Meta-analysis of wildfire science applying network analysis and topic modeling

The increasing threat posed by wildfires has sparked considerable interest in diverse fields, from ecology to atmospheric science. Like fire itself, this body of research demonstrates complex but identifiable trends, divergences, and heterogeneities. In this meta-analysis, we applied mathematical techniques from network analysis and topic modeling to investigate the current state of research on wildfires, presenting as a case study the network structure around several novel models of fire. Structures of citations, collaborations, and language-use were analyzed using data accessed from arXiv's API (R package aRxiv). Of particular interest were the boundaries ("firebreaks") dividing disciplines, sub-disciplines, and topics. Findings were interpreted with two main goals: identifying productive avenues for interdisciplinary research, and posing genuine questions to the community of fire researchers.