With conditional sampling, atmospheric fluxes are estimated from the differ
ence in mean concentration of a gas between upward and downward moving eddi
es and a measure of the standard deviation of the vertical wind velocity. T
wo basic components of a conditional sampling system are a sonic anemometer
to measure the direction and standard deviation of the vertical wind veloc
ity, and a valve to separate air sampled from updrafts and downdrafts. Use
of the Campbell Scientific CSAT3 three-axis sonic anemometer to measure the
vertical direction of the eddies and to trigger a splitter valve requires
building a physical delay into the air sampling line to counter the electro
nic delay that is built into the anemometer, This physical delay in the for
m of an intake tube creates a source of dispersion that may reduce concentr
ation differences and affect estimates of fluxes. Our objectives were to de
termine the proper match of electronic and physical delays to separate eddi
es, and then to measure the amount of dispersion from the combined mixing i
n the tube and the splitter valve. A conditional sampling system to measure
CO2 noses was constructed using a CSAT3 anemometer and a three-may valve.
A physical delay was created by adding tubing on the upstream side of the v
alve. A, simple timing circuit was used to control a highspeed valve that p
ulsed CO2 in front of the intake tube to mark the locations of eddy transit
ions. Differences in concentrations between updraft and downdraft lines wer
e measured as now rates and electronic delays were varied, Tube length base
d on electronic delay, now rate, and intake tube geometry was found to be a
dequate for separation of updrafts and downdrafts. Dispersion in the intake
tube was significant and predictable. The effect of the splitter valve on
dispersion was minimal. A correction factor to adjust concentration differe
nces in the lines for dispersion may be warranted, but only for eddy revers
al frequencies > 2 Hz.