Representation of Drought Events in the United Kingdom: Contrasting 200 years of News Texts and Rainfall Records

Abstract

This paper combines evidence from the analyses of large sets of newspaper material and long-term rainfall records to gain insights into representations of drought events in the United Kingdom, between 1800 and 2014. More specifically, we bring together two different, though complementary, approaches to trace longitudinal patterns in the ways drought events have been measured and perceived, focusing specifically on the duration, spatial extent, and intensity of each event. The power of the combined approach is demonstrated through three case studies (1911–1913, 1940–1945, and 1947–1949), in which we explore the available evidence in more detail and explore the impacts of the droughts upon the British population. Using corpus linguistics methods, we examined four sets of newspaper material: 1) the British Library 19th century newspaper corpus, 2) The Times 20th century corpus (both i and ii with over five billion words), 3) 4,986 texts from British broadsheet papers (3.8 million words) and 4) 2,384 texts from tabloids (1.1 million words). An independent analysis of meteorological drought was undertaken using three sources: 1) the England and Wales Precipitation (EWP) series (back to 1800), 2) a statistically reconstructed version of the EWP which is more reliable in the early record (pre-1870), and 3) a high resolution gridded dataset (back to 1862) which is aggregated to NUTS1 regional scales. Meteorological droughts were assessed using the Standardised Precipitation Index, which allowed us to appraise drought intensity, severity and duration. We found an overwhelming agreement between the corpus data and meteorological records. For the very few cases of disparity between the corpus and rainfall data, there were in most cases plausible explanations. All in all, the present study demonstrates the power of the combined approach, presenting evidence on a scale that would not otherwise be possible, thus contributing to a better understanding of how drought is perceived, in addition to how it is traditionally “measured”.

Droughts and Media: when and what do the newspapers talk about the droughts in England?

“…Such an absence of media coverage, despite climatic anomalies favourable for drought conditions in our analysis, is interesting because the UKCEH also confirmed the drought-prone conditions in these years with reports of unusual hydrological events published for the years of 2003-2006, 2010-2012. A similar mismatch between media coverage and drought-prone meteorological conditions was also identified by Dayrell et al (2022) who explored the text corpora about droughts over the last 200 years in the UK. Dayrell et al (2022) suggested that such mismatched derived from the newsworthiness of events: even if the meteorological conditions are favourable for droughts, they may not generate media coverage due to the lagging effect of rainfall shortage to newsworthy drought events.…”

“…We then explored how drought in spring 2012 and summer 2022 were presented in media contents performing close reading of newspaper headlines (c.f. Dayrell et al 2022). We also visualized the topic compositions for each month for these two events, to show the prevalence of topics.…”

“…Recent studies have upscaled the volume and range of data from digital text archives encompassing historical and contemporary corpora and analysing links between media coverage and meteorological conditions. Dayrell et al (2022) examined UK newspaper corpora over the last 200 years to explore the correlation between a drought index (i.e., SPI) and media coverage. O'Connor et al (2022) examined the relationship between drought indices (i.e., SPI and SSI) and historical newspaper articles in Ireland and found the best predictor of drought-related articles to be SPI-3.…”

Reassessing and Extending the Composite Rainfall Record of Manchester, Northwest England: 1786–Present

“…The prolonged period of inhomogeneity in the Manchester and HadNWEP comparison (Figure 4b), marked by statistically significant breakpoints in 1933 and 1948, occurred during a 16-year period of the Whitworth record. Notable droughts were recorded within the UK in both 1933-35 [45,46] and 1946-49 [47]; however, the latter was poorly documented and reported on in the national press because of wartime reporting restrictions, as food rationing was in place. The station metadata provides no information clarifying the source of the detected inconsistency.…”

Relating drought indices to impacts reported in newspaper articles

“…(2017) also used newspaper collections to verify the occurrence and duration of historical droughts. The utility of newspaper articles as a source of information on drought impacts has also been demonstrated in the UK (e.g., Dayrell et al ., 2022) and elsewhere (e.g., Llasat et al ., 2009; Linés et al ., 2017; Brázdil et al ., 2019).…”