World Data Centre for Greenhouse Gases (WDCGG)

Category	Air sampling observation
Sampling Type	flask
Sampling Height/Depth
Sampling and Analysis Frequency	Weekly
Sampling Environment
Measurement Method	NDIR
Current status and history of Instruments	NDIR
Description of Instruments	SIemans U3
Time Zone	UTC
Data Period	1975-07-01 - 2005-12-28
Data Type	event, daily, monthly
Parameter detail
Other Descriptions for Sampling and Analyses	Environment Canada initially used evacuated 2 L glass flasks fitted with 6 mm bore high vacuum stopcocks (lubricated with Apiezon-N grease) at Alert, Sable Island and Cape St. James. A new flask type with a glass barreled valve with VitonR o-ring seals was introduced at Alert and Estevan Point in 1992. It is more robust for field sampling and easier to open in extreme cold temperatures than the greased stopcock flasks. EC introduced a third flask type into its sampling network at Alert and Estevan Point in 1993 and at Sable Island in June 2003. This flask has dual glass barreled valves with VitonR o-ring seals. Using a pumping unit, this flask is flushed and filled inside the laboratory to 15 psi with air drawn in from a sampling line extending to the top of nearby towers. In 1988, EC assumed responsibility for the flask programs at Alert, Cape St. James, and Sable Island from the Insitute of Ocean Sciences, British Columbia. The analyses were initially made using a Hartman and Braun URAS-3E analyzer. In October 1989, the analyzer was changed to a Maihak UNOR-6N. All analyses were made using CO₂-in-air reference gases. Three CO₂ standard gases (340 ppm, 355 ppm and 365 ppm) and a reference gas (330 ppm) were used on the analysis system. The mixing ratios of the standard gases were calibrated at EC. The Environment Canada CO₂ flask analysis system was designed to analyze a maximum of eight flasks per run. An 8-port stream selection valve was used to switch between flasks while a vacuum pump (KNF Neuberger pump, Model UN022-SV1) was used to extract the air from the flasks. The three calibration gases and eight flasks were switched manually. A Campbell Scientific CR21x data logger was used to collect the raw data. All measurements were made on a dry air basis, with both sample and reference gas air streams passing through stainless steel traps (t inch diameter) submerged in a methanol bath at t55o C. Only one aliquot was extracted and analyzed from each flask. A typical flask run sequence consisted of the calibration gases (C1 to C3), followed by eight flasks, and again followed by the calibration gases (C1 to C3). For each run, the voltages of the calibration gases, before and after the flasks, were averaged. The averaged voltages and mixing ratios assigned to the calibration gases were used to generate a quadratic system response function for the run. The mixing ratios for the flasks were calculated by substituting the flask voltages into the system response function for each flask run. Minimal or no drifts were observed in the bracketing voltages of the calibration gases. Changes were made to the flask analysis system in April 1991. The UNOR-6N analyzer was replaced with a Siemens Ultramat III analyzer. The change in analyzer was necessary because the UNOR-6N was sensitive to subtle changes in sample flow, which occurred even though mass flow controllers were used. Because the Siemens analyzer had a larger sample cell volume (100 cm3) compared to the UNOR-6N analyzer (8 cm3), it was necessary to increase the flushing volume from each flask. The system was modified for ¡Èstatic¡É measurements to compensate for the extra volume required for flushing. To achieve static measurements, the system is flushed with gas and then a solenoid valve shuts off the gas flow to the sample cell. The gas in the sample cells equilibrates to ambient pressure before a measurement is taken. Static measurements require less gas from a given flask. Changes to the system included the addition of a 10-port stream selection valve, a target gas (350-360 ppm), two additional calibration gases (340-380 ppm), and a solenoid valve upstream of the sample cell. The reference gas, five calibration gases, target gas, and the outlet of the flask stream selection valve were connected to the 10-port valve. At present, a typical flask run has the following sequence: reference, calibration gases (C1-C5), reference, target, flasks (F1-F8), reference, calibration gases (C1-C5), reference and target. Prior to running the flask analysis system on any given day, the regulators on the reference tanks, calibration tanks and target tanks are flushed with about 3 L of gas. The gas is also passed through the system to stabilize the analyzer. To extract air from the ambient pressure flasks, an automated sequence is programmed into the data logger to control 1) the switching of the stream selection valves, 2) the opening and closing of the solenoid and 3) turning the pump on and off. The data logger also collects the raw data and calculates the average and standard deviation voltages for each sample.
Scientific Aim
Supporting Contributor(s)
Last update	2006-10-10
Situation

Category	Parameter	Country/Territory	Contributor

Updated in
the last 365 days

This site is maintained by the Japan Meteorological Agency
in cooperation with the World Meteorological Organization
(Created : 2001/07/02 Modified : 2010/09/02)

WMO World Data Centre for Greenhouse Gases
c/o Japan Meteorological Agency
1-3-4, Otemachi, Chiyoda-ku
Tokyo 100-8122, Japan

Tel:	+81-3-3287-3439
Fax:	+81-3-3211-4640
E-mail: