from Lake Kanono, Northland, New Zealand indicate an annual to sub-annual accumulation rate at a 500 µm Itrax µ-XRF scanning resolution. This fast accumulation rate provides the opportunity to produce quasi-annual µ-XRF time series data. Here we present twelve ca. 2200 year long elemental and element ratio µ-XRF time series including P, S, Ti, Ca, K, Si, Fe, Mn/Fe, (Fe+Ti)/K, Sr/Ca, Br/Cl, and Inc/coh. These twelve variables were interpreted as environmental proxies using a combination of principal component analysis (PCA) and cluster analysis, then verified with comparison to regionally averaged empirical rainfall and wind climate station data from Northland. Our results show that the wavelet patterns align with the PCA results allowing the µ-XRF time series to be classified into: Group I (detrital) and Group II (biological productivity and normalized climate proxies). The normalized Group II µ-XRF time series wavelet analyses displayed periodicities in the 2-16 year frequency, likely associated with ENSO, from ca. 237 BCE – 1330 CE. The data shows clear evidence of from both Polynesian and European settlement phases in this part of northern New Zealand as increased detrital influx. There is indication that Polynesian settlement impact was coeval with changes in ENSO intensity and a phase shift in SAM ca. 1350 CE, however European settlement is less associated with climatic shift. The Medieval Climate Anomaly (MCA) and the Little Ice Age (LIA) appear in the µ-XRF time series data as separate clusters. This data shows that the MCA is associated with windy/dry conditions with intermittent storms. During the LIA, the 2-16 year periodicity associated with ENSO decreased and centennial length periodicities increase, which may be an indication of an underlying SAM signal within the data. The combination of the methods of PCA, cluster analysis, and wavelet analysis indicates that sediment loading after ca. 1330 CE may be accentuating the climate signal in the Group I (detrital) proxies, and so interpretations after this time must be made with caution.
Description
Keywords
Verfification, Lakes, Sediment, X-ray fluorescence anaylsis, Polar-cap absorption, Data, Climates, Quaternary period, Winds, New Zealand, Rain
Citation
Evans, G. Augustinus, P., Gadd, P., Zawadzki, A., & Ditchfield, A. (2019). Verification of wavelet analysis interpretation of µ-XRF paleoclimate data from lake sediments using PCA, cluster analysis and empirical climate data. Poster presented to the 20th INQUA Congress 25th - 31st July 2019, Dublin, Ireland. Retrieved from: https://virtual.oxfordabstracts.com/#/event/public/574/submission/827