Redistributions of 15N highlight turnover and replenishment of mineral soil organic N as a long-term control on forest C balance

A large-scale N-15 tracer study was initiated at the Harvard Forest in 1991 in two forest types (red pine and mixed hardwoods) as a means to test hypotheses concerning long-term dynamics in ecosystem-level N cycling and carbon-nitrogen interactions. Here we describe the application of a biogeochemical process model TRACE, with the ability to simulate 15N tracer redistributions, to help interpret the field study and explore its ramifications. We had three main goals: (1) to compare field results of 8-year time series in (NH4)-N-15 and (NO3)-N-15 redistributions against previous model predictions; (2) to gain insight into ecosystem C/N interactions through an iterative set of model changes and direct model-data comparisons; and finally (3) to forecast temporal dynamics in the future effects of elevated N inputs on altered C storage in the regionally representative hardwood forest. Model interpretations of field-observed N-15 redistributions indicated that mineral soil organic matter contains a fraction that retains illuviated N-15 rapidly (within 1 year), then releases some of this 15 N for plant uptake through the following 5-8-year period. Our simulations also suggested that the mineral soil supplied a long-term source of N for the aggrading pools of N in vegetation and the O horizon over the course of stand development. The model structure that best fits the decadal-scale field data for pools and fluxes of C, N, and N-15 forecasted an elevated C storage relative to elevated N inputs that is much lower than published estimates based on ecosystem stoichiometry. TRACE forecasted a maximum differential C storage in N-amended plots of 725 g C m(-2), occurring largely in living and dead wood, peaking 30 years after the start of N amendment treatments of +5 ga N m(-2) per year (a cumulative + 150 g N m(-2)). This amounts to a ratio of elevated C storage to cumulative, elevated N inputs of less than 5:1 over the 30-year period. These results imply that mineral soil supplies much of the N needed for forest aggradation, partially regulating changes in ecosystem C storage, and that elevated N deposition may cause relatively small amounts of elevated C storage after a time lag of decades. (C) 2004 Elsevier B.V. All rights reserved.