Disentangling seasonal signals in Holocene climate trends by satellite-model-proxy integration

Abstract

[1] Past sea surface temperatures (SSTs) are routinely estimated from organic and inorganic remains of fossil phytoplankton or zooplankton organisms, recovered from seafloor sediments. Potential seasonal biases of paleoproxies were intensely studied in the past; however, even for the two most commonly used paleoproxies for SST, U37K′ and Mg/Ca ratios, a clear global picture does not yet exist. In the present study we combine Holocene SST trends derived from U37K′ and Mg/Ca ratios with results from idealized climate model simulations forced by changes in the orbital configuration, which represents the major climate driver over the last 10 kyr. Such changes cause a spatiotemporal redistribution of incoming solar radiation, resulting in a modulation of amplitude and phasing of the seasonal cycle. Considering that the climate signal recorded by a plankton-based paleoproxy may be affected by the seasonal productivity cycle of the respective organism, we use the modern relationship between SST and marine net primary production, both obtained from satellite observations, to calculate a seasonality index (SI) as an independent constraint to link modeled SST trends with proxy data. Although the climate model systematically underestimates Holocene SST trends, we find that seasonal productivity peaks of the phytoplankton-based U37K′ result in a preferential registering of the warm (cold) season in high (low) latitudes, as it was expected from the SI distribution. The overall smoother trends from the zooplankton-derived Mg/Ca SSTs suggest a more integrated signal over longer time averages, which may also carry a seasonal bias, but the spatial pattern is less clear. Based on our findings, careful multiproxy approaches can actually go beyond the reconstruction of average climate trends, specifically allowing to resolve the evolution of seasonality.