ADVANCED GLOBAL ATMOSPHERIC GASES EXPERIMENT
MACE HEAD, IRELAND.
Final Report
April 2003
Contract No. EPG 1/1/130
International Science Consultants
39 Avon Castle Drive
Ringwood
Hants. BH24 2BB.
1. INTRODUCTION
During the course of the current contract May 1st, 2001-April 30th, 2003 overall performance of the Mace Head instrumentation, which records high frequency observations of the radiatively active gases, CFCs, HFCs, HCFCs and other halocarbons, has been very good with few disruptions. In particular, the automated gas chromatograph-multidetector (GC-MD) has performed extremely well with relatively few days of loss data due to minor mechanical failures.
Due to the added complexity of the automated gas chromatograph-mass spectrometer (AGAGE GC-MS) there has been a more mixed performance with a number of mechanical problems in 2001, whereas in 2002/03 there were very few periods of instrumental downtime and consequently minimal loss of data. In general the majority of the problems were resolved as they occurred with the GC-MS being serviced and thoroughly overhauled during the regular site visits. In spite of the relatively modest loss of data for some species we have acquired a consistently good dataset over the 24-month period of the contract.
Several publications have resulted from this research, and we have recently submitted a paper on the observations of methyl bromide (CH3Br) and methyl chloride (CH3Cl) to the Journal of Atmospheric Chemistry. At Mace Head both gases have well-defined seasonal cycles with similar annual decreases of 3.0% yr-1 (CH3Br) and 2.6% yr-1 (CH3Cl). Similar downward trends have also been observed at the AGAGE Cape Grim, Tasmanian station. The fact that both gases are steadily decreasing in the atmosphere at both locations implies that a change has occurred which is affecting a common major source of both gases (possibly biomass burning) and/or their major sink process (destruction by hydroxyl radical).
Because Mace Head receives both clean oceanic air characteristic of the background troposphere as well as polluted air masses from Continental Europe we take great care to sort the data to obtain baseline monthly means. In addition to a statistical filtering process, which is routinely performed as part of the AGAGE programme by Georgia Institute of Technology, our colleagues at the Met. Office also select baseline data using their “NAME” particle dispersion model. From all of the individual species datasets we can make a number of general observations. Atmospheric mixing ratios of the major greenhouse gases CO2, CH4, N2O, and ozone are currently at their highest historical Northern hemispheric concentrations. The rate of accumulation of CFC-12 has slowed considerably from a rate of about 10 ppt/yr in the 1990s, although it is still growing in the atmosphere with an average annual growth rate from 1999-March 2003 of only 0.8 ppt/yr. Also during the 1999-2003 period the two principal ozone-deleting CFCs, (CFC11, and CFC-113), have both declining steadily in the atmosphere, at rates of 1.66 ppt/yr and 0.64 ppt/yr, respectively. Similarly, carbon tetrachloride continues to decline at about 0.81 ppt/yr. In contrast methyl chloroform, which was the largest contributor to ozone depletion prior to the implementation of the Montreal Protocol, is still decreasing but has recently slowed to an annual average rate of 7.5 ppt/yr.
For the main HFCs and HCFCs, which have replaced the CFCs in refrigeration and foam blowing applications, we continue to observe substantial growth rates through sustained emissions. Annual average growth rates have been calculated from the baseline-filtered monthly means over the period 1994-2003. For example, HFC-134a is growing at an annual average rate of 3.28 ppt/year, while the HCFCs 22, 141b, and 142b are increasing at 5.6 ppt/yr, 1.78 ppt/year, and 1.06 ppt/year, respectively. A paper is currently in preparation describing the growth of these compounds, together with estimates of their emissions. Of special interest are the decreasing trends in, methyl bromide (0.53 ppt/year), methyl chloride (12.6 ppt/year), and methylene chloride (1.5 ppt/year). The decline in methyl bromide emissions appears to be in line with its phase-out under the terms of the Montreal protocol and its amendments, while a reduction in the many solvent applications of methylene chloride is consistent with its decreasing atmospheric abundance. More problematic is the concurrent decline in methyl chloride, which implies a change in either its major source (biomass burning) or principal sink (OH destruction). Carbon monoxide is another important atmospheric constituent where the baseline selected data continues to show a downward trend of about 1.5 ppb/year averaged over the 1987-March 2003 time frame.
We continue to enjoy a very productive collaboration with our colleagues at the Met. Office (D. Derwent, D. Ryall, and A. Manning), where the Mace Head data has been used successfully to derive European emission estimates. Through participation in a parallel monitoring programme to AGAGE called SOGE (System for Observing Greenhouse Gases in Europe), we have transferred the automated GC-MS technology developed at the University of Bristol to the three other European SOGE sites at Ny Alysund, Norway; Jungfraujoch, Switzerland; and Monte Cimone, Italy. The first two years of measurements from these stations are currently being analysed by various modelling groups in Europe to provide additional emissions estimates of the key HFCs and HCFCs over a wider European scale.
2. INSTRUMENTAL PERFORMANCE
As noted above the AGAGE GC-MD instrument has performed well with only minor disruptions. Poor sensitivity of the hot mercuric oxide reduction gas detector (RGA3), which specifically measures CO and H2, resulted in the loss of data from 30th May – 5th June, and again from 24 - 28th September 2001. All of the other species measured on the AGAGE GC-MD were unaffected by the RGA3 detector problem and have been acquired on an essentially continuous basis over the past 24 months. The ozone analyser, which has collected data continuously since 1987, also performed well with only minor losses of data associated with routine servicing and re-calibrations. However, data quality in recent years has declined progressively with lower overall precision mainly due to the age of the instrument, which has been in service for more than 16 years. The ozone analyser was therefore retired in April 2003 and replaced with a new ozone analyser (Thermoelectron), which was generously supplied by the Air Quality Division of DEFRA. This new instrument will also be regularly serviced and calibrated as part of the UK ozone-monitoring network. The carbon dioxide analyser has experienced a number of mechanical failures over the past two years, again principally due the fact that it is a 12-year old instrument. Data was lost from 28th May – 4th June, 2002 and again from 11th July – 13th August, 2002. Our French collaborators at CEA/CNRS replaced the instrument in late 2002 with another CO2 analyser. Furthermore, they intend to install a completely new CO2 instrument in the summer of 2003.
Operation of the AGAGE GC-MS instrument has been more problematic with a number of mechanical failures, particularly in 2001. A summary of these problems follows: Firstly it was noted that the peak shape and hence data quality of HFC-125 was compromised by an interference peak on its tail. A decision was taken to replace the microtrap on 25th May, even though the chromatography of all other compounds was excellent. The change of microtrap improved peak shape of HFC-125 dramatically, all HFC-125 data before the microtrap change was flagged. In June 2001, the efficiency of the Peltier thermoelectric cooler decreased as the minimum temperature attained dropped to –42°C, rather than the optimal set point of –50oC. During this period the data became increasingly noisy. Although the Peltier unit and valve V2 were replaced on 20th June 2001, the data still appeared to be noisy hence the microtrap was replaced on 21st June. On 4th July the new Peltier mysteriously malfunctioned and was again replaced by another Peltier unit on 12th July. The GC-MS instrument worked well until the 29th August, when the Agilent GC oven vent flapper motor started to jam intermittently. The flapper motor, which regulates the GC oven temperature, was replaced on 10th September. This was the second time a GC flapper motor had to be replaced, the first failure occurring in June 1999. It was then noted that the microtrap was heating erratically, this was due to ingress of water into the Peltier cells. A new Peltier unit and microtrap (046/C04/C36/260301) was installed on 28th October, 2001. Unfortunately, we still continued to experience problems with noisy data, and after a number of exhaustive tests, it was determined that one of the Peltier cells was not working efficiently, in addition a small amount of Carboxen adsorbent material had been released from the microtrap into the rotor of valve V3. The effect of this was twofold, firstly the Carboxen “dust” scored the valve rotor creating cross-port leakage, secondly, the dust acted as a secondary site for re-adsorbing compounds of interest after they had been desorbed from the microtrap, this in turn caused peak tailing which was compound specific. On the 18th November 2001 an extensive service of the ADS was undertaken. A new valve V3 and standoff were installed along with a new Peltier unit and a new style microtrap (056/C04/C36/CT4/151001), which contained 4mg of Carbotrap adsorbent added to the inlet of the microtrap. This weaker adsorbent material should enable the less volatile compounds to be desorbed more easily. The ADS has performed exceptionally well since this overhaul.
The only other unrelated problem to occur since the major GC-MS service was when the AADCO zero air generator unit failed on 11th December 2001; this in turn compromised the efficiency of the Nafion dryer, since dry air from the generator is used to purge the Nafion. This allowed wet ambient air to enter the microtrap and caused a build up of ice in the trap. The AADCO unit was replaced immediately.
3. INTERCOMPARISONS AND INTERCALIBRATIONS
Data intercomparisons and intercalibrations are an ongoing activity within the AGAGE program. The data collected at common sites by different laboratories are useful as a sensitive diagnostic test of data quality for the laboratories involved and they enable data sets to be merged more accurately for use by AGAGE scientists and other researchers. Intercomparisons have been carried out between AGAGE in situ data and flask and/or in situ data collected by laboratories such as the Commonwealth Scientific and Industrial Research (CSIRO), the National Institute for Environmental Studies (NIES), the National Oceanic and Atmospheric Administration-Climate Monitoring & Diagnostics Laboratory (NOAA/CMDL), the Scripps Institution of Oceanography (SIO), and the University of East Anglia (UEA). The AGAGE GC-MS data for comparison with data from other laboratories are selected within a window of ± 3 hours of the relevant flask data. Table 1, lists where comparisons have been carried out between AGAGE in situ GCMS data and other laboratory data (in situ, flask, instrument). For most intercomparisons the difference is substantially < 5%, with the largest difference between AGAGE GC-MS and UEA flask data of about 12%. These differences are most likely due to different absolute calibration scales.
Table 1. Comparisons between AGAGE in situ GC-MS data and other laboratory data at the same site; ratio = (AGAGE/other laboratories-1)*100 ± 1 standard deviation, n = number of comparisons.
Species / Site / Comparison toAGAGE in situ /
Comparison period
/% Ratio ± 1 sd, (n)
HFC-134a / Mace Head / NOAA flask, MS / Oct 1998 – Sep 2002 / 0.57±1.78, (24)Cape Grim / NOAA flask, MS / Mar 1998 – Sep 2002 / 2.12±3.31, (139)
Cape Grim / NIES flask, MS / Sep 2001 – Oct 2002 / 3.27±4.48, (19)
Cape Grim / UEA flask, MS / Mar 1998 – Dec 2001 / 12.55±3.66, (25)
HCFC-22 / Mace Head / NOAA flask, MS / Feb 1999 – Sep 2002 / -0.49±0.68, (18)
Cape Grim / NOAA flask, MS / Mar 1998 – Sep 2002 / 0.29±0.82, (136)
Cape Grim / UEA flask, MS / Mar 1998 – Aug 2001 / 0.02±1.25, (19)
Cape Grim / SIO flask, ECD / Mar 1998 – Oct 2001 / 0.11±0.83, (43)
HCFC-141b / Mace Head / NOAA flask, MS / Oct 1998 – Sep 2002 / 0.064±1.484, (24)
Cape Grim / NOAA flask, MS / Mar 1998 – Aug 2002 / 0.739±1.041, (119)
Cape Grim / NIES flask, MS / Mar 2000 – Oct 2002 / 2.031±2.586, (42)
Cape Grim / UEA flask, MS / Mar 1998 – Aug 2001 / -3.731±1.556, (22)
HCFC-142b / Mace Head / NOAA flask, MS / Oct 1998 – Sep 2002 / 4.73±1.48, (23)
Cape Grim / NOAA flask, MS / Mar 1998 – Sep 2002 / 4.77±1.49, (139)
Cape Grim / NIES flask, MS / Feb 2000 – Oct 2002 / 5.09±2.93, (44)
Cape Grim / UEA flask, MS / Mar 1998 – Aug 2001 / 1.14±3.39, (20)
4. DATA ANALYSIS
Summaries of the monthly mean “baseline” concentrations for the two principal CFCs (CFC-11 and CFC-12) are illustrated in Figure 1. Each of the monthly mean datasets was fitted with a polynomial function to obtain average trends over the 16-years of observations. However, over the more recent period 1999-March 2003, CFC-11 has been decreasing at an average rate of 1.66 ppt/year, while in contrast CFC-12 continues to increase at an annual average rate of 0.8 ppt/year, mainly due to the bank of old refrigeration units. Moreover, this rate of increase continues to slow year-on-year as is also evident in the time series of all of the observations plotted by the Georgia Institute of Technology in Figure 2. Notice that in addition of the obvious slowing of the trend the magnitude of the pollution events (elevated red lines) have also substantially decreased over time, but have not yet disappeared.