To separate the sources of respired CO2, a series of chambers are often installed and
CO2 effl uxes monitored at frequent intervals from soil, stem, branches, and leaves. Alternatively,
respiration sources can be identifi ed by monitoring the isotopic composition of
carbon (d13C) and oxygen (d18O) in CO2 diffusing into the turbulent transfer steam. The
scientifi c value of full system analyses with eddy correlation techniques has proved
immense, particularly when conducted over a series of years (Goulden et al., 1996). Eddyfl
ux installations therefore serve for testing the underlying assumptions and accuracy of
stand-level ecosystem models.