Author
|
Submitted to: Agricultural and Forest Meteorology
Publication Type: Peer Reviewed Journal Publication Acceptance Date: 4/12/1998 Publication Date: N/A Citation: N/A Interpretive Summary: There is a critical need within the research community and in environmental monitoring/protection arena, for simple yet accurate methods to measure the exchange of gases and particulates between surface and atmosphere. The most accurate method is eddy covariance, but it requires sensors that can measure the change in concentration of a substance at high frequencies (5 to 10 Hz) and there are no sensors capable of this for most gases of interest. A new method, conditional sampling, uses the principles of eddy covariance but does not require a rapid-response sensor. It uses a sonic anemometer to measure the vertical windspeed, w, at high frequency and to divide sampled air into two sample lines, depending on the sign and magnitude of w. The transport of the trace gas of interest is then calculated by measuring the difference in mean concentration between these two lines. This measurement can be made at low frequency, and hence a wide variety of gases can be measured. However, adoption has been slow because there have been questions about its theoretical basis and its accuracy. Here we show the derivation of the method and its direct linkage to eddy covariance. We describe a possible limitation of the method, its susceptibility to systematic error due to random noise in the measurement of vertical windspeed, but we also demonstrate that this error is insignificant for commercially available sonic anemometers. Finally, we show that the accuracy of the method can be optimized by a simple technique involving the use of a sampling deadband. The information presented should be useful to both scientists and regulatory agencies interested in quantifying the transport of gases and particulates to or from the earth's surface. Technical Abstract: There is a need for simple, robust methods for measuring exchange of trace gases and particulates. Conditional sampling is a new method that has received increasing attention, because it is related to theoretically attractive eddy covariance, but does not require a rapid response sensor for the covariate. It does require rapid measurement of the vertical windspeed, w, and sorting of sampled air into two separate lines on the basis of w. As originally proposed, the flux was then calculated as the product of the mean difference in concentration between the upward and downward moving eddies, the standard deviation of the vertical windspeed, and an empirical coefficient, b. Subsequent exposition showed that b was derivable from the statistics of joint gaussian distribution, although field experiments have consistently found values somewhat lower than the theoretical expectation. Here we reexamine the method, and show that if the eflux is instead expressed as the product of the regression-estimated slope of the concentration-windspeed relation and the variance of w, then it is exactly equivalent to eddy covariance with no need for an empirical coefficient. The aim of conditional sampling then becomes proper estimation of b1as DC/DW. We show that this quantity has a consistent positive bias when samples are sorted simply into positive and negative excursions from mean w. Inclusion of a sampling deadband, symmetric about the mean w, improves the accuracy of the slope estimate and decreases its variance as well. We conclude that conditional sampling is a maturing method, with increasing evidence indicating that the underlying relationships between scalar concentration and windspeed are sufficiently robust to support widespread use. |
