Torm õhus ja maa peal – tugevad tormid Balti provintsides 19. sajandi teisel poolel ajakirjandusteadete ja ilmaandmete järelanalüüsi võrdluses
Storms in the Air and on the Ground: A Comparison of Severe Storms in the Baltic Provinces in the Second Half of the 19th Century Based on Press Reports and Weather Data Analysis

Summary/Abstract: Historical climatology deals with narrative sources as well as historical observation data. Observation data are, for example, series of temperature and air pressure measurements recorded in early modern observatories, which can be analysed after calibration and verification in a similar way to modern meteorological measurements. Narrative sources, on the other hand, are mainly used for the study of extreme phenomena, not only because of their detail, but also because of the difficulty in detecting this detail in the observation data. Although narrative sources can also be quantified by coding them, there is as yet no consistent systematic method for comparing them with data series, since objective meteorological data cannot be unambiguously derived from subjective descriptions. In this article, we tried to solve this problem by comparing two databases. As a meteorological phenomenon, we considered strong storms in the Baltic provinces in the mid and second half of the 19th century, combining both narrative sources and historical observation data. The Tallinn University Historical Storms Database, based on the Estonian Historical Climate Database, provided narrative reports of storms collected from the 19th century press. These were compared with data from the Northern Hemisphere 19th-21st Century Weather Data Post-Analysis Project compiled by the US National Oceanic and Atmospheric Administration (NOAA). We selected 19 strong storms ranked from press reports and searched for their matches in the NOAA database, attempting to reconstruct a synoptic description of the events where possible. The juxtaposition of two different datasets, collected by methods of humanitarian and natural science, proved to be viable and promising. It was generally possible to model the ‘journalistic’ storm in a wider climatic context, explaining its origin and nature (Appendix 2). At the same time, the method draws attention to the question of which storms were perceived as particularly strong in the 19th century and why, i.e. which storms were reported in the newspapers. Convective or summer storms, in which rising air currents form thunderclouds, were an example. Powerful convective systems, on the other hand, are accompanied by sudden thunderstorms, storms, tornadoes, etc., which create a ‘moment of surprise’ and cause major destruction. It should be noted, however, that such convective storms cannot be detected or distinguished by postanalysis models. However, we can see if preconditions in air pressure maps were favourable for convective storms to from. Narrative information from journalistic reports therefore provides an important complement to the historical observation data. Based on NOAA data, we identified the three most powerful model storms per year between 1836 and 1899 (Appendix 1). As can be seen, the most powerful storm(s) modelled in the post-analysis data may not have generated a large news flow. There are probably several reasons for this. Firstly, it is worth remembering how news of storms got into the newspapers in the first place, or how and under what headings they were reported. An extreme event in weather terms does not necessarily have to be so in journalistic terms. If, for example, the storm did not result in destruction or loss of life, but only in the threat of flooding, the coverage is likely to be modest, and vice versa. In addition, particularly severe storms may simply have been overshadowed by the storm season. Additionally, the strongest winds might have blown in the sea or outside our study area. In one case there were no journalists, in other case there was no data for this study. We need to analyse past storm trajectories in more detail to analyse their possible impact precisely. Additionally, combining modelled data with newspaper data might help to understand storm parameters more clearly

• Details
• Contents