Flux partitioning in an old-growth forest: seasonal and interannual dynamics.

Turbulent fluxes of carbon, water and energy were measured at the Wind River Canopy Crane, Washington, USA from 1999 to 2004 with eddy-covariance instrumentation above (67 m) and below (2.5 m) the forest canopy. Here we present the decomposition of net ecosystem exchange of carbon (NEE) into gross primary productivity (GPP), ecosystem respiration (R(eco)) and tree canopy net CO(2) exchange (DeltaC) for an old-growth Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco)-western hemlock (Tsuga heterophylla (Raf.) Sarg.) forest. Significant amounts of carbon were recycled within the canopy because carbon flux measured at the below-canopy level was always upward. Maximum fluxes reached 4-6 micromol m(-2) s(-1) of CO(2) into the canopy air space during the summer months, often equaling the net downward fluxes measured at the above-canopy level. Ecosystem respiration rates deviated from the expected exponential relationship with temperature during the summer months. An empirical ecosystem stress term was derived from soil water content and understory flux data and was added to the R(eco) model to account for attenuated respiration during the summer drought. This attenuation term was not needed in 1999, a wet La Niña year. Years in which climate approximated the historical mean, were within the normal range in both NEE and R(eco), but enhanced or suppressed R(eco) had a significant influence on the carbon balance of the entire stand. In years with low respiration the forest acts as a strong carbon sink (-217 g C m(-2) year(-1)), whereas years in which respiration is high can turn the ecosystem into a weak to moderate carbon source (+100 g C m(-2) year(-1)).