Social Media and Climate Change

Abstract

Climate change is an omnipresent issue in politics, industry, science, and society. Scientist knew of its existence for several decades, but only recently has the perception of it shifted from a solely environmental problem to a matter of international security. Climate change is thought to impact the likelihood of extreme weather events, intensity of droughts, rising sea levels and more. In this way climate change contributes to ongoing conflicts around the world by destroying resources or forcing people out of habitats, creating tensions among groups of people.

As the impacts of climate change are being felt more directly, calls for bolder climate action are getting louder. At the same time, there is also resistance by climate skeptics in industry, government, and among voters. Although most people appear aware and concerned with the development of climate change, the attitude towards climate action is extremely broad. Political rhetoric and disinformation in the media compound the disagreement on climate change, causing polarization. Although some polarization is beneficial for democratic deliberation and for mobilizing voters, if polarization divides the electorate into mutually distrustful groups, it can trap societies in a stifling gridlock or gradually erode democratic institutions. These pernicious effects are what necessitates closer study to monitor societal polarization, prevent conflict and improve climate action efforts.

Social media offers huge amounts of textual data to study discourse on climate change, which this study leverages to make first steps in understanding what topics make up the public discourse, how it
varies across linguistic regions (English, German, Dutch) and how it changes considering climate-related events, namely floods and a climate protest. Using both topic modelling and sentiment analysis
on Twitter data, it could be identified that activity spiked around major events: namely the 2019 Climate Strike and the 2021 Flood in Western Europe. However, reaction to events appears to be
mainly regional with reaction to floods seen only in the Dutch and German tweets. The English Twittersphere seemed to be mainly skewed towards events in America. Polarization appeared low as
most tweets were neutral. Examining only tweets with high sentiment, a different behavior was detected in terms of news coverage as there existed more high negative sentiment tweets, but almost
no high positive sentiment tweets. During the 2019 Global Climate Strike, extreme sentiment tweets went both positive and negative, possibly indicating more intense discourse and a more polarized
atmosphere.

Overall, results show that conflict-triggering climate events are very likely to be domestic events. Thus, from a security perspective, it suggests on the short-term to focus on the monitoring of domestic
actors. On the long-term however, policies need to be implemented which address underlying factors that make society susceptible to polarization, such as socioeconomic inequalities and exclusion,
disinformation in media or corruption. Lastly, from an academic perspective, more research on other dimensions of polarization needs to be conducted. Polarization is a complex phenomenon, whose full
implications cannot be captured by the sentiment metrics. Thus, research moving forward should strive towards a more refined multi-method approach to further investigate climate change polarization and its conflict potential. A better understanding could benefit practitioners in many areas such as security and social policy, for strategic planning regarding climate change.