RESUMO
Policies and management decisions in the marine environment are driven in part by public sentiment which can grow more intense during hazard events like Harmful Algae Blooms (HABs). The public conversations on social media sites like Twitter (before X) reveal the polarized nature of HABs through nuanced language and sentiment. This article uses mixed methods of machine learned topic modeling and inductive qualitative coding to describe the ways the long-term 2017-2019 Karenia brevis "red tide" bloom were politicized across Florida's South West coast. It finds that there are topical differences in keywords related to place (e.g. beach, Florida, coast), agent (individual or organization), and epistemic values (reliance on scientific and/or media reports). These topical differences demonstrate different levels of politicization and partisanship in qualitative analysis. Conceptually, this research demonstrates the ways different dimensions of a long-duration marine hazard can be polarized. Regarding management, this research provides insights to political and organizational stakeholders and the gaps in the discourse shaping marine hazards which can be used to strategically guide future social media engagement to manage politicization. What if all the careful work that resource and environmental managers do can be undone by simple, seemingly uncontroversial words? In an era of increased environmental and marine distress-coupled with short format communication-the ways environmental managers choose their words is crucial, even between ostensibly inconsequential nouns like "red tide" or "algae bloom." Policies and management decisions in the marine environment are driven in part by public sentiment which can grow more intense during hazard events like Harmful Algae Blooms (HABs). The public conversations on social media sites like Twitter (before X) reveal the polarized nature of HABs through nuanced language and sentiment. This article relies on mining social media posts, and uses mixed methods of machine-learned topic modeling and human-driven inductive qualitative coding to describe the ways the long-term 2017-2019 Karenia brevis "red tide" blooms were politicized across Florida's South West coast. It finds that there are topical differences in keywords related to place (e.g. beach, Florida, coast), agent (individual or organization), and epistemic values (reliance on scientific and/or media reports). These topical differences demonstrate different levels of politicization and partisanship in qualitative analysis. Conceptually, this research demonstrates the ways different dimensions of a long-duration marine hazard can be polarized. Regarding management, this research provides insights to political and organizational stakeholders and the gaps in the discourse shaping marine hazards which can be used to strategically guide future social media engagement to manage politicization.
