Diving into the Linguistic Melting Pot: Using GIS to Identify Limited English Proficiency Communities Across NWS Warning Areas

The United States is known as “The Great Melting Pot”, for the wide variety of communities, cultures and languages from around the world present in the country. While this has many benefits, it also comes with challenges. Across the United States, approximately 25 million (or nearly 1 in 10) Americans are considered to have Limited English Proficiency (LEP) (i.e., they speak English less than “very well”). However, knowledge of where these populations are located is quite limited. In the case of the National Weather Service (NWS), removing barriers to messaging is one of the most prevalent challenges in weather operations today. While the NWS translates some weather information from English to Spanish and Samoan, as well as providing outreach material in Spanish at a handful of offices, it does not cover all languages and the translation work is done manually. This paradigm leaves the NWS with a substantial service coverage gap. The NWS cannot accomplish its goal of a “Weather-Ready Nation” without understanding the communities they serve and how they utilize information to empower their response.
The NWS Automated Language Translation Project is currently investigating how to leverage language models trained by artificial intelligence (AI) to translate English NWS products at scale. This technology can then be incorporated into NWS operations to address this major gap in service to multilingual communities across the nation. To help with these efforts, we used Geographic Information Systems (GIS) to identify where these multilingual populations are located to contribute to the strategic expansion plan for translation technology implementation across NWS Weather Forecast Office (WFO) County Warning Areas (CWAs).
In our presentation, we will present maps of the linguistic representation of NWS CWAs in several languages, including but not limited to Spanish, Vietnamese, Mandarin, Tagalog, among others. Using methodologies in geography, we explore fascinating spatial analysis of multilingual “hot spots” and note regions that are in need of further resources and support due to their higher vulnerability and underserved nature, caused by language barriers as well as other socioeconomic disadvantages. We will also provide a first look at the GIS Dashboard and show how the NWS can take advantage of this tool for local NWS community engagement and outreach across the country.