Abstract
Carbon (C) fluxes through roots are the most uncertain of all C exchanges between the atmosphere, plants, and soil. Yet the three dominant methods to characterize root C fluxes (minirhizotron, sequential coring, and isotopes) yield significantly different estimates of temperate forest root mortality turnover times. We contend that these discrepancies result from limitations in interpreting these very distinct types of observations. In this study we used a whole-ecosystem 14C label to develop, parameterize, and test a model (Radix1.0) of fine-root mortality and decomposition. Radix simulates two live roots pools (one with structural and non-structural C components), two dead root pools, non-normally distributed root mortality turnover times, a stored C pool, seasonal growth and respiration patterns, a best-fit to measurements approach to estimate model parameters, and Monte Carlo uncertainty analysis. We applied Radix at a temperate forest in Oak Ridge Tennessee using 14C measurements from two root size classes (<0.5 mm and 0.5−2.0 mm) and three soil depth increments (O horizon, 0−15, and 30−60 cm). Predicted root lifetimes were 0.1-0.9 y and 11-14 y for fast and slow live root pools respectively, and 0.1-4 y and 11-14 y for fast and slow dead root pool decomposition turnover times, respectively. We estimated that C fluxes through fine roots <2 mm diameter are ~40, 220, and 90 g C m-2 y 1 in the O horizon, 0−15 cm, and 30−60 cm depth intervals, respectively. We conclude that accurate characterization of C flows through fine roots required a model with two live fine-root pools, two dead fine-root pools, and root respiration. Further, root turnover times on the order of a decade imply different response times in biomass and growth than are currently predicted by models with a single annual turnover pool.