Changing fire regimes are rapidly and extensively reshaping dry forest ecosystems across the Western US. Understanding the implications of contemporary fire regimes for biodiversity is necessary both for the direct conservation of species and to make informed decisions about the costs and benefits of potential forest management actions. However, the hypothesis that atypically large, severe fires in the Sierra Nevada are influencing overall biodiversity – beyond just a few priority species – has proven difficult to test. The confluence of three technologies have created new possibilities: durable, low-cost recording hardware enable landscape-scale monitoring, the machine learning algorithm BirdNET enables the extraction of bird diversity data from the resulting audio, and the Bayesian statistics program Stan enables the implementation of population models capable of accommodating complex avian community data. We conducted a Before-After, Control-Impact study of the effects of the 129,000 ha North Complex Fire on 67 species of diurnal birds. This fire caused site extinction probabilities to increase significantly for 26 species and caused site colonization probabilities to increase significantly for eight species. Thus, this large, high-severity fire has caused both a change in community composition and a net loss of avian biodiversity.