Logging is a larger cause of adult tree mortality in northeastern U.S. forests than all other causes of mortality combined. We used Forest Inventory and Analysis (FIA) data to develop statistical models to quantify three different aspects of aggregate regional forest harvest regimes: (1) the annual probability that a plot is logged, as a function of total aboveground tree biomass, (2) the fraction of adult tree basal area removed if a plot was logged, and (3) the probability that an individual tree within a plot was removed, as a function of the fraction of basal area removed at the plot level, the species of tree, and its size. Results confirm that relatively frequent partial harvesting dominates the logging regimes, but with significant variation among different parts of the region and different forest types. The harvest regimes have similarities with natural disturbance regimes in imposing spatially and temporally dynamic mortality that varies predictably as a function of stand structure as well as tree species and size.