Do forests receive occult inputs of nitrogen?

The nitrogen (N) cycle of forest ecosystems is understood relatively well, and few scientists expect that major revisions will be necessary; most current work on N cycling focuses on improving the precision estimates of pools and fluxes, or measuring the magnitudes of well-known pools in response to management or disturbances. However, in the past few decades more than a dozen articles in refereed journals have claimed very high rates of N input, far beyond the rates expected for known sources of N. In this review, we summarize the literature on N accretion rates in forests that lack substantial contributions from symbiotic N-fixing plants. We critique each study for the strength of the experimental design behind the estimate of N accretion and consider whether unexpectedly large inputs of N really occur in forests. Only 6 of 24 estimates of N accretion had strong experimental designs, and only 2 of these 6 yielded estimates of >5 kg N ha^-1 y^-1. The high accretion estimates with a strong experimental design come from repeated sampling at the Walker Branch Watersheds in Tennessee, where N accretion rates ranged from 50 to 80 kg-N ha^-1 y^-1 over 15 years after harvesting. At the same location, an unharvested stand showed no significant change. We conclude that there is no widespread evidence of high rates of occult N input in forests. Too few studies have carefully tested for balanced N budgets in forests (inputs minus outputs plus change in storage), and we recommend that at least a few of these studies be undertaken on soils that permit high precision sampling.