Implications of weather-induced tree mortality on forest carbon dynamics based on remeasured forest inventory plots in the Eastern United States

Forest tree mortality plays an important role in the global carbon budget through so-called 'background' mortality rates and larger, less frequent mortality events. The actual mortality turnover rates of forest biomass are not well understood and can vary with forest type, stand characteristics, and environmental conditions. Different agents, such as fire, insects, disease, and weather, operate on different time scales and have different effects on ecosystems. These differences make it difficult, but important, to determine a continuum of return frequencies for agent-specific mortality, especially when making projections of forest carbon balance. Some regional and global ecosystem models include a separate fire component to account for burn emissions, but events such as hurricanes can also influence carbon dynamics and are not simulated. Thus, the effects of potential changes in hurricane frequency and intensity over time would not be captured by existing models. Furthermore, many regional and global ecosystem models assume a single, non-fire mortality rate for all forests, which likely introduces bias to projections of forest carbon balance. Using the United States Forest Service (USFS) Forest Inventory Analysis DataBase (FIADB) we estimated historic (~1970 - 2010) mortality rates for Eastern United States forests. We present spatially-explicit estimates of total mortality and of agent-specific mortality due to insects, disease, fire, weather, and harvest. These estimates show that uniform mortality rates in ecosystem models might be improved if varied spatially. The relative contribution of weather-induced mortality indicates that it results from smaller, more frequent events in addition to the effects of more extreme events such as hurricanes. Evidence of relatively high, hurricane-induced mortality suggests that the effects of extreme weather events should be explicitly modeled.