Multimodal narratives of climate denial: A novel, visual-first methodology for analysing conspiracy theory discourse on Instagram

• climate denial discourse is represented on social media multimodally. • dominant visual signatures are isolated using an unsupervised machine learning technique. • the chemtrails conspiracy theory describes the most prevalent visual signature. • chemtrails conspiracy and climate denial discourses intertwine with other conspiratorial discourses through visuals, hashtags, and text. Problematic information about climate change in online spaces can impact public awareness about climate change’s causes, impacts, and mitigation strategies, contributing to declining trust in science and diminished effectiveness of climate policies. Problematic climate discourses often overlap with conspiratorial material in denial communities; members may hold low levels of trust in social institutions. On social media, misinformation travels further and faster when it is multimodal, yet this area of scholarship remains relatively unexamined, particularly on the distinctly multimodal platform, Instagram. This paper uses a corpus of Instagram posts labelled with the hashtag #climatechangehoax to develop a visual-first methodology for analysing problematic climate change content on social media. Combining unsupervised machine learning, co-hashtag analysis, and multimodal critical discourse analysis (MCDA), we identify dominant visual clusters and examine how the semiotic resources present within a cluster’s posts can articulate meaning. The most salient visual signature that we identified aligns climate denial and the chemtrails conspiracy theory; this signature predominantly comprises photographs of skies containing clouds or condensation trails. Within this cluster, other modalities entreat the viewer to use their own senses for proof, together representing an interdiscursive formation where embodied perception (“Look to the skies”) is privileged over scientific authority. Here, visual cues are reconfigured as evidence of elite deception. Hashtags in post text reinforce these conspiratorial meanings and, through co-hashtag networks, structurally align climate denial within an assemblage of anti-elite, post-truth discourse, and the broader conspiratorial community.