Constraining the Nineteenth-Century Temperature Baseline for Global Warming

• Published in: Journal of climate Vol. 36; no. 18; pp. 6261 - 6272
• Main Authors: Schneider, L., Konter, O., Esper, J., Anchukaitis, K. J.
• Format: Journal Article
• Language: English
• Published: Boston American Meteorological Society 01.09.2023
• Subjects: 19th century; Archives & records; Bias; Climate; Climate change; Climate policy; Datasets; Estimates; General circulation models; Global warming; Intergovernmental Panel on Climate Change; Mean temperatures; Measurement techniques; Northern Hemisphere; Paris Agreement; Remote regions; Summer; Temperature; Temperature distribution; Temperature effects; Temperature fields; Tree rings; Uncertainty; Urban heat islands; Warm seasons; Europe
• ISSN: 0894-8755; 1520-0442
• DOI: 10.1175/JCLI-D-22-0806.1

Since the Paris Agreement, climate policy has focused on 1.5° and 2°C maximum global warming targets. However, the agreement lacks a formal definition of the nineteenth-century “pre-industrial” temperature baseline for these targets. If global warming is estimated with respect to the 1850–1900 mean, as in the latest IPCC reports, uncertainty in early instrumental temperatures affects the quantification of total warming. Here, we analyze gridded datasets of instrumental observations together with large-scale climate reconstructions from tree rings to evaluate nineteenth-century baseline temperatures. From 1851 to 1900 warm season temperatures of the Northern Hemisphere extratropical landmasses were 0.20°C cooler than the twentieth-century mean, with a range of 0.14°–0.26°C among three instrumental datasets. At the same time, proxy-based temperature reconstructions show on average 0.39°C colder conditions with a range of 0.19°–0.55°C among six records. We show that anomalously low reconstructed temperatures at high latitudes are underrepresented in the instrumental fields, likely due to the lack of station records in these remote regions. The nineteenth-century offset between warmer instrumental and colder reconstructed temperatures is reduced by one-third if spatial coverage is reduced to those grid cells that overlap between the different temperature fields. The instrumental dataset from Berkeley Earth shows the smallest offset to the reconstructions indicating that additional stations included in this product, due to more liberal data selection, lead to cooler baseline temperatures. The limited early instrumental records and comparison with reconstructions suggest an overestimation of nineteenth-century temperatures, which in turn further reduces the probability of achieving the Paris targets.