Late Cenozoic paleotemperatures reconstructed from the lignin-methoxy hydrogen isotope composition of sub-fossil wood in the Canadian Arctic

Abstract:

Proxy-based reconstructions of Cenozoic warm climates are a valuable data source for constraining climate sensitivity and understanding what a future, warmer world may look like. Such insights are especially critical in the terrestrial Arctic where the fastest rates of warming are underway and likely to continue. Here we estimate ca. late Miocene and Pliocene temperatures from hydrogen stable isotopes of lignin-methoxy groups (δ²H_LM) of Pinaceae sub-fossil wood at several Canadian Arctic Archipelago sites (74-80°N) – Banks Island (late Miocene and late Pliocene), Prince Patrick Island (late Pliocene), Meighen Island (late Pliocene), Bylot Island (late Pliocene) and Ellesmere Island (late Pliocene). Previous studies have demonstrated great potential of tree-ring δ²H_LM values as a proxy for mean annual precipitation δ²H (δ²H_MAP). However, this proxy has not yet been applied to late Cenozoic wood from the Canadian Arctic. The δ²H_LM of multi-year tree-ring increments was measured from 5-10 wood specimens per site and used to estimate a site-average δ²H_LM and δ²H_MAP (accounting for a -213‰ apparent fractionation). Reconstructed δ²H_MAP values at all sites were more positive than the site-specific modern δ²H_MAP values _{(amount-weighted)}, indicating reduced Rayleigh distillation compared to today. Miocene and Pliocene reconstructed δ²H_MAP values were higher than modern values by 88±18‰ (n=10) and 57±20‰ (n=33), respectively. δ²H_MAP gradients across the Arctic are strongly influenced by mean air temperature, with a modern Δδ²H_MAP-T sensitivity of 3.9‰·°C^-1 (r² = 0.93; Canadian GNIP network >50°N). Assuming a similar Δδ²H_MAP-T sensitivity in the past, the relatively positive Miocene and Pliocene δ²H_MAP values can be explained by a mean climate that was warmer than today by 22.6±4.6°C and 14.7±5.1°C, respectively. These are conservative estimates as they do not account for ²H-depleted Miocene and Pliocene ocean water. These data, in combination with previously published paleo-temperature and -ecological data from some of the same sites, document a significantly warmer Arctic during the Miocene and Pliocene. Attribution of this warming to changes in atmospheric, vegetation and land-ocean boundary conditions is the subject of continued study and may hold important lessons for modelling future Arctic change.