1. 1 Trends and Anomalies in Southern Hemisphere OH Inferred from 12 Years of 14 CO Data Martin Manning, Dave Lowe, Rowena Moss, Gordon Brailsford National Institute of Water & Atmospheric Research (NIWA) New Zealand with acknowledgements to: Bill Allan (NIWA) Rodger Sparks, Institute of Geological and Nuclear Sciences, New Zealand Carl Brenninkmeijer, Max Planck Institute fuer Chemie, Mainz, Germany Research supported by the New Zealand Foundation for Research Science and Technology under contract C01X0204.

2. 2 CO sources CO vs 14 CO sources Total source ~ 10 14 Mole yr -1 Lifetime ~ 8 to 10 weeks Inventory ~ 1.7 x 10 13 Mole 14 CO sources Total source ~ 570 Mole yr -1 Lifetime ~ 10 to 12 weeks Inventory ~ 120 Mole

3. 3 14 CO cycling in the atmosphere 14 CO (52%) 14 CO (35%) direct production transport cosmic ray neutrons 14 CO 2 OH Stratosphere Troposphere In the extra-tropical southern hemisphere, recycled 14 CO accounts for 15 to 20% of the total in the troposphere. 14 CO (13%) recycled Carbon uptake by biosphere and oceans

4. 4 Model derived 14 CO distributions Tropospheric distributions of 14 CO are not very sensitive to details of production distribution pattern…. … but are sensitive to the location and strength of cross- tropopause transport, and… …show large latitudinal gradients in the winter hemisphere But observed gradients in the high latitudes are less than simulated in models! Source: Joeckel, P.; “ Cosmogenic 14 CO as tracer for atmospheric chemistry and transport”. PhD Thesis, Mainz, 2000.

5. 5 The role of the variable Sun Electrically charged material ejected from the Sun interacts with the magnetic field around the solar system. During periods of high solar activity a larger proportion of cosmic rays are deflected away from the solar system. Changes in sunspot numbers track the variation in solar activity - but observed neutron fluxes are a more direct indicator of 14 C production.

6. 6 Cycles in 14 C production Monthly sun spot numbers. NOAA NGDC web site. 14 C production rates (molec cm -2 s -1 ) derived from neutron count rate data. Lowe and Allan (in press).

7. 7 14 CO data analysis - I New Zealand 14 CO measurements Antarctic 14 CO measurements Storage correction for 14 CO production in sample cylinders Lowe et al (in press) Inverse model CO sources (Bergamaschi et al) Expected 14 C/C ratios for recycled CO Observed CO concentration Subtract recycled 14 CO compare sites Direct 14 CO

8. 8 Direct 14 CO data - New Zealand and Antarctica Note: Large annual cycle relative to mean concentration. No strong gradient between New Zealand and Antarctica.

9. 9 Note: Secular trend following solar cycle of estimated 14 C production rates. Direct 14 CO data - New Zealand and Antarctica

10. 10 14 CO data analysis – I (continued) New Zealand 14 CO measurements Antarctic 14 CO measurements Storage correction for 14 CO production in sample cylinders Lowe et al (in press) Inverse model CO sources (Bergamaschi et al) Expected 14 C/C ratios for recycled CO Observed CO concentration Subtract recycled 14 CO merge 14 C production rate Scale by production rate (2, 3 or 4 month lag) Normalized direct 14 CO data series

11. 11 Normalized 14 CO data Normalized direct 14 CO concentrations have a fairly regular cycle over 12 years (using 3-month lag from production). However, there is some residual inter- annual variability.

12. 12 Calculate apparent net removal rate R (for constant source) Analysis of 14 CO dynamics (with minimal reliance on models) Normalized direct 14 CO data series Extratropical southern hemisphere uniformity suggests behavior as a well mixed box Tropospheric production + transport from stratosphere Removal rate k [OH] Smooth and calculate derivatives with error analysis estimate S 0 so that mean value for R = 6 yr -1 Determine average annual cycle in R. NB this includes seasonality in OH and cross tropopause transport Hypothesis: Variations in R about climatological average values are most likely due to [OH] variations Where is this

13. 13 Derived apparent net removal rates Apparent net removal rate R: monthly values (red band) and average seasonal cycle (blue line) Net removal rate determined with a mean value of 6 yr -1. Varies by a factor of ~3 from winter to summer.

14. 14 Apparent removal rate vs calculated [OH] values Phasing and seasonal amplitude agree closely with [OH] values derived by Spivakovsky et al for southern hemisphere mid latitudes. Lower apparent removal rates in September to December period may reflect higher stratosphere troposphere exchange at this time.

15. 15 Variations in Effective Removal Rates Monthly ratios of apparent net removal rate to climatological average value. Upper and lower estimates based on data errors (purple lines) plus smoothed values (12 month window). No significant trend in removal rate from 1990, but two “events” with variations of > 20% over time scales of 3 to 6 months. Pinatubo eruption? Kalimantan fires?

16. 16 Variations in cross- tropopause transport may be related to QBO in the stratosphere or other dynamical effects. Schauffler and Daniel proposed that Pinatubo eruption caused an increase in stratosphere troposphere exchange. However, various indicators suggest that such effects are small. Anomalies in cross-tropopause transport ? Daily anomalies in potential vorticity at 350 K isentropic surface over 30 to 70 o S Daily anomalies in total column ozone at 45 o S

17. 17 Summary n Tropospheric 14 CO concentrations appear to scale linearly with our estimates of 14 C production rate n The lag between production and concentration appears consistent with model estimates of 3 months n The apparent net removal rate derived from the data is very similar in phase and seasonality to that expected for mid latitude OH n There appears to be no significant trend in southern hemisphere OH over the 1989 – 2001 period (agrees with the AGAGE analysis of methyl chloroform data) n However, two “events” show major anomalies in apparent removal rates of > 20% over 3 to 6 month time scales n We propose that the main cause of these variations is change in OH concentrations

18. 18 What I said in abstract Radiocarbon (14C) is produced in the lower stratosphere and upper troposphere by cosmic ray neutrons where it promptly forms 14CO. This naturally occurring radioactive form of CO is removed predominantly through oxidation by OH with an average lifetime of about 3 months (Brenninkmeijer et al, 1992; Joeckel, 2000). Concentrations of 14CO in the lower troposphere are typically 5 to 20 molecules/cm3(STP) and can be measured by extraction of CO from whole air samples followed by accelerator mass spectrometry to measure 14C/12C ratios (Brenninkmeijer, 1993). In the extra-tropical southern hemisphere about 15% of atmospheric 14CO is due to surface emissions of CO and this “recycled” component can be estimated reliably from the total CO concentration. Changes in the remaining “prompt” component of 14CO reflect variations in production or removal rates or in transport effects. The predominant variations are due to the seasonal cycle in OH and the 11-year cycle in solar activity, which modulates cosmic ray intensities and hence 14C production. We present an analysis of 12 years of 14CO measurements from clean air sites at Baring Head, New Zealand, and Scott Base, Antarctica. After removal of the seasonal cycle in 14CO remaining decadal variations are closely consistent with estimated changes in 14C production rates and there appears to be no significant trend in southern hemisphere OH, as found in analyses of methyl chloroform trends for the same period (Prinn et al, 2001). However, we find evidence for significant anomalies in OH over time scales of a few months. In particular transient reductions in OH of 20% or more occur at the time of the eruption of Mt Pinatubo in 1991 and during large scale biomass burning in Indonesia in 1997