When knowing is not enough: weather literacy, alert comprehension and the limits of warning-driven behaviour change.

The effectiveness of weather warning systems depends not only on forecast accuracy but on the full communication chain — from alert issuance to public understanding to protective action. This paper presents findings from a survey (2.020-respondents) conducted in 2025 among users of the LaMMA meteorological service of Tuscany Region, Italy, which hosted a section aimed at examining how citizens receive, interpret and respond to official weather alerts, and what role meteorological literacy plays across each of these stages.

The survey population is self-selected and weather-engaged, consulting forecasts daily and preferring alert updates through messaging platforms such as WhatsApp and Telegram rather than institutional channels they must actively seek out. The analysis examines how meteorological literacy relates to alert comprehension, confidence in personal risk assessment, and preparedness to act — and whether these cognitive dimensions translate into actual behaviour change across different hazard types and alert levels, from yellow to red. The study also explores the role of alert fatigue, drawing on both quantitative patterns and open-ended responses, to understand what erodes willingness to respond even among motivated users.

Underlying these patterns is a structural tension that extends well beyond Italy. Warning communication remains predominantly framed around meteorological processes rather than human impacts, and alert language reflects this: when asked to choose between technical terminology and plain impact-oriented alternatives, fewer than one percent of respondents prefer current institutional wording. Yet shifting toward impact-based communication is not simply a matter of rewording. Impacts-based warning communication requires spatially detailed knowledge of exposed elements — infrastructure, population, land use — that is rarely available at the resolution needed to match meteorological forecasts, and even more rarely integrated into operational warning workflows. The practical result is that impact descriptions tend to be generic lists, repeated across events and territories, that citizens quickly learn to recognise as routine rather than informative. This dynamic reinforces alert fatigue: when consequences feel disconnected from local reality, warnings lose credibility regardless of their meteorological accuracy. Closing the gap between comprehension and protective action will therefore require not only clearer language but a deeper integration between meteorological services, territorial risk knowledge and civil protection — so that warnings speak not just to what the atmosphere will do, but to what it will specifically mean for each place and its people.