Studying attention to IPCC climate change maps with mobile eye-tracking.

Many visualisations used in the climate communication field aim to present the scientific models of climate change to the public. However, relatively little research has been conducted on how such data are visually processed, particularly from a behavioural science perspective. This study examines trends in visual attention to climate change predictions in world maps using mobile eye-tracking while participants engage with the visualisations. Our primary aim is to assess engagement with the maps, as indicated by gaze metrics. Secondary analyses assess whether social context (as social viewing compared to solitary viewing) affects these trends, the relationship between projection types and visual attention, compare gaze metrics between scientific map and artwork viewing, and explore correlations between self-reported climate anxiety scores and attention patterns. We employed wearable, head-mounted eye-tracking to collect data in relatively naturalistic conditions, aiming to enhance ecological validity. In this research, participants engaged with ten world maps displaying near- and far-term climate projections across five data categories, adapted from the online interactive atlas provided by the International Panel on Climate Change (IPCC). To compare scientific information processing with aesthetic perception, participants also viewed two large-scale artworks. Responses to the Climate Change Anxiety Scale (CCAS) were also collected. Participants viewed the displays alone (single-viewing condition, N = 35) or together with a partner (paired-viewing condition, N = 12). Results revealed that the upper parts of the maps, particularly the continental Europe, received significant attention, suggesting a Euro-centric bias in viewing patterns. Spatial gaze patterns were similar between single and paired conditions, indicating that the visual attributes of the maps predominantly shaped attention locations. Although dwell times were comparable, the paired condition showed higher fixation counts, shorter average fixation durations, and longer scanpaths, suggesting a potentially dissociable viewing strategy and more exploratory viewing patterns influenced by social interaction. No substantial differences were observed in attention across projection timeframes or types, although individual variations were noted. Artwork viewing exhibited notably shorter average fixation durations compared to climate map viewing, potentially reflecting different visual engagement styles. Despite positive linear correlations among the four CCAS subscales, there was no apparent correlation between CCAS scores and main gaze metrics, indicating a lack of a direct relationship between self-reported anxiety and gaze behaviour. In summary, visual attention to climate change visualisations appears to be mainly influenced by the inherent visual attributes of the maps, but the social context may subtly influence visual attention. Additionally, the comparison with aesthetic viewing highlights relatively distinct attentional patterns in scientific versus aesthetic engagements.