The Sensitivity of North American Terrestrial Carbon Fluxes to Spatial and Temporal Variation in Soil Moisture: An Analysis Using Radar-Derived Estimates of Root-Zone Soil Moisture

This study examines the impact of variation in root-zone soil moisture (RZSM), a key component of the Earth's hydrologic cycle and climate system, on regional carbon fluxes across seven North American ecosystems. P-band synthetic aperture radar-derived RZSM estimates were incorporated into the ecosystem demography (ED2) terrestrial biosphere model through a model-data blending approach. Analysis shows that the model qualitatively captures inter-daily and seasonal variability of observed RZSM at seven flux tower sites (r = 0.59 +/- 0.26 and r = 0.70 +/- 0.22 for 0-10 and 10-40 cm of soil layers, respectively; P < 0.001). Incorporating the remotely sensed RSZM estimates increases the accuracy (root-mean-square deviations decrease from 0.10 +/- 0.07 and 0.09 +/- 0.06 m(3)center dot m(-3) to 0.08 +/- 0.05 and 0.07 +/- 0.03 m(3) center dot m(-3) for 0-10 and 10-40 cm of soil layers, respectively) of the model's RZSM predictions. The regional carbon fluxes predicted by the native and RZSM-constrained model were used to quantify sensitivities of gross primary productivity, autotrophic respiration (R-a), heterotrophic respiration (R-h), and net ecosystem exchange to variation in RZSM. Gross primary productivity exhibited the largest sensitivity (6.6 +/- 10.7 kg center dot cm(-2)center dot year center dot theta(-1)) followed by R-a (2.9 +/- 7.3 kg center dot cm(-2)center dot year(-1)center dot theta(-1)), R-h (2.6 +/- 3.1 kg center dot cm(-2)center dot year(-1)center dot theta(-1)), and net ecosystem exchange (-1.7 +/- 7.8 kg center dot cm(-2)center dot year(-1)center dot theta(-1)). Analysis shows that these carbon flux sensitivities varied considerably across regions, reflecting influences of canopy structure, soil properties, and the ecophysiological properties of different plant functional types. This study highlights (1) the importance of improved terrestrial biosphere model predictions of RZSM to improve predictions of terrestrial carbon fluxes, (2) a need for improved pedotransfer functions, and (3) improved understanding of how soil characteristics, climate, and vegetation composition interact to govern the responses of different ecosystems to changing hydrological conditions.