Rethinking the Holocene temperature conundrum

Abstract

Recent scholarship argues for more research to resolve the ‘Holocene temperature conundrum’, an apparent discrepancy between decreasing proxy-reconstructed and increasing model-simulated long-term temperature trends during the late Holocene. Here, we argue that the observed proxy–model offset likely results from inappropriate comparisons of different seasonal and spatial signals in the reconstructed and simulated palaeo-data. Since proxy archives have been used to reconstruct global annual mean temperatures, they have been compared against model simulations of the same seasonal and spatial domains. However, we suggest that most of the proxy-based large-scale reconstructions are biased towards Northern Hemisphere summer temperatures, and as such model comparisons have predominantly focused on the wrong target data. Further to advancing our understanding of long-term temperature trends, we recommend prioritising the refinement of proxy networks and climate reconstructions to preserve the full spectrum of naturally forced, interannual to multi-millennial variations needed to contextualise recent anthropogenic changes against past Holocene ranges.