1. 1 Cosmic rays and climate Ilya G. Usoskin Sodankylä Geophysical Observatory, University of Oulu, Finland

2. 2 Cosmic Ray Induced Ionization (CRII) CRII is an important factor of the outer space influences on atmospheric properties. CR undergo nuclear interactions with the air  Nucleonic, electromagnetic, muon cascade  ionization of the ambiente air

3. 3 Modelling Direct ionization by primaries: Analytics – “Thin” and “thick” target model. Significant contribution from other sources. Cascade: Monte-Carlo

4. 4 CRII models: basic information Monte-Carlo: Oulu CRAC:CRII (CORSIKA+FLUKA)Bern model ATMOCOSMIC (GEANT-4): Usoskin et al., J. Atm. Solar-Terr. Phys, (2004).Desorgher et al., Int. J. Mod. Phys. A, (2005) Usoskin, Kovaltsov, J. Geophys. Res., (2006, 2010).Scherer et al. Space Sci. Rev. (2006). Physics behind: Monte-Carlo simulation of the cascade, all species and processes included Accuracy: - below 100 g/cm 2 (15 km) -10% Output of the models Cosmic ray induced ionization (CRII) rate, i.e. the number of ion pairs produced in 1 cm 3 (g) per second. Equilibrium concentration depends on the recombination processes and ion mobility and forms an independent task.

5. 5 CRII: details The results of Monte-Carlo simulations of the atmospheric cascade for primary protons with energy 0.2 GeV, 10 GeV and 100 GeV.

6. 6 CRII: ionization function CRII is defined as an integral product of the ionization yield function Y and the energy spectrum of GCR J. The most effective energy of CRII depends on the atmospheric depth – from ≈1 GeV/nuc in the stratosphere to about 10 GeV/nuc at the sea-level.

7. 7 CRII: altitude vs. latitude

8. 8 Comparison with measurements

9. 9 Spatial distribution of CRII (cm -3 sec -1 ) Ground level12 km altitude Solar maximum Solar minimum

10. 10 CRII: last decades CRII since 1951 at the atmospheric depth x=700 g/cm 2 (about 3 km altitude) for polar regions and to the equator.

11. 11 Long-term CRII Computed CRII at x=700 g/cm 2 (about 3 km altitude) in a sub-polar region, based on cosmic ray flux reconstruction [Usoskin et al., 2002]. This is consistent with a ~0.15 %yr -1 decerase of the air conductivity (measured in Europe) between 1910’s and 1950’s [Harrison & Bennett, 2007]

12. 12 CR time profile for January 2005 Time profile of Oulu NM hourly count rate for January 2005 (http://cosmicrays.oulu.fi).

13. 13 Ionization effect of GLE 20-01-2005 (A)Differential fluence of solar protons from the 20-01-2005 event and the daily fluence of GCR protons for the day of 20-01-2005, including the effect of the Forbush decrease (dotted line). The dashed curve depicts the average GCR proton fluence for January 2005. (B) The vertical profile of the daily averaged CRII in the polar region from GCR (dotted curve) and SEP (solid curve) separately, for the day of 20 January 2005.

14. 14 Ionization effect of GLE The relative ionization effect of GLE 20-Jan-2005 as function of the geomagnetic cutoff rigidity Pc and atmospheric depth h.

15. 15 How often do SEP events occur? The probability of a SEP event to occur (McCracken et al., JGR, 2001)

16. 16 Climate forcing Climate Internal forcing (volcanos, oceans) Orbital forcing Direct solar forcing Indirect Solar forcing ? Anthropogenic!!!

17. 17 SA/CR  climate Sunspot number reconstruction (Usoskin et al., 2007) along with the climate shifts in Europe to cold/wet conditions (Versteegh, 2005) for the last 6500 years: 14 cold spells – vs – 15 Grand minima  12 coincide.

18. 18 Earth’s radiative equilibrium absorbed solar radiation = emitted infrared radiation: Black-body temperature 255 K Real temperature 288 K

19. 19 CR vs. climate This affects the amount of absorbed radiation, even without invoking notable changes in the solar irradiance. Clouds play an important role in the radiation budget of the atmosphere by both trapping outgoing long wave radiation and reflecting incoming solar radiation. This affects the amount of absorbed radiation, even without invoking notable changes in the solar irradiance. Two possible mechanisms have been proposed: Ions + molecules  complex cluster ions (aerosols) grow by ion-ion recombination or ion-aerosol attachment  cloud condensation nuclei (Svensmark & Co; Yu 2002) The atmosphere is not a Wilson chamber! SA GCR + SA  global electrical circuit ( vertical currents)  Ice in super-cooled water  enhanced precipitation (Tinsley & Co.) Hard to distinguish from other SA-driven effects.

20. 20 CLOUD+SKY experiment Duplissi et al. (2010), Pedersen et al. (2011) A 10-fold increase of ionization (typical changes 25%)  3-fold formation rate, BUT… A 10-fold increase in SO2 (typical)  1000-fold formation rate change. Temperature, humidity also affect…

21. 21 Marsh and Svensmark, JGR, 2003 Solar cycle This result has been disputed.

22. 22 Inter-annual scale CR – cloud relation is not homogenous but depicts a clear geographical pattern (Marsh & Svensmark, 2003; Palle et al., 2004; Usoskin et al., 2004; Voiculescu et al., 2006), including three major regions: NE Atlantics + Mediterranean (Europe); S Atlantic + W Indian; NW Pacific; and almost no relation in other parts of the world. Correlation significance map – Palle et al. (2004) Correlation map – Usoskin et al. (2004), Voiculecu et al. (2006)

23. 23 Atmospheric responses to daily CR variations Pudovkin & Veretenenko, JASTP, 1995: Change of the mean cloud cover (ground-based observations) after FD at high latitudes (> 60 o N) Roldugin & Tinsley, JASTP, 2004: changes in atmospheric transparency associated with FD at high latitudes (> 55 o N) Kniveton, JASTP, 2004: Todd & Kniveton, 2004; Zonal mean total cloud anomalies associated with FD – polar and equatorial regions. Stozhkov et al. ( Geom.Aer., 1996; J.Phys.D, 2003): Precipitation changes related to FD/SPE.

24. 24 BUT... Calogovic et al. (2010) and Kristjansson et al. (2008) : No notable effect of Forbush decreases in the cloud cover globally or in some regions. (Kristjansson et al.  if the effect exist, then only in South Atlantic) Statistical results are inconclusive.

25. 25 Direct evidence: 20-01-2005 Mironova et al. (GRL, 2008) Decrease of the Aerosol index at Antarctic stations on 2-nd day after the event – columnar density. Mironova et al. (ACPD, 2011) Formation of CCN-size aerosols in the same region – altitude range 15-20 km. Severe SEP event  barely noticable effect. Serious limitation on the direct CR effect.

26. 26 Conclusions Cosmic rays form the main source of atmospheric ionization and related physical-chemical changes in the low-mid atmosphere. This is well understood and modelled. The direct effect of CR on clouds is unclear, most likely it is small (experiments + a case study of a severe SPE). Indirect effect (top-down dynamic strato-troposphere coupling) – hard to distinguish from SI effects. CR may play a role on the long-term scale, but difficult to distinguish from solar irradiance variations. It is not CR per se but its variability, that may affect climate.

27. 27 FAQ YES  Do CR affect climate  probably YES, but we don’t know how... YES  Would we see clouds if there were no cosmic rays?  YES YES  Would climate be somewhat different if there were no CR?  Probably YES  Are we sure in a strong solar/CR-climate relation?  NO NO, but a bulk of indirect evidences + no solid contra-proof  Do we know exact mechanism of such a relation?  NO NO, but some qualitative ideas and indirect experiments  Is the present GW a result of the solar activity?  YES NO partly YES (before 1950-1970) and probably NO (since 1970)

28. 28