2.7. AUTOMATED AIRCRAFT SAMPLING

The aircraft sampling project has been in continuous operation at the Carr, Colorado [40.9°N, 104.8°W] site since November 1992. Until April 1995, profiles of 20 samples each were taken on a biweekly basis. After April, profiles were scheduled on a weekly basis to investigate improved signal to noise from the increased data rate and to gain experience with the logistics of more frequent flights. Approximately 30 flights were made in 1995 alone. An additional change was the use of a Cessna Model T210 aircraft beginning in June 1995 that increased the profile height to 7.9 km typical (9.1 km maximum). An example of a smoothed-curve fit to a profile altitude interval from the last 3 years of data is shown in Figure 2.24. Although the plot shows a good seasonal cycle and trend, the irregularities in the fit and outliers in the data are a reminder of the difficulty in measuring continental air masses and the complexities of individual profiles.

Fig. 2.24. Carbon dioxide in samples obtained between 3-4 km above Carr, Colorado.

Also in 1995, collaboration began with the Aviaecocentre group in Moscow, Russia, to fly the CCG automated sampling system on a Russian Air Force twin-turboprop AN-24. The project acquired six profiles between June and the end of the year at altitudes from 0.3 to 6.1 km at a location about 100 km southeast of Moscow (54.9°N, 35.5°E). Meteorology and air mass back-trajectory data were also recorded. A typical profile from the Moscow site is shown in Figure 2.25. The Russian collaborators are working to develop a climatology for the site that can be correlated with the long-term carbon data. It is expected that this work will continue at least through August 1996.

Fig. 2.25. Vertical profiles of trace species over Yuchonov, Russia, near Moscow, on November 14, 1995.

All of the sampling was performed with the automated sampling package developed by the Carbon Cycle Group. This package continues to evolve with experience and the testing of various components. A newer version of the package uses a more compact valve-actuator system so that each flask can be equipped with two valves and flushed more thoroughly and reliably than evacuated flasks. An internal Global Positioning System receiver was also included in the package for fully automatic operation.

As part of this research and development effort, year-long sample storage tests were performed on our "standard" 2.5-L glass flasks equipped with Teflon o-rings, to measure the trend of trace gas variability. Over a 12-month period, the tests show a 0.4 ppm loss of CO, a 0.5 ppb gain of CH (not significant), a 7.2 ppb gain of CO, a 5.7 ppb gain of H2 (not significant), and a 1.1 ppb loss of NO. These trends imply that the trace gas mixing ratio drift within the sample flask is easily less than the analysis limits if the samples are stored for a month or less, which is our normal operating procedure.
br>

[back] [contents] [next]