Using an ecosystem process model to examine effects of increased nitrogen deposition on soil carbon storage through decreased decomposition

Background/Question/Methods

Increasing anthropogenic nitrogen (N) deposition affects ecosystem carbon storage through changes in community composition, net primary productivity, and soil carbon storage.  Effects of N deposition on soil carbon storage in experimental studies depend on the litter chemical composition and the form, timing, and amount of N addition. We modeled carbon and nitrogen cycling in four sugar maple stands in Michigan using an ecosystem process model (TRACE) and compared model results to field data from a eighteen-year N deposition experiment within each stand. At each stand, simulated N deposition in the field led to increased soil carbon (C)storage corresponding to changes in microbial activity and community composition with no changes in above our below ground litter production. We calibrated C and N cycling within TRACE (Tracer Redistributions Among Compartments in Ecosystems) using field data from one of the four sugar maple stands. We then compared modeled biogeochemical processes within all four stands to field data under both ambient and increased N deposition addition.

Results/Conclusions

TRACE accurately modeled ecosytem pools and fluxes of C and N under ambient N deposition within each stand.  However, TRACE did not initially reproduce changes in soil C storage under simulated N deposition. Using data from previous forest litter decomposition studies, we determined that N additions decrease lignin decomposition rates by ~30% and increased cellulose decomposition by ~9%. These relationships were not affected by litter lignocellulose index or the amount of N addition. After adding these changes in decomposition rate to TRACE, model results did not accurately reproduce increased soil C storage observed in the field. Using the same set of litter decomposition studies to examine changes in decomposition limit values, we determined that N addition increased the amount of litter entering the humus pool. This relationship was dependent on the litter lignocelluloses index, but not the amount of N addition. When the amount of decomposed litter entering the humus pool was increased under simulated N deposition, TRACE accurately modeled soil Cstorage under simulated N deposition. Our results indicate that increases in the amount of litter entering the humus pool before decomposition, rather than slower rates of decomposition, appear to cause greater soil C storage under simulated N deposition.