Large language models reveal big disparities in current wildfire research

This study uses a large language model (LLM) to analyze over 60,000 wildfire research papers published between 1980 and 2022. The analysis reveals significant geographical and thematic disparities in wildfire research. Western United States, despite having a small percentage of global burned area, accounts for a disproportionately large share of publications. Conversely, regions with extensive burned areas, such as Siberia and Africa, are underrepresented. Similar imbalances exist across various wildfire components, such as fuel types, ignition sources, and climatic drivers. The findings highlight the need for increased transdisciplinary collaboration and AI-aided research to address these disparities and ensure sustainable wildfire management.

Zhengyang Lin, Anping Chen, Xuhui Wang, Zhihua Liu, Shilong Piao