Despite the enormous size of the organic nitrogen (N) pool contained in mineral subsoils, rates of N cycling and soil exoenzyme activities are rarely measured in soils below 10 or 20 cm depth. Furthermore, assumed relationships between N mineralization rates and the activities of various decomposition exoenzymes are poorly characterized. We measured rates of gross and net N mineralization and nitrification as well as the potential activities of hydrolytic and oxidative enzymes at five soil depths (forest floor to 50 cm) in Spodosols at three hardwood forests of varying age (45 and 100 years post-harvest and old growth) at and near the Hubbard Brook Experimental Forest in New Hampshire, USA. As expected, rates of N cycling and potential enzyme activities per unit soil mass correlated strongly with soil carbon (C) concentration, and these parameters declined exponentially with increasing soil depth. After normalization per unit soil organic matter, N cycling rates and specific enzyme activities generally decreased little with depth within the mineral soil. Gross N mineralization rates correlated with specific activities of those enzymes that hydrolyze cellulose (β-glucosidase, cellobiohydrolase) and N-rich glucosamine polymers (N-acetylglucosaminidase), but not those that degrade protein or more complex C compounds. Hence, gross N cycling appear associated with the N released during microbial N recycling, rather than from decomposition of soil organic matter. Across the three stands, the youngest had a larger ratio of N- to-phosphorus-acquiring enzyme activities, indicating a greater N demand in younger than older forests. For all three stands, mineral soil below 10 cm contributed 30–53% of total gross and net N cycling per unit area to 50 cm depth. Overall, even though microbial N cycling and enzyme activities per unit soil mass decreased with depth, microbial processes in subsoils contributed substantially to ecosystem-scale gross N fluxes because of the sustained microbial activity per unit soil organic matter at depth and the large size of the organic matter pool in the mineral soil. These results support the inclusion of often-ignored mineral subsoils and microbial N recycling in both ecosystem N budgets and in model simulations, due to their contribution to soil N fluxes and the importance of microbial N dynamics in forest stands.
