ABSTRACT
In this chapter, we use over five billion words of data from eight newspapers taken from the British Library Newspapers collection to explore occurrences of nineteenth-century drought events. Records of dry spells and water shortage in the UK are sparse before the twentieth century so environmental scientists, particularly those interested in climate change, are keen to build up a fuller picture of the frequency and severity of extreme weather events before this time. Nineteenth-century newspapers, with their daily renewal of news stories and their preoccupation with the weather, not least in relation to the wellbeing of the harvest, offer real opportunities to access the physical realities of past droughts. However, a corpus composed solely of historical newspapers must be approached cautiously. Although it is unlikely that weather reporting was subject to deliberate manipulation by contemporary journalists, the digitised versions of the newspapers that we worked with were produced by running optical character recognition (OCR) software across old and often damaged newspapers and are thus prone to a varying but often significant degree of error.
The use of triangulation in this chapter is therefore an essential means of both corroborating our findings and overcoming the challenges within the data itself. This triangulation takes the form of close reading of the news texts in order to establish the meaning and quality of OCR texts; the comparison of our findings with those already offered by historical hydrological studies; and the application of concordance geoparsing, which allows us to focus in on UK-based drought events. In concert, these methods allow us to confidently identify a series of nineteenth-century droughts, some of which have been hitherto overlooked by hydrologists, and provide a narrative of these dry spells that takes into account their impact on local communities as presented by newspapers at the time.
