Towards Real-Time Assessment of Heatwave Risk via Information-Seeking

We investigate dynamic changes in heatwave-related risk across European regions by leveraging digital social sensing data, specifically Google search interest for heat-related topics. We do this by analyzing high temperature events at national and weekly scales from 2010 to 2019, categorizing them based on high versus low search interest, and contrasting functional temperature-mortality relationships across these event types. This approach allows us to assess how vulnerability evolves not only before but also during high temperature events, moving beyond static representations most common in previous analyses. Given the increased frequency, intensity, and duration of heatwaves due to climate change, mitigation strategies across Europe have evolved. However, residual risk remains, particularly with regard to inefficiencies in communication and behavioural responses. This highlights the need for a better understanding of the dynamic relationships and interactions among risk drivers, particularly the vulnerability component. We employ all-cause mortality data from Eurostat and temperature data from the E-OBS, we focus on NUTS-level regions across Europe to evaluate the potential of information-seeking indicators in capturing real-time shifts in societal risk to extreme heat.

Preliminary findings reveal divergent patterns in all-cause mortality outcomes for similar temperatures but given differences in the intensity of concurrent information-seeking behaviour. This is found across all considered information themes and across climatic and socio-demographic gradients. Notably, regions with lower population density tend to have higher mortality rates during periods of high information-seeking behaviour compared to periods of low information seeking. The opposite is observed for areas with higher population density. This suggests the importance of potential mediating contextual factors, such as urbanisation and adaptive capacity. Further testing of the influence of pre-event information-seeking patterns revealed generally weak and non-significant effects. These results highlight the importance of regional factors and emphasise the value of real-time, during-event information-seeking patterns. Overall, our results emphasise the need to consider dynamic public awareness and population-level information-seeking behaviour in heat risk assessments. The use of social-sensing data emerges as a promising approach to capture these processes, offering actionable, open insights for sustainable resilience strategies in response to heatwaves and other hazards.