Greenhouse gas analyses

Sampling from December 1970 to March 1972 was from Makara, Wellington (41.15 S, 174.41 E, 280 m), and from December 1972 has been at Baring Head (41.41 S, 174.87 E, 80 m).

Determinations of atmospheric CO₂ mixing ratios are made using a Siemens Ultramat-3 non‑dispersive infrared (NDIR) gas analyzer. Ambient air is dried in series using a glass trap in a refrigerator set at 2°C then through a metal trap immersed in an ethanol bath set at -80°C. The NDIR CO₂ analyzer is connected via a gas manifold consisting of stainless steel tubing and computer-controlled solenoid switches to 12 gas cylinders and 2 sample air lines that are continually purged.

Four working standard gases are measured every few hours. Eight online WMO standard gases, calibrated at a WMO Central Intercalibration Laboratory, are used to calibrate the working gases approximately every fortnight. The system operates automatically, running cycles of calibration gases and ambient air, and has the facility to be monitored or controlled remotely from another site.

Baseline CO₂ concentration data are selected when the wind is from a predominantly southerly direction and the standard deviation for the recorded data is less than 0.1 ppmv for 6 hours or more.

Air samples are collected in pairs of 5 L evacuated glass flasks during southerly wind baseline conditions. The flasks are returned to the Scripps Institution of Oceanography for measurements of ¹³CO₂ and CO₂ concentrations.
 

Sampling from 1954 to June 1987 was from Makara, Wellington (41.15 S, 174.41 E, 280 m), and from July 1988 has been at Baring Head (41.41 S, 174.87 E, 80 m).  The samples have all been collected by static absorption of atmospheric CO₂ into a solution of carbonate–free NaOH. From 1954 to 1997 the collection method used an exposed tray, which was generally exposed for intervals of 1 – 2 weeks. In 1999 the collection method changed to using an exposed bottle. The CO₂ is extracted from the exposed NaOH solution by acidification followed by cryogenic distillation. From 1954 until May 1995, the ¹⁴C in the extracted carbon dioxide was determined by gas proportional counting. From May 1995 onwards the samples have been analysed by accelerator mass spectrometry. Considerable fractionation occurs during absorbtion into the NaOH solution, and the standard fractionation correction is used to determine a D¹⁴C value corrected to d¹³C = -25 per mil (‰).

Sample collection is typically with southerly (on-shore) winds. Between 60 and 100 samples are typically collected each year.

Dry whole air samples are collected in light gauge 22 and 70 L stainless steel tanks, using a KNF Neuberger diaphram pump (model N400.3 ANL) and dried using a 1 L stainless steel trap containing 3 mm 13X molecular sieve pellets. Samples are also collected in 2 L light gauge stainless steel cans using a rubber diaphragm pump without drying.

Since 1996 samples have also been collected in 2 L glass flasks with magnesium perchlorate drying, and pumped with a metal bellows pump (model MB 158).

The tanks and flasks are pumped using an automated sampling system that samples in conditions when the wind comes from the clean southerly sector at speeds greater than 7.5 m/s and when CO₂ has not changed by more than 0.02 ppm in the previous 4 hours.

Methane concentration is determined by gas chromatography fitted with flame ionisation detection and compared against standards calibrated at NOAA.

For the isotopic analyses, methane in the air samples is quantitatively converted to CO₂ in the laboratory over a platinum catalyst after the removal of interfering species like CO₂, N₂O, CO and non-methane hydrocarbons (NMHC). This CO₂ is measured on a stable isotope ratio mass spectrometer for ¹³C against an international light barium carbonate standard distributed by the IAEA in Vienna, Austria. ¹⁴C is determined by converting the CO₂ to a graphite target and analysis by accelerator mass spectrometry versus international radiocarbon standards.

CO concentrations are determined by gas chromatography fitted with a cold vapour atomic fluorescence spectrophotometer on the same air samples as those used for CH₄ concentration measurements. Between 60 and 100 samples are collected each year, typically in southerly (on-shore) wind conditions. The measurements are compared against standards purchased from NOAA.

Large dry whole air samples (1000-1200 L) are also collected for isotopic analysis. Typically 15-20 of these samples are collected each year. CO is extracted from the whole air samples by first removing CO₂ with very high efficiency cryogenic trapping, then oxidising the CO and removing the resulting CO₂ (Brenninkmeijer, 1993). As the extraction procedure is quantitative, a manometric measurement of the CO concentration can be obtained from pressure measurements of the extracted CO₂ in a calibrated volume. The CO-derived CO₂ is measured on a ratio isotope mass spectrometer for stable isotopes of ¹³C and ¹⁸O and then diluted with “dead” CO₂, containing virtually no ¹⁴C, to provide approximately 0.2 - 0.3 mg of carbon for ¹⁴C measurement by accelerator mass spectrometry.

An automatic air sampling and gas chromatograph system was established at Baring Head from December 1991 to September 1993; measurements were made at approximately 4-hourly intervals. The system was based on a Siemans Sichromat with cryogenic sample trapping by liquid nitrogen.

From 1993 to 2000, flask samples were collected from Baring Head. The flask samples were analysed on a GC system based on a Varian 2740-20, with a two stage cryogenic trapping and transfer system.

In addition to the light alkanes and alkenes, the system was sensitive to benzene, toluene, xylenes, ethyl benzene and methyl chloride. Detection limits were about 2 parts per trillion (10¹²) (ppt) for benzene and up to 10 ppt for the lighter non methane hydrocarbons.

Calibration was by comparisons with reference mixtures containing about 30 NMHCs in ppb amounts. To establish a reliable and stable calibration, several air samples were exchanged between NIWA and KFA Julich. Local working standards were prepared from mixtures of light hydrocarbons supplied by BOC Gases in New Zealand and NIST in the United States. The lab also participated in the international intercomparison exercise for NMHCs, NOMHICE (Apel et al., J. Geophys. Res., 99, 16651-16664, 1994).

A Dasibi UV ozone monitor (Model 1003-PC) was used to measure surface ozone at Baring Head from 1991 to 2004. In 2005 the Dasibi was replaced with a Thermo Electron Corporation Model 49i. Air for these measurements is drawn from 5 metres above the ground through a teflon tube with a 0.5 mm teflon filter on the inlet to exclude aerosols.

The concentration of CN in all winds typically follows the solar cycle with a peak in CN concentration around midday during steady wind conditions. The Environment One monitor (Environment One Model Rich 100 CN monitor measuring particles >2.5 nm in diameter according to its specification.) makes a bulk measurement and is not accurate at low concentrations. There is not sufficient accuracy to determine any seasonal cycle in southerly unpolluted conditions.

In unpolluted air at Baring Head, the major source of both non-sea-salt sulphate  (nssSO₄) - and the only source of methanesulphonic acid  (MSA) is thought to be the atmospheric oxidation of dimethylsulphide gas. The total atmospheric sulphate loading is about 1.2 mg/m³ in onshore winds and 1.0 mg/m³ in offshore winds. The order is reversed for nssSO₄ - with an average concentration of 0.4 mg/m³ and 0.6 mg/m³ in offshore and onshore winds respectively. In onshore winds, there is evidence for a seasonal cycle in nssSO₄  and MSA concentrations rising around November and falling again in March. This cycle is likely to be associated with the seasonal cycle in dimethylsulphide precursor which has also been measured at this site. Similar trends in sulphur components have been observed at Leigh and Karamea where aerosol measurements have also been made.

Measurement of O₂/N₂ ratios in clean southern hemisphere marine air was started at Baring Head in 1991 in collaboration with the University of Rhode Island (Prof Michael Bender, Dr Taylor Ellis). Flask pairs are returned to URI for measurement using a mass-spectrometric isotope ratio method developed by Prof Bender and colleagues. This uses repeated measurements of the difference between mass 29 to mass 32 ratios in sample and standard air. Results are reported as the relative difference between sample and standard in parts per million (quoted as “per meg”). The annual cycle of about 80 per meg in dO₂ is due largely to the release of O₂ by marine photosynthesis in the southern oceans during the spring and summer. The apparent decline in O₂ levels is related to the ongoing combustion of fossil fuels.

The work was carried out for the US Department of Energy. The remote atmospheric measurements program (RAMP) covers 41 sites globally. The sampling at Baring Head involved continuous sampling of atmospheric radionuclides onto filters that were sampled for the duration of a week.