Identifying bias in cold season temperature reconstructions by beetle mutual climatic range methods in the Pliocene Canadian High Arctic.

Abstract Well-preserved beetle elytra from the fossil and subfossil record are used by palaeoclimatologists to estimate past temperatures. Beetle-derived estimates of temperature across the Pliocene Arctic are consistently lower than those derived from other palaeoclimate proxies. Here we test if that pattern is attributable to either the modern dataset or specific mutual climatic range method (MCRM). We used a global dataset of observations (Global Biodiversity Information Facility) for modern beetles as opposed to a North American regional dataset, and a consistent MCRM for comparison to vegetation-derived estimates, termed CRACLE (Climate Reconstruction Analysis using Coexistence Likelihood Estimation), rather than different MCRM methods for each analysis. Beetle-derived estimates from one site, Beaver Pond, are consistently much lower than the vegetation-derived estimates, stable isotope results from flora and fauna, and summer temperatures from bacterial tetraethers. The reanalysis of beetle data produced a small reduction in proxy estimate mismatch overall. We found good agreement among proxy estimates of maximum temperature of the warmest month for two of the three sites in the Canadian Arctic, but increasingly poorer agreement for warmest quarter, coldest quarter and minimum temperature of the coldest month. This pattern may be due to beetles' avoidance mechanisms and a climate space different from modern calibration sets during the Pliocene. Beetle derived temperature estimates, especially cold season, should be interpreted with caution when reconstructing palaeoclimates. Highlights • Beetle-derived estimates of Pliocene Arctic winter temperature appear cold-biased. • Global dataset and consistent method improved warm season estimate mismatch. • Cold season temperature mismatch persists. [ABSTRACT FROM AUTHOR]