Contributions of atmospheric circulation variability and data coverage bias to the warming hiatus

Abstract

The warming hiatus shows a strong seasonal and geographical asymmetry, with cooling in the Northern Hemisphere winter, especially over land, and warming elsewhere and in the other seasons. We show that the characteristics of the Northern Hemisphere winter cooling in 1998–2012 can mostly be explained by missing observations and by internal variability in the atmospheric circulation of the Northern Hemisphere extratropics. Estimates of the annual and seasonal temperature trends in 1998–2012 obtained by considering the concurrent effects of unforced natural variability and of coverage bias are much closer to the corresponding long‐term trends. Reanalyses suggest that the coverage bias was exceptionally pronounced during recent years and that an area of strong warming was missed due to the incomplete observational coverage. Coupled Model Intercomparison Project Phase 5 climate models indicate that trends in atmospheric circulation during the hiatus period did not occur as a response to anthropogenic forcing.

Robust comparison of climate models with observations using blended land air and ocean sea surface temperatures

“…The recent divergence between the models and the observations occurs after 1998, the period commonly associated with the so‐called global warming “hiatus” [ Fyfe et al , ; Fyfe and Gillett , ; Tollefson , ]. Several contributory factors to the divergence have been identified, including an increase in moderate volcanic eruptions [ Solomon et al , ; Ridley et al , ; Santer et al , , ], a reduction in solar activity, a decrease in stratospheric water vapor concentration [ Solomon et al , ], internal variability [ Meehl et al , , ; Trenberth and Fasullo , ; Kosaka and Xie , ; Mann et al , ; Steinman et al , ; Dai et al , ], and a bias due to the omission of the Arctic, which is warming more rapidly than projected by the models [ Cowtan and Way , ; Saffioti et al , ]. The contribution of internal variability to the remaining discrepancy between the models and the observations is beyond the scope of this analysis.…”

“…When comparing against spatially incomplete records, the model temperature fields may be masked to reduce coverage to match the observations or make the assumption that the observed regions are representative of the unobserved regions. This assumption may not hold for the last two decades of accelerated Arctic warming [ Simmons and Poli , ; Saffioti et al , ]. Although in some cases the model simulations were masked for coverage, most studies have used the surface air temperature field from models rather than blended land‐ocean temperatures, with the notable exception of Marotzke and Forster [] and some attribution studies, e.g., Knutson et al [].…”

The global warming hiatus has faded away: An analysis of 2014–2016 global surface air temperatures

“…Furthermore, recent studies have attributed the hiatus to bias in global temperature reconstructions due to the sparsity of observational stations, especially over polar regions (Cowtan and Way, 2014;Karl et al, 2015;Richardson et al, 2016). Such sparsity results in a cold bias in temperature and a reduction in warming trends (Saffioti et al, 2015). However, other study have suggested that this effect is marginal, in the order of a few hundredths K/decade (Gleisner et al, 2015), and therefore, not large enough to affect the global mean surface temperature (Jones, 2016).…”

The relative contributions of tropical Pacific sea surface temperatures and atmospheric internal variability to the recent global warming hiatus

“…While GMST is a fundamental metric for defining the hiatus, obeying global energetic constraints and facilitating conceptual understanding, it masks important spatial and seasonal variation for which a more detailed mechanistic explanation is needed. As many studies have shown, the hiatus is primarily a boreal winter phenomenon of the Northern Hemisphere (NH) continents, with cooling over Eurasia making the dominant contribution [i.e., Cohen et al, 2012;KX13;Trenberth et al, 2014;Saffioti et al, 2015;Li et al, 2015;Huang et al, 2017]. The proximate cause of the Eurasian winter cooling is a trend in the large-scale atmospheric circulation toward enhanced anticyclonic conditions accompanied by more frequent submonthly blocking episodes [previous references and Luo et al, 2016].…”