Statistical reconstruction of daily temperature and sea level pressure in Europe for the severe winter 1788/89

Abstract

Abstract. The winter 1788/89 was one of the coldest winters Europe had witnessed in the past 300 years. Fortunately, for historical climatologists, this extreme event occurred at a time when many stations across Europe, both private and as part of coordinated networks, were making quantitative observations of the weather. This means that several dozen early instrumental series are available to carry out an in-depth study of this severe cold spell. While there have been attempts to present daily spatial information for this winter, there is more to be done to understand the weather variability and day-to-day processes that characterised this weather extreme. In this study, we seek to reconstruct daily spatial high-resolution temperature and sea level pressure fields of the winter 1788/89 in Europe from November through February. The reconstruction is performed with an analogue resampling method (ARM) that uses both historical instrumental data and a weather type classification. Analogue reconstructions are then post-processed through an ensemble Kalman fitting (EnKF) technique. Validation experiments show good skill for both reconstructed variables, which manage to capture the dynamics of the extreme in relation to the large-scale circulation. These results are promising for more such studies to be undertaken, focusing on different extreme events and other regions in Europe and perhaps even further back in time. The dataset presented in this study may be of sufficient quality to allow historians to better assess the environmental and social impacts of the harsh weather.

Early meteorological observations in West Africa during the 18th century

“…The volcanic eruptions of Lakagígar (Iceland) in June 1783 and Etna (Italy) in July 1787 have been considered causal agents of climatic variations in the European continent (Písek and Brázdil, 2006) and the Iberian Peninsula (Rodrigo, 2019) during the 1780s. In our data, we do not find information that can be linked to this causal mechanism, but Prelong's comment (1793, p. 269), according to which “Le grand hiver de 1788 à 1789 se sit sentir dans la zone torride… En comparant mes notes de 1789 à celles de 1788, je trouve seudement que le mois de mars 1789 fut plus froid que le même mois de l'année précédente.” The winter of 1788–1789 was one of the coldest winters on the European continent in the past 300 years (Pappert et al ., 2022) as a result of the combination of several factors, such as the predominance of the negative phase of the North Atlantic Oscillation (NAO), the effects of the Laki volcanic eruption in 1783, as well as fluctuations of the El Niño–Southern Oscillation (ENSO) (Kington, 2009). The fact that thermal conditions during the 18th century (Figures 5 and 9) are very similar to those of the modern reference period may indicate the absence of significant volcanic eruptions during much of this century (Jones and Briffa, 2006).…”

Comment on cp-2024-40

“…For the analyses of daily weather, we not only used the raw data, but reconstructed daily pressure fields over Europe from the pressure observations using a simple analog approach (see also Pappert et al, 2022). For that we used the ERA5 reanalysis (Hersbach et al, 2020(Hersbach et al, ) from 1940(Hersbach et al, -2023 We extracted sea-level pressure at the 1740 observation locations, deseasonalized and standardized the data in the same way as described above (using the entire period) and then determined, for each day in 1740, the closest analog day in ERA5 within a window of ±60 calendar days of the target day.…”

A 258-year-long data set of temperature and precipitation fields for Switzerland since 1763

“…When sufficient data are available, fields can be reconstructed based on interpolation methods such as for the UK for daily precipitation fields dating back to 1890 (Keller et al, 2015). Single weather events have been reconstructed by resampling present-day meteorological fields (Pappert et al, 2022;Flückiger et al, 2017) and by dynamically downscaling historical reanalysis data (e.g. Stucki et al, 2015;Brugnara et al, 2017).…”

“…However, only a few attempts have been made to reconstruct weather back to this time using quantitative data (e.g. Pappert et al, 2022). For this period, only a few data records have so far been available because it is a tremendous amount of work to find, image, digitize, quality check, and homogenize these data.…”