Chemical and Dynamical Influences on Decadal - FMI...
115
Chemical and Dynamical Influences on Decadal Ozone Change (CANDIDOZ) EVK2-CT-2001-00133 Final report (1.4.2002-30.9.2005) Coordinator: E. Kyrö Project home page: http://fmiarc.fmi.fi/candidoz/
Transcript of Chemical and Dynamical Influences on Decadal - FMI...
Chemical and Dynamical Influences on Decadal Ozone Change (CANDIDOZ) EVK2-CT-2001-00133 Final report (1.4.2002-30.9.2005) Coordinator: E. Kyrö Project home page: http://fmiarc.fmi.fi/candidoz/
Management Report
Participant information Institution Street name
and no. Postal code
Town Country Title Familyname
First name
Tel. fax e-mail
1.Finnish Meteorological Institute-Arctic Research Centre (FMI)
• Tähteläntie 62 •
FIN-99600
• Sodankylä
• Finland
Professor Kyrö Esko +358 40 552 74 38
+358-16-619 623
2. Swiss Federal Institute of Technology (ETH-Z)
ETH Hoenggerberg HPP
CH-8093
Zuerich Switzerland Professor Thomas Peter +411 6332756
+411 632 1058 [email protected]
3. Centre for Atmospheric Science, Department of Chemistry, The University of Cambridge (UCAM)
Lensfield Rd CB2 1EW
Cambridge
United Kingdom
Professor Pyle John A +44 1223 336473
+44 1223 336362
4. Service d'Aéronomie, Université Pierre et Marie Curie (UPMC)
4 Place Jusseau 75252 Paris, Cedex5
France Doctor Godin Sophie +33 1 44 27 47 67
+33 1 44 27 37 76
5. National and Kapodistrian University of Athens (NKUA)
6, Christou Lada Str.
10561 Athens Greece Professor Zerefos Christos +30 108071910
6. Institute of Atmospheric Physics of the Academy of the Sciences of the Czech Republic (ASCR)
Bocni II, 1401 14131 Prague Czech Republic
Doctor Lastovicka Jan +420 2 67 103055
+420 2 72 763745
7. The Czech Hydrometeorogical Institute (CHMI)
Hvezdarna 456 50008 Hradec Kralove
Czech Republic
Doctor Vanicek Karel +420 495260352
+420 49 5264127
8. Danish Meteorological Institute (DMI)
Lyngbyvej 100 2100 Copenhagen
Denmark Doctor Knudsen Bjoern +45 39157416
+45 39 157461 [email protected]
9. Alfred Wegener Institute for Polar and Marine Research (AWI)
PO BOX 600149 Telegrafenberg A43
14473 Potsdam Germany Doctor Rex Markus +49 3312882127
+49 331 2882178
10. Department of Geophysics, University of Oslo (UiO)
PO BOX 1022 0315 Oslo Norway Professor Isaksen Ivar +47 2285 5822
+47 2285 5269 [email protected]
11. Norwegian Institute for Air Research (NILU)
PO BOX 100 Instituttveien 18
2027 Kjeller Norway Doctor Brathen Geir +47 63898000
+47 638 98050 [email protected]
12. Institute of Environmental Physics, University of Bremen (IUP-UB)
PO BOX 330 440 28334 Bremen Germany Doctor Weber Mark +49 421 218 2362
+49 421 218 4555
2
Management Report
1. Management and Resource usage summary for extension period 1.4-30.9.2005 1.1 Objectives
Work package 1: Northern mid-latitude long-term trend analysis • Completion of OHP ozone and temperature statistics • Completion of Neural model analysis • Documentation and publication of CATO data set (CANDIDOZ Assimilated Three-
dimensional Ozone)
Work package 2: Polar ozone, polar vortex and effects to mid latitude • The re-evaluation of the remaining Svalbard data • Completion of Arctic ozone sonde analysis
Work package 3: Changes in ozone transport: Residual circulation and coupling to the troposphere
• Further development of the regression model and application of the final model to Dobson/Brewer global network and satellite measurements.
• Completion of long term data sets of diabatic subsidence and strength of the residual circulation.
• Synthetize the role of climate patterns in stratosphere-troposphere coupling and their impact on ozone column.
Work package 4: Long-term ozone simulation • Completion of the remaining CTM integrations • Quantification of the contribution of chemical and dynamical processes of CTM ozone trend • Contribution of the WP4 results in the final WP5 Synthesis • Analysis of the Unified Model experiments’ output to assess the relation between ozone
depletion and climate change. In addition to above specific points all Work packages 1-4:
• Writing last manuscripts for the scientific publications • Contributing to the synthesis and final CANDIDOZ report.
Work package 5: Synthesis Main effort in extension period was devoted to the 4 tasks of the synthesis WP5:
• Attribution to causes of the past ozone trends in the Northern hemisphere: A review of the results of the individual trends studies of ozone times series (Ground based, satellite based and 3-dim global data set CATO) to assess the statistical significance and explanatory power of proxies and ultimately the importance of the Dynamica/Chemical processes behind the proxies.
• Evidence for turnaround of ozone trends as a consequence of the Montreal Protocol:
Assessment of the goodness (in explanatory power and statistical significance) of the proxies used to describe long term change due to homogeneous chemistry: Classical linear ramp vs. EESC (estimated equivalent stratospheric chlorine) or “hockey stick”.
3
Management Report
• Critical factors affecting future trends of stratospheric ozone: Review of the results of University of Cambridge Unified Model run into future with realistic alternative climate and EESC scenarios.
• Trend study of CTM ozone output using process oriented statistical model.
4
Management Report
5
1.2 Scientific/Technical progress made in different work packages according to the planned time schedule
Project planning and time table; (time-work flow-chart). Changes to original chart indicated plusses (Shift ahead) and uses (delay of start). Most changes are due to delay in delivery of ERA-40. Red crosses refer to inputs/meetings associated
with Synthesis and final reporting
Year First year 2002/2003 Second year 2003/2004 Third year 2004/2005 Extention 2005
Month: A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S
Month Nº 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
Coord. + + + + + +
WP 1
Task 1.1 + +
Task 1.2 + +
Task 1.3 + +
Task 1.4 + + + + +
Task 1.5 + +
WP 2
Task 2.1 - - - - + + + + + +
Task 2.2 + + + + +
Task 2.3 + + + + - - - - - - + +
Task 2.4 - - - - - - + + + + + + + +
WP 3
Task 3.1 - - - - - - - + + + + +
Task 3.2 - - + +
Task 3.3 + + - - + +
Task 3.4 + + - - + +
Task3.5 + +
WP 4
Task 4.1 + + + + + + + +
Task 4.2 + + + + +
Task 4.3 + + + + +
WP 5
Task 5.1 + + + + + +
Task 5.2 + + + + + +
Task 5.3 + + + + + +
Task 5.4 + + + + + +
Publication of three ERA-40 data streams
- - 1, 2, 3
Meetings - + + - + - + + - + +
Reporting - + - + + - + +
Managem
ent Report
6
Table 1. The actual costs by category of the extension period 1.4.2005-30.9.2005 derived from the cost statements. Negative numbers represent adjustment to the previous periods. The original plan did not include extension period, therefore no planned costs are available.
Partner Personnel Durableequipment
Subcontrac-ting
Travel and subsistence
Consumables Computing Protectionof
knowledge
Other specific costs
Overheads TOTAL
Coordinator4 FIMI.ARC
10155,16 1526,00 8327,94 20009,10ETH.AS
UCAM.DC
3834,29 766,86 4601,15UPMC.SA/
CNRS.SA
21673,04 681,50 2923,28 10836,52 36114,34NKUA
-14,60 -566,43 -110,79 -691,82CSIAP.CA
3 252,00 1 042,00 1 106,00 4 343,00
CHMI.SO 491,00 305,00 862,00 1658,00
DMI.RDD. MAR 16 542,58 3 308,52 19 851,10
AWI.FP 43 788,26 3 638,33 1 520,55 35 030,61 83 977,75
UOSLO.DGP 26 899,70 5 379,90 32 279,60
NILU 25 596,61 600,21 863,61 17 787,47 44 847,90 UBRM. FPE. IUP 955.88 -1630.96 191.18 -483.90
Total Project
Management Report
Person power table Person power hours are derived from the extension period 1.4.-30.9.2005 cost statements. The distribution of hours among various work packages are roughly estimated by coordinator from the distribution of remaining work after the 3rd year because no WP-wise accounting exists. All the hours in this table represent deviation from the original work plan which covered 3 years.
CANDIDOZ 1.4.2004-31.3.2005
Partner Coordination
WP1 WP2 WP3 WP4 WP5 Total
FMI 189 189
ETHZ 300 485 785
UCAMB
UPMC/CNRS 304 604 100 1008
NKUA -
ASCR -
CHMI -
DMI 305 100 405
AWI 1008 170 1178
UiO 500 183 683
NILU 281 100 381
IUP-UB
Total
7
Management Report
1.3 Deliverables and milestones obtained during extention period Deliverable No Deliverable title Delivery month
7 WP1 Task 3 Temperature trends using different techniques on the 1979-2003 period.
42
9 WP1 Task 4 Application of neuronal models for simulation of total ozone in the European region
42
15 WP2 Task 5 Update and extension of the trend analysis of selected surface ozone series and comparison with results of trend analysis of long-term satellite data
42
25 WP3 Task 1 Validated long term data set of diabatic subsidence rates inside the NH polar vortex
42
26 WP3 Task 1 Long-term dataset of strength of the residual circulation
42
27 WP3 Task 1 Decadal statistics of polar subsidence and residual circulation
42
30 WP3 Task 3 Correlation between EP-flux and residual circulation
42
39 WP4 Task 4 Improved understanding of role of stratospheric climate in ozone recovery during next 50 years
42
40 WP5 Task 1 Attribution of Northern hemispheric ozone trends to chemical and dynamical causes
42
41 WP5 Task 2 Critical assessment of early indication of the turnaround of Northern hemispheric ozone trends as consequence of the Montreal protocol and its amendments
42
42 WP5 Task 3 Evaluation of the most critical factors affecting future trends of stratospheric ozone
42
1.4 Deviations from the work plan Due to delay of the delivery of the key data set, ERA-40, and specific problems with the data the project completion delayed six months compared with the original schedule. The extension period helped achieve all the technical deliverables. Scientifically, ERA-40 problems (quality of some products) probably hampered some aspects of the originally planned science.
8
Management Report
1.5 Coordination information
Coordinating meeting Vienna, 27 April 2005. The meeting took place during European Geosciences Union General Assembly in Vienna, 27 April 2005. Only issue was planning and time table of the final report. Details available from the meeting minutes at Candidoz web-page http://fmiarc.fmi.fi/candidoz Phone conference September 22, 2005 The final reporting phone meeting took place September 22, 2005, just before the end of the project. It resulted in an outline of the synthesis report (WP5). After that the processing of the Final report was coordinated through e-mails and Candidoz www-page. 1.6 Difficulties encountered at management and co-ordination level proposed/applied solutions Many of the final results were considerable refined and scientific manuscripts produced during the extension period so that the extension period itself was extremely useful. In final reporting phase, however, there were more delays than in earlier reporting phases. On the other hand, the cumulative delay of final report less than 2 months is reasonable in the light of intense involvement of key contributors in new projects, especially SCOUT and its Darwin campaign.
9
Management Report
Section 2. Executive summary of main activities and results Contract n EVK2-CT-2001-00133 Reporting period: 1.4.2005-30.9.2005 Title Chemical and Dynamical Influences on Decadal Ozone Change Objectives:
1 Reviewing and finalizing the results of the trend studies obtained in the original project period 1.4.2002-31.3.2005.
2 Completion of the remaining CTM Decadal integrations
3 Analysis of the Unified Model experiments’ output to assess the
relation between ozone depletion and climate change.
4 Completing the synthesis of the project final results: Attribution to causes of the past ozone trends in the Northern hemisphere, Evidence for turnaround of ozone trends as a consequence of the Montreal Protocol. Critical factors affecting future trends of stratospheric ozone. Trend study of CTM ozone output using process oriented statistical model.
Scientific achievements:
- Finalization of new ozone data products: Global data sets (satellite and ground based) and re-evaluated local ozone data series.
- Final form of an advanced, process oriented multilinear statistical model to estimate changes and trends and contributing factors in long term ozone time series.
- Final trends in the NH mid latitudes and in the Arctic based on the optimized statistical model and new or re-evaluated data sets.
- Final integrations of decadal model runs using CTMs and Chemistry Climate models: Global decadal 3-D data fields of ozone and other trace gases, Quantification of the relative contributions to ozone change from dynamics and chemistry and estimated polar and mid-latitude ozone loss for different years. Improved understanding of role of stratospheric climate in ozone recovery during next 50 years.
- Attribution of Northern hemispheric ozone trends to chemical and dynamical causes as obtained from statistical studies of long term data sets.
- Assessment of NH long term trend change from negative to less negative and factors behind the change.
- Evaluation of the most critical factors affecting future trends of stratospheric ozone.
10
Management Report
Conclusions: The overall outcome from the individual analyses (listed below) shows clearly one common feature in the NH mid latitudes and in the Arctic: Almost monotonic negative trend from late 1970s to mid 1990s followed by a relative recovery after that. An inflection point seems to appear around mid 1990s, somewhat later in the Arctic than in the NH mid latitudes. A similar change can be also seen in the trends of the Arctic vortex characteristics. Furthermore, all individual analyses point to the changes in dynamical drivers, such as residual circulation represented by EP- or heat-flux from troposphere to stratosphere in mid-/high latitudes, playing a key role in the observed turnaround together with closely associated heterogeneous ozone chemistry represented by PSC volume. Equivalent latitude proxy representing synoptic scale changes in lowermost stratosphere was also a powerful driver of recent recovery locally. In most long term studies where the comparison was made EESC, which since mid 1990s has been leveling off as a consequence of Montreal protocol and its amendments, was observed to represent homogeneous ozone loss better than the simple linear trend. Some influence in recent ozone recovery was also attributed to the solar cycle no. 23. It was shown in several studies that the polar ozone loss signal effectively spreads out to summer mid latitudes. Because of the correlation between them the impact of the residual circulation- and polar ozone loss proxies are not easily separable in the statistical models. More direct way to calculate dilution is by tracking the depleted air masses by trajectory calculation. This study also shows significant impact of vortex depletion dilutions to NH mid latitudes in April and May, typically 10-40 DU depending on location after cold Arctic winters. Similar results were obtained from the CTM integrations (SLIMCAT with simple PSC chemistry) which showed that the northern middle latitudes 35-60 N can experience reduced total ozone in spring via dilution up to -35 to - 40 DU in cold arctic winters. The trajectory calculation shows further that the dilution may explain about 29 % of the trend in the period 1979-1997 and 33 % of the trend in the period 1979-2002. Since, EESC will remain sufficiently high to sustain massive polar ozone loss in the cold stratosphere for the nearest few decades the dynamics of the NH winter will largely determine both the interannual variability and the development of the long term trend of ozone in the NH mid- and high latitudes. In this connection also the effect of climate change to the stratospheric temperatures must be kept in mind. Climate/chemistry (with simple ozone scheme) simulations indicated that the observed ozone loss does not affect (through the radiative feedback) the circulation in the NH middle latitudes. In relation to the greenhouse gases radiative feedback, a present ozone and CO2 levels scenario produced an enhanced planetary wave forcing, supporting the CTM’s suggestion of a dynamically-driven turnaround. Keywords: Ozone, ozone trends, ozone loss, ozone depletion, stratosphere, Arctic, Montreal protocol, 3D mode.
11
Management Report
2.1 Publications Peer-reviewed articles: Authors Date Title Journal ReferenceLastovicka J. 2002 Very strong negative trends in
laminae in ozone profiles. Phys. Chem. Earth 27 (No. 6-
8), 477-483 Maeder, J.A. 2003 Haupteinflussfaktoren auf das
stratosphärische Ozon in der nördlichen Hemisphäre
PhD thesis ETH-Zürich
Orsolini, Y. J. and F.J. Doblas-Reyes
2003 Ozone signatures of climate patterns over the Euro-Atlantic sector in the spring
Quart. J. Roy. Meteor. Soc.
129, 3251-3263
Sinnhuber B.-M., M. Weber, A. Amankwah, and J.P. Burrows
2003 Total ozone during the unusual Antarctic winter of 2002
Geophys. Res. Lett 30, 1850, doi:10.1029/2002GL016798, 2003.
Weber M., S. Dhomse, F. Wittrock, A. Richter, B.-M. Sinnhuber, and J.P. Burrows
2003 Dynamical Control of NH and SH Winter/Spring Total Ozone from GOME Observations in 1995-2002
Geophys. Res. Lett 30, 1853, doi:10.1029/2002GL016799
Coldewey-Egbers, M., M. Weber, M. Buchwitz, and J.P. Burrows,
2004 Application of a modified DOAS method for total ozone retrieval from GOME data at high polar latitudes
Adv. Space Res 34, 749-753
Hadjinicolaou, P and J.A. Pyle
2004 The Impact of Arctic Ozone Depletion on Northern Middle Latitudes: Interannual Variability and Dynamical Control
J. Atm. Chem. 47 (1), 25-43
Karpetchko A, and Nikulin G.
2004 Influence of early winter upward wave activity flux on midwinter circulation in the stratosphere and troposphere
J. Climate v.17, N22, pp4443-4452
Krizan P., J. Lastovicka
2004 Definition and determination of laminae in ozone profiles.
Studia geoph. et geod. 48 (4), 777-789
Orsolini, Y., 2004 Seesaw Fluctuations in Ozone between the North Pacific and North Atlantic
J. Meteor. Soc. Japan 82, No 3, 941-949
Rex, M., Salawitch, R.J., Gathen, P. von der, Harris, N.R.P., Chipperfield, M., Naujokat, B.
2004 Arctic ozone loss and climate change
Geophys. Res. Lett., 31, L04116, doi:10.1029/2003GL018844
Andersen, S. B. et al.
2005 Comparison of recent modeled and observed trends in stratospheric total column ozone
J. Geophys. Res. accepted
12
Management Report
Brunner, D., J. Staehelin, H.-R. Kuensch, and G. Bodeker
2006 A Kalman filter reconstruction of the vertical ozone distribution in an equivalent latitude - potential temperature framework from TOMS/GOME/SBUV total ozone observations
J. Geophys. Res in press
Dhomse S., M. Weber, I. Wohltmann, M. Rex, J. P. Burrows
2005 On the possible causes of recent increases in NH total ozone from a statistical analysis of satellite data from 1979 to 2003
Atmos. Chem. Phys. Discuss.
Page(s) 11331-11375. SRef-ID: 1680-7375/acpd/2005-5-11331
Hadjinicolaou, P., J. A. Pyle, and N. R. P. Harris
2005 The recent turnaround in stratospheric ozone over northern middle latitudes: A dynamical modeling perspective,
Geophys. Res. Lett. 32, L12821, doi:10.1029/2005GL022476
Hansen G. and T. Svenøe
2005 Multi-linear regression analysis of the 65-year Tromsø total ozone series
J. Geophys. Res Vol. 110, No. D10, D10103 10.1029/2004JD005387.
Karpetchko, A., E. Kyrö, and B. M. Knudsen
2005 Arctic and Antarctic polar vortices 1957–2002 as seen from the ERA-40 reanalyses
J. Geophys. Res. 110, D21109, doi:10.1029/2005JD006113,
Krizan P., J. Lastovicka
2005 Trends in positive and negative ozone laminae in the Northern Hemisphere
J. Geophys. Res doi: 10.1029/2004JD005477
Lastovicka J., P. Krizan
2005 Trends in laminae in ozone profiles in relation to trends in some other middle atmospheric parameters
Phys. Chem. Earth accepted
Nikulin G. and A. Karpechko
2005 The mean meridional circulation and midlatitude ozone buildup,
Atmos. Chem. Phys. 5, 3159-3172,
Wohltmann, I., Rex, M., Brunner, D., Maeder, J
2005 Integrated equivalent latitude as a proxy for dynamical changes in ozone column
Geophys. Res. Lett 32, 9, L09811, doi: 10.1029/2005GL022497
Submitted and planned future publications (peer reviewed) Authors Date Title Journal Status Andersen S. B. and B. M. Knudsen
2005 The influence of polar vortex ozone depletion on NH midlatitude ozone trends in spring
Atmos. Chem. Phys. Discuss.
submitted
Hadjinicolaou P. and Michaelides
2005 Long-term changes of stratospheric ozone and
Ann. Geoph. submitted
13
Management Report
S. tropopause properties over the South-East Mediterranean
Jrrar A., Braesicke P.,Hadjinicolaou P., and Pyle J.A.
2005 Trend analysis of CTM-derived total ozone using self-consistent proxies: How well can we explain dynamically induced trends ?
Quart. J. Roy. Meteorol. Soc.
submitted
Krizan, P., J. Lastovicka: J.
2005 Ozone laminae: comparison of the Southern and Northern Hemisphere, and explanation of trends.
Atmos. Solar-Terr. Phys., submitted
Vogler C., Brönnimann S., and G. Hansen
2005 Re-evaluation of the 1950-1962 Svalbard total ozone record
Atmos. Chem. Phys. submitted
Zanis P., E. Maillard, J. Stahelin, C. Zerefos, E. Kosmidis, K. Tourpali
2005 On the turnaround of stratospheric ozone trends deduced from the re-evaluated Umkehr record of Arosa, Switzerland
J. Geophys. Res submitted
Hadjinicolaou et al.
2006 Multi-decadal simulations of stratospheric ozone chemistry with a 3D chemical-transport model
Atmos. Chem. Phys.
Hadjinicolaou et al.
2006 The influence of polar ozone depletion on middle latitudes: inter-annual variability and trends
Geophys. Res. Lett
Hadjinicolaou et al.
2006 Diagnosis of the dynamically driven stratospheric ozone changes
J. Geophys. Res
Kivi, R., et al. 2006 Ozonesonde observations in the Arctic during 1989-2003: ozone variability and trends in lower stratosphere and free troposphere
J. Geophys. Res
Orsolini, Y.J., F. Doblas-Reyes, D. Stephenson
2006 The role of large-scale flow patterns on the occurrences and clustering of ozone miniholes
J. Geophys. Res.
Wohltmann I., et al.
2006 A process-oriented regression model for column ozone
J. Geophys. Res
Non-refereed literature: Andersen, S.B. 2002 The influence of vortex ozone
depletion on Arctic and midlatitude trends
Nordic Ozone Group Meeting, Boras, 12-13 April, 2002
Andersen S. B., and B. M. Knudsen
2003 The influence of polar vortex ozone depletion at midlatitudes
Geophysical Research Abstracts, EGS-AGU-EUG Joint Assembly, Nice, 06-11 April 2003
05183, Vol. 5, 2003
Vaníček K., Dubrovský M. and Staněk M.
2003 Evaluation of Dobson and Brewer total ozone observtaions from Hradec Králové, Czech
Publication of the Czech Hydrometeorological Institute
Prague, ISBN 80-86690-10-5
14
Management Report
Republic, 1961-2002 Staehelin,J., R.D. Evans, J.B. Kerr, and K. Vanicek
2003 Comparison of total ozone measurements of Dobson and Brewer spectrophotometers and recommended transfer functions
GAW/WMO Report No.149, Geneva, 2003
Krizan P. 2003 Trend of ozone amount in positive ozone laminae in high latitudes of the Northern Hemisphere.
Proceedings of 6th European Symp. on Ozone, Göteborg, 2-6 September, 2002.
pp261- 264
Weber M. et al.: 2003 Dynamical control of ozone transport and ozone chemistry from satellite observations and CCMs
Workshop on CCM validation, Grainau, Germany, November
Rex M et al., 2003 An approach to diagnose polar ozone loss in CCM calculations,
Workshop on CCM validation, Grainau, Germany, November
Orsolini Y. and Francisco J. Doblas-Reyes
2003 Signatures of Climate Patterns upon Stratospheric Ozone
European Geophysical Symposium, Nice, April 7-11, 2003
Vanicek K. 2003 Differences between Dobson and Brewer total ozone observations in Hradec Kralove - Preliminary results
The Brewer Biannual Workshop, El Arenosillo, Spain
Hadjinicolaou P., A. Jrrar, J. A. Pyle
2003 Multi-decadal simulations of stratospheric ozone,
Roy. Met. Soc. Conference, U.E.A.Norwich, Uk, 1-5 Sept.
Vanicek K., 2003 Differences between Dobson and Brewer total ozone observations in Hradec Kralove - Preliminary results achieved under the project CANDIDOZ.
The Brewer Biannual Workshop, El Arenosillo, Spain, 2003
Damski, J., Backman, L., Thölix, L. and Kaurola, J.
2003 A Chemistry-Transport Model Simulation of Middle Atmospheric Ozone from 1980 to 2019 Using Coupled Chemistry GCM Winds and Temperatures
International Conference on Earth System Modelling, 15-19 September 2003, Max Planck Institute for Meteorology, Hamburg, Germany,
Wohltmann I. et al.
2004 Control of stratospheric ozone transport and chemistry by tropospheric climate variability.
EGU 1st GeneralAssembly, Nice, April 2004.
Hadjinicolaou P., Pyle J.A., Braesicke P., and Harris N.R.P.
2004 Dynamical contributions to decadal ozone changes in the stratosphere
4th Annual Meeting of the European Meteorological Society, Nice, France, 26-30 Sept., 2004
Vanícek K. 2004 Consistency of Long-Term Total Ozone Observations in Hradec Kralove - Outputs from the Project CANDIDOZ.
Publications of the Institute of Geophysics, Polish Academy of Science
64(371), Warsaw, p.83-87, ISBN-83-88765-43-4, 2004
15
Management Report
Andersen, S. B. 2004 Ozone recovery: comparison of recent modeled and observed trends in stratospheric total column ozone.
NOAA, Boulder, Colorado, May 11, 2004
Andersen S. B. and B. M. Knudsen
2004 Why is the Mid-Latitude Ozone Depletion in Spring worst over Europe and Russia ?
Proceedings of 6th European Symp. on Ozone, Göteborg, 2-6 September, 2002.
pp. 35-38
Metelka L., Kliegrová S. and Vaníček K.,
2004 Development and Testing of Neural Models for Simulation of Total Ozone in the European Region. Report of the project CANDIDOZ - the Deliverable No.9.
The Czech Hydrometeorological Institute, Prague, March 2004
Hadjinicolaou P. and Michaelides S.
2004 Long-term changes of stratospheric ozone over Cyprus
Proceedings of the 7th Panhellenic (International) Conference on Meteorology, Climatology and Atmospheric Physics, Nicosia, Cyprus, 28-30 Sept., 2004
Backman, L., Ojanen, S.-M., Thölix, L. and Damski, J.
2004 Multi-Year stratospheric simulations using FinRose-CTM, comparison with observations
Nordic ozone group meeting, 15-16 April, 2004, Helsinki, Finland.
Proceedings of Stratospheric Ozone Workshop, Zürich, 16.-17. March 2004.
Stähelin J. 2004. Stratospheric Ozone Depletion: History, Present Knowledge and Present Questions,
pp6-7.
Harris N. 2004 Polar Ozone Delpetion p8 Rex M. 2004 The Dynamical Contribution to
Interannual Ozone Variability p12
Andersen, S.B., and E. C. Weatherhead
2004 Predicted column ozone recovery rates at northern mid-latitudes
p14
Backman, L., Damski, J., Ojanen, S.-M. and Thölix, L.
2004 Multi-year chemistry transport model simulation of middle atmospheric ozone using ERA-40 meteorological data: Comparison with observations
Vanicek K. 2004 Evaluation of long-term Dobson and Brewer total ozone observations from Hradec Kralove, Czech Republic.
pp24-25
Metelka L. 2004 Development and testing of neural models for simulation of total ozone in the European region.
p17
Dubrovsky M., 2004 On the difference between total ozone measured by Dobson and Brewer spectrophotometers.
p26
16
Management Report
Hadjinicoleau P. and J.Pyle
2004 Numerical Simulations of Stratopheric Ozone
p13
Proceedings of the XX Quadrennial Ozone Symposium, Kos, Greece, Editor C. Zerefos., IOC, 2004
Stähelin J., J.Mäder, and D. Brunner
2004 Trend Analysis using 65 Ground Bases Total Ozone Series of the Northern Hwemisphere
pp29-30
Brunner D., J.Stähelin, and J.Mäder
2004 Trends in Total Ozone and in Vertical Ozone Profiles Reconstructesd from Homogenized TOMS/GOME Total Ozone Data Set by Data Assimilation
pp37-38
Backman L., Damski, J., Thölix, L. and Kaurola, J.
2004 A Chemistry-Transport Model Simulation of Middle Atmospheric Ozone from 1980 to 2019 Using Coupled Chemistry GCM Winds and Temperatures
Vol.II, pp. 729-730, 2004
Hansen G. and T.Svenø
2004 Re-Evaluation and Analyusis of the 65-year Tromsø Total Ozone Series
pp69-70
Karpetcko, A., E.Kyrö, and B.Knudsen
2004 Variability of Arctic Polar Vortices in 1957-2002 as seen from ERA-40 data base
pp823-824
Godin-Beekman S., T.Song, and A.Haucherone
2004 Short-Term and Long-Term Variability of Starospheric Ozone Vertical Distribution from Systematic Lidar Ozone Measurements in the South of France
pp67-68
Zerefos, C., P. Zanis, E. Kosmidis, and K. Tourpali:
2004. Statistical trend analysis of UMKEHR data at selected northern hemisphere stations
pp1201-1202
Krizan P., J.Lastovicka, and J.Kryscin
2004 Ozone Laminea and Teleconnection Patterns
pp421-422
Metelka L. and K. Vanicek
2004 Development and testing of neural models for simulation of total ozone in the European region
p40
Kivi, R., E. Kyrö, and T. Turunen
2004 Stratospheric ozone observations at Sodankylä during 1989-2003
pp377-378
Weber M., et al. L.N. Lamsal, M. Coldewey, K. Bramstedt and K. Vanicek
2004 Validation of GOME total ozone retrieved using WF-DOAS
p127
17
Management Report
Andersen S.B., et al.
2004 Comparison of modeled and observed stratospheric ozone springtime maxima.
vol I, pp 155-156.
Hadjinicolaou P., Pyle J.A., Jrrar A., and Harris N.R.P.
2004 Dynamical contributions to decadal ozone changes in the stratosphere
vol. II, pp 755-756
Jrrar A., Braesicke P., Hadjinicolaou P., Pyle J.A.
2004 Model study of ozone signatures of climate patterns over the northern hemisphere
vol. II, pp821-822
Geophysical Research Abstracts, EGU, General Assembly, Vienna, 24-29 April 2005
Zanis P., C. Zerefos, E. Maillard, E. Kosmidis, J. Stahelin
2005 Ozone variability and long-term trends deduced from the step-corrected Umkehr record of Arosa, Switzerland
EGU05-A-04473, vol.7
Andersen, S.B. 2005 Comparison of recent modelled and observed trends in total column ozone.
EGU05-A-10106, Vol. 7
Kivi, R., E. Kyrö, and T. Turunen
2005 Ozone profile observations at Sodankylä, Finland
EGU05-A-07735, Vol. 7
Hadjinicolaou P., Pyle J.A., and Harris N.R.P.
2005 The recent turnaround in stratospheric ozone over middle latitudes: a dynamical modelling approach
EGU05-A-08666, Vol. 7
Hadjinicolaou P., Pyle J.A., and Harris N.R.P.
2005 Past changes in stratospheric ozone over northern middle latitudes: modelling the chemical and dynamical impacts
EGU05-A-08846, Vol. 7
Vanicek K. 2005 Activities on monitoring and
research of atmospheric ozone in the Czech Republic.
Report of the Sixth Meeting of the Ozone Research Managers of the Parties to the Vienna Convention. Vienna 2005, UNEP/WMO/GAW
Vanicek K. 2005 Comparison of Brewer and recent satellite total ozone observations at selected NH mid and high latitude stations.
The Brewer Biannual Workshop, Delft, The Netherlands, 2005.
Metelka L. and Vaníček K.
2005 Application of neural models for simulation of total ozone in the European region.
Publication of the Czech Hydrometeorological Institute, Prague
ISBN:80-86690-30-X, June 2005
Hadjinicolaou P., Pyle J.A., and Harris N.R.P.
2005 Past changes in stratospheric ozone: modelling the chemical and dynamical impact
Royal Meteorological Society Conference, Exeter, UK, 11-16 Sept. 2005
18
Final Report
3. Final report 3.1 Work Package 1 3.1.1 General Objectives
The main objective of WP 1 is to perform long-term ozone trend analysis using variety of measurements. These include ground-base and satellite total ozone datasets, Umkehr and lidar ozone profile data. As a part of dataset preparation, WP1 also includes homogenisation of almost 40 years of Dobson and Brewer total ozone measurements at Hradec Kralove (Czech Republic). WP1 consists of five tasks:
Task 1.1: Development of the statistical models. Task 1.2: Analysis of Umkehr measurements of the Northern Hemisphere. Task 1.3. Analysis of lidar and other ozone measurements at Haute-Provence
Observatory (OHP). Task 1.4. Homogenization of ~40 year long total ozone series of Hradec Kralove
and analysis of chemical ozone losses by ‘neural models’. Task 1.5. Trend analysis of long-term ground based and satellite ozone
measurements. 3.1.2 Methodology and scientific achievements
Task 1.1 Development of Statistical Models (Contributors: ETHZ)
This task was mainly completed during first two years of the project (see 1st and 2nd Annual Reports). Detailed description of the method can be found in PhD dissertation of Jörg Mäder (Mäder, 2003). The attempt has been made to identify the most suitable explanatory variables on total ozone using multiple regression models and a large number of explanatory variables have been tested by an elimination process. Furthermore the effect of the Montreal protocol (and its amendments) has been studied by comparing models in which a linear trend or Equivalent Effective Stratospheric Chlorine (EESC) is used to fit the measurements.
The multiple regression models used here and adopted to the individual stations have the form:
TOZ = intercept + ak +b1X1 + b2X2 + .. + bjXj.+ .... + bmXm+ ε (1)
-where: TOZ (target variable): measured total ozone value. - ak: effect of month, k, k=1,2,….12, describing the seasonal variation - X1, X2, ..., Xj,.., Xm: explanatory variables - b1, b2,...,bj,...,bm: coefficients describing the influence of the respective
explanatory variable on total ozone - m: number of explanatory variables of the model
19
Final Report
- ε: random noise
station network
180 W 120 W 60 W 0 60 E 120 E 180 E
30 N
60 N
90 N
C Central Asia
E Europe
F Far East
P Polar
N North America
T Tropical
Figure 1.1 Stations used in the study
The study is based on monthly mean values of the target and the explanatory
variables. The total ozone measurements are extracted from WOUDC. Only stations with long-term series started before 1975 have been used (see Figure 1.1). A large variety of explanatory variables including many climate indices were tested. The method used to select an appropriate model includes the following steps: (1) Start with a multiple regression model including all reasonable explanatory variables. This leads to a statistical model which is strongly over-determined. (2) Eliminate the least important variable based on values of the F-statistic. The variable with highest p value (the least formally significant variable) is removed from the model. (3) Step (2) is repeated until only one explanatory variable is retained. (4) Calculate the ranks of the variables Xj, defined as the reversed order of the elimination: The variable eliminated first obtains rank m, the variable remaining in the last model, rank 1. (5) The ranking tables of the individual stations are averaged over the selected groups representing the different regions. Standard deviations of the ranks within the groups are determined in order to characterize the degree of agreement of the rankings. To obtain the most comprehensive model including the entire Northern hemisphere we applied step (5) based on the 6 regions (instead of the individual stations). Results of the trend analysis are further discussed in Task 1.5.
Task 1.1 resulted in Deliverable №2.
Task 1.2: Statistical analysis of Umkehr data (Contributors: NKAU)
During the 1st project year, the Umkehr datasets have been updated using the latest alg-99 (Bojkov et al., Meteorol. Atmos. Phys., 79, 127-158, 2002). Subsequently, trend analysis of the re-evaluated Umkehr ozone profiles has been performed. Table 1.1 lists all the stations used in the study. In order to attribute unambiguously the anthropogenic component to the observed long-term ozone changes, the contribution of various natural processes affecting stratospheric ozone has been taken into consideration in a multiple
20
Final Report
regression model with autoregression (AR=1) by using a number of explanatory variables such as: a) Indices from known climatic oscillations including the 11-year solar cycle, QBO, SOI, AO, and NAO, b) dynamical proxies including the tropopause pressure (from NCEP reanalysis), the EP-flux at 100hPa averaged over 45-75 latitudes (from ERA-40) and an equivalent latitude proxy, c) chemical proxies such as the Equivalent Effective Stratospheric Chlorine (EESC). The tropopause pressure and the equivalent latitude seem to be important predictors throughout the year in the lower stratosphere while AO and NAO are significant predictors for lower stratospheric levels during the cold season only (Figure 1.2). EESC explains part of the variability in the upper stratosphere throughout the year and in the lower stratosphere during spring. EP-flux has little correlation with Umkehr layers for all the Umkehr stations. Especially for Arosa, all the calculations have been additionally carried out for the recently re-evaluated Umkehr record.
Table 1.1: Updated Umkehr data sets in NH mid-latitudes using the algorithm 99. Station (Umkehr) in NH mid-latitudes
From (month/year)
To (month/year)
Arosa (035) 1/1956 1/2005 Aswan (245) 4/1985 9/2002 Belsk (068) 4/1963 10/2000 Boulder (067) 2/1978 6/2004 Kagoshima (007) 3/1958 12/2002 Mauna Loa (031) 12/1983 8/2002 Naha (190) 7/1974 12/2002 New Delhi (010) 9/1957 2/1998 OHP (040) 9/1983 6/2004 Sapporo (012) 3/1958 12/2002 Tateno (014) 8/1957 8/2004 Trend calculations were carried for all Umkehr stations out for the period 1970-2003 (and 1979-2003) but emphasis was especially given in the issue of “turnaround”. Different methods were used to study the problem of the turnaround of stratospheric ozone trends: (i) a statistical method using two linear trends including a linear downward trend starting in 1970 (or 1979) and an additional trend starting in 1996 and (ii) the method of cumulative sum (CUSUM) of white noise residuals after removing the linear trend based on the period 1/1970-12/1995. In order to attribute unambiguously any observed long-term changes to the various natural and anthropogenic processes that affect the ozone variability in the stratosphere, the climatic and dynamical proxies discussed in the previous paragraph have been introduced as explanatory variables in the multiple linear autoregression model. The double trend model method allows a change in trend at 12/1995 calculating two time coefficients, a pre-turnaround trend ω1 and a change in trend coefficient ω2 after turnaround. The overall trend estimate ω = ω1 + ω2 is a measure of the overall ‘‘recovery’’ effect of linear trend in the statistical model after the turning point T0. A positive change in the direction of a preexisting downward trend, that is, ω2 > 0, is a necessary (but not sufficient) requirement to affect an overall trend recovery, ω > 0. The vertical distribution of year round ozone trend estimates in %/decade for three
21
Final Report
stations Arosa, Boulder and Tateno using the double trend model are shown in Figure 1.3.
Figure 1.2: Time-height cross-section of the correlation coefficient of ozone Umkehr layers at Arosa during 1956-2004 with a) equivalent latitude (EL), b) tropopause pressure (TP), c) NAO index, and d) EESC for every individual month. The solid black at 0.3 and –0.3 lines indicate statistically significant correlation at 95% significance level. The layer 3 represents the sum of layers L2 and L3 while the layer 8 represents the sum of layers L8, L9, and L10.
Figure 1.3 shows for Arosa a notable positive change in trend after 1995 (ω2>0)
both at the lower stratosphere (below 23.5 km) and upper stratosphere (above 32.6 km). However this change in trend (ω2) is statistically significant at 95% confidence level only for the lower stratosphere. For Boulder we see a positive change in trend after 1995 (ω2>0) for the whole profile but not statistically significant at 95% confidence level. Finally for Tateno we see a positive and statistically significant at 95% confidence level change in trend both for middle and upper stratosphere but not for lower stratosphere. For all three stations we note negative ozone trends at the upper stratosphere for the pre-turnaround period. Additionally we fitted the measurements with Equivalent Effective Stratospheric Chlorine (EESC) instead of linear trend(s). The EESC model describes conceptually better the reality than the double trend model as it includes (besides proxies for natural related ozone variability) a time evolution of ozone depletion, which is more
22
Final Report
realistic than two linear trends. This EESC method showed statistical significant after turnaround trends in the upper stratosphere at Arosa and Boulder.
Arosa - Double trend model with qbo+LF107+TP
1
2
3
4
5
6
7
8
9
10
-10 -5 0 5 10 15 20
trend (%/decade)
Um
kehr
laye
rs pre-turnaround trend ω1
change in trend ω2
1
2
3
4
5
6
7
8
9
10
-10 -5 0 5 10 15 20
trend (%/decade)
Um
kehr
laye
rs
pre-turnaround trend ω1
change in trend ω2
Tateno - Double trend model with qbo+LF107+TP
1
2
3
4
5
6
7
8
9
10
-10 -5 0 5 10 15 20
trend (%/decade)
Um
kehr
laye
rs
pre-turnaround trend ω1
change in trend ω2
Boulder - Double trend model with qbo+LF107+TP
Arosa - Double trend model with qbo+LF107+TP
1
2
3
4
5
6
7
8
9
10
-10 -5 0 5 10 15 20
trend (%/decade)
Um
kehr
laye
rs pre-turnaround trend ω1
change in trend ω2
1
2
3
4
5
6
7
8
9
10
-10 -5 0 5 10 15 20
trend (%/decade)
Um
kehr
laye
rs
pre-turnaround trend ω1
change in trend ω2
Tateno - Double trend model with qbo+LF107+TP
1
2
3
4
5
6
7
8
9
10
-10 -5 0 5 10 15 20
trend (%/decade)
Um
kehr
laye
rs
pre-turnaround trend ω1
change in trend ω2
Boulder - Double trend model with qbo+LF107+TP
Figure 1.3: Vertical distribution of year round ozone trend estimates in %/decade for three stations Arosa, Boulder and Tateno using the double trend model. The pre-turnaround (set as year 1995) trend estimates ω1 are in red while the change in trend estimates ω2 are in blue. The double trend model includes QBO, solar cycle and tropopause pressure as explanatory variables to account for solar, QBO and dynamical related ozone variability. The Error bars denote 2-sigma standard errors. The layer 3 represents the sum of layers L2 and L3 while the layer 8 represents the sum of layers L8, L9, and L10.
However, the lack of statistical significance for the upper stratosphere at Arosa is
also indicated using the CUSUM method (see Figure 1.4). The lack of significant trends using the CUSUM method contradicts earlier results derived from upper stratospheric satellite measurements. It is presently not clear whether this difference originates from the inappropriate length of satellite measurements which are only available since 1978 and a previously not documented upper stratospheric ozone fluctuation of a period of 15-20 years.
23
Final Report
Figure 1.4: Cumulative sum of white noise residuals at Arosa during 1970-2004 for layers a) L8910, b) L7, c) L4 and d) L23 after removing the linear trend based on the period 1/1970-12/1995. The linear trend has been calculated after applying the regression models R1 (black line) which includes QBO and solar cycle, R2 (red line) which includes solar cycle, QBO and tropopause pressure (TP), R3 (green line) which includes solar cycle, QBO and equivalent latitude (EL) and R4 which includes solar cycle, QBO and NAO (blue line). The upper part of the 95% confidence envelope is denoted with a solid dashed line.
Although the positive change in trends after 1995 both for upper and lower stratospheric ozone is in line with the reduction of the emissions of ozone depleting substances from the successful implementation of the Montreal Protocol and its amendments, we recommend due to lack of significance for the upper stratospheric trends repeating this analysis in a few years in order to overcome ambiguous results for documentation of the turnaround of upper stratospheric ozone.
Task 1.2 resulted in Deliverable №3,10,11.
Task 1.3: Statistical analysis of stratospheric ozone and temperature data (lidar and other measurements of Haute-Provence Observatory (OHP). (Contributors: CNRS/UPMC) Estimation of the influence of different processes on the ozone series and trend analysis
Three series of ozone profile measurements at OHP: Lidar, Umkehr and also SAGE II have been statistically analysed by means of multiple linear regression. A discrepancy
24
Final Report
is observed in the time-series of ozone anomaly among different measurements at the highest altitudes. This discrepancy led to different correlation with explanatory parameters, especially those having a long term cycle like as 11-year solar cycle.
Regression model includes following explanatory parameters: QBO, solar cycle,
NAO, EP-flux, aerosol, EESC (Equivalent Effective Stratospheric Chlorine), tropopause pressure and teleconnection patterns (East Atlantic Oscillation, Scandinavian pattern and EU-blocking pattern). These parameters have been tested and the relative priority has been derived among them. For the annual time-series of ozone profiles, QBO, solar cycle and tropopause pressure are the dominant explanatory parameters in the whole range, the highest and middle altitudes (27~37km) respectively. As regards the seasonal time-series, QBO and solar cycle are of importance as well. NAO and EP-flux explained significantly the ozone variation in the summertime and wintertime (more than 50% in the middle stratosphere) respectively.
Inclusion of recent years in ozone time-series resulted in a switch of the ozone
trends in the positive direction at all altitudes. However, because of the variation of measurement precision as a function of altitude and of the discrepancy in trends analysis depending on the measurement method, it is premature to draw a definitive conclusion on the turnover of stratospheric ozone recovery from the regression of annual ozone profiles.
Figure 1.5 Trend of ozone vertical distribution as a function of altitude from lidar
measurements in the period 1985-2000 and 1985-2003 and from SAGE measurements in the period 1984-2003.
25
Final Report
Figure 1.6 Influence (%) of polar (winter) and subtropical (winter and summer) air
mass at OHP (from left to right). Average of 7 isentropic levels (450, 475, …, 600 K) for winter and 5 levels (500, 525, …, 600 K) for summer.
Assessment of polar and subtropical air-mass influence on OHP ozone amount
The fraction of polar/subtropical air masses in ozone observation is quantified by transport analysis using Potential Vorticity (PV) field. High resolution PV fields at different isentropic levels are modelled by means of the MIMOSA model (Modèle Isentropique de transport Mésoéchelle de l'Ozone Stratosphérique par Advection) and ERA40 data. PV distribution as a function of equivalent latitude allowed to determine the borders of polar and subtropical air mass. More than 40 % of air masses at OHP are advected from subtropical region during summer. The average influence of subtropical air mass is about 3 % during winter with a maximum of 7 % in 1990. The influence of polar air mass during winter is more significant (approximately 7 % on the average) than that of subtropical air mass and its inter-annual evolution is quite variable (Figure 1.6). During the last project year, this analysis has been extended for the entire mid-latitude band using SAGE II measurements. The overall fraction of polar and subtropical air masses increases as the isentropic level increases and the occurrence of polar air mass is more frequent from 1995. The influence of polar air mass is discontinuous and significant at the middle stratosphere in the wintertime. In mid-latitudes (43.9±5N), Pacific range shows, in particular, a lower polar fraction and a higher subtropical fraction compared to the other regions. It is concluded that this could lead to a regional difference in the ozone trends.
Temperature trends at OHP Long-term temperature lidar measurements are performed according to the
backscatter lidar technique at OHP since 1979. These measurements were used to retrieve the temperature trends in the mid to upper stratosphere [30 –50 km]. Due to interferences with aerosols, the backscatter lidar technique can provide temperature measurements only above 30 km in the mid-latitude stratosphere. In addition, temperature measurements from radiosoundings performed in the nearby town of Nîmes are used to derived temperature trends in the lower stratosphere in the 15 – 28 km range. The temperature trends are computed using a classic multilinear regression approach similar to that used for the ozone profile analysis. In Figure 1.7 are displayed temperature anomalies from the lidar time series for the period [1979 – 2004] in the upper panel and from the sondes time series for the period [1985 – 2004] in the lower panel. Monthly temperature anomalies correspond to the difference at each altitude range between the monthly temperature
26
Final Report
mean and the climatological temperature average for the corresponding month. Figure 1.7 shows that temperature anomalies are larger in the mid-stratosphere than in the lower stratosphere. In this region, anomalies can exceed 10 K in absolute value particularly in winter, while they are in the range of ± 2 K in the lower stratosphere. The temperature anomalies are generally coherent over 10 to 15 km in the former region. Particular warm anomalies are visible in the winter 1984/85 with values reaching 14 K in January in the 30-40 km region. In general, the 1980-1990 decade was warmer than the following one. The period 1995-2000 was particularly cold as compared to the period 1985-1990. More recently the year 2003 has been rather warm as compared to the previous decade. Linear Temperature trends have been computed from these times series and are shown in Figure 1.8 as a function of altitude. Although expressed in K/decade, the linear trends do not correspond to the same time period for both time series. Temperature trends at OHP show a cooling of the stratosphere reaching –1.2 K/decade at 45 km.
Figure 1.7 Temporal evolution of the monthly temperature anomalies (K) from
lidar measurements (upper panel) and from sondes measurements (lower panel) as a function of altitude at OHP
27
Final Report
Figure 1.8 Temperature trends at OHP in K/decade as a function of altitude, from
the lidar and radiosounding measurements. Error bars correspond to 2 sigma uncertainty.
Task 1.3 resulted in Deliverables №4, 5, 6, 7. Task 1.4: Homogenization of the 40 year record from Hradec Kralove. (Contributors: CHMI)
In this task, two main objectives have been met: i) homogenisation of the 40-year record of total ozone observations at Hradec Kralove observatory, and ii) development and application of ‘neural models’ for estimation of reduction total ozone due to chemical processes in the European Region.
Homogenisation of the 40-year record of total ozone observations at Hradec Kralove observatory
52.162 hand-taken Dobson total ozone observations performed at the Solar and Ozone Observatory of CHMI in Hradec Kralove (SOO-HK), CR, have been re-calculated by standard procedures and by means of re-defined calibration constants of the D074 ozone spectrophotometer of the period 1961-2002. Also, 87.774 total ozone measurements carried out at SOO-HK by the automated Brewer B098 spectrophotometer have been re-evaluated using homogenised calibration constants of B098. The DS and ZS observations were compared and accuracy of updated zenith polynomials investigated and assessed. The re-evaluated Dobson and Brewer total ozone observations taken with D074 and B098 spectrophotometers in Hradec Kralove have been defined as D074-V2003 and B098-V2003 data sets and have been re-deposited into the World Ozone and UV Data Center (WOUDC), Toronto in 2005. The methodology of re-evaluation and results achieved are summarized and published in (Vanicek et al , 2003).
Parallel Dobson and Brewer total ozone observations show an evident seasonal
course of differences between both data series. Amplitudes of the differences are more pronounced for DS daily averages and they have a similar magnitude like differences from other mid-latitude stations with collocated Dobson and Brewer spectrophotometers
28
Final Report
(Staehelin et al., 2003). If strictly simultaneous observations (+-10min.) were selected and corrected for total SO2 and for ozone effective temperature then the amplitudes of differences have been reduced by about a half but certain residual seasonal oscillations still persist - see Figure 1.9 These are expected to be originated by the uncertainty in definition of temperature dependency of ozone cross sections related to UV wavelengths really selected by spectrophotometers during routine measurements. The comparison of D074 and B098 observations also shows a sudden shift (about -1%) between both data series in June/July 1997 after intercomparisons of the spectrophotometers towards traveling references. The offset has been confirmed by next ICs realized in later years. The most probable explanation is that this change has appeared due to a shift between calibration levels of the traveling references in the second half of the nineties. Therefore, this phenomenon needs to be investigated at more stations.
Differences between simultaneous D074 and B098 total ozone observations (original, SO2-corrected, TO3-corrected) Hradec Králové, 1994-2002
-10
-8
-6
-4
-2
0
2
4
6
8
10
I-94 I-95 I-96 I-97 I-98 I-99 I-00 I-01 I-02 I-03
Year
100*
(D-B
)/D (
%)
Original+TO3cor Original Original+TO3cor+SO2cor
2.434 pairs
Figure 1.9 Differences between individual simultaneous (+- 10 min) DS total ozone observations measured with D074 and B098 instruments. Original, SO2 and TO3 corrected values.
The re-evaluated total ozone measurements from SOO-HK were compared with
satellite overpass observations TOMS-V7. This comparison gave a very good (below 1 %) agreement and consistency with D074-V2003. Thus a high quality of the re-calculated Dobson measurements taken in the period 1978-1993 has been validated. Comparison of D074-V2003 and B098-V2003 towards EP-TOMS and GOME data sets confirmed the systematic bias between D074 and B098 observations after July 1997 that was mentioned in the previous paragraph. TOMS-8 and GOME-WFDOAS satellite data sets do not have the same relation to the ground observations. The best fit appears between B098 and GOME (0.3 %) while TOMS is lower than B098 by 1.5 % and the differences are seasonally dependent. Since 2002 the TOMS observations have been shifted towards ground data by about-4 % and could not be used for comparisons after this date. The GOME-WFDOAS measurements give a good fit and time-consistent relation with B098
29
Final Report
while for D074 an evident seasonal variation of differences exists. This supports the conclusion that the differences between D074 and B098 occur mainly due to seasonal influence of ozone effective temperature on the Dobson observations. The analyses indicated that TOMS and GOME data sets are not fully consistent with each other over the whole year in Hradec Kralove. Their validation by Brewer data from high-latitude stations (Sodankyla and Tromso) confirmed a better fit for GOME (0.2%) than for TOMS (-1.8%) and the -4% offset of TOMS after 2001.
It is concluded that:
• Uncorrected differences between Dobson and Brewer data can introduce an additional statistical signal into long-term seasonal trends of total ozone if Dobson observations are replaced/continued by Brewer measurements at a particular station. Therefore, combined D074 and B098 data series are recommended to be used for investigation of this problem that could imply a general impact in the global GAW network.
• The seasonal differences between Dobson, Brewer total ozone observations and data from particular satellite missions are highly correlated with ozone effective temperature at Hradec Kralove. This phenomenon should be investigated for more stations to assess/avoid future unrealistic shifts in decadal ozone changes. In the Northern high latitudes a special attention should be paid to the analysis of the relation between Brewer and satellite observations.
Development and application of ‘neural models’ for simulations of total ozone in the European Region.
The main goal of this study was to develop the neural model which would be able to simulate relation between the CANDIDOZ model proxies (heights and temperatures at selected tropospheric and stratospheric levels, stratospheric wind, NAO-index, aerosol optical depth, solar activity) and total ozone to estimate chemically induced ozone losses in the European region. Detailed description of the methodology and results can be found in Metelka et al (2005).
Single-station methodological tests were carried out by the neural model within the
first project year. The methodology was then applied on the data from several selected stations located in the European region (Belsk, Hradec Kralove, Arosa, Uccle, Lerwick, Oslo, and Tromso). The model was then run with ERA-40 stratospheric data and monthly total ozone values for the period IX/1957 – XII/1999 at ERA-40 grid points (region 5W-25E, 45N-75N, step 2.5°) were obtained. Two versions of the neural model were trained and applied - the time-dependent model that includes a chemical component for the ozone hole period and the time-independent model that fixes influence of stratospheric dynamics on total ozone prior the 80-ties. Residuals between both models then estimated the magnitude and time evolvement of impacts of the EESC. The overall model performance was tested by direct comparison between measured total ozone data and modeled data at the nearest grid point. This comparison indicated good agreement between observed and modeled values.
30
Final Report
Figure 1.10 5-yeas means of adjusted estimate of the chemical factor throughout the region and 95% confidence intervals of the mean values
The modeled values indicate that latitudinal dependency of the estimated chemical component is more pronounced than the longitudinal one (Figure 1.10). In the ERA- 40 period its contribution to overall ozone change varies from 25 to 40 D.U (6 -12% with respect to the late 50s and early 60s) in high latitudes to almost zero in southernmost parts of the region. The prevailing part of total ozone changes occurred due to increasing influence of EESC in the recent decades. Only about 5 - 10 D.U. (1 – 3%). can be explained by systematic long-term changes of proxies, i.e. by processes of the non-chemical origin. Results also indicate large changes in annual course of the chemical component. In D,J,F,M,A in high latitudes it contributes by about 30 - 40 D.U. (10-15%) and rapidly decreases southward. In late summer and early autumn the changes of annual course of EESC do not exceed 10 D.U. (about 3%) even in the northernmost parts of the region. The chemical component significantly contributes to the decadal total ozone change in the European region. In high latitudes (60-75 deg N) total ozone has been reduced due to EESC by about 5 - 10% with respect to the late 50s and early 60s. In the southern parts the chemical factor does not play important role in the long-term ozone change. The neural model simulations show that the magnitude of the chemical losses in the European region has become stable in the recent decade but no tendency to decreasing has been identified yet (Figure 1.11).
31
Final Report
Figure 1.11 Time evolution of mean, maximum and minimum values of the adjusted chemical factor throughout the region, month by month
Both time-dependent and time-independent technologies can be also used for
validation or homogenization (e.g. filling up gaps in interrupted observations) of long-term ground total ozone data series.
Task 1.4 resulted in Deliverables №8, 9, 12. Task 1.5: Trend analysis of long-term ozone series of selected ground based stations, trend analysis of satellite measurements. (Contributors: ETHZ, UIP-UB)
Trend analysis of ground based and satellite total ozone series has been performed. A new algorithm to retrieve total ozone from GOME measurements (GOME WF-DOAS) has been developed and applied to obtain improved data series. In addition, 3-D ozone data set has been created based on satellite measurements and ERA-40 PV fields.
Ground-based total ozone trend analysis
Ground based total ozone series have been analysed using model developed in Task 1.1. A detailed discussion of the results can be found in PhD thesises of Jörg Mäder (see Mäder, 2003). It was shown that most variability of total ozone can be explained by the seasonal variation (described by m in this approach), local variables (such as temperature
32
Final Report
at 50 hPa, tropopause pressure, or PV400), aerosols and EESC. If the local variables are excluded m, EESC and aerosols are still within the highest ranking explanatory variables, but the ranks of the other explanatory variables are much more dependent on the different regions. With the same type of elimination process we tested whether a linear trend or EESC are more suitable to describe the long-term evolution of total ozone in the northern hemisphere. The results can be viewed as (qualitative) evidence for the effectiveness of the Montreal Protocol. Further analysis aiming at the attribution of the trends to chemical and dynamical factors is presented in synthesis package WP5. New algorithm for GOME ozone retrieval
A new algorithm to retrieve total ozone from nadir observations of the Global Ozone Monitoring Experiment has been developed. By fitting the vertically integrated ozone weighting function rather than ozone cross-section to the sun-normalized radiances, a direct retrieval of vertical column amounts is possible. This new weighting function DOAS approach takes into account the slant path wavelength modulation that is usually neglected in the standard DOAS approach using single air mass factors. Further improvements have been achieved by taking into account molecular ozone filling-in as part of the so-called Ring effect. For the first time several auxiliary quantities directly derived from the GOME spectral range such as cloud top height, cloud fraction and effective albedo are used in combination as input to the ozone retrieval. The precision of the total ozone retrieval is estimated to be better than 3% for solar zenith angles below 80°.
Results from the algorithm were analysed by comparing them with correlative
measurements of total ozone from ground, GOME Version 3.0 (GDP V3.0), and TOMS V8. Station data from the World Ozone and Ultraviolet Radiation Data Center (WOUDC) and Network for the Detection of Stratospheric Changes (NDSC) were selected for the global comparison. They comprise of Dobson, Brewer and SOAZ spectrometers, but Dobsons are in majority. The collocation radius was set to less than 300 km between satellite and the correlative data and both are from the same day. At mid-latitudes (30°-60°) excellent agreement between WF-DOAS and WOUDC data has been reached (Figure 1.12). Seasonal variation in the observed differences are generally below ±0.5%, while GDP V3 and TOMS V8 show larger annual variations on the order of ±1.5% and ±1%, respectively. No significant solar zenith angle (SZA) dependence is observed in the WFDOAS-station differences. GOME V3 underestimates (1 to 2%) in the tropics, while WF-DOAS agrees to within 1%.
33
Final Report
Figure 1.12 Comparison of WF-DOAS (black and blue) and V3 GOME total ozone (red) with 9 western European stations at mid-latitudes as a function of the day of year (1996-1999). WFD-MI stands for the retrieval using mid-latitude ozone profile shapes (black, default retrieval) and WFD-HI using high-latitude profile shape (blue, testing the profile shape effect) from the TOMS V7 profile climatology (Wellemeyer et al., 1997). Black dotted lines are the 2σ scatter of the differences (WFD-MI). Similar results (only shifted according to season) are observed when comparing with SH mid–latitude stations.
At polar latitudes WFDOAS differences increase on average to 2 to 5% at high
solar zenith angles particularly for those stations which are near/inside the polar vortex. Under ozone hole condition WF-DOAS differences can be as high as 6% to 8% A strong positive bias is also observed with GOME V3 and TOMS V8 with respect to ground data at high solar zenith angle in polar region. This observed satellite-station bias may come from (1) stratospheric ozone temperature correction that is not accounted for in the retrieval technique of Dobson/Brewer measurements but in the GOME retrieval, (2) enhanced stray light problem associated with both satellite and ground measurements, and (3) other sources originating from instrumental differences and differences in retrieval wavelengths. Nevertheless, the accuracy of the WFDOAS results are to within the uncertainty of the ground-based measurements at high latitudes, see for instance results from the TOMS3-F campaign at Fairbanks (Staehelin et al., 2003), where improved straylight and ozone temperature correction lead to differences of +3 to +4% in the ground data compared to the WMO-GAW standard retrieval.
Figure 1.13 shows the monthly mean differences between WF-DOAS and
collocated Dobson (both direct sun and zenith-sky) from Lauder (45°S) for the period 1996-2003. Particularly striking is the long-term stability of the GOME retrieval, despite the fact that increasing optical degradation of the GOME scan mirror is observed since 2000 (Tanzi et al., 2001). For comparison the time series of the TOMS V8 – Lauder differences are also shown. In general, TOMS V8 tends to underestimate ozone during winter/spring thus showing a distinct seasonal cycle in the differences to ground data.
34
Final Report
Figure 1.13 Monthly mean differences between satellite and Lauder Dobson data. Top panel: WF-DOAS, bottom panel: TOMS V8 (until 2000). Orange curves are the results from fitting a sine time series to the daily data.
Separating the validation into comparisons with Dobsons and Brewers at the same
location and over an extended period is quite instructive. Such separate comparisons have been carried out with data from Hradec-Kralove (50°N) and Hohenpeissenberg (48°N). Generally, the agreement is better with the Brewer measurements than with the Dobson. The differences between Dobson and Brewer daily averages from the same day have a distinct seasonal signature varying from –3% (winter) to 0% (summer) as discussed in Vanicek (1998).
The agreement between WF-DOAS and Dobson significantly improves if the proper stratospheric ozone temperature is applied in the Dobson and Brewer retrieval. From ozone sonde ascents at Hohenpeissenberg from the same day an ozone profile weighted stratospheric temperature was derived that lead to an average increase of +3 DU in the Brewer and up to +8 DU in the Dobson retrieval during winter (Vanicek et al., 2003). Taking these corrections into account the WF-DOAS comparison particularly with the Hradec-Kralove Dobson improved considerably with no seasonal cycle signature left in the observed differences.
The developed WF-DOAS algorithm is attractive for application to SCIAMACHY
and future instruments like OMI and GOME2. WF-DOAS version of daily GOME data has been compiled and trend analysis has been performed. Results are discussed in the synthesis package WP5.
Statistical analysis of the satellite ozone data During the reporting period, trend analysis of the data set CATO (CANDIDOZ
Assimilated Three-dimensional Ozone) was completed. CATO is a quasi-three dimensional ozone data set with daily resolution for the period 1979 - 2003 reconstructed
35
Final Report
from a combined TOMS/GOME/SBUV satellite total ozone data set using meteorological information on short-term adiabatic advection and adopting a Kalman filter for the sequential assimilation of the measurements (Brunner et al., 2006a, in press). This unique data set is an important addition to existing ones having lower spatial and temporal coverage.
From this reconstructed data set, time series of monthly values of a) total ozone as a
function of equivalent latitude, and b) ozone partial pressures as a function of equivalent latitude and pressure have been created for detailed trend analysis. We have employed a state-of-the-art multiple linear regression model using explanatory variables describing the influence of the solar cycle, the Quasi-Biennial Oscillation (QBO), volcanic stratospheric aerosols, the strength of the Brewer-Dobson circulation, and polar stratospheric ozone depletion. A linear response of ozone to changes in these variables is assumed. The proxies are selected to model as closely as possible the most relevant processes influencing stratospheric ozone variability.
Adopted multiple linear regression model is similar to that used in previous studies and can be formulated as: ,
where Y(t) is the target variable, i.e. the measured time series of total ozone or ozone partial pressure, and Xj(t) are time series of N different explanatory variables and c
∑ =+⋅+⋅+= N
j jj ttXtcttbtatY1
)()()()()()( ε
j their regression coefficients. Table 1 presents a list of the different explanatory variables and the source of the data. Time t (number of months since beginning of the time series) may be used as additional explanatory variable to describe the linear trend in ozone not explained by the other proxies of natural variability. Its coefficient b can thus be interpreted as representing the influence of human activity (i.e. the release of CFCs). In some simulations we have replaced time by Equivalent Effective Stratospheric Chlorine (EESC) to represent the anthropogenic influence more directly. The coefficient a is the seasonally varying offset and ε(t) is the autocorrelated residual time series. All regression coefficients a, b, and cj are chosen to be time dependent as they are allowed to vary with season. This time dependence is modelled by 12-month, 6-month, and 4-month sine and cosine harmonic series, i.e.
)12
2sin()12
2cos()(1 12,2,1,
tkctkcctc Kk kjkjjj ⋅⋅+⋅⋅+= ∑ = + ππ , K=3
and the same formula is used for a(t) and b(t).
36
Final Report
a) QBO@30 hPa b) QBO@10 hPa c) Volcanic aerosols d) Solar Cycle e) EP-flux f) VPSC x EESC g) EP-flux (no seasonal offset) h) VPSCxEESC(no seas.) Figure 1.14 Contributions to variability in total ozone as a function of season and equivalent latitude. To show all figures in comparable units the regression coefficients are multiplied by one standard deviation of each proxy time series. Values thus represent the change in total ozone (DU) per one standard deviation change in the proxy. Shading indicates statistically insignificant values at the 95% significance level. Panels g and h are taken from a simulation where no seasonal cycle of the offset coefficient was included (i.e.coefficient a is the same for all months).
The analysis of trends and variability in total ozone as a function of season and
equivalent latitude is presented in Figures 1.14 and 1.15. If not stated differently the results are obtained from a model including all proxies (with either time t or EESC describing the anthropogenic influence) and covering the period Jan 1979 - Dec 2003. The Quasi-Biennial Oscillation is represented by two separate components, the zonal wind at 30 hPa (Figure 1.14a) and at 10 hPa (Figure 1.14b). The phase of the QBO is approximately π/2 apart at these levels. Since any phase lag φ of a wave sin(ωt+ φ) can be reproduced by a linear combination of a sine and a cosine function the model is able to
37
Final Report
adjust the time lag of the ozone response automatically. The QBO signal makes a significant contribution to ozone variability both in a narrow band in the tropics (throughout the year) and in the extratropics between 25° and 55° (from winter to spring on the respective hemisphere). A prominent signal is also seen over the North Pole but it is not significant at the 95% confidence level.
The effect of volcanic eruptions is most evident at mid to high northern latitudes
during the winter and spring months (Figure 1.14c). The volcanoes El Chichon (17°N) and Pinatubo (15°N) are both located in the NH and their eruptions in 1982 and 1991, respectively, have more strongly perturbed the stratospheric trace gas composition in the NH even though aerosol surface area changes were comparable in the SH. Some effect on total ozone is nevertheless also seen in the SH shifted by 6 months compared to the NH. Ozone depletion is strong in the SH due to processes in the cold Antarctic vortex and the additional aerosols from volcanic eruptions appear to add relatively little to this. Some ozone depletion following the eruptions appears to have occurred also at low latitudes (mainly 10-20°) in both hemispheres. Total ozone is generally enhanced during periods of maximum solar activity (Figure 1.14d). The effect is strongest at subtropical latitudes throughout the year but it appears to be also present at high latitudes during winter and spring.
The Brewer-Dobson circulation has a profound impact on total ozone
concentrations at high northern latitudes between January and May (Figure 1.14e). It explains a dominant fraction of the generally large variability over this region. An intensified circulation is associated with high total ozone values over the poles and reduced values at low latitudes, with a local minimum of the effect between 30° and 40°N. The pattern is similar in the SH but shifted by 6 months.
Finally, integrated PSC volume is a good proxy for total ozone loss within the
Arctic polar vortex (Rex et al., 2004). Since no similar data set was available to us for the SH we focus our analysis on the NH only. We have multiplied the PSC volume by EESC to account for the modulation of polar ozone loss by long-term changes in stratospheric chlorine loading. As expected, the signal is strongest between February and April north of 70°N. Interestingly, the signal spreads to lower latitudes (down to nearly 30°N) between April and July suggesting a significant impact of polar ozone destruction in mid-latitudes during summer.
The EP-flux signal disappears surprisingly rapidly between May and July despite
the fact that total ozone values in late summer are closely connected to those in spring as shown by Fioletov et al., (2003). The model attributes this memory effect at least partly to polar ozone loss rather than to the Brewer-Dobson circulation but the two processes can not well be separated because they are closely correlated. In a simulation excluding VPSCxEESC the EP-flux proxy takes over much of the signal of VPSCxEESC and in that case its influence extends much longer into summer and early autumn (not shown).
Figures 1.14g and 1.14h are included to illustrate the difference between a
simulation with and without including a seasonal cycle of the offset (model coefficient a). The annual cycle in ozone, in particular at high latitudes, can be largely explained by the annual cycle of the Brewer-Dobson circulation which in turn is dominated by wave
38
Final Report
forcing during winter (Fioletov et al., 2003). Therefore the model has difficulties in separating a seasonal cycle from the influence of the EP-flux proxy. Figures 1.14g and 1.14h suggest that the influence of EP-flux and polar ozone depletion at high latitudes can be more realistically described in a model without seasonal cycle coefficient. In this regression model the influence of EP-flux extends further south in the NH and also later into the year (until September instead of June). In addition, the model attributes a larger fraction of ozone variability over the Arctic in spring to polar ozone loss and less to EP-flux.
Figure 1.15 compares linear trends in total ozone (anthropogenic influence)
estimated for two different time periods, 1979-1995 and 1979-2003. The linear downward trend of ozone is considerably smaller for the longer time period (Figure 1.15b) ending in 2003 compared to that ending in 1995 whereas the significance levels are generally higher due to the longer time span. Figure 1.15c additionally shows total ozone changes due to changes in EESC. The pattern of EESC is very similar to that of linear trend 1979-1995 with strongest decreases at high latitudes in both hemispheres maximizing in Feb/Mar in the NH and in Oct/Nov in the SH. However, significant downward trends also exist throughout the year polewards of about 30°. There are indications for a maximum in northern mid-latitudes between January and April which appears to be detached from the maximum ozone loss over the pole. a) Linear trend 1979 – 1995 b) Linear trend 1979 – 2003 c) EESC 1979 – 2003 Figure 1.15 Model estimates of the anthropogenic influence on ozone trends. Panels a and b show the linear trend component (coefficient b) in percent per decade for two different time periods (in panel a the model was applied only to the period 1979-1995). Panel c shows the influence of changes in EESC from a simulation where the variable time has been replaced by EESC in the regression model.
Figure 1.16 shows the different annually averaged contributions to the variability of
ozone partial pressures as a function of equivalent latitude and pressure. As before, the full model with EESC was used. The vertical structure of the signals provides additional insight into the total ozone responses. QBO, for instance, has a profound influence on stratospheric ozone in the entire stratosphere (Figure 1.16a and 1.16b). However, the response can be at the same time positive and negative at different levels. Summing over the levels the contributions may cancel out each other. This cancelling is responsible for the bands of minimum response of total ozone to QBO variations seen in Figures 1.14a and 1.14b at about 10° and 60°.
39
Final Report
a) QBO @ 30 hPa b) QBO @ 10 hPa c) Volcanic aerosols
d) Solar cycle e) EP-flux f) VPSC x EESC g) EP-flux (no seasonal offset) h) VPSC x EESC (no seas.)
Figure 1.16 Annual average contributions to variability in ozone partial pressure as a function of equivalent latitude and pressure. Panels g and h are taken from a simulation where the seasonal cycle of the offset coefficient a was suppressed. See Figure 1.14 for further details.
The ozone response to volcanic eruptions is negative below about 24-25 km
altitude and extends down into the lowermost stratosphere poleward of about 40° (Figure 1.16c). Between 25 and 35 km the response is positive, in good agreement with findings of Hofmann et al. (1994) based on ozone sonde observations at Boulder, USA. The solar cycle signal is relatively week and only significant and positive between about 25 and 35 km altitude with largest amplitude around 30° north and south (Fig. 1.16d). There are also signals over polar regions and in the lower stratosphere but these are less certain due to the high variability of ozone over these areas. EP-flux again dominates the variability at high latitudes (Figure 1.16e). An increased EP-flux leads to strongly enhanced values
40
Final Report
from the tropopause to 30 km altitude north of about 60°N and to lower ozone concentrations between 17 and 25 km between the equator and the mid-latitudes. The influence is minimal in between at about 50° in both hemispheres. Finally, arctic ozone depletion (Figure 1.16f) is largest north of about 65°N but contributions are also seen in the lower stratosphere (13-17 km) in midlatitudes. As suggested by Figure 1.14f, the latter is probably due to southward transport of polar air depleted in ozone between April and July following the break-up of the polar vortex. The signal around 22 km extending from 60°N towards the equator is probably due to the imperfect separation between EP-flux and VPSC effects. In a simulation excluding VPSC the EP-flux signal tends to become positive in this area whereas it is mostly negative in a simulation including VPSC. This is a clear indication for the competition between these proxies in the model. Figures 1.16g and 1.16h are taken from a simulation excluding a seasonal cycle of the offset. They highlight the problem of the competition of the seasonal cycle with EP-flux and, to a lesser extent, with VPSC.
a) Linear trend 1979 – 2003 b) EESC 1979 – 2003 Figure 1.17 Same as Figure 1.15 but model estimates of the anthropogenic influence on annual mean O3 partial pressures as a function of equivalent latitude and pressure.
Figure 1.17 shows the vertical distribution of the annual mean linear trend (Figure
1.17a) and influence of changes in EESC (Figure 1.17b), which can be compared to Figure 1.15 showing the horizontal and seasonal distribution of total ozone trends. Largest downward trends (1979-2003) of more than 20%/decade are seen over the Antarctic south of 60°N below 40 hPa (or 25 km). Much lower though still significant annual mean downward trends of up to 10%/decade are seen over the Arctic. The maximum negative trend is located at about 19 km altitude but is less well confined in latitude as over the Antarctic. In both panels a vertically well confined band of significant downward trends is seen near 20 km altitude extending from about 40°S to 40°N, which is a feature yet unexplained. Another region of large negative trends in the upper tropical troposphere is only seen in panel 1.17a. This is not a robust feature as it changes significantly between different model versions. Ozone partial pressures are low in this region and therefore the large percentage changes concur with only small absolute changes. Note also that trends in the upper stratosphere (around 2 hPa) are underestimated in this version of CATO because at these levels the assimilated ozone distribution was relaxed towards a climatology.
Task 1.5 resulted in Deliverables №1, 13, 14, 15, and 16.
41
Final Report
3.2 Work Package 2 3.2.1 General objectives
The main objectives of Work Package 2 are to assess polar ozone variability, its dependence on polar vortex, and effects to mid-latitudes. WP2 also includes homogenization of the long-term total ozone series in the Arctic. The work package consists of the following tasks:
Task 1: Decadal variations of polar vortices Task 2: Statistical analysis of polar ozone data including re-evaluation of historical
total ozone series. Task 3: Impact of polar ozone depletion on midlatitude trends. Task 4: Isentropic exchange: ozone laminae and mini-holes.
3.2.2 Methodology and scientific achievements Task 2.1 Decadal variations of polar vortices (Contributors: FMI, AWI)
The main objectives of this task were to extend statistical vortex studies over the period covered by new ECMWF 40-year reanalysis and also to provide indicators of polar ozone loss suitable for the statistical models. The results are summarized in Karpetchko et al (2005). Using algorithms developed during the project such parameters of the Arctic and Antarctic vortices as vortex area, strength, and longevity, vortex mean temperature, as well as area with T<Tnat (PSC area) have been calculated for the whole period of the ERA40 data base. The vortex parameters have been calculated on several isentropic levels from 395 to 850K. Some issues of the ERA40 stratospheric data quality related to the vortex have been identified. These are: cold bias (with respect to the other existing data sets) during early winter in the Arctic, too strong and too cold vortex during the pre-satellite period in the Antarctic, vertically oscillating temperatures over the Arctic since 1998 and over the Antarctic during the whole period. A study of interrelationship between vortex characteristics revealed significant qualitative differences between the Arctic and Antarctic. Using winter averaged vortex parameters, it has been found that in the Arctic, a larger vortex is usually colder and stronger whereas in the Antarctic winter such relationship is not established. These findings were confirmed by the NCEP/NCAR reanalyses. Long-term changes in the vortex parameters have been assessed. Results at selected isentropic levels are presented in Table 2.1. Due to suspicious ERA40 data in pre-satellite period, trend analysis of Antarctic vortex was restricted to the period after 1979. It is found that the Arctic PSC area has increased during 1958-2002 period (barely significant at the 95% significance level) but no statistically significant trends in size, coldness or longevity of the Arctic lower stratospheric vortex since 1979 are found as opposed to earlier studies. In contrast, for the period 1979-1997, positive trends statistically significant at 95% level are found in vortex longevity and sizes of March vortex area and March PSC area in the lower stratosphere. For the period 1979-1997, positive trend in March vortex strength is significant at 99%. The Antarctic spring vortex
42
Final Report
has become stronger, colder and breaks up later during the period 1979-2001. However, the Antarctic vortex cooling has not affected the October vortex area, which shows only little change for the same period. It is found that the area of the Antarctic vortex during late winter and spring depends on the planetary wave propagation to the stratosphere in the preceding period whereas the corresponding relationship between the waves and the PSC area in October is destroyed by the trends in the PSC area.
Table 2.1: Trends and two-sigma uncertainties in vortex parameters.Trends significant at 95% level are marked in blue
NH SH Isentropic level, K
(1958-2002) (1979-2002) (1979-1997) (1979-2001)
475 0.8±3.8 5.0±11.0 15.1±12.7 9.8±5.7 600 1.1±4.1 3.7±12.2 11.2±13.9 9.1±5.7
Break-up date, (days/decade)
850 -0.9±3.6 7.5±9.1 10.7±12.4 5.1±5.8 475 0.3±0.7 1.0±2.0 3.7±2.2 4.1±2.0 600 1.2±1.6 2.8±4.7 8.2±5.7 5.7±3.3
March/October vortex strength ((m/s)×
(PVU/deg) /year) 850 -0.6±0.9 0.9±2.0 2.9±2.9 2.1±1.7 395 2.2±3.0 1.7±8.7 14.4±10.0 0.1±2.7 475 0.7±2.1 0.8±5.8 7.5±7.0 -1.8±2.7
March/October vortex area
(106 km2/decade) 530 0.5±2.1 0.5±5.6 6.8±6.7 -1.6±2.8 395 0.16±0.15 0.4±0.5 0.9±0.7 3.8±1.8 475 0.15±0.15 0.3±0.5 0.8±0.7 3.1±1.1
March/October PSC area
(106 km2/ decade) 530 0.06±0.07 0.1±0.3 0.4±0.4 0.2±0.5 The PSC areas and PSC volumes (VPSC) have been provided to the partners for
using in the statistical models as a proxy for chemical ozone destruction. Task 2.1 resulted in Deliverables № 18, 19.
Task 2.2 Statistical analysis of polar ozone data (Contributors: NILU, FMI)
Analysis of polar ozone data included creation of regression models of total ozone and ozone profiles from ozonesondes. Existing data were complemented by new total ozone series from Tromsø and Svalbard which were homogenized and re-evaluted in frame of the project. Analysis of high-latitude ozone sonde data
Within the project the interannual and longer term variations in ozone profiles over the Arctic from 1989 to 2003 using ozonesonde observations from high northern latitude stations were studied. The stations of long-term ECC ozone sonde records located north of 60º N and used in the CANDIDOZ analysis are listed in Table 2.2 below:
43
Final Report
Table 2.2 ECC ozone sonde stations located north of 60º N and used in the analysis
Station Country wmo # Lat Lon Data recordResolute Canada 024 74.7º N 95.0º W 1979-Alert Canada 018 82.5º N 62.3º W 1987-Sodankylä Finland 262 67.4º N 26.6º E 1989-Ny-Ålesund Svalbard 089 78.9º N 11.9º E 1989-Lerwick UK 043 60.1º N 1.2º W 1992-Eureka Canada 315 80.0º N 85.9º W 1992-Scoresbysund Denmark 717 70.5º N 22.0º W 1993-** additional data in Feb-May 1989 and Nov 1991-April 1992
The ozonesonde data were carefully examined and made as internally consistent as possible, with corrections applied for operational changes. The results of a series of test flights made in Sodankylä were used to apply altitude dependent corrections on sonde data.
According to the sonde data set of 1989-2003 the largest long-term and interannual changes have occurred in the late winter/spring period. In January-April season significant negative trends of -2.9±1.1 %/year were found at the altitude of 150-40 hPa and a smaller decrease at 40-10 hPa layer: -1.5 ±1.1 %/year for the time period 1989-1996. However trends since then are positive: 4.7±2.6 %/year between the tropopause and 150 hPa, 3.1±2.0 %/year between 150 and 40 hPa and 1.3±1.2 %/year at higher levels (40-10 hPa). In order to explain the trends and interannual changes, a multilinear regression model was applied to the ensemble ozone sonde data including proxies for dynamical and chemical processes. It was found that a model using the following explanatory variables: average tropopause height, the calculated volume of polar stratospheric clouds, 100 hPa Eddy Heat Flux averaged over 45-70º N, and the mean aerosol backscatter in 200-100 hPa range, can explain 65-95 % of the observed variance throughout the lower stratosphere depending on the altitude in January-April. The given proxies account for the changes in the synoptic scale dynamical processes, the vortex ozone depletion, the ozone transport through meridional circulation, and the Pinatubo aerosol effect, respectively. At altitudes between 50 and 70 hPa it can be estimated that chemical polar ozone depletion accounted for up to 50 % of the March ozone variability. Statistical model suggests that negative trends in lower stratosphere prior to 1997 can be attributed to the combined effect of dynamical changes, impact of Pinatubo aerosols and to winters of relatively large chemical ozone depletion. Since 1996-1997 the observed increase in lower stratospheric ozone can be attributed primarily to dynamical changes.
Figure 2.1 shows monthly deviations of total ozone from sondes (1989-2003) in comparison to the 1979-1988 mean using TOMS v.8 satellite data. Running mean of monthly sonde data and the mean values of March-September satellite observations (TOMS v.8 and GOME VFDOAS v.1 data sets) over given sonde locations per each year during 1979-2003 are also shown. Figure suggests that in the region of 60-82º N (sonde data set) the long-term minimum in total ozone is not as well defined as in global analysis, where minimum in year 1993 becomes evident. The two lowest January to
44
Final Report
April and annual total ozone values according to Arctic sonde data set are observed in years 1996 and 1997.
1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004-35
-30
-25
-20
-15
-10
-5
0
5
10
15
20D
evia
tion
(%)
Soundings, monthly meansGOME/TOMS, Mar-Sep meansSoundings, 13 month running means
Figure 2.1 .Monthly deviations of total ozone from sondes in the region of 60-82º N (1989-2003) in comparison to the 1979-1988 mean by TOMS v.8 satellite data (grey line). Also shown is the 13-month running mean of the monthly ozone sonde data (red line) and the combination of TOMS v.8 and GOME VFDOAS v.1 data of the period 1979-2003 as mean March-September deviations from the 1979-1988 period (blue line).
Homogenization and re-evaluation of historical total ozone data from Tromsø and Svalbard
The focus of this task was on stations at around 70 and 80º N, which so far have not been included in the international ozone data bases, in particular the Tromsø series which dates back to the 1930s, and the Svalbard series, where measurements started in the 1950s.
The problem with the Tromsø series is that the series are discontinuous both in time
and with respect to measurement modes/technical state and operation schedule. The main parts of the series are described in Table 2.3.
Besides the major gap from summer 1972 to end of 1984, the series also contains
two shorter gaps due to instrument re-construction (June 1949 - July 1950) and repair after a water damage (August 1993 – May 1994). In order to close these gaps and to quality-assess the partial series from Tromsø, a comparison was performed using
45
Final Report
Nimbus-7 TOMS data from the period 1979 – 1993 at the sites Arosa, Reykjavik, Murmansk, Sodankylä, and Andøya.
The Arosa data are the only ozone series, which is longer in time than the Tromsø
series. For this reason it has played an important role to quality-assess the Tromsø ozone data before 1972. This was necessary as it turned out that the composed monthly means (direct sun, zenith sky, all wavelength pairs) especially from the 1950 – 1972 partial series are significantly lower especially in summer than the values after 1984. Average monthly mean ratios were calculated based on the 1985-2000 series from both stations. These were then compared with average ratios of the same months in the early period (1950-1972), separated according to measurement mode. It turned out that the CC’ measurements from this period show very similar ratios, while the AD direct sun measurements, which usually are regarded as the standard reference method, deviate significantly from the 1985-2000 ratios. The problem was solved by removing AD and CD measurements from the data set, since these were obviously wrong and the number of data significantly lower than C wavelength measurements. Table 2.3 Overview Tromsø ozone measurements
Time period Instrument Measurement Mode
Comments
1935 – 1940 Ferry spectrograph direct sun detector: photographic plates
1939 – 1949 Dobson # 14 CC’ zenith sky no calibration data available
1950 – 1972 Dobson # 14 CC’, AD, CD measurements, zenith sky, direct sun
no absolute calibration data available
1985 – 1990 Dobson # 14 as previous no calibration or inter-comparison since 1977
1990 – 1999 Dobson # 14 as previous regular calibrations and inter-comparisons
1994 – 1999 Brewer # 104 direct sun, focused sun, zenith sky
regular calibrations
2000 – 2003 Brewer # 104 as above, new global irradiance method
instrument moved to ALOMAR (130 km WSW of Tromsø)
With respect to the 1939 –1949 partial series, problems occurred with the
homogenisation of direct-sun (DS) and zenith-blue (ZB) measurements one side and zenith-cloudy (ZC) measurements on the other. For this reason, to have a complete series available at the end of project year 2, we decided to consider only DS and ZB data in the official record, but to keep the option of extending the series later including ZC data which is highly desirable because of the large number of additional data.
The original observations (on paper hardcopies) from Longyearbyen, Svalbard,
from the period 1950 – 1962 have been collected, and the re-evaluation in cooperation with ETHZ was started in summer 2004.Until spring 2003, all direct sun and zenth-blue observations have been re-calculated and quality-assessed. Based on the experience with the Tromsø series, special emphasis was put on the selection of the most appropriate
46
Final Report
observation mode. A comparison baseline was established from TOMS observations at Svalbard and Tromsø in the period 1979-1990, and then applied in comparison of the two sites in the period 1950-1962. It turned out that, as in Tromsø, the C wavelength measurements are clearly most reliable and thus suited for the establishment of the new data series. In total, 2819 single measurements covering 821 days in the 12-year period were recovered and quality-assessed. During the extension period of the project, also the comprehensive data set of zenith-cloudy and moon measurements was re-evaluated and quality-assessed, as far as feasible with the limited information available. With respect to zenith cloudy data, it turned out that a division into only two classes of cloud conditions was reasonable: high/thin clouds and variable/low/thick clouds. Including these data, the number of ozone measurements increases to 4837, covering 1676 days. The moon measurements cover another 137 days; that the complete re-evaluated data set statistics is given in Table 2.4.
Table 2.4 Number of measurements per month in the re-evaluated Svalbard data set. 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962Jan. 0 5 3 7 3 4 0 2 0 0 0 8 6 Feb. 0 5 7 1 2 5 2 6 0 0 0 2 8 March 0 23 28 17 19 6 4 6 0 0 0 16 16 April 0 27 27 21 25 29 27 29 0 0 0 30 29 May 0 27 27 26 30 28 30 28 0 6 0 27 23 June 0 26 30 28 25 30 29 27 23 10 16 16 20 July 0 27 31 25 25 30 31 23 23 21 30 13 0 August 0 30 26 23 30 28 26 23 29 16 12 10 0 Sept 25 7 17 0 26 25 25 17 22 23 0 0 13 Oct. 8 0 1 0 8 7 3 2 0 1 0 0 0 Nov. 5 0 0 4 0 4 2 0 0 0 0 0 0 Dec. 7 4 8 1 3 4 3 0 0 0 9 5 0
The newly evaluated Svalbard data, combined with data for the period 1963-1968
stored in WOUDC, TOMS data and recent ground-based measurements (since 1994), have been analyzed statistically in the same manner as the Tromsø series. The analysis showed that, when extrapolating the components derived from the TOMS and the recent Dobson data back to 1950, there is a very good agreement with the data derived recently. This further confirms the quality of the re-evaluated data from 1950-1962 and their consistency with more recent data sets. The analysis also yielded very interesting results on the parameters influencing total ozone in Svalbard; these are to some degree consistent with the results from Tromsø, but also in line with meridional gradients noticeable in the comparison between Arosa and Tromsø.
Multi-linear statistical analysis of the complemented Tromsø ozone series The 65-year total ozone record from Tromsø, has been analyzed using multiple
linear regression. Parameters included are modified linear depletion term, Northern Hemisphere teleconnection patterns, local temperature at the 30-mbar level, stratospheric aerosols, Quasi-Biennial Oscillation (QBO), and solar activity. Because of the
47
Final Report
Figure 2.2 Multiple linear regression analysis of the Tromsø total ozone monthly means. Black line/diamonds: measurements; light blue: regression fit; histogramme: number of measurements. Bold coloured lines: contributions from modified linear depletion (green), stratospheric aerosols (dark blue), T@30 mbar (orange), QBO (yellow), sum of
48
Final Report
teleconnection patterns (red), solar flux (turquoise). Contributions are vertically offset for clarity. availability of most input parameters, the analysis was limited to the period 1950–2001. All multiple correlation coefficients found are above 0.7, most even larger than 0.8. The series reveals a linear depletion term of about 9% decrease/decade in March/April since the late 1970s, decreasing to few percent in summer. The depletion term is insensitive to the envelope function (linear versus equivalent effective stratospheric chlorine), but highly sensitive to the modification term, forcing the contribution to zero in years with early vortex breakup. Aerosol-induced depletion is found throughout late winter and spring, while the QBO is of importance in summer. Mid-stratospheric conditions, parameterized by the 30-mbar level temperature, are dominant throughout the winter half year from September to February. Teleconnection patterns reveal a complex contribution, with no single pattern being persistently dominant over several months. The Polar-Eurasian pattern is most prominent in midwinter, the North Atlantic Oscillation pattern in March. In May, the East Atlantic Jet pattern is dominant, while in June the Pacific Transition pattern is the most prominent parameter in the analysis. This variability of teleconnection pattern contributions is probably caused by the changing influence of air masses on the local tropopause altitude, with a dominant influence of North Atlantic and Arctic patterns in winter/spring, and of lower-latitude patterns in summer and early autumn. Figure 2.2 shows the linear regression results for all months from February to November. Task 2.3 Impact of polar ozone depletion on midlatitude trends (Contributors: DMI)
Vortex depletion and mid-latitude ozone trends Chemical destruction of the ozone inside the polar vortex can influence the mid-
latitude ozone by dilution of the polar air into the mid-latitudes after vortex breakup (Knudsen and Andersen, 2001). The main goal of this task is to assess the impact of polar ozone dilution by including it into the trend calculations as well as to explain longitudinal differences in the mid-latitude trends. Ozone trends for the spring period april-may were calculated for the merged TOMS/SBUV data set. As explanatory variables were used vortex ozone depletion ‘dilution’, geopotential height, solar cycle and QBO. Calculations were made from January 1979 through December 2002. Vortex ozone depletion ‘dilutions’ were calculated for 1993, 1995, 1996, 1997 and 2000 including all years with major polar ozone depletion and thus reliable height resolved ozone depletion estimates (Figure 2.3). In 1997 the total depletion was largest and the vortex broke up latest, so that the mixing to lower latitudes was weakest.
April-May ozone trends are shown in Figure 2.4. Using a multiple linear regression
model we find that for April-May the dilution may explain 39 % of the trend in the period 1979-1997 and 54 % of the trend in the period 1979-2002. Dilution helps explain the strong downward ozone trends over Scandinavia and Russia. The geopotential height trends explain that the downward ozone trend is weak south of Greenland and strong over the UK. Combination of vortex ozone depletion ‘dilution’ and geopotential height may explain 80% of the longitudinal variation in trends at midlatitudes in the period 1979-1997 and 62 % in the period 1979-2002. The residual trends shown in Figure 2.4 (e) can
49
Final Report
be caused by e.g. local ozone depletion outside the vortex. By calculting the effect of simply removing the dilutions we get the possibly more reliable estimates that that
50
Final Report
Figure 2.3 April-May averages of the ozone depletion originating from the polar vortex in 1993, 1995, 1996, 1997, and 2000. 5° latitude-longitude boxes were used.
Figure 2.4 The April-May ozone trend for the periods 1979-1997 (left) and 1979-2002 (right). (a) Total trends (%/decade). (b) The ozone trend due to the trend in the 250 hPa height. (c) The ozone trend due to the vortex depletions in 1993, 1995, 1996, 1997, and (only on the left panel) 2000. (d) The sum of b and c. (e) The residual trend
a)
b)
c)
d)
e)
51
Final Report
dilution may explain 29 % of the trend in the period 1979-1997 and 33 % of the trend in the period 1979-2002.
a) M arch
b) April
c) M ay
d) M arch
e) April
f) M ay
Figure 2.5: Percentage occurrence frequency of the vortex (remnants) for 1993, 1995, 1996, 1997, and 2000 in a) March, b) April, and c)May, and for other years 1979-2000 in d) March, e) April, and f) May. Contours at 5, 15, 25, ... %. Latitude circles 30, 40, 60, and 80ºN. Greenwich meridian at the bottom.
a) d)
b) e)
c) f)
52
Final Report
Frequency distribution of the occurrence of the vortex and vortex remnants To see whether the preferred positions for the vortex and its remnants found in the
dilution calculations are valid also for a longer period and to see what the influence is in March, the vortex occurrence frequency at any given point is calculated for 1993, 1995, 1996, 1997, and 2000, which are the years with strongest vortex ozone depletion for which the dilution calculations are made, and also for other years 1979-2000. Figure 2.5 shows the frequency of the occurrence of the vortex and its remnants in March, April, and May. In all 3 months the most likely occurrence of the vortex (remnants) in the midlatitudes is Europe and Eastern Russia both during the years with most vortex depletion and during the other years. Due to the low ozone inside the vortex (remnants) these years this helps explain the larger ozone trends over Europe and Eastern Russia. Eastern Russia is at the zonal climatological ozone maximum so the ozone depletion there might not matter as much as over Northern Europe, where the climatological minimum is.
Task 2.3 resulted in Deliverables № 17, 20.
Task 2.4 Isentropic exchange; occurrence of polar ozone laminae and ozone mini-hole (Contributors: ASCR, NILU) Trends in ozone laminae characteristics
In this task, fast isentropic transport of the ozone which shows up in such phenomenon as ozone laminae and mini-holes has been investigated. Results of study on the trends in ozone laminae are summarized in Krizan and Lastovicka (2005) and Krizan and Lastovicka (submitted). The main finding is that data from all stations northward of 35oN display for both the positive and negative laminae a principal change of trend in the overall ozone content in laminae per profile and the number of laminae per profile from a strong negative to a well–pronounced positive in the mid-1990s (Figure 2.6). There are indications that the main reason of that change is changes in dynamics. At lower northern latitudes (below 35oN) the trends in laminae are very weak and questionable without a clear indication of change of trends, and the number of laminae is evidently lower. The Southern Hemisphere laminae occurrence is much lower than that in the Northern Hemisphere for both middle and high latitudes. No evident trends in laminae and no detectable change of trends have been observed in the Southern Hemisphere. Understanding the variability in occurrences of ozone miniholes
Intense, low-ozone events associated with anticyclonic conditions are commonly referred to as “mini-holes”. The ozone column decrease during such events is caused by a combination of uplift of the air masses above the anticyclone and meridional advection of low latitude ozone-poor in the lower stratosphere. Extreme events are found in conjunction with a displaced polar vortex aloft, with the ozone-poor mid-stratospheric polar air causing additional lowering of the ozone column. Orsolini and Limpasuvan [2001] linked more frequent occurrences of ozone mini-holes over the North Atlantic in the nineties than in the eighties to the prevalence of the positive phase of the North Atlantic Oscillation (NAO). Here, we examine the influence of 4 leading Euro-Atlantic climate patterns derived from ERA-40 upon the occurrences of the ozone mini-holes.
53
Final Report
a) b)
Figure 2.6 (a) The overall ozone deficit in negative laminae per profile and (b) the overall ozone content in positive laminae per profile for the European middle latitude stations Hohenpeissenberg (dotted line), Legionovo (heavy full line), Lindenberg (dashed-dotted line), Praha-Libus (medium full line), Payerne (thin full line) and Uccle (dashed line), 1970-2003.
Figure 2.7: Composite differences between positive and negative phases of the 4
leading Euro-Atlantic patterns for the number of days with ozone minihole conditions (in days per month).
Minihole days conditions have been calculated in various phases of the patterns for
each winter (DJF). The conditions are defined when ozone column drops by 40 DU on a synoptic time scale [Orsolini et al, 1998]. It has been found that the positive phases of the
54
Final Report
NAO and East-Atlantic pattern (E-ATL), and the negative phases of the European Blocking (EU-BLOCK) and Scandinavian (SCAN) patterns do favour occurrences of ozone miniholes over Northern Europe (Figure 2.7).
The dominant effect is from the NAO though. Positive phases of the other climate
patterns are associated with leading anticyclonic centers of action, which are regions void of minihole occurrences. This can be understood as slow, monthly-averaged changes in the jet stream around these anticyclonic centers channel the location where ozone mini-holes do occur. A strong extension and poleward tilt of the jet stream over the North Atlantic (e.g. NAO+ or E-ATL+) favour occurrences of mini-holes in that region, contrarily to blocking patterns (e.g. EUBLOC+ ) that inhibit the eastward extension of the jet over the north Euro-Atlantic sector and favours mini-holes over the subtropical Atlantic and southernmost regions of Europe.
Task 2.4 resulted in Deliverables № 22, 23, 24.
3.3 Work Package 3 3.3.1 General objectives
The main objective of WP 3 is to study the relation between variability and long-term changes in stratospheric dynamics and the ozone distribution. This also includes influence of the troposphere on the stratospheric dynamics and total ozone. WP3 contributes to the other work packages by providing dynamical proxies for statistical ozone modeling. The specific objectives of the work package are:
Task 1: Compilation multiannual time series of residual circulation from ERA40 Task 2: Calculation EP flux and compile time series of proxies of planetary wave
activity Task 3: Relation between planetary waves and residual circulation Task 4: Influence of residual circulation and tropospheric wave activity on ozone
distribution Task 5: Stratosphere-troposphere coupling, its relation to tropospheric climate
pattern and impact on ozone 3.3.2 Methodology and scientific achievements:
Task 3.1 Compilation of multiannual time series of residual circulation from ERA-40 (Contributors: AWI)
A measure for the strength of the residual circulation during an individual northern hemispheric winter is the degree of subsidence at high latitudes. Generally, the vertical residual transport velocity is given by the net diabatic heating rates (in an atmosphere in equilibrium). We have used an existing state of the art radiation transfer model to calculate the three dimensional distribution of diabatic heating rates over the last 40 years and the whole globe. We use a standalone version of the model presently in use by ECMWF (but with 4 spectral bands in the shortwave range, Morcrette et al., ECMWF
55
Final Report
Tech. Memo 252, 47 pp., 1998). We performed a detailed intercomparison with other radiative transfer codes, which yielded a good agreement between the different models if the same input data were used.
A precise calculation of the radiative transfer requires high quality input data of
temperature, water vapor and ozone. Temperature data were taken from ECMWF ERA-40 reanalysis data and operational analysis data, depending on the relative quality of the data sets. It turned out that large temperature differences (e.g. caused by large vertical temperature oscillations in ERA-40) existed between the data sets, which in turn led to large differences in heating rates and to unphysical behavior in some years (e.g. large subsidence in the lower stratosphere and small subsidence in the upper stratosphere). Hence, we performed detailed comparisons of these temperature data sets with each other, NCEP reanalysis data and independent data to choose the optimal data set for each year. In addition, ozone and water vapor fields of ERA-40 are relatively uncertain. We used ozone profiles of the CATO assimilated GOME/TOMS dataset of the ETH Zürich (Brunner et al., 2006a, in press) as input data for the calculations. The vertical distribution of ozone is reconstructed from the columns with a statistical method here. Water vapor was taken from SAGE II satellite data.
Figure 3.1 Interannual variation of the averaged diabatic descent of air parcels on trajectories starting at 550 K (orange), 520 K (cyan), 490 K (blue), 460 K (green), 430 K (red) and 400 K (black) inside the polar vortex on 10th March and calculated backwards to the 1st December of each winter. The overall subsidence was derived from cooling rates that are the results of global radiation transfer calculations with ERA-40 data. The vertical axis shows the change of potential temperature [K] due to diabatic subsidence for the indicated years and initial levels.
56
Final Report
Since the residual circulation is the zonal mean Lagrangian transport circulation, all calculations were performed in a zonal mean Lagrangian sense (as an alternative approach to the Transformed Eulerian Mean (TEM) formulation of the residual circulation). A sophisticated trajectory model was developed and optimized for long-term stability to be used in a Reverse Domain Filling trajectory run containing thousands of single trajectories and lasting several winter months each year. A binning of the trajectories into equivalent latitude and potential temperature bins allows detailed statements on the spatial and temporal development of subsidence, as shown in Figures 3.1 to 3.3. We verified the results of our radiative transfer calculations and trajectory calculations by comparison with subsidence rates derived from atmospheric tracers like N2O (Figure 3.2).
Figure 3.1. shows the interannual variability of the overall diabatic subsidence in the indicated northern hemisphere winters averaged over the polar vortex for several potential temperature surfaces (as starting points of the backward trajectories). Results have been averaged over trajectories. Figure 3.3 shows the development over the course of one winter. Both radiative transfer calculations and trajectory runs have been completed for the ERA-40 time frame. Figures showing these results for all winters are given in Tegtmeier, PhD thesis, 2006.
Figure 3.2 Comparison of calculated diabatic heating rates with N2O tracer measurements. The plot shows the average potential temperature of backward trajectories starting on 10th March at 380 K (black), 400 K (blue), 420 K (green), 450 K (red) and 470 K (yellow) inside the polar vortex as a function of time for the year 2000, calculated up to the 1st December (dashed lines). The solid lines show diabatic descent inferred from tracer measurement of N2O (Greenblatt et al., JGR, 107, 8279, 2002).
57
Final Report
Figure 3.3 Diabatic descent in K inside the Arctic vortex calculated with backward trajectories starting on 16 isentropic surfaces for one chosen winter (10.3.1995-1.12.1994).
Task 3.1 resulted in Deliverables № 25, 26, and 27.
Task 3.2 Compilation of multi-annual time series of EP flux using ERA40 (Contributors: AWI)
NCEP (1948-2004), ERA40 (1957-2002), ERA15 (1979-1993) and UKMO (UK Met. Office 1992-2004) datasets are used to compile long term EP flux time series. Although it has been proven that EP flux and its divergence are useful to study the dynamical influence on the ozone field, these quantities are highly derived and vary in magnitude dependent on the datasets used. Figure 3.2 shows EP flux and its divergence from different meteorological analysis scheme on 22nd September 2002, on where a record high heat flux was observed in the southern hemisphere, which led to the first observed breakup of the southern polar vortex in this year. ECWMF dataset clearly shows higher EP flux divergence than NCEP and UKMO analysis, and also the peak in the EP flux from NCEP is at higher altitude than that from ECMWF and UKMO.
58
Final Report
Figure 3.4 EP flux (vectors) and its divergence (contours) derived from NCEP, ECMWF and UKMO analysis schemes on 22nd September 2002, when a record high wave activity was observed in the southern hemisphere resulting in the first major stratospheric warming event ever observed in the SH.
On the monthly scale, it has been found that all the datasets agree reasonably well during the winter season (high wave activity period) and also the summer season (low wave activity period), as can be seen in Figure 3.3. The interannual variability in winter EP flux, which determines the interannual variability in the development of the polar vortex and regulates the spring build up in ozone and the magnitude of heterogenous chemistry, is shown in Figure 3.4. Trends in EP flux are important for the future development of polar ozone loss, as they are connected with trends in vortex strength and polar temperature. As the quality of all datasets during the pre-satellite era is questionable (especially in the SH), the trend calculation is done only for satellite period (1980-2003). Results for the northern hemisphere have been summarised in Table 1 for eddy heat fluxes and in Figure 3.5 for EP fluxes.
Figure 3.5 Eddy heat flux time series from different meteorological analysis for February in NH (left) and for September in SH (right). The heat flux has been averaged between 40°-70° at 100 hPa in each hemisphere.
Maximum decadal trend (~2 sigma level) has been found in September in the SH using both datasets ERA40 (2.6%/decade) and NCEP (3.0%/decade) whereas there is no statistically significant trend in monthly heat flux in the NH during the winter months (above the 90% confidence level) except for November and NCEP. Positive trends are observed in November and December, while January to March show negative trends.
59
Final Report
This was also noted by Randel et al. (2002), however, this trend (within 2σ) is not statistically significant. These trends may be connected to polar ozone depletion (change of radiative equilibrium, strengthening of the vortex alters the waveguide for planetary waves). Averaged over the winter (Sep-Mar) no significant trend in heat flux is observed in either hemisphere for the period 1980-2002.
Task 3.2 resulted in Deliverables № 28, and 29.
A−O J F M A M J J A S O N D−0.2
−0.15
−0.1
−0.05
0
0.05
0.1
0.15
0.2
Month
Tre
nd [1
04 kg/
s2 /yea
r]
Monthly trends 1979−2003
Figure 3.6 Monhtly trends in EP fluxes for NH (left) and SH (right) from ERA40. The errorbars assume a statistical error of 5% in the EP fluxes. The leftmost bar shows the trends in winter EP flux (O-M is October to March and A-O is April to October).
Table 3.1 NH trends derived for the monthly 100hPa heat flux averaged from 40°N-70°N from ERA40 and NCEP dataset for the 1980-2002 period.
O−M J F M A M J J A S O N D−0.2
−0.15
−0.1
−0.05
0
0.05
0.1
0.15
0.2
Month
Tre
nd [1
04 kg/
s2 /yea
r]
Monthly trends 1979−2003
Month ERA40 Trend 1σ Confidence level
NCEP Trend 1σ Confidence level
[Km/s2/y] [%] [Km/s2/y] [%] Jan -0.09 0.14 47 -0.1 0.14 50 Feb -0.18 0.2 61 -0.17 0.21 57 Mar -0.12 0.12 67 -0.1 0.13 53 Apr 0.03 0.08 26 0.02 0.07 28 May 0.07 0.04 90 0.07 0.04 92 Jun 0.01 0.01 37 0.01 0.01 46 Jul 0.02 0.01 97 0.02 0.009 99 Aug -0.02 0.01 82 -0.02 0.01 82 Sep 0.03 0.02 87 0.03 0.02 89 Oct -0.04 0.04 75 0.009 0.04 2 Nov 0.12 0.07 89 0.14 0.075 93 Dec 0.19 0.14 80 0.19 0.13 82
Sep-March 0.02 0.03 36 0.02 0.03 52
60
Final Report
Task 3.3 The relation between planetary waves and residual circulation
Vertically propagating planetary waves that enter the stratosphere through the tropopause and break at higher levels drive the residual circulation. In northern winter, breaking waves deposit their momentum in the stratosphere and decelerate the westerly zonal wind, which leads to imbalance between Coriolis force and pressure gradient and net northward motion. The additional mass transported to higher latitudes leads to adiabatic heating, which in turn is followed by diabatic cooling. Hence, the strength of wave activity and wave breaking (expressed by the EP flux and its divergence) determines the general behavior of many stratospheric variables. We examined the correlation of the parameters temperature, geopotential height, zonal wind and diabatic subsidence with the EP flux entering the stratosphere. Figure 3.7 and 3.8 show these correlations for stratospheric temperature and the diabatic heating rates from Task 3.1.
Task 3.3 resulted in Deliverable № 30.
0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2200
205
210
215
220
225
79
80
81
8283
84
85
86
87
88
89
90
91
92
93 94
95
96
97
98
99
00
01
02
Fz [105 kg/s2]
T [K
]
Fz (Jan) − T (Feb) at 50 hPa
ERA r=0.64NCEP r=0.63
0 30 60 900
5
10
15
20
25
30
35
40
45
Latitude [deg]
Alti
tude
[km
]Correlation Fz−T Dec−Mar Means
−1
−0.8
−0.6
−0.4
−0.2
0
0.2
0.4
0.6
0.8
1
Cor
rela
tion
Figure 3.7 Correlation of temperature and EP flux. The left plot shows the correlation of mean stratospheric polar temperatures in February (averaged zonally and north of 62.5° N, at 50 hPa) with the January EP flux through the tropopause (averaged 45-75° N at 100 hPa) for both NCEP and ERA-40 reanalysis data of the indicated years. The legend shows the correlation coefficients. The right plot shows the correlation of the EP flux (averaged Dec-Mar and 45-75° N at 100 hPa) with zonally averaged temperature (averaged Dec-Mar) at every latitude-pressure grid point of ERA-40. Dashed lines show 95%, 99% and 99.9% confidence intervals. Task 3.4 Influence of residual circulation on ozone distribution (Contributors: AWI, IUP-UB) Connection between the EP flux and the ozone distribution
Meteorological data from ERA40 and NCEP and ozone column measurements from TOMS/SBUV or GOME are used to study the differences between winters with high wave activity (warm polar winters) and with low wave activity (cold polar winters).
61
Final Report
Figure 3.8 Correlation between diabatic descent of backward trajectories starting at 460 K inside the polar vortex and EP flux trough the tropopause (45-75° N, 100 hPa), both with ERA-40 data. The temporal average time frame for both quantities is 1.12-10.3. The correlation is 0.75.
Figure 3.9: Correlation of winter EP fluxes (100 hPa, 45°-75° N, October-March) from NCEP with the NH zonal mean column change of GOME (left) or TOMS/SBUV (right) north of 62.5° in the same period. Cyan dots with years show points from the eighties, blue dots show points from the nineties, crosses are the ECHAM E39C model (“Year 2000” timeslice) for comparison. Lines show the regression for the eighties (dotted), all years (solid) and the model (dashed). Correlation coefficients are given in the legend.
0.7 0.8 0.9 1 1.1 1.2 1.3 1.460
70
80
90
100
110
120
130
140
150
160
170
80
8182
83
8485
86
87
88
89
90
91
92
93
94
9596
97
98 99
00
01
02
typical error
Fz [105 kg/s2]
∆ O
zone
[DU
]
NCEP − TOMS (North)
NCEP r=0.67NCEP 90s r=0.86ECHAM r=0.51
0.7 0.8 0.9 1 1.1 1.2 1.3 1.4
60
70
80
90
100
110
120
130
140
150
160
170
96
97
98
99
00
01
0203
typical error
Fz [105 kg/s2]
∆ O
zone
[DU
]
NCEP − GOME (North)
NCEP r=0.98ECHAM r=0.51
62
Final Report
The interannual variability in the winter EP flux (integrated EP flux for winter months, averaged over 45° to 75° N/S latitude at 100 hPa) is closely connected with the change in zonal mean ozone column at most latitudes in the same period. Higher wave activity leads to an increase in ozone transport, as well as to higher temperatures in polar regions, and in turn decreases ozone loss due to heterogeneous chemistry. Figures 3.6 and 3.7 show the correlation between zonal mean column change in fall/winter (October to March in NH, April to October in SH) with the winter EP flux, integrated over the same period. Correlations are better in the nineties than for the whole period 1979-2003 perhaps due to changes in halogen loading and data quality. The increased halogen loading in the nineties amplifies the sensitivity of ozone to changes in EP flux by a factor of about two compared to halogen free conditions. In the southern hemisphere, correlations are worse due to the complete destruction of ozone in some altitude layers and some years, which destroys the correlations. In warm winters like 1998/1999 the chlorine activation (here shown as winter averages of the daily mean OClO vertical column observed by GOME at 90° solar zenith angle) is lower than in cold winters, which also leads to correlations between OClO and EP flux, as is shown in Figure 3.8. The exceptional Antarctic winter of 2002 represents an intermediate case between the typical Antarctic winters and the cold Arctic winters, which can be seen in Figure 3.7, right panel.
Figure 3.10: Left: Same as Figure 3.6 for the southern hemisphere TOMS data and EP fluxes averaged over April to October and 45°-75° S. Right: Correlation of winter eddy heat flux (100 hPa, 45-75 N/S, Sep-Mar (N), Mar-Sep (S)) with the zonal mean ozone column ratio ([Mar-Sep]/Sep for NH and [Sep-Mar]/Mar in SH) from GOME
The relationship of EP flux and ozone column holds for most latitudes. Figure 3.9
shows the same correlation applied to two different latitude bands and the time period 1979-2003 using TOMS data. It can be seen that the relationship holds quite good for high and low latitude bands, with the exception of the SH data in the eighties in high latitudes.
0.4 0.5 0.6 0.7 0.8 0.9 1
−80
−60
−40
−20
0
20
40
6079
808182
83
84
85
86
87
88
8990
91
92
93
979899
typical error
Fz [105 kg/s2]
∆ O
zone
[DU
]
NCEP − TOMS (South)
NCEP r=0.15NCEP 90s r=0.56ECHAM r=0.54
63
Final Report
Figure 3.11 Winter integrated daily OClO mean vertical columns as a function of winter heat flux (updated from Weber et al., 2003).
The separation of the dynamical and the chemical change of ozone column as a
function of the residual circulation is important for the future prediction of ozone levels. We approached this by means of two different methods: 1. Calculation of the change by transport as the difference between the observed column change (e.g. by sondes) and the observed chemical change (e.g. by Match or vortex average methods). 2. Calculation of the different terms in the Transformed Eulerian mean continuity equation for ozone (i.e. residual circulation term, meridional mixing term, chemical term and observed change). Results of the first approsach can be seen in Figure 3.10. It can be seen that the sensitivity of ozone to the chemical and the transport effect of the residual circulation are about the same.
Figure 3.12 Correlation between winter heat flux and spring/fall ozone ratio from TOMS/SBUV (1979-2003) for a high latitude band (left, 60°-70 °) and a low latitude band (10°-20°, right) .
Thus, the overall sensitivity of ozone column to changes in EP flux was doubled in the nineties compared to halogen free background conditions. Results for the second
64
Final Report
approach for the winter of 2000 and ozone north of 62.5 and between 200-10 hPa are: 307 DU residual change, -71 DU meridional mixing, chemical change -213 DU and observed change 24 DU (ERA-40 meteorological data and assimilated TOMS ozone). This approach was hampered by the too rapid residual circulation in ERA-40, but the amount of chemical change and dynamical change are about the same in this approach, too.
Figure 3.13. Dynamical and chemical contribution to the correlation between ozone change through the winter and EP-flux during winter. The chemical loss has been inferred from ozone loss measurements. The dynamical supply (blue) is calculated as the difference between observed change (black) and chemical loss (red). Study of the dynamical contributions on the long-term ozone trend
We developed a statistical model for ozone column data from Dobson spectrometers, including variables for the residual circulation, short-term dynamics, aerosols, QBO, solar cycle, ENSO, homogeneous and heterogeneous chemistry, which we have published in [Wohltmann et al. 2005], which will be followed by a more detailed publication on the same topic (Wohltmann et al., 2006, to be submitted to JGR). The residual circulation was represented by the EP flux proxy (deliverables 31, 32), which is available over the CANDIDOZ web page. The proxy is given by the vertical component of the EP flux at 100 hPa and 45-75 deg north, integrated from October to the month of measurement. Short-term dynamics were represented by a newly developed proxy based on the equivalent latitude profile at a given station location, which proved to be extremely successful in explaining the short-term variability of ozone column. But it also turned out that these proxy, which represents vertical displacements of isentropes and horizontal advection, is the most prominent dynamical impact on the long-term trends of ozone (Figure 3.14). We also developed a regression model for zonal mean ozone column
65
Final Report
data from TOMS, SBUV and GOME in collaboration with the University of Bremen (Dhomse et al., 2005). Explanatory variables are EP flux, EESC, aerosols, QBO, solar cycle, PSC area and ENSO.
Task 3.4 resulted in Deliverables № 31, and 32.
1 2 3 4 5 6 7 8 9 10 11 12−1.4
−1.2
−1
−0.8
−0.6
−0.4
−0.2
0
0.2Europe 70
Month
Tre
nd [D
U/y
ear]
EqlatHorizontalVerticalEP fluxHom. ChemistryHet. ChemistryAerosolSolarQBOFitOzone
1 2 3 4 5 6 7 8 9 10 11 120
5
10
15
20
25
30
35
40
45
50Europe 70
Month
σ [D
U]
EqlatHorizontalVerticalEP fluxHom. ChemistryHet. ChemistryAerosolSolarQBOFitOzone
Figure 3.14 Trend and variability between 1970-2003 of several explanatory variables used in a multiple regression model of Dobson ozone column measurements (average over 8 European stations). The solid grey line shows the observed variability and trends of the ozone column and the dashed grey line is the fitted trend and variability of the model. The colored lines show the trends and variabilities of the explanatory time series as fitted by the model (trends and variability of the original time series multiplied by the regression coefficients to give values in DU/year and DU). The names of the time series are given in the legend. Task 3.5 Relation between total ozone and tropospheric climate patterns (Contributor: NILU)
Influence of low-frequency tropospheric dynamics upon column ozone inter-annual variability derived from EOF analysis of NCEP/NCAR and ERA40 reanalyses geopotential fields.
The objective is to better understand the coupling between tropospheric climate patterns and stratospheric variability, and the stratospheric ozone distribution in particular. The focus is on the estimation of the dynamically-induced ozone trend over Europe in late winter and spring, and its attribution to meteorological phenomena.
The influence of low-frequency tropospheric dynamics upon column ozone inter-
annual variability is estimated in the spring season over the Euro-Atlantic sector. This dynamical variability of tropospheric origin is examined in terms of leading climate patterns. These patterns have been derived from an empirical orthogonal function (EOF) analysis of the National Center for Environmental Prediction (NCEP) 500mb geopotential height analyses. They have also been calculated using the ECMWF ERA-40 fields.
66
Final Report
In order to fingerprint the spatial and temporal ozone signatures of these patterns, the Total Ozone Mapping Spectrometer (TOMS) satellite observations of column ozone during the last two decades are used. We extract by linear regression the geographical ozone signatures of these climate patterns, and reconstruct their time evolution.
Geopotential anomalies associated with four leading patterns of variability, namely
the North Atlantic Oscillation, the Scandinavian pattern, the East-Atlantic pattern and the European blocking pattern, induce column zone anomalies in the range of 5-15 Dobson Units per standard deviation of the pattern index. A hitherto little-known lifting of the Arctic tropopause by the polar night jet is also shown to imprint upon the ozone signatures of some patterns. We have identified this effect as a promising mechanism to be further investigated in case studies.
The ozone statistical signatures of climate patterns originating in the troposphere obtained in this study can readily be understood in terms of geopotential anomalies in the troposphere, and have been quantified. A major result is that it proves important to incorporate several patterns, in addition to the NAO pattern. More localized patterns, such as the European blockings, also influence regional ozone distributions.
Figure 3.11 Ratio of the trend from the reconstructed ozone time series to the observed ozone trend derived from TOMS observations. The reconstructed ozone series are made of anomalies induced by the four leading patterns superposed on a climatological mean.
The impact of these combined Euro-Atlantic climate patterns upon the regional ozone trends and inter-annual variability is also estimated. A significant part of the springtime ozone trend over given regions of the Euro-Atlantic sector has been accounted by dynamical trends in a series of climate patterns. First, we use the leading patterns to reconstruct the dynamically induced spring ozone anomalies over the northern hemisphere in the years 1979-2000. When the first pattern (i.e. NAO) induced ozone anomaly is added to the mean climatological ozone, and subsequently the Scandinavian pattern , the East-Atlantic pattern and the European blocking pattern contributions, the reconstructed year-to-year variations in spring ozone tend to resemble the curve derived from TOMS observations. The correlation coefficient increases from 0.42 (1 EOF) to
67
Final Report
0.54 (2 EOFs), and then to 0.66 (3 EOFs), and is not changed by the 4th EOF (EU-BLOCK). Using this procedure, we next estimate the amount of inter-annual variability and trend that is explained by the four combined patterns of tropospheric origin over the Northern Hemisphere. The ratio of the reconstructed trend to the observed trend (Figure 3.11) displays large regional variations. Naturally, it is over the Euro-Atlantic sector that a larger portion of the trend is explained. The explained trend is nevertheless significant over Central and Northern Asia, where some patterns have a strong influence through their modulation of the polar vortex. A maximum contribution ranking at up to 70% of the observed trend, is seen over the Iberian Peninsula, but the European average is closer to 30-40%, in line with other studies of the dynamical contribution to ozone trends. The series of regional maxima and minima in the dynamical ozone trend across the Euro-Atlantic sector (e.g. Knudsen and Andersen, 2001) can hence be confidently attributed to the 4 above-mentioned climate patterns. In order to reconstruct the regional ozone trend and variability, it hence proved important to incorporate several patterns, in addition to the dominant North Atlantic Oscillation.
Results can be summarized as follow:
• A series of leading patterns of variability, namely the North Atlantic Oscillation, the Scandinavian pattern, the East-Atlantic pattern and the European blocking pattern, induce column ozone anomalies in the range of 5-15 Dobson Units per standard deviation of the pattern index, with strong regional effects.
• There can be an indirect signature through the coupling of the patterns with the polar night jet.
• The impact of these combined Euro-Atlantic climate patterns upon the regional ozone trends and inter-annual variability is also estimated. Over the Iberian Peninsula, up to 70% of the observed ozone trend in spring, can be attributed to these four patterns, but the European average is closer to 30-40%, in line with other studies of the dynamical contribution to ozone trends (e.g. Knudsen and Andersen, 2001).
Influence of the North Pacific circulation over the Euro-Atlantic:
In addition the remote influence of the Pacific circulation upon ozone over the Euro-Atlantic sector has been considered. There is recent evidence that climate variations over the North Pacific and Atlantic sectors are coupled in late winter. Honda et al. [2001] show that, in February and March, there exists an inter-annual seesaw between the strength of the Aleutian Low (AL) and the Icelandic Low (IL). These two climatological features are the major wintertime, surface low-pressure cells in the northern hemisphere. Hence, they showed that the AL and the IL do not fluctuate independently, but rather show inter-annual out-of-phase variations. The Aleutian-Icelandic Index (AII) is defined as the normalised sea-level pressure anomaly over the AL minus the one over the IL. The AL-IL seesaw modulates the upward propagation of wintertime stationary planetary waves and wave activity fluxes into the stratosphere. The AII+ phase, with its deeper than normal IL, corresponds to a more pronounced planetary wave-2, and to a diminished upward wave flux. Honda et al. [2001] demonstrated that the seesaw gives rise to marked changes not only in the stationary flow patterns but also in the synoptic travelling weather systems in both oceanic basins. We characterised the ozone signature of the AII using
68
Final Report
TOMS observations spanning the last two decades. The climatological mean sea-level pressure from the re-analyses ERA-40 is used, covering the years 1959-2002.
We show in Figure 3.12 the ozone regression map associated to the AII index, and
for comparison, the Arctic Oscillation-regressed ozone map for February. The AII-regressed map shows a broader and stronger region of influence over the North Atlantic. Consequently, the AII index is able to capture more of the year-to-year ozone variability in Feburary over the AL and IL sectors than the AO. This is demonstrated in Fig. 3.13, where eddy ozone averaged over both the IL and AL sectors is shown for February 1979 to 2000. Ozone is normally higher over the AL sector reflecting the background planetary wave-1. The anti-correlation of the two curves is –0.43. The usefulness of the AII in capturing ozone year-to-year fluctuations over the AL and IL sectors is seen from the correlation of observed eddy ozone to an eddy ozone reconstructed by adding the AII-induced anomaly to the climatological mean (green line). The correlation is high over both sectors : 0.73 and 0.78 respectively. Using NAO or AO indices instead of the AII (dashed and dot-dashed black lines), one finds some correlation over the IL but not over the AL sector. The reconstructed curves based on NAO or AO hence capture less of the observed year-to-year variability over both sectors.
Figure 3.12 Ozone regression maps in February against the AII and the AO index (see text for definition). These maps represent an ozone anomaly associated with a one standard deviation anomaly of the pattern.
Low ozone over the North Atlantic and high ozone over the North Pacific in February are associated with the negative phase of the AII, and a stronger than normal AL. In the opposite phase, ozone is higher over the North Atlantic, as the weakened Euro-Atlantic ridge is elongated to the north-east. This ozone variability mirrors fluctuations in planetary waves in the UTLS region. For example, the AII+ is characterised by a reinforced wave-2 and weaker wave-1. It is well known there exists a local correlation between geopotential height in the UTLS region and column ozone on synoptic to seasonal time scales, our results nevertheless suggest that such late winter ozone fluctuations can emerge from a planetary-scale teleconnection between the distant geographical regions of the North Pacific and the North Atlantic. Using the AII, one can reconstruct late winter ozone time series over both the AL and IL sectors that are well
69
Final Report
correlated with observations, while corresponding AO or NAO-based time series were less correlated with observations over the IL sector, and not correlated over the AL sector. While the AII explains a significant amount of the ozone variability in February, it has no significant trend, and hence would not contribute directly to ozone decadal trends.
Figure 3.13 February-mean eddy ozone averaged over the IL and AL sectors for the years 1979-2000 (TOMS : blue lines). There are no data in 1995-1996. The 3 other lines are ozone reconstructed from the contributions of the AII (green), the NAO (black dashed line) or the AO (black dot-dashed line).
Influence of climate patterns on various dynamical quantities in the troposphere-stratosphere coupled system
We have examined the influence of climate patterns on various dynamical quantities in the troposphere-stratosphere coupled system (e.g. wave activity fluxes into the stratosphere, jet stream strength) using the Era-40 re-analyses. We used linear regression techniques and our focus was the Euro-Atlantic region in winter. All four leading patterns had significant signatures of the upward component of wave activity flux with some strong regional effects, but little correlation with vortex-related quantities, such as mean vortex temperature or PSC area.
The downward propagation of stratospheric anomalies and their influence on
climate patterns has also been examined in the Era-40 data. We focues on situations of weakened zonal flow, i.e. stratospheric sudden warmings. To identify SSW events, the stratospheric zonal index (SZI) was defined as the first principal component of the daily zonal mean zonal wind anomalies at 50 hPa during the winter season (DJFM). When the SZI stays one standard deviation below its long term mean for minimum 10 days a stratospheric sudden warming (SSW) event takes place. Altogether, 31 observed SSW events were identified.
Limpasuvan et al. (2004) recently performed a composite analysis of SSWs using
re-analyses from NCEP, showing a persistent, but weak Arctic Oscillation signature following SSWs, and hinting at precursory blockings over Northern Europe in the onset stage. We also find the occurrence of a negative NAO following SSWs, hence
70
Final Report
Figure 3.14 Geopotential height anomalies at 50 hPa (left column), 500 hPa (mid-column) and SLP at 1000 hPa (right column) during the composite life cycle of ERA40 SSWs. Each row represents a phase in the life-cycle, from onset (top) , to growth, mature, decline and final decay (bottom).
71
Final Report
corroborating covariability of the stratosphere and the troposphere (Baldwin and Dunkerton, 2001). The negative NAO was also shown to be associated with anomalies in the tropospheric jet stream, with an Atlantic storm track shifted south, a finding that could lead to medium-range forecast improvements. We found evidence for anomalous blockings over Northern Europe before SSWs, on average. The signal was weak but showed consistently in geopotential and wind speed patterns. Hence, the actual source of stratospheric variability and, ultimately, prediction resides in the troposphere, as argued by Polvani and Waugh [2005].
To construct composites we divided the composite in 10-day increments, reflecting
the onset , growth, mature, decline and final decay phases. Figure 3.14 shows the composite SSW life cycle of the anomalous geopotential height field at 3 pressure levels throughout these five phases. In the stratosphere there is a wave number 1 disturbance evident during the onset stage. It is associated with an anomalous cyclone over Eurasia and an anomalous anticyclone over Alaska. At the surface there are weak indications of a blocking situation. As the warming intensifies the anomalies become more zonally symmetric with anomalously high geopotential heights over the polar region and anomalously low geopotential heights over middle latitudes in both the stratosphere and troposphere. In the mature phase and after the warming the anomalies are AO-like, even at the surface (1000mb).
Figure 3.15 Time evolution of the composite NAO index related to SSWs (thick ble line), and NAO index related to individual SSWs (thin, black line) for a) model data and b) observations.
The daily NAO index during the composite life cycle of a sudden warming event is shown as the thick line in Figure 3.15. The ensemble of thin lines illustrates that the long-lasting negative NAO (or AO) signature in the troposphere following SSWs are a mean behavior, but that there are large differences between the individual sudden warming events, a fact that it not emphasised much in current litterature.
Task 3.5 resulted in Deliverable № 33, and 34.
72
Final Report
3.4 Work Package 4 3.4.1 General objectives
The objective of WP4 is to assess the relative roles of chemistry and transport in the long-term changes of ozone in the Northern Hemisphere and understand the impact of stratospheric variability on the expected ozone recovery in the next 50 years. These objectives are achieved through the following tasks:
Task 4.1: Multi-annual integrations of four chemical-transport models (CTMs)
(with simple and full chemistry) using the ERA-40 meteorological analyses. Task 4.2: Analysis of the role of long-term dynamical changes and quantification
of the contribution of chemical and dynamical processes to the observed ozone trend Task 4.3. Ozone response to changing emissions and the green-house effect with
climate model studies. 3.4.2 Methodology and scientific achievements:
Task 4.1 Multi-annual CTM integrations with ERA-40 analyses (Contributors: FMI, UCAMB, UPMC, UiO )
A series of multi-annual and multi-decadal integrations have been performed from the four participating CTMs. All models were forced by the ERA-40 ECMWF analyses, a global, long-term, high-resolution assimilated meteorological field of the troposphere and stratosphere (Uppala et al., 2005), suitable, in principle, for decadal chemical-transport modelling simulations to study ozone trends. However, there were problems in the ERA-40 vertical velocity field (van Noije, 2004; Uppala et al., 2005) and, due to the different treatment in their advection modules, some of the CTMs had a too strong Brewer-Dobson circulation which resulted in overestimation of ozone and trace gases. Therefore, despite the copious efforts with a range of solutions to overcome these problems (see CANDIDOZ annual reports), not all model’s ozone output was appropriate for the long-term trend analysis of task 4.2. Nevertheless, work in this task i) did result in new, global, decadal-scale ozone time-series useful (mainly from SLIMCAT) in task 4.2, ii) advanced model development and iii) provided valuable feedback to the ECMWF analyses community (A. Simmons, personal communication).
Emission sensitivity 1990-2000 OSLO CTM2 simulations The main objective for University of Oslo has been to make multi-year simulations
using Oslo CTM2 (ref) to study the chemical/dynamical effects on ozone. Three model simulations have been performed for the time period 1990 trough 2000, forced by the ERA-40 analyses. The first simulation used surface emissions from the EU project ‘POET’ taking into account year-to-year variation. The second simulation used the same emissions for every year, i.e. emissions of ozone precursors and ozone depleting substances were fixed at 1990 levels. In both these simulations the heterogeneous chemistry package by Carslaw et al. (1995) was used. The comparison between simulation 1 and 2 gave an estimate of the chemical and dynamical contributions to
73
Final Report
ozone change. A third run has been performed using the comprehensive heterogeneous chemistry package by de Zafra and Smyshlyaev (2001).
Figure 4.1 Zonal-mean total O3 in high- (80°N to 85°N) and mid-latitudes
(45°N to 50°N) as modelled by Oslo CTM2, cyan with changing emissions and pink without.
Comparing the two first runs where the heterogeneous chemistry of Carslaw is
used, we can see the effect of changes in emissions on ozone. In our calculations only minor changes are seen in the zonal-mean total ozone column at the selected latitude bands (Figure 4.1). The effect of emission changes is shown in Figure 4.2 for April 2000. Pronounced signals are modeled in the Equatorial region in the troposphere and the lower stratosphere in mid to high southern latitudes. In the tropics an increase of 8% is modeled in the lower troposphere, while in the Antarctic fall we obtain a decrease of about 6%. The increase in the troposphere reflects the increase in ozone precursor emissions, while the Antarctic reduction is due to the further increase of stratospheric halogen levels during the 1990s.
Figure 4.2 Zonal-mean O3 as modelled by Oslo CTM2 as monthly for April 2000. The figure shows the effect of the emission change.
74
Final Report
Figure 4.3: Zonal-mean total O3 in high- (80°N to 85°N) and mid-latitudes (45°N to 50°N) as modelled by Oslo CTM2 with Carslaw heterogeneous chemistry (cyan), de Zafra and Smyshlyaev heterogeneous chemistry (blue), and TOMS observations (green).
In the last period of the project the comprehensive heterogeneous chemistry
package by de Zafra and Smyshlyaev was included in the model. Shown in Figure 4.3 is a comparison of model simulations 1 and 3 at mid and high latitudes of the Northern Hemisphere. Both simulations use changing emissions, but simulation 3 features the new heterogeneous chemistry package. Simulation 1 (with the Carslaw heterogeneous chemistry) is shown in cyan and simulation 3 (de Zafra and Smyshlyaev heterogeneous chemistry) in blue. In simulation 1 Oslo CTM2 overestimates zonal-mean total ozone (as was reported in the second annual report of Candidoz). When the heterogeneous package by de Zafra and Smyshlyaev was included, less overestimation was calculated.
Comparisons with TOMS observations (green line in Figure 4.3) show that the
seasonal and day-to-day variations in the model are rather well simulated and that the level at high latitudes is realistically reproduced. At mid-latitudes the overestimation of ozone in simulation 3 is less than in the simulations using the Carslaw package.
Full chemistry 1958-2004 REPROBUS simulation A 47-year simulation of the evolution of stratospheric chemistry was integrated
using REPROBUS/ERA40 3-hourly forecast at 6°×6° horizontal resolution. 3-hourly forecasts from ERA40 and Operational of ECMWF are used to drive the advection calculation from 1958 to 2004. The monthly average fields resulted from the interactive global 2-D chemistry-aerosol model (Bekki et al. 1992) run with the scenario of source gases at the surface based on WMO 1998 was used to initialize the chemical field in 1958 and to update every month the source gases at the lowest levels. H2SO4 field is monthly updated by its zonal average distribution given by the 2-D chemistry-aerosol model.
Figure 4.4 shows that show a realistic day-to-day and inter-annual variation of total ozone, but the absolute amounts are generally overestimated. Use of low resolution results in an under-estimation of residual circulation strength, which leads to an insufficient transport of source gases and subsequently to an incorrect chemical equilibrium.
75
Final Report
Figure 4.4: Temporal evolution of zonal total ozone 1958-2004 by REPROBUS/ERA40 3-hourly forecast with low horizontal resolution of 6°x6°
Early “ozone holes” over South Pole from middle 1960s to early 1970s are due to extremely low ERA-40 temperature in the middle stratosphere. Extremely high total ozone amounts in Southern Hemisphere in 1975 and 1976 are related to high temperatures in the upper stratosphere which implies an amplified circulation. Low wintertime ozone amounts 1975-1976 in the Northern Hemisphere high latitudes are related to high temperatures in the upper stratosphere.
Higher horizontal resolution (2°x2°) results show more realistic contrast of total
ozone between low and middle-high latitudes. A 2°×2° resolution model run did give improved results but a longer simulation was performed only until the year 1987.
Full chemistry 1958-2005 FINROSE simulation A stratospheric simulation spanning from September 1957 to August 2002 was
done as a continuous run using ERA-40 analysis data (Figure 4.5). In addition, the period January 2002 to April 2005 was run using operational analysis winds and temperatures. The vertical velocity is calculated in the model from the divergence. The model was run at a horizontal resolution of 10x5 deg (lonxlat). Data from a 2D model by Bekki et al was used as input for the sulphuric acid zonal mean and lower boundary condition data for some of the constituents.
76
Final Report
Figure 4.5: Daily and zonally averaged total ozone (1957-2002) from FinRose-CTM output (Note the reversed time-axis).
The model output has to some extent been compared to ozone soundings, satellite data and climatological data. From these comparisons the ozone distribution and variability seem reasonable, at least in a qualitative sense. However, some problems are evident - Difficulties in the vertical distribution of ozone (sudden high mixing ratios in the UT/LS region), seen as total ozone values locally up to 600 DU in the winter/spring at high northern latitudes - An underestimation of the ozone column in equatorial regions. The problems are at least partly due to a too strong meridional circulation and noise in the ERA-40 fields. The stratospheric age-of-air in the FinRose does only slightly exceed 2 years. Some other peculiarities can be seen e.g. an increase in Arctic total ozone going from the 60s to the 70s, which is connected to a change in the circulation in the meteorological data. This can also be seen in the stratospheric age-of-air, which decreases with a few months to a minimum in the mid 70s. Furthermore, it seems that there are some early years with low ozone in the Antarctic. It may be due to unusually low temperatures in the SH vortex but it may also be partly due to dynamics.
Simple chemistry 1958-2004 SLIMCAT simulation A 46-year simulation was performed by University of Cambridge using the
SLIMCAT model (version 2.2) with the Cariolle scheme, a simple parameterisation of ozone chemistry (Hadjinicolaou and Pyle, 2004). The global time-series from 1958 to 2002 of the model column ozone (with the Cariolle scheme plus a PSC parameterisation where the ozone loss term is scaled in time according to the EESC trend, see CANDIDOZ 2nd year report for details) are shown in Figure 4.6. The model reproduces well the main latitudinal and temporal ozone characteristics, especially the inter-annual variability of the N.H.
77
Final Report
Figure 4.6 Latitude-time section of the ozone column in Dobson Units between 1958 and 2001, modelled by SLIMCAT using the Cariolle scheme and a parameterization for the polar depletion scaled by the EESC loading for every year.
The model run was validated against long-term total ozone measurements from
different stations. In the high latitudes of the N.H. the model shows exactly the same inter-annual variability and magnitude during winter (not shown) and spring compared with observations over Tromso. In the middle latitudes the model underestimates the observed total ozone over Arosa in winter (figure 4.7), where during the 1970s and early 1980s the comparison is not good. This must be (partly) caused by the ERA-40 forcing analyses which in the mid-1970s and early 1980s are reported to suffer from poor assimilation of the satellite data (Uppala et al., 2005). Nevertheless, the inter-annual variability is reproduced correctly, allowing the use of this model output to the trend analysis in task 4.2. Full chemistry 1958-2004 SLIMCAT simulation
A 46-year simulation was performed by University of Cambridge using the SLIMCAT model (version 2.2) with a comprehensive full stratospheric chemistry scheme (Chipperfield, 2003). The model bottom boundary at 340K was forced by time-dependent halogens and CH4, N2O fields (annual cycle taken from Bekki et al. 1992 2D model run and then scaled inter-annually according to WMO 1998/2003 scenario) which affect the ozone chemistry. A stratospheric time-dependent sulphuric acid field is also used to account for the heterogeneous chemistry after the 3 major volcanic eruptions (Agung 1963, El Chichon 1982, Pinatubo 1991, S. Bekki 2D model). A correction was applied to the S.H. lower stratospheric temperatures from the ERA-40 in order to reduce the strong
78
Final Report
effect on model ozone (a severe ozone hole was otherwise produced) of the abnormally cold austral winters and springs in the 1960s and 1970s.
Figure 4.7: Comparison of February SLIMCAT (with Cariolle scheme) total ozone against N.H. observations over Arosa.
Figure 4.8 shows the latitudinal and time evolution of key chemical species from
1979 until 2004. The growth of CFCs in the lower stratosphere is modelled well (with concentrations peaking after the mid-1990s) while the long-lived N2O temporal behaviour in the N.H. suggests a weakened Brewer-Dobson circulation until the early 1990s and an increased one towards the recent years. The chlorine species involved in ozone depletion show an increased abundance in the N.H. after 1980, with the HCl having a positive trend throughout the period.
As seen in figure 4.9, despite the temperature correction, pockets of very low ozone
in the Antarctic are still seen in some early years (1961, 1969, 1970, 1974-76) but the deep ozone hole (<200 DU) is more evidently seen after 1980, in agreement with observations. Ozone in the N.H. shows a large inter-annual variability with higher values in 1960s and 1970s. Less ozone appears later, in the 1990s. The effect of the El Chichon and Pinatubo eruptions can be seen in the lower ozone in 1983 and 1992, 1993 respectively, attributed, in this simulation, to the direct chemical and the indirect radiative/dynamical effects from the enhanced aerosol. A comparison of the modelled total ozone with ground-based measurements over Arosa in figure 4.10 shows that the full chemistry SLIMCAT ozone captures exactly, not only the sign of the inter-annual changes but also the absolute magnitude of the ozone column every year.
Task 4.1 resulted in Deliverables № 35 and 38.
79
Final Report
Figure 4.8: Zonal mean mixing ratio evolution of key species from the full chemistry SLIMCAT run.
80
Final Report
Figure 4.9: Zonal mean latitudinal and temporal variation of total ozone from SLIMCAT full chemistry run.
Figure 4.10: Comparison of February SLIMCAT (with full chemistry scheme) total ozone against N.H. observations over Arosa.
81
Final Report
Task 4.2 Chemical and dynamical contributions to past ozone trends (Contributors: UCAMB) Dependence of SLIMCAT simple chemistry ozone on dynamical parameters in the Northern Hemisphere
The dependence of the modelled zonal mean ozone O33 (PSC with EESC trend) from the simple chemistry experiments with ERA-40) on zonally averaged dynamical parameters (from ERA-40) was analysed for the period 1958 to 2001 (figure 4.11): The zonal wind at 500 K between 60o-70oN (which is a measure of the strength of the vortex) is positively correlated (r > 0.6) with the model ozone at high latitudes during December-March (figure 4.11 upper panel). Similar correlations hold when the model ozone O31 (gas-phase, no EESC) is considered, suggesting that these relations are connected to large-scale transport and are not chemically induced by the impact of the vortex strength and cold conditions on the PSC parameterization. The potential vorticity at 380K between 40o-60oN is positively correlated (r ~ 0.5) with the model ozone at middle latitudes from November to March (figure 4.11 middle panel), highlighting the effect of the injection of low-latitude, upper tropospheric, low-ozone/low PV air into the middle latitudes, caused by Rossby wave-breaking events near the tropopause. The temperature at 460K between 60o-70oN is positively correlated (r > 0.6) with the model ozone at high latitudes from December to March (not shown), in a similar way to the zonal wind/ozone relation. The temperature at 460K between 0o-20oN is positively correlated (r > 0.6) with the tropical model ozone (figure 4.11 lower panel) throughout the year. The vertical EP-flux at 100 hPa between 45o-75oN, a proxy for the strength of the Brewer-Dobson circulation, is positively and highly correlated with the high-latitude (r ~ 0.8) and mid-latitude (r > 0.5) model ozone, when it is averaged over the 3 months preceding the ozone value (not shown) confirming that the extra-tropics are influenced by the large-scale descent which brings higher ozone from the photochemical source region. When the EP-flux, when is lagged only by 1 month (or not at all), is also negatively and highly correlated (r < -0.8) with the model tropical ozone. Impact of SLIMCAT simple chemistry polar ozone depletion in middle latitudes
The monthly anomalies in N.H. middle latitudes are plotted in figure 4.12 for the model ozone column and the satellite observations from 1980 to 2002. The two model ozone schemes (with and without PSC chemistry) have very similar anomalies, suggesting that trends in ozone loss at polar latitudes do not have a large impact on middle latitudes in the model on long-term timescales. Many of the features of the temporal variation of the observations are well reproduced by the model, like the post-Pinatubo very low ozone in 1993, the recovery in the 2nd half of the 1990s and the lower ozone in 2000. There are differences in the early 1980s where the model underestimates the observed; this could be due to poor assimilation in the ERA-40 analyses for that period. The overall agreement of the model ozone changes (which are predominantly
82
Final Report
meteorologically driven) with the observations, suggest that dynamical processes can account for a good part of the long-term ozone variations.
Figure 4.11: Northern Hemisphere latitude-month variation of the correlation coefficients for 1958-2001 between zonal mean SLIMCAT ozone with simple chemistry and ERA-40: a) Zonal wind (U) at 500K averaged over 60o-70oN, b) Potential Vorticity (PV) at 380K averaged over 40o-60oN and c) Temperature at 460K averaged over 0o-20oN.
83
Final Report
Figure 4.12: Monthly mean column anomalies of SLIMCAT ozone O31 (gas-phase, blue) and O33 (PSC, red) and merged TOMS and SBUV observations (black) between 30o-60oN. The anomalies have been smoothed with a 24-month running mean to average both over the annual cycle and the QBO.
The Arctic ozone loss in the model can be quantified from the difference between the O31 gas-phase and the O33 ozone (with PSC parameterization). This modelled ozone depletion and its impact in the middle latitudes is shown for the N.H. from 1971 to 1985 (figure 4.13 upper panel) and 1986 to 2001 (figure 4.13 lower panel). The ozone depletion in the Arctic shows a considerable inter-annual variability and it is markedly stronger in the last 15 years (more than 25% of the total column depleted in some years), due to the increasing halogen loading (EESC peaks in 1996) and the frequent very cold northern winters. The middle latitudes are affected by the export of the polar depletion which after 1986 every year reaches 5% and in the colder years exceeds 10%.
84
Final Report
Figure 4.13: Time-series of percentage total ozone depletion in the N.H. calculated by SLIMCAT for a) 1971 to 1985 and b) 1986 to 2001.
The middle latitudes are affected by the anomalous polar chemistry via the dilution of ozone depleted air from higher latitudes. The key dynamical parameter in this process, where both meteorology and chemistry are important, is the extra-tropical vertical EP-flux at 100 hPa averaged between 45o-75oN and its relation with the polar vortex properties. We have shown in previous studies (from analysis of the UKMO and the ERA-15 data) that the strength of the wave drive is anti-correlated to i) the temperature of the polar vortex (and therefore the chlorine activation and the ozone depletion in high latitudes) and ii) the strength and persistence of the vortex (and thus the efficiency of mixing and export of ozone depleted air out of the vortex). This link is also confirmed with our current analysis of the ERA-40 simulation, and illustrated in figure 4.14 with the correlation of the relative mid/high-latitude ozone depletion during April and May with the winter EP-flux during January to March.
85
Final Report
Figure 4.14 1958-2001 time-series of April-May SLIMCAT ozone depletion (fraction of 40oN/80oN, solid triangles) and ERA-40 January-March vertical EP-flux at 100 hPa averaged over 45o-75oN (open squares).
Chemical and Dynamical Contributions to past ozone trends from SLIMCAT simulations
The chemical and dynamical contributions to long-term ozone changes and trends from 1979 to 2003 are estimated from monthly time-series of the SLIMCAT ozone output from the simple (Cariolle – dynamically-driven) and full (chemically and dynamically-driven) chemistry integrations. The monthly anomalies in N.H. middle latitudes are plotted in figure 4.15 for the model ozone column and the satellite observations from 1980 to 2002. The two model ozone schemes (gas-phase Cariolle (dynamically-driven) and full (chemically and dynamically-driven) chemistry) have very similar anomalies until 1992. The positive anomalies in the late 1980s (and similar) in both model schemes (in contrast to the observed) imply a strong influence of transport brought about by the ERA-40 forcing analyses (with the known shortcomings in that period). Generally the full chemistry model scheme has more negative anomalies than the simple chemistry scheme, as expected (for example in 1993 magnitude and the timing of the post-Pinatubo low ozone is captured exactly by the full chemistry which includes the volcanic influence on chemistry). In the period after 1993 the anomalies from the full chemistry model scheme remain negative and large with no sign of turnaround.
86
Final Report
Figure 4.15: Monthly mean column ozone anomalies of SLIMCAT simple (Cariolle gas-phase, blue open circles) and full chemistry (blue solid circles) and merged TOMS and SBUV observations (red) between 35o-60oN. The anomalies have been smoothed with a 24-month running mean to average both over the annual cycle and the QBO.
A linear regression analysis with 2 time terms before and after 1994 was performed
using the SLIMCAT output from the simple (Cariolle) and full chemistry schemes in order to assess the role of meteorological and chemical changes on the past ozone trends and the recently observed turnaround. Figure 4.16 (left plot) shows the calculated trends for 1979-1993 from these simulations, together with the trends for the satellite measurements. The SLIMCAT full chemistry year-round trend can explain around 2/3 of the observed trend in the N.H. middle latitudes (north of 30oN). The SLIMCAT trend with the simple (Cariolle) chemistry can explain 1/3 of the observed trend. We conclude that in the N.H. middle latitudes, for the period 1979-1993, the SLIMCAT trend is caused 50% by chemical and 50% by dynamical changes. Figure 4.16 (right plot) shows that the SLIMCAT dynamically-driven ozone trend during 1994-2003 can explain the entire observed trend. The SLIMCAT full chemistry trend is less negative in 1994-2003 than in 1979-2003 for the latitudes north of 45oN. These comparisons suggest that there is still very strong chemical depletion occurring in the N.H. and that the apparent recent turnaround is solely dynamically induced.
87
Final Report
Figure 4.16: Zonal mean 1979-1993 and 1994-2003 year-round trends for total ozone from SLIMCAT simulation with simple chemistry and full chemistry and satellite data.
Past ozone/climate interactions from UM simulations
The Met. Office Unified Model (UM) (58-level version) was used to study aspects of chemical/dynamical coupling.
All integrations started in December 1978 and calculate ozone using the Cariolle
scheme (see SLIMCAT runs). They differ only in the ozone used in the radiation scheme. In one experiment the Cariolle ozone is used; in two further integrations it is specified as detailed below. CO2 is held constant. The analysis is based on the twenty years from January 1980 to December 1999. Note that the dates only refer to the SST climatology used. Thus, the three runs are:
Experiment CO2 SSTs Ozone in Radiation Control Present day, const. observed Interactive Climatological Present day, const. observed Clim. from Control Loss Present day, const. observed Imposed O3 loss
i) Control: the model is integrated using observed SSTs and the ozone produced by
the Cariolle is used in the radiation scheme. This run is used to produce the zonal and monthly mean ozone climatology used in the other two integrations.
ii) Climatological: again, the model is integrated using the observed SSTs and ozone is calculated using the Cariolle scheme. However, in contrast to the Control
88
Final Report
integration, the ozone used in the radiation scheme is the zonal and monthly mean ozone climatology from the Control.
iii) Loss: this run is similar to the Climatological run, but with an additional imposed idealised ozone anomaly in middle latitudes. We have chosen a Gaussian distribution in latitude, pressure and time. Ozone loss is applied between January and June, from 40oN to 80oN. As figure 4.17a shows, the largest loss occurs at 60oN, 50 hPa during March (up to 1 ppmv or around 20% averaged over a layer from 100 to 30 hPa, see Braesicke and Pyle, 2003, for details).
Figure 4.17 The difference in a 5d mid-March average of zonal-mean zonal wind between the experiment using a superimposed mid-latitude ozone loss and the climatological integration using interactive ozone. The dashed line indicates the area where the ozone deficit exceeds 5%. b) The difference in March mean zonal-mean partial pressure of ozone for the experiment using a superimposed mid-latitude ozone loss and the control integration with interactive ozone. c) The difference in March mean zonal-mean partial pressure of ozone within the control integration partitioned using the strength of the mid-winter vortex (weak versus strong). Adapted from Braesicke and Pyle (2003).
In Figure 4.17a, the run with imposed ozone (compared to the Climatological run) loss shows very small dynamical changes in middle and low latitudes. Larger, more significant changes are found upwards and northwards of the location of imposed loss in the region of the polar vortex in March. In Figure 4.17b the ozone deficit caused by the imposed ozone loss mirrors the dynamical changes, and is only important in the lower vortex. Much larger ozone deficits in mid-latitudes are obtained for the Control run if we stratify results using the strength of the mid-winter vortex (Figure 4.17c). The inter-
89
Final Report
annual variation in dilution/transport of depleted polar ozone is evidently more important for the mid-latitude loss in spring than any possible subsequent positive feedback on the in-situ loss. Based on this experiment the in-situ sensitivity to the imposed mid-latitude ozone loss is mostly weak. We conclude that the (imposed) middle latitude anomaly does not have an important feedback on middle latitudes.
Figure 4.18: Time series of the Niño 3 index (red, left scale) and the equatorial lower stratospheric ozone column (blue, right scale). Graph B: Lag correlations between the Niño 3 index and the lower stratospheric ozone column in different latitude bands. The Niño 3 index is shifted forward in time. Graph C: Scatter plot between the Niño 3 index and the equatorial lower stratospheric ozone column at a lag of three months (as indicated in Graph B). From Pyle et al. (2005).
The ozone response to different climate regimes (El Niño/La Niña) was investigated using the output of the Control run with the observed SSTs and the modelled ozone interacting with the radiation scheme. Figure 4.18.a shows the anti-correlation between Niño 3 index and lower stratospheric column ozone index (derived by averaging the deseasonalised partial ozone column above 100 hPa from 20oS to 20oN and applying a running mean of 6 months, blue curve, right scale), with tropical lower stratospheric ozone being anomalously low during El Niño events. Closer inspection of the two graphs reveals a small time lag involved. Fig. 4.18.b shows lag-correlations between the Niño 3 index and the stratospheric ozone index for different latitude bands. A lag of three months (El Niño leading the equatorial ozone index by three months) yields the strongest correlation (y-axis inverted) for the equatorial ozone index (red line, as shown in Fig. 4.18.a. For higher latitudes in the northern and southern hemisphere the correlations are much weaker. There is a weak implication of slightly decreased LS ozone in the northern
90
Final Report
hemisphere (blue line) and a general tendency of slightly increased LS in the southern hemisphere (black line) after El Niño. Note that these are just tendencies and that the results in middle latitudes are not statistically significant. Fig. 4.18.c shows the detailed relation between the three-month-shifted Niño 3 index and the LS ozone column. The correlation appears tight for positive Niño 3 indices. There is a large spread for ‘‘neutral’’ conditions (close to zero) and a bias towards higher ozone values than explained by the linear correlation for negative Niño 3 indices. Lagged correlations between total ozone and the Niño 3 index (not shown) reveal banded structures of positive (but low) correlations around 30o N and 50o S. These structures maximise in areal extent (reaching to the poles) when the Niño 3 index is shifted by five to six months. This seems to support the findings of Zeng and Pyle (2005) that there is a weak tendency for increased ozone in higher latitudes after an El Niño event.
Task 4.2 resulted in Deliverables № 36 and 37.
Task 4.3 Ozone – climate interaction studies (Contributors: UCAMB)
The Met. Office Unified Model (UM) (64-level version) was used to study the
effect of changing CO2 and changing O3 on circulation and ozone distribution, with a set of idealized experiments, summarized in Table 4.1:
Table 4.1 UM model Run Scenarios Climatology in Radiation 352 ppm CO2 704 ppm CO2 1980 Ozone 1A 2A 2000 Ozone 1B 2B
The radiation is calculated using ozone climatologies with typical “1980” or “2000”
values (recurring monthly and zonal means, Fortuin and Langematz, 1995, updated) under low present-day (352ppmv, 1xCO2) and high (704ppmv, 2xCO2) CO2
concentrations. The main difference in the ozone climatologies are the polar ozone deficits which developed substantially between “1980” and “2000”. Seasonal-varying trends in other latitude bands are smaller but non-zero. Sea-ice and sea-surface temperature climatologies are the same for all four integrations and have an annual cycle but no inter-annual variability. The implication of this for the double CO2 experiments, 2A and 2B, is that we are focussing on the radiatively-driven stratospheric changes.
Figure 4.19 shows the mean annual cycle of ozone at 70hPa as a function of month and latitude for experiment 1A (upper left panel) and differences with respect to experiment 1A for all other experiments. The largest negative anomaly in the northern hemisphere during midwinter is found in experiment 2B. Weaker changes of opposite sign can be found during early winter in experiments 1B and 2A. In experiment 1B this would hint towards some kind of “self healing effect”, where an imposed ozone loss causes a response in the modelled ozone of opposite sign.
A clear impact on the evolution of ozone is visible in the SH in all experiments. In
all runs the modelled ozone decreases toward the end of the ozone-hole season, implying a prolonged and enhanced ozone hole as a response to either form of climate change (i.e.
91
Final Report
CO2 increase or ozone decrease). Anomalies on the northern hemisphere are only marginally significant. The complex behaviour of the northern hemisphere because of the larger internal variability compared to the southern hemisphere is discussed below.
Figure 4.19: Annual cycle of modelled ozone at 70 hPa as a function of latitude and month for all four experiments. Absolute values are plotted for experiment 1A; the percentage change with respect to experiment 1A is plotted for experiments 1B, 2A and 2B. From Braesicke et al. (2005).
Analysis of the zonal mean wind zonal wind at 60oN reveals (not shown) that the
weakest vortex is modelled for “2000” ozone with 1xCO2. Nearly identical means are obtained for “1980” ozone with 1xCO2 and “2000” ozone with 2xCO2. This demonstrates the possibility of compensating effects in the model, in particular the subtle balance between direct radiative and indirect dynamical forcings influenced by changes in the two radiatively active gases.
Generally, a CO2 increase is associated with tropospheric warming and
stratospheric cooling. Here, this is not necessarily apparent at high latitudes in the northern hemisphere during winter. In our runs doubling CO2, using “1980” ozone actually results in a weaker and warmer mid-winter vortex, whereas doubling CO2 using “2000” ozone shows the more expected behaviour leading to a colder and stronger vortex.
92
Final Report
Figure 4.20: Mid-latitude 100hPa heat flux for experiment 1A (magenta, left axis) as a function of month. Differences of mid-latitude heat fluxes between the idealised climate change experiments (right axes, see legend box for details). The yellow shading indicate the “midwinter” time window discussed here.
Comparing the mid-winter mid-latitude monthly mean heat fluxes at 100hPa (the
“tropospheric forcing”) between the four integrations provides further evidence for the competition between dynamics and the direct radiative effect in determining the wintertime vortex response to changes in ozone under different CO2 concentrations. Figure 4.20 shows the absolute time series for experiment 1A and the relative time series for all other experiments. Comparing the model response to ozone changes from “1980” (A) to “2000” (B) ozone under different CO2 concentrations (1B-1A for 1xCO2; 2B-2A for 2xCO2; colour bars), it is obvious that the changes in tropospheric forcing in mid-winter are of opposite sign, with increased tropospheric forcing in 1B leading to a generally weaker and warmer vortex compared to 1A and vice versa under doubled CO2. Interestingly, even though the signal is opposite during the winter months, in spring the signal is in phase. Another point is that in absolute heat flux values 1A (“1980” ozone and 1xCO2) has the smallest tropospheric forcing of all four experiments (see line-graphs) during January. The dynamical responses in scenario 1A (2000 ozone loss and 1xCO2) could be considered a present day scenario and are somewhat consistent with the results of SLIMCAT studies in 4.2, where a stronger circulation and a weaker and warmer vortex may have caused the apparent turnaround in N.H. ozone in recent years (see also Hadjinicolaou et al., 2005; Schnadt and Dameris, 2003).
Task 4.3 resulted in Deliverable № 39.
93
Final Report
3.5 Work Package 5 Synthesis of the EU-project CANDIDOZ (Chemical and dynamical Influences on Decadal Ozone Change) 1. Introduction
The possible depletion of the ozone layer was raised in the early 1970s (Crutzen, 1971; Johnston, 1971; Molina and Rowland, 1974; Stolarski and Cicerone, 1974). In the mid to late 1980s decreasing ozone amounts were observed at polar and middle latitudes which were related to the release of man-made Ozone Depleting Substances (ODS) such as chlorofluorocarbons and Halons. Meanwhile, in response to the threat of ozone destruction the Vienna Convention was signed in 1985. The Montreal Protocol which limits the emissions of ODS was signed in 1987 and was subsequently tightened up several times (Table 5.1). The implementation of the Montreal protocol and its amendments has successfully reduced the TROPOSPHERIC emissions of ODS at the end of the 1980s (Figure 5.1, middle panel). This has resulted in a recent decline of the effective stratospheric halogen loading by about 5% after peaking in the late 1990s as expressed by the time series of the equivalent effective stratospheric chlorine loading (EESC) shown in the top panel of Figure5.1. ______________________________________________________________________ Table 5.1: Global regulations to limit the emissions of ozone depleting substances ______________________________________________________________________
• (1977 World action plan (for ozone layer)) • 1985 Vienna convention • 1987 Montreal Protocol • 1990 London Adjustment and Amendment • 1992 Copenhagen Adjust. and Amendment • 1995 Vienna Adjustment • 1997 Montreal Adjustment and Amendment • 1999 11th Meeting of the Parties (China) • etc. needs update • 2005 10 years Vienna convention
______________________________________________________________________
The world’s longest total ozone series (Arosa, Switzerland) shows the typical features of ozone in the Northern mid-latitudes (Figure 5.1, bottom panel). The total ozone decreased from the early 1970s until the mid-1990s. After the record low ozone values in the early 1990s (related to the eruption of Mt Pinatubo in 1991) total ozone at northern mid-latitudes has increased for more than a decade. The decline up to the mid-1990s has been commonly attributed to chemical ozone depletion caused by the increasing concentrations of ODS. Now that the peak in EESC has passed, it is important to know for both scientific and political reasons whether the implementation of the Montreal Protocol has been effective. However, the attribution of ozone trends to changes in ODS emission is a difficult task because many factors contribute to ozone variability and trends, in particular at mid-latitudes. They include:
94
Final Report
• Large volcanic eruptions; • Arctic ozone depletion; • Long-term climate variability; and • Changes in the stratospheric circulation.
Figure 5.1: Top: Ability of ODS to deplete stratospheric ozone (Equivalent Effective Stratospheric Chlorine - EESC) in comparison with a linear trend starting in 1970. EESC is an overall measure of chemical ozone depletion taking into account the lifetimes and the chemical ozone depleting potentials of the individual chemical species. Middle: Global production of ODS. Bottom: Annual mean values of the total ozone series of Arosa (Switzerland) and relevant processes influencing total ozone at Northern mid-latitudes.
Analysing the existing measurement record in order to quantify how the Montreal protocol and its amendments have affected the ozone layer is thus hard and requires great care. However it is important in its own right and because improved understanding of these factors is needed to provide reliable predictions of stratospheric ozone.
95
Final Report
In particular, the quantification of dynamical influences on stratospheric ozone changes was highlighted as an outstanding issue in ozone research, for which the level of scientific understanding was quoted as medium or medium-low in WMO 2002 (Chipperfield and Randel, 2003 – see Table 4.5). This problem strongly restricts the interpretation of the past evolution of the ozone layer and reduces our confidence in predictions of its future evolution.
The EU-project CANDIDOZ (Chemical And Dynamical Influences on Decadal Ozone Change) aimed to address these issues. A major objective was to separate and quantify the individual factors contributing to past ozone variability and trends of the northern extratropical ozone shield. This report synthesizes the main findings of CANDIDOZ with respect to the policy-related questions outlined above. It does not attempt to describe all the work performed during the project which is described in the CANDIDOZ final report and in published papers, although a brief summary is given in Section 2 to show the importance of this ‘underpinning’ work to the headline CANDIDOZ results. Section 3 summarizes the results most relevant identifying and understanding of the effect of the Montreal Protocol on the ozone layer and Section 4 provides a summary.
2. Approach of the CANDIDOZ project The CANDIDOZ project adopted a multi-faceted approach with different research
groups involved in the: • production and re-evaluation of ozone data sets; • use of ERA-40 re-analyses to investigate dynamic changes in the atmosphere; • use of ERA-40 re-analyses in long Chemical Transport Models (CTMs)
simulations; and • development of process-based statistical approaches to describe past ozone
changes. 2.1. Ozone data sets
The available knowledge concerning the long-term evolution of the ozone layer comes from ground-based measurements based on the Dobson and Brewer instruments coordinated in the WMO’s Global Atmospheric Watch programme and from measurements by satellite instruments. The near-global coverage provided by the satellite instruments started in 1979 and complements the longer-term information from the ground-based measurements. The two systems are quasi- independent and they now provide essential quality control for each other.
In CANDIDOZ groups worked on the following topics (see Table 5.2): • Re-evaluation of the long-term total ozone series; • Re-evaluation of Arctic ozone sonde records; • Re-evaluation of ESA GOME record; • Comparison between satellite and ground-based measurements; and • Production of assimilated ozone data set with complete global coverage.
96
Final Report
Table 5.2 New ozone data sets from CANDIDOZ (produced by re-evaluation or data processing) ________________________________________________________________________ Re-evaluated ground-based total ozone observations
• Tromso, Norway (70ºN, 19ºE), 1935-1972, based on measurements from a Ferry spectrograph and from a Dobson spectrophotometer (Hansen and Svenoe, 2005)
• Hradec Kralove, Czech Republic (50ºN, 16ºE), 1962-2003, based on careful re-evaluation of measurements of Dobson and Brewer spectrophotometers (Vanicek at. al., 2003)
• Svalbard, Spitzbergen, Norway (79ºN, 12ºE), 1950-1962, based on measurements of a Dobson instrument (Vogler et al., 2005)
Profile ozone measurements from light balloons Eight Arctic stations from Europe and Canada (60º-80ºN), 1989-2003, homogenized taking into account e.g sensing solution concentrations and scaled to satellite (TOMS and GOME) total ozone. Satellite GOME WFDOAS, 1996-2003: Ozone measurements of satellite GOME produced by a new retrieval algorithm (Weighting Function DOAS algorithm WFDOAS (Coldewey-Egbers et al, 2004, 2005; Weber et al., 2005). Ozone data set CATO (CANDIDOZ Assimilated Three dimensional Ozone data set): Quasi three dimensional data set with daily resolution for the period 1979-2004 reconstructed from combined TOMS/GOME/SBUV satellite total ozone measurements (NIWA, provided by G. Bodeker) using meteorological information on short-term adiabatic advection (based on ERA-40 data) and adopting Kalman filter for the sequential assimilation of the measurements. This provides a new unique data set with complete global and temporal coverage of both total column ozone and the vertical distribution (Brunner et al., 2006a). Comparison of ground-based total ozone measurements (from Dobson and Brewer instruments (Vanicek at. al., 2003)) as well as between ground-based and satellite data (GOME and TOMS, Weber et al., 2005). ________________________________________________________________________ 2.2. Use of ERA-40 reanalyses to investigate dynamic changes in the atmosphere
The reanalysis of global meteorological data (ERA-40) by the European Centre for Medium Range Weather Forecast (ECMWF) was completed shortly after CANDIDOZ started. The availability of these reanalyses, with their reasonably high degree of internal consistency, allowed us to investigate dynamic influences on decadal timescales. A particular focus was on the dynamical influences on the Arctic stratosphere (Karpetchko et al., 2005) and the relation between northern hemisphere total ozone fields and meteorological patterns such as the Arctic Oscillation (Orsolini and Doblas-Reyes, 2003; Orsolini, 2004). While it has long been recognised that low-frequency tropospheric dynamics contributes to ozone column variability in the northern hemisphere, early studies focused mostly on the impact of the North Atlantic Oscillation or the Arctic Oscillation. More recent studies, based either on ground-based ozone time series analysis [Hansen and Svenoe, 2005; Bronnimann et al., 2000; Steinbrecht et al., 2001] or TOMS
97
Final Report
satellite observations [Orsolini and Doblas-Reyes, 2003] have shown that a variety of climate patterns and teleconnections leads to strong regional signatures over the Euro-Atlantic sector. In addition, Orsolini [2004] and Bronnimann et al. [2004] showed the remote influence of meteorological phenomena over the North Pacific, such as fluctuations in the Aleutian Low or El-Nino episodes, upon ozone variability in the Euro-Atlantic sector. The former study showed that inter-annual ozone variations between the North Pacific and Northern Europe were associated with an Aleutian-Icelandic seesaw-like variability, and were considerably larger in February than those associated with the Arctic Oscillation. Finally, in addition to the studies using ERA-40, changes in the occurrence of laminae over Europe were investigated using a combination of ozonesonde and meteorological data (Krizan and Lastovika, 2004; 2005). 2.3. Use of ERA-40 re-analyses in long Chemical Transport Models (CTM) simulations
The ERA-40 data were also used to run CTMs for more than 40 years. This period covers a decade of the chemically undisturbed stratosphere, the period of increasing chemical ozone depletion and the last decade in which ozone in northern mid-latitude increased. Long numerical simulations have the potential to quantify the importance of the different chemical and dynamical processes affecting the ozone layer. However the delay in the production of the ERA-40 data combined with unforeseen technical problems in using ERA-40 to run stratospheric CTMs meant that less progress in using models to understand past ozone trends was made than originally hoped for. Nevertheless long-term simulations with both simplified and full chemical descriptions were made (Table 3) and initial interpretations have been performed (Hadjinicolaou et al., 2005). More results from long runs are expected in future, e.g. in the SCOUT-O3 project. Table 5.3 Long term simulations performed by global models in CANDIDOZ. Model Description of simulation FINROSE (L. Backman, FMI) 1958-2005; studying effect of model
formulation on stratospheric circulation; model validation in Arctic
OSLO CTM 2 (I. Isaksen, UiO) 1990-2000; full chemistry; investigating chemical and dynamical influences and testing heterogeneous chemistry schemes
REPROBUS (F.Lefevre, UPMC/CNRS) 1958-2004; full chemistry; studying effect of model formulation on stratospheric circulation
SLIMCAT (P. Hadjinicolaou, UCAMB) 1958-2004; (i) simplified chemistry to investigate dynamic influence; (ii) full chemistry to study chemical effect
UM-UCAM (J. Pyle, UCAMB) 1980-2000; investigating radiative feedback of O3 change on dynamics; sensitivity studies of stratospheric changes under doubled CO2 conditions.
98
Final Report
2.4. Development of process-based statistical approaches to describe past ozone changes
Multiple linear regression has been the most common approach used to determine trends in ozone measurements (Chipperfield and Randel, 2003 and references therein). Up until a few years ago, the underlying assumption was that any long-term change was linear and caused by chemical ozone depletion resulting from the steadily increasing concentrations of ODS. Two factors have caused a revision in this approach. First, ODS concentrations stopped increasing. Second the influence of dynamic influences on ozone on decadal timescales became more apparent (e.g., Hood et al, 1997; Steinbrecht et al., 1998). However while correlations between ozone and a number of meteorological quantities (such as tropopause height) were statistically significant, it was impossible to link these unambiguously to particularly physical processes (Chipperfield and Randel, 2003). A real effort was made in CANDIDOZ to produce process-based indicators of dynamic and chemical influences that could be used in statistical models in order to aid the attribution of past ozone changes. Again, the availability of the ERA-40 analyses was of great benefit as long, internally consistent time series of these dynamic quantities are required for the statistical analyses (Wohltmann et al., 2005). Three particular proxies were routinely used in the analyses described in the next section and are described further here.
First, the volume of polar stratospheric clouds (PSC) correlates closely with the accumulated chemical ozone loss in the Arctic vortex in each winter (Rex et al., 2004). This quantity shows large interannual variability depending on the specific conditions in the polar vortex each winter. Over the last decades the potential for chemical ozone depletion has changed significantly (top panel in Figure 5.1) and so the PSC volume was scaled by the EESC in order to investigate the multi-decadal influence of ozone depletion in the Arctic vortex.
Second, a new proxy was developed to describe short-term meteorological variability resulting from horizontal and vertical movements (Wohltmann et al., 2005). This proxy was calculated using equivalent latitude (i.e the real latitude adjusted to allow for the displacements of the jet stream, etc.) and potential temperature in the vertical (to allow for vertical expansion or compression of air) and the ozone climatology from CATO (see Tab. 2).
Third, the strength of the stratospheric circulation, i.e. the relatively slow movement of air from the tropics to higher latitudes, determines how much ozone is present above mid- and high latitudes in each late winter and spring, and so is a strong influence on the interannual variability in ozone (Fusco and Salby, 1999; Randel et al., 2002) and polar ozone chemistry (Weber et al., 2003). Time series of the measure of this circulation, the Eliassen-Palm (EP) flux, were calculated from ERA-40 and NCEP data and used in a number of studies. Finally, a neural network approach was used for trend analysis of seven European stations with long total ozone records. This new technique was to our knowledge not applied in ozone trend analysis before (Metelka et al., 2005).
3. What have we learnt about long-term ozone trends? Three main issues of interest to the public and to policy-makers have been
addressed in CANDIDOZ which relate to long term changes in ozone:
99
Final Report
• The causes of past ozone changes and the impact of the Montreal Protocol to date; • Long-term changes in the Arctic stratosphere; and • The implications for future ozone.
Different aspects of these have been addressed in different sub-projects in
CANDIDOZ and here we attempt to pull together the results. Each issue is addressed in turn. Before doing so, it is worth discussing the concept of recovery to show just how complex an issue it is.
The term ‘recovery’ has been used in a number of different ways (Hofmann and Pyle, 1999; Weatherhead et al., 2000; Reinsel et al., 2002) and there is on-going disagreement and/or confusion about what is its most suitable definition, even among specialists. Should it be defined according to halocarbons, ozone or UV radiation? Should it be defined as a return of any or all of these to pre- 1980 ozone levels? Should it be defined as a clear reversal of past trends? Or as a clear reduction in the magnitude of past trends? Does it matter what the cause of any change in the trend of ozone or UV radiation is? Should it be linked to the actions taken under the Montreal Protocol? Does it matter that the atmosphere is unlikely to return to its pre-1980 condition? To complicate matters further, recovery (particularly for ozone and UV radiation) may occur in some regions or at some times of year, and not at others. For example, much of the early work on recovery focused on the regions where the signs of recovery might first be observed – polar regions, high altitudes (e.g. Hofmann and Pyle, 1999).
From a political perspective it makes sense to consider recovery in terms of ozone recovery. In this light, the first stages of recovery, namely reductions in ODS emissions and the subsequent turnaround and decrease in EESC are already in the past. The next stages – the slowing of the ozone decline and the reducing chemical impact of ODS (often called ‘turnaround’) – are strongly debated and are discussed below. Our scientific understanding tells us that we must be passing through these stages around now, but we need to demonstrate it unequivocally. The final stage, the full recovery of ozone, is clearly in the future. The main point of current discussion about this is what level of EESC this will correspond to. The commonly used threshold is 2000 pptv (Cl equivalent), but this is an arbitrary choice which coincides with the start of the modern satellite record and our scientific understanding tells us that ozone depletion must have occurred at lower EESC levels. However that is a point about how future scenarios are presented and not the central concern of CANDIDOZ.
In the following discussion, we concentrate on whether the influence of the reduced ODS emissions can be seen in the observational ozone record. To minimize potential confusion, we attempt to be precise about the conclusions of each analysis. In particular the use of process-based explanatory variables in the statistical analyses allows us to attribute past changes in ozone to particular processes, at least to some degree. The statistical methods applied in CANDIDOZ are summarized in the Box. A summary of the results from this section is presented in Section 4.
100
Final Report
________________________________________________________________________ Box: Concepts for description of long-term changes by statistical modeling ________________________________________________________________________ In linear multiple regression analysis the measured (monthly mean) ozone values (Y(t)) Y(t)=a(t)+b(t)·t+ +ε(t) ∑ =
⋅N
jjj tXtc
1)()(
are fitted by a constant (a(t)) and terms describing natural variability using explanatory variables (Xj) (see Section 2.4) and taking into account autocorrelations in the residuals e(t) which is important to determine realistic confidence intervals. For description of trends (b(t)) the following options are used:
1. Single linear trend vs. time (“simple linear trend”): This approach was commonly used in earlier ozone assessments and the determined trends were usually attributed to anthropogenic ozone depletion. This description is still useful to characterize the long-term changes, but this concept cannot describe the influence of the time evolution of ODS after the middle of the 1990s when EESC started to deviate from a (almost) linear increase (Fig. 5.1)
2. Two linear trends (“hockey stick”): The first linear downward trend starts at the beginning of the series (or at the beginning of the 1970s), an additional linear trend is introduced aiming to describe the influence of the Montreal Protocol. The second slope term includes information on (i) whether the second linear slope significantly deviates from the first one and (ii) whether the second period shows a significant upward trend. However, this approach is rather empirical and the series to fit the second linear trend is short.
3. The ozone time series can be fitted by EESC instead of linear trends (“EESC model”). This approach is based on a physical quantity and the fit of explanatory variables is based on the entire series. However it should be remembered that ozone loss is not simply linearly dependent on EESC and depends on other factors such as aerosol surface area, PSC existence.
4. In the first two cases, the trend term represents that part of the ozone tendency which can not be explained by other (natural) factors, and which is therefore interpreted as being due to anthropogenic activity.
________________________________________________________________________ 3.1. The causes of past ozone changes and the impact of the Montreal Protocol to date
Observations of total ozone over Arosa show a decrease until the mid-1990s followed by a smaller increase (Figure 5.1). The statistical trend analyses in CANDIDOZ show that the change in trend from the earlier period is significant for selected months over large areas in the extratropics, similar to results presented by Reinsel et al. (2005). This type of statistical analysis which was proposed to describe the turn-around in ozone does not in itself explain why the total ozone trends have changed. Much of the work in CANDIDOZ has been addressing this issue.
101
Final Report
month Figure 5.2 Seasonal cycle of monthly ozone changes in DU/year for various processes as derived from eight European Dobson stations (197X-2003). A large contribution of the observed changes are related to the equivalent latitude proxy (cyan line, see also Section 2.4) describing changes in dynamics (Wohltman et al., 2005).
A full understanding of the upward trends in the 1990s requires a quantitative understanding of all known processes and factors that can contribute to the observed decrease until around 1996 and the subsequent increase. The significance of this increase is harder to evaluate, although it is outside the variability expected using the ‘traditional’ statistical models analyses. The results of CANDIDOZ indicate that the trend change in the northern hemisphere is primarily related to the processes other than ODS emission decrease. Analysis of the long-term total ozone records at 8 European stations shows the influence of dynamic processes on the calculated trends (Figure 5.2). Ozone trends over Europe appears to have been particularly affected by dynamical changes (see also Staehelin et al., 2001)
Analyses of total ozone measurements by satellites (Dhomse et al., 2005; Brunner et al., 2006b) also show that (i) in general the chemical turn-around in gas-phase chemistry is not the dominant contribution (in some cases not statistically significant), when dynamical processes are properly accounted for in the model, and ii) changes in ozone transport (or in polar chemistry) and the solar cycle have larger contributions to the observed total ozone increases in northern mid-latitudes.
In Dhomse et al. (2005) both the use of monthly linear trend terms for the whole
period 1979-2003 and, alternatively, the EESC term leads to identical goodness of fits to the observational record. From the EESC fit a modest recovery of 3 to 4 DU/decade (1-2%) is significant during winter and spring north of about 35 deg. If one replaces the dynamical proxies such as the eddy heat flux terms with the PSC volume proxy (polar ozone loss) the EESC contribution (gas phase chemisty) becomes insignificant within 2sigma as shown in Figure 5.3 (Dhomse et al. 2005). Replacing EESC or linear trend terms by the hockey stick, the trend change after 1996 becomes statistically insignificant (not shown). In nearly all analyses the contribution from the rising solar cycle after 1996 to the recent observed ozone increases is important.
102
Final Report
Figure 5.3 Monthly linear trends for the whole period (1979-2003) (left) and, alternatively, EESC fit (right) derived from SBUV V8 total ozone. When dynamical proxies as used in the left panel are replaced by the PSC volume proxy (right panel) the EESC fit (here expressed as a linear trend up to 1995) standing for changes in gas-phase chemistry is statistically insignificant (Dhomse et al,. 2005). Yellow and green shadings indicate 1sigma and 2sigma statistical significance.
In Brunner et al. (2006b), only one analysis of satellite measurements indicates a
significant increase for a few months and selected latitudes when the processes described in 2.4 are included in the statistical model. This can be seen by comparing the left and right panels of Figure 5.4 where the area of significant change is markedly reduced when the additional factors are included. Overall, while the influence of declining halogen levels on total ozone is becoming discernible in the observational record, the signal is still somewhat ambiguous and is dominated by other influences in the recent past.
The conclusions of these statistical analyses of observations are supported by a
SLIMCAT modelling study which compared total ozone observations with total ozone for a simulation driven by ERA-40 winds and temperatures, but without chemical ozone depletion from ODS (Hadjinicolaou et al. (2005); Figure 5.5). The ozone trends at northern mid-latitudes prior to the mid-1990s were found to be ~30% by dynamic changes, leaving ~70% caused by ODS chemistry (Figure 5.6). The dynamic influence on the trend during this first period does not seem to be due to a radiative feedback resulting from the chemical ozone loss itself (Braesicke and Pyle, 2003). Since the mid-1990s effectively all the increase can be explained by dynamic changes. The companion simulation with full chemistry (not shown) shows no evidence for a chemical recovery due to the turnaround in ODS. The magnitude of the observed increase in total ozone is too large for it to have been caused by the small decrease in EESC after 1995. These results are consistent the conclusions of a separate simulation of full chemistry with SLIMCAT (Chipperfield et al., 2005) and with 2D and 3D GCM model simulations (e.g., Andersen et al., 2005; Chipperfield and Randel, 2003) which show little anticipated effect of the decreasing EESC to date.
103
Final Report
Figure 5.4 Results of statistical analyses of the data set CATO, depicted as change in rate from a linear decreasing downward trend (see Section 2.4). Three different sets of explanatory variables were used: Left side: QBO, solar cycle and aerosols; middle: meridional circulation, QBO, solar cycle, aerosols right side: meridional circulation, polar ozone chemistry (see Table 5.3), QBO, solar cycle, aerosols. Shading indicates that the change in trend is not significant at that latitude and month at the 95% significance level (Brunner et al., 2006b).
Figure 5.5 Monthly mean column anomalies of SLIMCAT ozone (simple gas-phase, blue; also including simple PSC, red) and merged TOMS and SBUV observations (black) between 30o-60oN. The anomalies have been smoothed with a 24-month running mean to average both over the annual cycle and the QBO (adapted from Hadjinicolaou et al., 2005).
104
Final Report
Figure 5.6 Zonal mean trends (year-round) of the modelled and observed ozone trends from 1979-1993 and 1993-2004 calculated after removal of QBO and solar cycle. The error bars show 95% confidence limits (Hadjinicolaou et al., 2005).
Different levels in the vertical distribution of ozone are affected to varying degrees by dynamics and chemistry. In the upper stratosphere ozone trends are expected to be less influenced by changes in dynamics and therefore more directly affected by ODS. Newchurch et al. (2003) analysed satellite and ground-based measurements of upper stratospheric ozone and found evidence for the ‘first stage of recovery’. In CANDIDOZ, Zanis et al. (2006) have analysed the long, revised Umkehr record from Arosa (1956-2004) using similar techniques. They find that the differences between trends in the upper stratosphere (Umkehr levels 7 & 8 in Figure 5.7) before and after 1995 are not statistically significant. Further, a better fit to the Arosa total ozone record is obtained when EESC instead of a single linear trend, but that a similar effect is not found in the upper stratosphere. They conclude that it is too early to say whether a turnaround in upper stratospheric ozone has occurred although the results are clearly consistent with the decreasing EESC. It is possible that the lack of significance in these results is due to the use of a record (albeit a long one) at a single site.
Orsolini and Reyes [2003] demonstrated that, in the spring, up to 70% of the local ozone negative trends over Europe were due to changes in four leading climate patterns. Inclusion of these four leading patterns captured springtime decrease in ozone over Western Europe (defined as 40N-60N, 15W-15E) over the years 1979-1997, and increase in the late nineties, following the absolute minimum in spring 1997, which was caused by an anomalous East-Atlantic pattern (Figure 5.8).
105
Final Report
Figure 5.7 Vertical distribution of year-round ozone trends (%/decade) before and after 1995 calculated from the homogenized Umkehr measurements of Arosa (Switzerland) using the EESC model and other proxies: QBO and solar cycle in (a1); QBO, solar cycle and tropopause height in (a2); and QBO, solar cycle and equivalent latitude proxy in (a3). (From Zanis et al., 2006).
So, what has caused the recent increase? The answers from multiple regression
models of CANDIDOZ are not completely clear: Dhomse et al. (2005) report enhanced ozone transport into higher latitudes as part of the residual circulation and at the same time reduced frequency of cold Arctic winters with enhanced polar ozone loss while Brunner et al. (2006b) find different results when fitting with EESC (“EESC model”, see Table 5.3) or by the hockey stick method with contributions from different processes which varies from month to month.
106
Final Report
Table 5.4 Summary of long-term trend analysis of CANDIDOZ Authors; method
Measurements Explanatory variables
Result turn-around
Ground-based total ozone measurements Wohltmann et al. 2005 - Multiple regression
8 ground-based European Dobson stations Period: 1970-2003
Equiv. latitude, QBO, solar cycle, aerosols, EP flux PSCvol.xEESC Linear trend, EESC
not significant
Hansen and Svenoe, 2005 - Multiple regression
Total ozone series of Tromso
not attempted
Metelka et al., 2005 - Neural network
7 European Dobson Series
NAO, solar cycle, aerosols, stratospheric wind, vertical H and T profiles; some preprocessed by principal component analysis
no evidence for turn-around
Zanis et al., 2005 - Multiple regression
(profile) Umkehr series of Arosa
QBO, solar cycle, equivalent lat./ Tropopause altitude/NAO
evidence for upper stratospheric turn-around only when EESC-model (see Tab 5.3) is used
Satellite measurements Dhomse et al., 2005 - Multiple regression
SBUVTOMS V8 until 94 GOME 1995-2003 Period: 1979-2003
QBO, ENSO, solar cycle (MgII) EP flux or PSCvol.xEESC Linear trend or EESC
not significant
Brunner et al., 2006b - Multiple regression
CATO Period: 1979-2003
QBO, solar cycle, EP flux, PSCvol.xEESC, linear trend, EESC
coordinates : equival. Lati/ q some indication for turn-around
3.2 Long-term changes in the Arctic stratosphere
Ozone in polar regions can be rapidly destroyed by heterogeneous reactions occurring on polar stratospheric clouds (PSCs). These processes take place in the polar vortex, which forms every winter over the poles and the extent of the ozone loss is determined by a number of chemical and dynamical factors. Today we know that the strong interannual variability of Arctic ozone depletion depends on the following factors:
(i) The temperature of the polar vortex and formation (below a certain temperature) of PSCs;
(ii) The duration of the existence of PSCs each winter; and (iii) The longevity of the polar vortex into spring.
The knowledge about Arctic ozone loss has increased strongly over the last decade
or so through the large European campaigns and through improved monitoring in the Arctic region particularly with ozonesondes. The chemical ozone loss has been calculated for every winter since 1991 and a remarkably tight, linear relation is found between this and the volume of PSCs present that winter (Rex et al., 2004). For each cooling of 1 K, an additional 15 Dobson Units of depletion in the ozone column is anticipated, 3 times
107
Final Report
larger than calculated in existing atmospheric models. Further long term cooling in the Arctic winter stratosphere had resulted in a three-fold increase in PSCs, so that the stratospheric climate conditions had become significantly more favourable for ozone loss since the 1960s, an effect which considerably amplified the increasing halogen loading of the stratosphere. If these temperature trends continue into the future, Arctic ozone losses would increase until 2010-2015 ozone and decrease only slowly afterwards (Knudsen et al., 2004).
Figure 5.8 Spring ozone over Western Europe (40N-60N, 15W-15E) from TOMS data (thick line) for years 1979-2000; reconstructed spring ozone with anomalies from the four leading Euro-Atlantic patterns successively added to the climatological mean: i.e. including the NAO (thin), plus the Scandinavian pattern (dot), plus the East-Atlantic pattern (dot-dash) and the European blocking pattern (long dash). The combined curve (long dash) better tracks the observed ozone variations, albeit the last pattern brings little change over that region. [Adapted from Orsolini and Reyes, 2003]
Decadal variations in the polar vortex have been determined using ERA-40
reanalysis data from 1957 to 2003 (Karpetchko et al., 2005). In general, the large interannual variability is the dominant feature, obscuring any long-term changes. However some features can be distinguished. For example the average PSC existence in the Arctic increased from 1959 to 2002, but no statistically significant trends in size, temperature and duration were found for 1979-2002. Statistically significant increases (95% error probability) were found in vortex duration, vortex size and PSC area in the lower stratosphere for March for 1979-97 but not for the period as a whole. The lack of an overall trend in PSC existence is consistent with the analysis of Rex et al. (2004) who argue that, while there is no trend in most winters, more PSCs are now formed in the extremely cold winters.
Identification of long-term changes in the Arctic winter stratosphere is tricky due to multi-annual climate variability. The strong dependence on the length of the record being considered is also found in analyses of the re-evaluated Arctic ozonesonde record, which showed statistically significant decreases in stratospheric ozone for 1989-97, but not for the period from 1989-2003 (Kivi et al., 2006). The ozone decreases are strongly correlated with the tropopause height and this dynamical influence is the largest influence
108
Final Report
on stratospheric ozone in the Arctic. There is also a significant correlation with the measures for chemical polar ozone depletion (accounting for up to 50% of ozone variability in March) and the strength of the stratospheric circulation. In the Arctic troposphere a statistically significant increase of 12% is observed from 1989 to 2003, with the largest changes in January-April.
The strong correlation between the ozonesonde measurements and tropopause height is probably another manifestation of the influence of low frequency tropospheric patterns found in total ozone data (Orsolini and Doblas-Reyes, 2003). These patterns (e.g., the North Atlantic Oscillation) are associated with anomalies in the tropopause height and so contribute significantly to the ozone variability. Again, trends in the occurrence of these patterns and in ozone are found over the first part of record (up to the mid-1990s), but they become insignificant when recent years are included.
Ozone loss in the Arctic vortex can contribute to reduced ozone amounts over mid-latitudes when ozone-depleted air is transported from high to middle latitudes. The magnitude of this has been estimated in several of the trend studies described in Section 3.1 and modelling studies (Hadjinicolaou, and Pyle, 2004), but it is hard to develop a reliable process-based indicator which takes into account all the chemical and dynamical factors involved. In CANDIDOZ, this has also been investigated in a series of case studies of individual winters which concluded that dilution can explain 29% of the overall trend in total ozone from 1979-1997 (Andersen and Knudsen, 2006). It should be noted that this fraction depends quite strongly on the period being considered because (a) proxies for dynamical changes were not included in the statistical model and (b) Arctic ozone loss varies substantially from year to year.
3.3 Implications for the future The results in Section 3.1 and 3.2 illustrate the sensitivity of ozone (and hence UV)
to changes in dynamics as well as changes in chemistry. The total ozone column is sensitive to changes in the large-scale stratospheric circulation which is driven by the tropospheric forcing at high latitudes. If, as some models suggest, the speed of this circulation increases (Schnadt and Dameris, 2003; Braesicke et al., 2006), then there will be enhanced ozone columns at mid and high latitudes and reduced chemical ozone loss in the (warmer) Arctic stratosphere. However predicting the effect of changed tropospheric forcing is not simple, and it causes much of the disagreement between models (Andersen et al., 2005). It is also possible that a cooler lower stratosphere will result in a reduced circulation. Either way, the resulting impact on the ozone column could be several percent and be similar to the changes that have occurred as a result of ODS outside the regions of severe polar loss.
The past record also shows that physical changes in the lower stratosphere can have a strong influence on ozone. Again the main question is are any long-term changes likely to occur as a result of climate change, and again it is not possible to give categorical answers based on existing model studies. However analyses of past meteorological records show that the tropopause height has increased since the 1950s and that there has been an acceleration since the 1970s which is attributable to climate change (Santer et al, 2003). This effect is larger at higher latitudes and is likely to continue as greenhouse gas
109
Final Report
levels rise. The ozone column is sensitive to tropopause height and a rising tropopause will lead to reduced ozone levels.
Neither the individual nor the cumulative effects of the possible future dynamic changes can be quantified at present. The results in CANDIDOZ illustrate the sensitivity of stratospheric ozone to decadal changes in dynamics and transport and emphasise the importance of gaining a better understanding of what changes might actually occur.
4. Discussion and Conclusions The large dynamical influence on decadal changes in ozone over the last 25 years
means that any meaningful discussion of recovery has to be carried out with reference to the effects of the Montreal Protocol. The importance of dynamical influences on ozone trends has been recognised for some while. However the difficulties in quantifying the impact has led to past trends being generally attributed to halogen chemistry with an implicit assumption that any uncertainties resulting from interannual changes in dynamics are adequately represented in the error bars associated with the trend estimates. This assumption holds as long as the underlying dynamical behaviour (and variability) does not change over the period under consideration. However if changes are occurring on decadal timescales, then this no longer holds true.
The observational record shows unambiguously that total ozone over northern mid-latitudes decreased from 1980 to the mid-1990s. Since then it has increased by a similar amount. Statistical and modeling studies carried out in CANDIDOZ show that the main cause of this change in ozone trends results from changed dynamical behaviour. Further it seems likely that dynamical changes are occurring in both the large-scale circulation (as measured by the EP flux) and in processes in the lower stratosphere. It is not possible at this stage to say what has caused these dynamical changes and whether they are likely to continue.
The amount of halogen in the stratosphere (as measured by EESC) peaked in 1997 and by 2004 had fallen by about 6% as a result of the application of the Montreal Protocol. However this change is not sufficient (on current understanding) to explain all (or even most) of changed trend in total ozone over the last 10 years. We would expect the chemical ozone trend to have flattened in the late 1990s, and there is some evidence in the observational record that this has occurred. In the upper stratosphere, where gas-phase halogen chemistry should dominate the ozone trend, the observed ozone changes are well described by a statistical model using EESC as an explanatory variable. This is the best evidence to show that the Montreal Protocol is having a beneficial effect on stratospheric ozone.
110
Final Report
References (* indicates CANDIDOZ related work) *Andersen, S.B., A. Stevermer, E.C. Weatherhead, J. Austin, C. Brühl, E.L. Fleming, J.
de Grandpré, V.Grewe, I. Isaksen, G. Pitari, R.W. Portmann, B. Rognerud, J.E. Rosenfield, S. Smyshlayev, T. Nagashima, G.J.M. Velders, D.K. Weisenstein and J. Xia, Comparison of recent modeled and observed trends in total column ozone, JGR, accepted, 2005.
*Andersen, S.B. and B.M. Knudsen, The influence of polar vortex ozone depletion on NH midlatitude ozone trends in spring, submitted to ACPD, 2006.
Baldwin, M.P. and T. J. Dunkerton, 2001: Stratospheric harbingers of anomalous weather regimes, Science, 294, 581-584.
Bekki, S. and Pyle, J. A., Two-Dimensional Assessment of the Impact of Aircraft Sulphur Emissions on the Stratospheric Sulphate Aerosol Layer, J. of Geophys. Res., 97 (D14), 15839-15847, 1992
Berglen, T.F., T.K. Berntsen, I.S.A. Isaksen, J. K. Sundet, A global model of the coupled sulfur/oxidant chemistry in the troposphere: The sulfur cycle, J. Geophys. Res., 109 (19), doi:10.1029/2003JD003948, 2004.
*Braesicke, P. and J.A. Pyle, Changing ozone and changing circulation in northern mid-latitudes: Possible feedback?, Geophys. Res. Lett., 30, 1059, doi:10.1029/2002GL015973, 2003.
*Braesicke, P., M. Hurwitz and J.A. Pyle, Idealised climate sensitivity studies with the UM: implications for ozone, Meteor. Z., submitted, 2006.
Bronnimann, S., J. Luterbacher, C. Schmutz and H. Wanner, Variability of total ozone at Arosa, Switzerland, since 1931 related to atmospheric circulation indices, Geophys. Res. Lett., 27, 2213-2216, 2000.
Bronnimann, S., et al., Extreme climate of the global troposphere and stratosphere in 1940-42 related to El Nino, Nature, 431 (7011), 971-974, 2004.
*Brunner, D., J. Staehelin, H. R. Kuensch, and G. E. Bodeker, A Kalman filter reconstruction of the vertical ozone distribution in an equivalent latitude - potential temperature framework from TOMS/GOME/SBUV total ozone observations, J. Geophys. Res, in press, 2006a.
*Brunner, D., J. Staehelin, J. Maeder and I. Wohltmann, Trends in total and vertically resolved stratospheric ozone based on a 3D assimilated data set derived from satellite total ozone observations, to be submitted to ACPD, 2006b.
Carslaw K., Luo B., Peter, T., An analytic expression for the composition of aqueous HNO3+H2SO4 stratospheric aerosols including gas phase removal of HNO3. Geophys. Res. Letters, 22, p.1877-1880, 1995.
Chipperfield, M., A three-dimensional model study of long-term mid-high latitude lower stratosphere ozone changes, Atmos. Chem. Phys.,3, 1253– 1265, 2003.
Chipperfield, M.P. and W.J. Randel et al., Chapter 4 in Scientific Assessment of Ozone Depletion: 2002, WMO Global Ozone Research and Monitoring Project Report No. 47, Geneva, 2003.
Chipperfield, M.P., W. Feng, M. Dorf and K. Pfeilsticker, Studies of mid-latitude ozone changes using a 3D CTM and ERA-40 analyses, to be submitted to ACPD, 2005.
* Coldewey-Egbers, M., M. Weber, M. Buchwitz, and J.P. Burrows Application of a modified DOAS method for total ozone retrieval from GOME data at high polar latitudes, Adv. Space Res., 34, 749-753, 2004.
111
Final Report
* Coldewey-Egbers,M., M. Weber, L. N. Lamsal, R. de Beek, M. Buchwitz, J. P. Burrows, Total ozone retrieval from GOME UV spectral data using the weighting function DOAS approach, Atmos. Chem. Phys. 5, 5015-5025, 2005
Crutzen, P.J., Ozone production rates in an oxygen-hydrogen-nitrogen oxide atmosphere, J. Geophys. Res., 76, 7,311-7,327, 1971.
de Zafra, R. and S. Smyshlyaev, On the formation of HNO3 in the Antarctic mid-to-upper stratosphere in winter, J. Geophys. Res., 106, D19, 23115-23125, 2001.
*Dhomse, S., M. Weber, I. Wohltmann, M. Rex and J.P. Burrows, On the possible causes of recent increases in NH total ozone from a statistical analysis of satellite data from 1979-2003, ACPD, 5, 11331-11375, 2005.
Fioletov, V. E. and Shepherd, T. G., Seasonal persistence of midlatitude total ozone anomalies, Geophys. Res. Lett., 30, 1417, doi:10.1029/2002GL016739, 2003.
Fortuin J.P.F and U. Langematz, An update on the global ozone climatology and on concurrent ozone and temperature trends. – Proceedings of the Society of Photo Optical Instrumentation Engineers, 2311, 207-216, 1995.
Fusco, A.C. and M.L. Salby, Interannual variations of total ozone and their relationship to variations of planetary wave activity, J. Climate, 12, 1619-1629, 1999.
*Hadjinicolaou, P and J.A. Pyle, The Impact of Arctic Ozone Depletion on Northern Middle Latitudes: Interannual Variability and Dynamical Control, J. Atm. Chem., 47 (1), 25-43, 2004.
*Hadjinicolaou, P, J.A. Pyle and N.R.P.Harris, The recent turnaround in stratospheric ozone over northern middle latitudes: a dynamical modelling perspective, Geophys. Res. Lett., 32, L12821, doi:10.1029/2005GL022476, 2005.
*Hansen, G., and Svenoe, T., Multilinear regression analysis of the 65-year Tromso total ozone series, J. Geophys. Res., 110, D10103, doi:10.1029/2004JD005387, 2005.
Hofmann, D.J. and J.A. Pyle, Chapter 12 in WMO Scientific Assessment of Ozone Depletion: 1998, WMO Global Ozone Research and Monitoring Project Report No. 44, Geneva, 1999.
Hofmann, D. J., S. J. Oltmans, W. D. Komhyr, J. M. Harris, J. A. Lathrop, A. O. Langford, T. Deshler, B. J. Johnson, A. Torres, and W. A. Matthews, Ozone loss in the lower stratosphere over the United States in 1992-1993: Evidence for heterogeneous chemistry on the Pinatubo aerosol, Geophys. Res. Lett., 21, 65-68, 1994.
Hood, L. L., J. P. McCormack, and K. Labitzke, An investigation of dynamical contributions to midlatitude ozone trends in winter, J. Geophys. Res., 102, 13,079-13,093, 1997.
Hurwitz, M.M., Stratospheric chemistry-climate interactions: Scientific background and results from two sets of modelling experiments – CPGS Report, University of Cambridge, 2005.
Johnston, H., Reductions of stratospheric ozone by nitrogen oxide catalysts from supersonic transport exhaust, Science, 173, 517-522, 1971.
*Karpetchko A., E. Kyro, and B. Knudsen, Arctic and Antarctic polar vortices 1957-2002 as seen from the ERA-40 reanalyses, J. Geophys. Res., 110, D21109, doi:10.1029/2005JD006113, 2005.
112
Final Report
*Kivi, R., E. Kyrö, T. Turunen, N.R.P. Harris, P. von der Gathen, M. Rex, and S. B. Andersen, Ozonesonde observations in the Arctic during 1989-2003: ozone variability and trends in lower stratosphere and free troposphere, in preparation, 2006.
*Knudsen, B.M., N.R.P. Harris, S.B. Andersen, B. Christiansen, N. Larsen, M. Rex, and B. Naujokat, Extrapolating future ozone losses, Atmos. Chem. Phys. 4, 1849-1856, 2004.
*Krizan, P. and J. Lastovicka, Definition and determination of laminae in ozone profiles, Stud. Geophys. Geod., 48, 2004.
* Krizan, P. and J. Lastovicka, Trends in positive and negative ozone laminae in the Northern Hemisphere, J. Geophys. Res., doi: 10.1029/2004JD005477, 2005
Limpasuvan, V., D. W. J. Thompson, D.L. Hartman, 2004: The life cycle of the northern hemisphere sudden stratospheric warmings, J. Clim., 17, 2584-2596.
*Metelka, L., K. Vanicek and S. Kliegova, Application of neural models for simulation of total ozone in the European region, Publication of the Czech Hydrometeorological Institute, Prague, ISBN: 80-86690-30-X, 2005
Molina, M.J., and Rowland, F.S., Stratospheric sink for chlorofluoromethanes, chlorine atom catalysed destruction of ozone, Nature, 249, 810-812, 1974.
Newchurch, M.J., E.-U. Yang, D.M. Cunnold, G.C. Reinsel, J.M. Zawodny, and J.M. Russell III, Evidence for slowdown in stratospheric ozone loss: First stage of ozone recovery, J. Geophys. Res., 108, 4507, doi 10.1029/2003JD003471, 2003.
Orsolini, Y. D.B. Stephenson, and F.J. Doblas-Reyes, “Storm track signature in total ozone during northern hemisphere winter”, Geophys. Res. Letters, 25, 2413-2416, 1998.
Orsolini, Y. and V. Limpasuvan, "The North Atlantic Oscillation and occurrences of ozone miniholes", Geophys. Res. Letters, 29, 4099-4102, 2001.
*Orsolini, Y. J. and F.J. Doblas-Reyes, “Ozone signatures of climate patterns over the Euro-Atlantic sector in the spring, Quart. J. Royal Meteorol. Soc., 129, 3251-3263, 2003.
*Orsolini, Y. J., Seesaw Ozone Fluctuations between the North Pacific and Atlantic in Late Winter, J. Meteor. Soc. of Japan, 82, vol. 3, 941-948, 2004.
Polvani, L. and D. Waugh, Upward wave activity flux as precursor to extreme stratospheric events and subsequent anomalous surface weather regimes, J. Clim, in press, 2005.
Pyle J.A., Braesicke P., and G. Zeng, Dynamical variability in the modelling of chemistry-climate interactions – Faraday Discussion, 130, DOI: 10.1039/b417947c, 27-39, 2005.
Randel, W. J., F. Wu and R. Stolarski, Changes in column ozone correlated with the stratospheric EP flux, J. Met. Soc. Japan, 80 (4B), 849-862, 2002.
Reinsel, G.C., E.C. Weatherhead, G.C. Tiao, A.J. Miller, R.M. Nagatani, D.J. Wuebbles and L.E. Flynn, On detection of turnaraound and recovery in trend for ozone, J. Geophys. Res., 107, D10, 10.1029/2001JD000500, 2002.
Reinsel, G.C., A.J. Miller, E.C. Weatherhead, L.E. Flynn, R.M. Nagatani, G.C. Tiao and D.J. Wuebbles, Trend analysis of total ozone data for turnaround and dynamical contributions, J. Geophys. Res., 110, doi:10.1029/2004JD004662, 2005.
113
Final Report
*Rex, M., R.J. Salawitch, P. von der Gathen, N.R.P. Harris, M.P. Chipperfield and B. Naujokat, Arctic Ozone Loss and Climate Change, Geophys. Res. Lett., 31, LO4116, doi:10.1029/2003GL018844i, 2004.
Santer, B.D., R. Sausen, T.M.L. Wigley, J.S. Boyle, K. AchutaRao, C. Doutriaux, J.E. Hansen, G.A. Meehl, C. Roeckner, R. Ruedy, G. Schmidt and K.E. Taylor, Behavior of tropopause height and atmospheric temperature in models, reanalyses and observations: Decadal changes, J. Geophys. Res., 108, doi:10:1029/2002JD002258, 2003.
Schnadt, C. and M. Dameris, Relationship between North Atlantic Oscillation changes and stratospheric ozone recovery in the Northern Hemisphere in a chemistry-climate model, Geophys. Res. Lett., 30(9), 1487, doi:10.1029/2003GL017006, 2003.
Staehelin, J., N.R.P. Harris, C. Appenzeller, J. Eberhard: Ozone trends: A review, Rev. Geophys., 39, 231-290, 2001.
Steinbrecht, W., H. Claude, and U. Köhler, Correlations between tropopause height and total ozone: Implications for long-term changes, J. Geophys. Res., 103, 19183-19192, 1998.
Steinbrecht, W., H. Claude, U. Köhler, and Winkler, P., Interannual changes of total ozone and northern hemisphere circulation patterns, Geophys. Res.Lett., 28, 1191-1194, 2001.
Stolarski, R.S. and Cicerone, R.J., Stratospheric chlorine: A possible sink for ozone, Can. J. Chem., 52, 1610-1615, 1974.
Uppala, S., et al., The ERA-40 reanalysis, Q. J. R. Meteorol. Soc., in press., 2005. Vaníček., Dubrovský M. and Staněk M., Evaluation of Dobson and Brewer total ozone
observtaions from Hradec Králové, Czech Republic, 1961-2002 Publication of the Czech Hydrometeorological Institute, Prague, ISBN: 80-86690-10-5, 2003
van Noije, T. P. C., H. J. Eskes, M. van Weele, and P. J. F. van Velthoven, Implications of the enhanced Brewer-Dobson circulation in ERA-40 for the stratosphere-troposphere exchange of ozone in global chemistry-transport models, J. Geophys. Res., 109, D19308, doi:10.1029/2004JD004586, 2004.
*Vogler, C., Brönnimann S. and Hansen, G., Re-evaluation of the 1950-1962 Svalbard total ozone record, submitted to ACPD
Weatherhead, E.C., et al., Detecting the recovery of total column ozone. J. Geophys. Res., 105, 22,201-22,210, 2000.
*Weber et al. dynamic control of NH and SH winter/Spring Total Ozone from GOME Observations in 1995-2002, Geophys. Res. Lett., 30, 1853, doi10.1029/2002GL016799, 2003.
*Weber, M., L. N. Lamsal, M. Coldewey-Egbers, K. Bramstedt and J. P. Burrows, Pole-to-pole validation of GOME WFDOAS total ozone with ground-based data, Atmos. Phys. Chem., 5, 1341-1355, 2005.
*Wohltmann, I. M. Rex, D. Brunner and J. Mäder, Integrated equivalent latitude as a proxy for dynamical changes in ozone column, Geophys. Res. Lett., 32, L09811, doi:10:1029/2005/GL022497, 2005.
World Meteorological Organization, Scientific assessment of ozone depletion 2002, Rep. 47, World Meteorol. Organ., Geneva, Switzerland, 2003.
114
Final Report
*Zanis. P., E. Maillard, J. Staehelin, C., Zerefos, E. Kosimidis, and K. Tourpali, On the turnaround of stratospheric ozone trends deduced from the re-evaluated Umkehr record at Arosa, Switzerland, submitted to J. Geophys. Res., 2006.
115
