Air was sampled by semi-automatic filling of boro-silicate flasks stored inside purpose-built suitcases (called ‘programmable flask packages’), which contain an array of 17 0.7-litre flasks at SAN site and 12 0.7-litre flasks at TAB, ALF and RBA. The programmable flask packages are connected to a second suitcase containing batteries and two compressors in series (called ‘programmable compressor packages’), which is connected to an air inlet on the outside of the aircraft. More details of these packages are available at http://www.esrl.noaa.gov/gmd/ccgg/aircraft/sampling.html.

To fill the flasks at our set of pre-determined altitudes, the aircraft pilot initiates sampling by toggling a switch that initiates the pumps in the programmable compressor package and switches flask valves in the programmable flask package. Its manifold is first flushed with 5 litres of air, and then the flask valves are opened and flushed with 10 litres of air. The downstream flask valve is then closed and the samples are pressurized to 260 kPa before closing the upstream valve. The full set of 12 or 17 flasks are filled during one descending spiral profile from 4,420 m to 300 m a.s.l. From altitudes of 4,420 m down to 1,200 m we sampled every 300 m and from 1,200 m downwards we sampled every 150 m down to almost the canopy. Profiles were usually taken between 12:00 and 13:00 local time, because this is the time when the boundary layer is close to being fully developed. It is also the time at which the column average is most similar to the daily mean9.

Once a programmable flask package (that is, one vertical profile) has been filled with air, it is transported to the IPEN Atmospheric Chemistry Laboratory in Sao Paulo, where it is analysed by a replica of the NOAA/ESRL/GMD trace gas analysis system at Boulder, Colorado, USA. Air samples are analysed for CO2, CO and SF6 (as well as CH4, N2O and H2). CO2 is measured with a non-dispersive infrared analyser20, CO by gas chromatography followed by HgO reduction detection and SF6 by gas chromatography followed by electron capture detection. Reference gases for all species were obtained from NOAA/ESRL and are directly tied to the World Meteorological Organization official standard scales.

Once analysed, the flasks are prepared for sampling by flushing them with dry air, followed by synthetic air with 350 p.p.m. CO2. Because our approach depends on CO2, CO and SF6 measurements from both the IPEN (aircraft-based vertical profiles) and NOAA/ESRL laboratories (background site records at RPB and ASC) high accuracy (‘measurement trueness’) is crucial. The procedures followed to ensure high accuracy have been documented in ref. 20. Three methods of assessing inter-laboratory comparability are being routinely pursued. First, comparisons of CO2 mole fraction from ‘target tanks’ (calibrated air cylinders treated as unknowns) demonstrate the long-term repeatability (standard deviation, 1σ, of 20 analyses) of ±0.03 p.p.m. (here ‘long-term repeatability’ refers to our estimate of the stability of our implementation of the calibration scale at the IPEN laboratory over 5–10 years) and a difference between measured and calibrated values of 0.03 p.p.m. (the calibrated mole fraction at NOAA/ESRL was 378.60 ± 0.03 p.p.m. and at IPEN it was 378.57 ± 0.03 p.p.m.). Second, a comparison of air in flask pairs sampled weekly by IPEN and NOAA/ESRL (pairs taken within 20–30 min of each other) on the Atlantic coast and analysed by NOAA/ESRL and IPEN has been in operation since October 2006. Weekly samples have been collected at Arembepe, Bahia (2006–2010) and Natal, Rio Grande do Norte (2010 to present). Results show a mean difference of only +0.02 p.p.m. (IPEN minus NOAA). Finally, the international 5th World Meteorological Organization RoundRobin31 in which three target tanks were measured by numerous laboratories around the world showed differences between IPEN and the calibrated value of only +0.02–0.03 p.p.m.