NOTICE (March 2018): This website provides access to the CDIAC archive data temporarily. It will be gradually transitioned into data packages in the new ESS-DIVE archive. This site will continue to operate in parallel during and after the transition, and will be retired at a future date. If you have any questions regarding the data or the transition, please contact

image image image image

Atmospheric Carbon Dioxide Record from Alert

graphics Graphics   data Data


N.B.A. Trivett, V.C. Hudec
Atmospheric Environment Service, Downsview, Ontario M3H 5T4, Canada

C.S. Wong
Marine Carbon Research Centre,
Institute of Ocean Sciences,
Sidney, British Columbia, Canada

Period of Record



Air samples are collected approximately once per week, between 12:00 and 16:00 local time, in a pair of evacuated 2-L thick-wall borosilicate glass flasks. Samples are collected under preferred conditions of wind speed and direction (i.e., upwind of the main station and when winds are strong and steady). The flasks are evacuated to pressures of ~1 × 10-4 mbar or 0.01 Pa prior to being sent to the stations. The air is not dried during sample collection.

The Atmospheric Environment Service (AES) designed a new grab type flask in 1989. The new flasks are 2 L in volume and have single glass-barreled, high-vacuum stopcocks with double Buna-N o-ring seals. The flasks are made of borosilicate glass and have a UV protective plastic coating. Field tests are currently being carried out at Alert to assess the suitability of the new flasks. The new flask is being used during the same sampling period as the greased stopcock flask, and the results are being compared. It is hoped that the new flasks will soon replace the greased stopcock type flasks in the sampling network.

An alternate method of flask sampling, currently ongoing at Alert and Estevan Point, involves pressurizing a flask by using a pumping unit connected to a sample line. The flask has both an inlet valve and an outlet valve so that the flask can be completely flushed with sample air and any contamination can be removed from the flask before it is pressurized with sample air. Because of the possible advantages of this method, a new double-valved sampling flask was designed for use in a pressurized flask sampling program. The new 2-L flasks are composed of borosilicate glass and have two glass-barrelled, high-vacuum stopcocks with double Buna-N o-rings. The flasks are also plasticoated and have a specially designed internally anchored glass delivery inlet tube with outlet ports, to ensure complete flushing of the flask during sampling. A pressurized flask sampling program was implemented at Estevan Point in January 1993 and at Alert in June 1993. The pressurized samples will be directly compared with the evacuated grab samples taken during the same time period at the site.

From 1975 to 1988, flask samples were analyzed for CO2 concentrations at the Institute of Ocean Sciences (IOS) by using a URAS 2T nondispersive infrared analyzer (NDIR). From 1988 through 1991, AES analyzed samples on URAS 3E and UNOR 6N NDIR analyzers. Currently, a Siemans ULTRAMAT III NDIR is used for CO2 flask determinations. All samples are dried before entering the analyzer. Since March 1989, the calibration gases used have been CO2 in natural air mixtures. Before that time, CO2-in-nitrogen and CO2-in-synthetic air mixtures were used for calibration. All of the calibration gases are traceable to the Scripps Institution of Oceanography scales. More details about the sampling methods used in the Canadian flask-sampling program are provided in Trivett and Higuchi (1989).

To be considered indicative of background conditions, flask pairs must agree within 0.6 parts per million by volume (ppmv), and aliquots from the same flask must agree within 0.3 ppmv. Flasks that meet these criteria are retained for time series analysis to determine trends and seasonal cycles. For further details on data selection criteria, see Trivett and Higuchi (1989)

Map showing location of Alert, Northwest Territories, Canada

Alert, Northwest Territories, Canada
Tundra (snow covered for almost
three quarters of the year)
82°28' N, 62°30' W
210 m above MSL


Alert is located ~800 km from the North Pole on the northern tip of Ellesmere Island. Currently, both continuous and grab flask sampling programs are conducted at Alert. The flask data from Alert show an increase in the annual atmospheric CO2 concentration from 341.35 parts per million by volume (ppmv) in 1981 to 357.21 ppmv in 1991. Trivett and Higuchi (1989) reported that the mean annual rate of increase, obtained from the slope of a least squares regression line through the annual averages, for Alert was 1.49 ppmv per year.

Atmospheric CO2concentrations at Alert show a seasonal pattern: the annual minimum occurs during the late summer (August-September); the annual maximum, during the early spring (April-May).


  • Higuchi, K., and S.M. Daggupaty. 1985. On variability of atmospheric CO2 at station Alert. Atmospheric Environment 19:2039-44.
  • Trivett, N.B.A., and K. Higuchi. 1989. Trends and seasonal cycles of atmospheric CO2 over Alert, Sable Island, and Cape St. James, as analyzed by forward stepwise regression technique. In W.P. Elliott (ed.), The Statistical Treatment of CO2 Data Records. Air Resources Laboratory, Silver Spring, Maryland.
  • Trivett, N.B.A., D.E.J. Worthy, and K.A. Brice. 1989. Surface measurements of carbon dioxide and methane at Alert during an Arctic haze event in April, 1986. Journal of Atmospheric Chemistry 9:383-97.
  • Wong, C.S., Y.H. Chan, J.S. Page, R.D. Bellegay, and K.G. Pettit. 1984. Trends of atmospheric CO2 over Canadian WMO background stations at Ocean Weather Station P, Sable Island, and Alert. Journal of Geophysical Research 89:9527-39.

CITE AS: Trivett, N.B.A., V.C. Hudec, and C.S. Wong. 1994. Atmospheric CO2 records from sites in the AES air sampling network. In Trends: A Compendium of Data on Global Change. Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, U.S. Department of Energy, Oak Ridge, Tenn., U.S.A.