Actionable Weather Communication Recommendations for Hard-to-Reach Populations Using Topic Modeling and Thematic Analysis
Abstract Language barriers, geographic isolation, and limited digital access are some of the obstacles that hard-to-reach populations in the United States face when trying to access weather and climate information. These conditions restrict equitable access to the public benefits mandated by federal weather policy and public warning systems. We identified actionable, evidence-based strategies for strengthening hazard communication for these populations. We conducted topic modeling on 7,705 peer-reviewed article abstracts and an in-depth thematic analysis of 129 peer-reviewed studies documenting effective approaches for engaging groups with differentiated information needs. The thematic analysis produced four integrated recommendations: community-embedded communication, hybrid and redundant communication systems, cross-sector coordination, and inclusive and actionable communication. Across these themes, studies emphasized strategies such as mobile service delivery, low-tech communication channels, culturally specific messages, trusted messengers, and participatory design. The results show how broader federal priorities for improving forecast communication and public response can be advanced and hazard communication improved through practices validated in health and social services. When taken as a whole, these insights provide communicators, emergency managers, and legislators with scalable ways to create more resilient and inclusive hazard information systems.
Authors
- Melissa A. Kenney (ORCID: https://orcid.org/0000-0002-2121-8135)
- Satyandr Kumar Sharma (ORCID: https://orcid.org/0000-0002-3122-5726)
- Tara Conway
- Ashley Bergman Humphrey
Institutions
- University of Minnesota (US)
Publication Details
- Journal
- Natural Hazards Review
- Published
- 2026-09-30
- DOI
- https://doi.org/10.1061/nhrefo.nheng-2843
- Primary Topic
- Computational and Text Analysis Methods
- Type
- article
- Field-Weighted Citation Impact
- 0.00