Testing multiple paleoclimatic proxies in a Triassic marine record from China

Abstract:

Multiple proxies developed from sedimentary sequences enhance confidence in the characterization of paleoclimate and paleoenvironmental change. However, the relationships among different proxies and their relative strengths or weaknesses are rarely well understood, yet represent a wealth of untapped information for paleoclimate study.

To decipher the relationship among multiple proxies and to estimate their individual signal-to-noise levels, we propose two criteria for multiple proxy datasets: (1) For proxies affected by similar processes, time-dependent changes of the proxies should be similar, and (2) the paleoclimate proxy that is more sensitive to an external forcing, such as astronomical forcing, is more useful than one that is not. We introduce two methods to evaluate these criteria, i.e. hierarchy cluster analysis and power decomposition analysis.

We evaluate 16 high-resolution proxies collected from the Daxiakou section from South China, which was deposited in the Tethys region during the Early Triassic Epoch. The proxies are: spectral gamma ray (gamma-ray intensity, potassium, uranium, thorium, thorium/uranium and thorium/potassium), rock color series (L*, a* and b*), magnetic susceptibility (MS), anhysteretic remanent magnetization (ARM), water depth rank, simplified depth rank, non-carbonate fraction, carbonate thickness and bed thickness.

Hierarchy cluster analysis shows that the proxies cluster into groups likely affected by the same process. Milankovitch forced insolation drives seasonality. ARM and thorium/uranium reflect hinterland weathering cycles. Gamma-ray, MS and non-carbonate fraction refer to terrestrial input. L*, a* and depth rank indicate productivity, redox state, and relative sea-level, respectively. Power decomposition analysis shows that Milankovitch forcing of the gamma-ray, ARM, non-carbonate fraction proxies is very high which suggests that these proxies are ideal as paleoclimate proxies while the others are too noisy.