Forest density tracks the agent of disturbance more consistently than its risk across the United States

Preprint, not peer reviewed. Thinning to lower stand density is a primary tool for reducing forest disturbance risk, yet whether relative density is associated with how often forests are disturbed, by what agent, and how severely has not been tested at continental scale. Across 148,604 remeasured inventory conditions in the conterminous United States, we related relative density, crowding scaled to the site maximum, to all three. Density was a weak correlate of occurrence, moving its probability four to eight percentage points to a shallow peak near half of maximum density, and its direction differed among ecoregions. It was a consistent correlate of cause, as the insect share of disturbances rose about 7 to 12 percentage points and the fire share fell 5 to 12, while per-event severity changed little. Measuring density over the plot instead of the occupied area inflated the apparent effect. Within a growth-based range, density targets could follow the likely agent.