Dynamic redistribution of isotopically labeled cohorts of nitrogen inputs in two temperate forests

We compared simulated time series of nitrogen-15 (N-15) redistribution following a large-scale labeling experiment against field recoveries of (NH4+)-N-15 and (NO3-)-N-15 in vegetation tissues. We sought to gain insight into the altered modes of N cycling under long-term, experimentally elevated N inputs. The study took place in two contrasting forests: a red pine stand and a mixed deciduous stand (predominantly oak) at the Harvard Forest, Massachusetts, USA. We used TRACE, a dynamic simulation model of ecosystem biogeochemistry that includes N-15/N-14 ratios in N pools and fluxes. We simulated input-output and internal fluxes of N, tracing the labeled cohorts of N inputs through ecosystem pools for one decade. TRACE simulated the peaks and timing of N-15 recovery in foliage well, providing a key link between modeling and field studies. Recovery of tracers in fine roots was captured less well. The model was structured to provide rapid, initial sinks for (NO3-)-N-15 and (NH4+)-N-15 in both forests, as indicated by field data. In simulations, N in litter turned over rapidly even as humus provided a long-term sink for rapidly cycling N. This sink was greater in the oak forest. Plant uptake fluxes of N in these fertilized plots were on the same order of magnitude as net assimilation fluxes in forest-floor humus. A striking result was the small rate of incorporation of N in humus resulting from the transfer of litter material to humus, compared with large fluxes of N into humus and its associated microorganisms through direct transfers from pools of inorganic N in soils.