Landscape-level likelihood estimation of eastern spruce dwarf mistletoe (Arceuthobium pusillum) infestations in lowland black spruce (Picea mariana) forests of Minnesota, USA

Biotic disturbance agents are important factors influencing forest dynamics; incorporating them into management planning requires a detailed understanding of their distribution, prevalence, and effects on stand dynamics. However, this information can be difficult to collect in remote forest systems, such as lowland black spruce (Picea mariana (Mill.) B.S.P.) forests affected by eastern spruce dwarf mistletoe (Arceuthobium pusillum Peck, hereafter ESDM). In such cases, predictive modeling is often needed to inform management decisions and address forest health questions. Here, we used two publicly available datasets to predict areas where black spruce is more likely to be infested with ESDM in northeastern Minnesota, USA. Using random forest modeling and logistic regression, we found location, stand age, basal area, site index, average diameter, and metrics of species composition to be among the most important predictors of ESDM occurrence. Predictions showed two regions of greater likelihood of infestation with distinct ecological characteristics and ownership patterns. By understanding how stand structural characteristics relate to ESDM infestations, managers can improve monitoring and management of ESDM at the stand and landscape scales. Additionally, our approach of using multiple datasets and modeling methods can serve as a framework for decision-making on other forest health concerns.