PP045-10
Pollen archives reveal convergence of terrestrial and marine climate variability over centennial to millennial timescales in the Holocene
Pollen archives reveal convergence of terrestrial and marine climate variability over centennial to millennial timescales in the Holocene
Tuesday, 15 December 2020: 16:27
Virtual
Abstract:
Natural climate variability remains a major source of uncertainty for regional climate projections and it is thus necessary to better characterize it beyond instrumental timescales, using climate sensitive proxies over the Holocene. While marine temperatures where shown to exhibit strong variations on multidecadal and longer time-scales as evidenced from direct observations and indirect proxy data (Laepple & Huybers, 2014) such evidences remain anecdotal over land. It is commonly thought that in contrast to slow varying ocean temperatures, terrestrial variability behaves ‘weather like’ and averages out on longer time-scales, but the shortness of instrumental records and limits of proxy records inhibited any robust observational based inference on multidecadal and longer time-scales. Using a large database of pollen records covering the Holocene, we extend a continuous picture of regional temperature variability over land from the rather stable ‘weather like’ annual to decadal timescales, to increasingly variable ‘ocean like’ centennial to multi-millennial timescales. In contrast, climate model transient simulations over the Holocene remain rather stable up to multi-millennial timescales. Therefore, we conclude that important processes and feedbacks responsible for generating natural variability, or redistributing forced variability to other timescales, are missing from climate models, leading to a large deficit of simulated climate variability at longer timescales over land. As a result, the expected anthropogenic warming for the 21st century may be significantly different at the local scale felt by humans compared with the global trend, potentially leading to more heterogeneous spatial patterns of warming than the rather linear climate model projections (Hébert & Lovejoy, 2018).