Accuracy of tracer measurement of gas-desorption rates

Methods for measuring the gas-desorption rate for flowing water in streams involve the use of different gases (krypton-85, propane, and ethylene), injection procedures (slug and continuous), and computational approaches. For slug injection, the gas-desorption rate may be computed by determining the change in the peak concentration at upstream and downstream measurement locations (peak method) or by determining the change in the area under the gas concentration versus time curve at upstream and downstream measurement locations (area method). This paper summarizes the results of comparisons of paired measurements of the reaerati on-rate coefficient (K 2 ) utilizing the different tracer-gas methods as follows: 108 K 2 pairs from the propane- and ethylene-peak methods, 32 K 2 pairs from the propane- and ethylene-area methods, 43 K 2 pairs from the propane-peak and -area methods, 52 K 2 pairs from the ethylene peak and -area methods, 8 K 2 pairs from the propane-peak and krypton-85 methods, and 9 K 2 pairs each from of the krypton-85 and the propane-area and ethylene-peak and -area methods. The methods utilizing propane and/or ethylene yield unbiased estimates of K 2 . The mean absolute relative percent difference in the estimated K 2 values obtained from tracer-gas methods ranged between 11.3 and 20 percent. Sixty-seven percent of the paired K 2 measurements had relative percentage differences (RPD) with absolute values less than 20 percent and errors relative to the mean of the measurements less than 10 percent. On the basis of the data analyzed, the standard error of tracer-gas measurements of gas desorption is approximately 10 percent.