Modeling the flow of 15N after a 15N pulse to study long-term N dynamics in a semiarid grassland.

Many aspects of nitrogen (N) cycling in terrestrial ecosystems remain poorly understood. Progress in studying N cycling has been hindered by a lack of effective measurements that integrate processes such as denitrification, competition for N between plants and microbes, and soil organic matter (SOM) decomposition over large time scales (years rather than hours or days). Here I show how long-term measurements of 15N in plants, microbes, and soil after a one-time addition of 15N ("labeled" N) can provide powerful information about long-term N dynamics in a semiarid grassland. I develop a simple dynamic model and show that labeled-N fractions in plant and microbial-N pools (expressed as a fraction of total N in each pool) can change long after 15N application (> or = 5 years). These 15N dynamics are closely tied to the turnover times of the different N pools. The model accurately simulated the labeled-N fractions in aboveground biomass measured annually during five years after addition of 15N to a semiarid grassland. I also tested the sensitivity of five different processes on labeled-N fractions in aboveground plant biomass. Changing plant/microbial competition for N had very little effect on the labeled-N fraction in aboveground biomass in the short and long-term. Changing microbial activity (N mineralization and immobilization), N loss, or N resorption/re-translocation by plants affected the labeled-N fraction in the short-term, but not in the long-term. Large long-term effects on the labeled-N fraction in aboveground biomass could only be established by changing the size of the active soil-N pool. Therefore, the significantly greater long-term decline in the labeled-N fraction in aboveground biomass observed under elevated CO2 in this grassland system could have resulted from an increased active soil-N pool under elevated CO2 (i.e., destabilization of soil organic matter that was relatively recalcitrant under ambient CO2 conditions). I conclude that short- and long-term labeled-N fractions in plant biomass after a 15N pulse are sensitive to processes such as N mineralization and immobilization, N loss, and soil organic matter (de-)stabilization. Modeling these fractions provides a useful tool to better understand N cycling in terrestrial ecosystems.