“…Ultimately, FVS was chosen due to its (1) wide applicability and technical accessibility to a variety of audiences (i.e., growth and mortality models can be calibrated to specific geographic areas of the United States and twenty geographically-specific FVS variants exist), (2) ease at which it allows users to test alternative hypotheses related to different treatments and disturbances, and assess future forest conditions, and (3) long history of applications in forest resource assessment and planning (Crookston and Dixon, 2005). The FVS is not without its well-documented limitations, such as weaknesses in predicting crown and canopy fuel characteristics (e.g., Cruz and Alexander, 2010;Keyser and Smith, 2010), shortcomings when simulating responses to canopy gap-related disturbances (e.g., Arseneault and Saunders, 2012), inability to use spatially-explicit information or spatially-explicit predictions (e.g., Chivoiu et al, 2006), and general inaccuracies and imperfections in the underlying growth model (e.g., Ex and Smith, 2014;Petrova et al, 2014;Dickinson et al, 2019). Ultimately we selected FVS as the best choice of underlying model that adequately generalized forest stand development and its response to silvicultural treatments, as well as provided robust estimates that could be used to assess the risk of state transitions and predict the impact of management on an ecosystem's response to future disturbances.…”