Do the early nineteenth century eruptions strengthen evidence for volcanically-induced Eurasian winter warming?

Abstract

It is well-established that explosive volcanic eruptions typically lead to cooler surface temperatures in summer, but the picture in Northern Hemisphere winter is much more uncertain. Recent large, low-latitude eruptions have been followed by warm anomalies across Eurasia in winter and cold anomalies near Greenland, hypothesized to be part of a dynamical response to the volcanic forcing that drives a positive North Atlantic Oscillation (NAO). But the evidence for a dynamical, winter warming response is inconclusive because internal variability is large, many climate models do not simulate a dynamical response like this, and there are few such eruptions to study. New datasets that allow additional eruptions from the early 19th century to be studied are therefore particularly valuable and we will present new analyses of the winters following four large eruptions in 1809, 1815, 1831 and 1835 (alongside four later eruptions in 1883, 1902, 1982 and 1991). This analysis is made possible by a new gridded instrumental dataset combining marine and land air temperatures from the 1780s onwards developed in the ongoing GloSAT project. It is supplemented by analysis of an ensemble of historically-forced simulations with UKESM1.1 initialised in 1750, also from the GloSAT project, and by two reanalyses (20CRv3 from 1806 and ModE-RA from 1421). For the instrumental and reanalysis datasets, warming in Europe was found in the first post-eruption winter following six out of the eight cases studied, and in the second post-eruption winter in five. Similar results were found for cold anomalies near Greenland and for a positive winter NAO index.  The anomaly magnitudes for individual cases were mostly within the range of internal variability but the consistency of the response across eruptions and datasets was significant in comparison with non-volcanic winters. The UKESM1.1 simulations showed a significant response (with Eurasian winter warming, Greenland cooling and positive NAO) for only the largest eruption (Tambora), suggesting a response may require a minimum forcing strength to occur.