Effects of Galactic Cosmic Rays on the Atmosphere and Climate. Jón Egill Kristjánsson, Univ. Oslo

Transcription

1 Effects of Galactic Cosmic Rays on the Atmosphere and Climate Jón Egill Kristjánsson, Univ. Oslo

2 Overview of talk Hypotheses for Coupling between Galactic Cosmic Rays and Climate Observational studies Recent Results from Lab Studies (CERN) Recent Results from Modeling Studies Physical Mechanisms Unexplored issues

3 Galactic Cosmic Rays and the Atmosphere GCR are the dominant source of penetrating ionizing particle radiation GCR provide the sole source of ions away from terrestrial sources of radioisotopes GCR directly influence the global atmospheric electrical circuit Carslaw et al. (2002: Science)

4 GCR Ion Production vs. Height Solar Max. Solar Min. Max. at km (upper troposphere / lower stratosphere) Neher (1971: JGR)

5 Clouds and Climate Thin, high ice clouds: Warming Dense, low liquid clouds: Cooling Trapping of LW > Reflection of SW Trapping of LW << Reflection of SW

6 Observational Studies

7 Correlations between Galactic Cosmic Rays and Low Clouds (ISCCP IR data) Marsh and Svensmark (2000: Phys.Rev.Lett.)

8 GCR Low Clouds Climate Galactic Cosmic Ray Flux is Declining (3.5% during ) Fewer Low Clouds (6.1% during ) Marsh and Svensmark (2000: Space Sci.Rev.) Less Reflected Solar Radiation (assumed) Warmer Climate

9 Cosmic Ray Flux Kilde:

10 Low Cloud Amount (MODIS) Laken et al. (2012: J. Climate)

11 GCR Low Clouds Climate No trend in Galactic Cosmic Ray Flux (during ) More Low Clouds (MODIS: 0.9% during ) / Fewer Low Clouds (ISCCP) No Trend in Reflected Solar Radiation ( ) Warmer Climate

12 Forbush Decrease Events ~ Same amplitude as Solar cycle variation (Min minus Max)

13 5 Strongest FD events of Kristjánsson et al. (2008: Atmos.Chem.Phys.)

14 5 Strongest FD events of Globally averaged data: Signals in particle size, liquid cloud cover and cloud liquid water content 5-10 days after FD onset Svensmark et al. (2009: Geophys.Res.Lett.)

15 5 strongest FD events of : TOA Fluxes from CERES a) b) In addition to the GCR signals, non-cosmic ray signals of similar amplitude as the cosmic ray signals appear

16 Physical Mechanisms

17 Near-cloud mechanism: GCR ionization creates electrical charges on clouds Carslaw et al. (2002: Science)

18 Power Spectra: High-Pass Filtered GCR vs Cloud Base Height (Stratus) : 1.68 yr GCR signal : GCR ~ clouds GCR GCR Stratus Stratus Harrison et al. (2011: PRSLA)

19 Clear-air mechanism: GCR ionization aids particle formation Carslaw et al. (2002: Science)

20 A typical particle formation event 5 July 2006 at Hyytiälä, Finland Kulmala et al. (2010: ACP)

21 Mechanisms for aerosol nucleation Measurement sites Ion contrib. vs. Nucleation rate Kulmala et al. (2010: ACP) Ion contribution large only when nucleation rates are small => Overall ion contribution < 10%

22 Kirkby et al. (2011: Nature)

23 NEUTRAL GCR PBL GCR UPPER TROP. Kirkby et al. (2011: Nature)

24 Kirkby et al. (2011: Nature)

25 NH 3 = 150 pptv Dots: PBL measurements Lines: Lab measurements at room temp. 278 K NH 3 < 50 pptv 248 K NH 3 = 190 pptv Kirkby et al. (2011: Nature) - Ion-induced binary nucleation at a significant rate in the free troposphere at atmospheric [H 2 SO 4 ] 292 K - Binary nucleation within the PBL is negligible - In the PBL, ion-induced nucleation can not explain observed nucleation rates NH 3 < 35 pptv

26 Recent modeling studies

27 Solar Min minus Solar Max TSI signal by itself: W m -2 Negligible Impact on Top of Atmosphere Short Wave Radiation! Global Aerosol-Climate Model: ECHAM5-HAM Ionization by GCR based on the analytical theory of O Brien [2005] The GCR produced ions drive aerosol formation from the gas phase via charged nucleation of sulfuric acid (H 2 SO 4 ) and water (H 2 O) in the model Kazil et al. (2012: GRL)

28 Forbush Decrease events Snow-Kropla et al. (2011: ACP) Global CTM GEOS-Chem model extended with the TOMAS aerosol microphysics model 40 size bins from 1 nm to 10 µm diameter Sulfate, sea-salt, organic carbon, black carbon, mineral dust Ion-mediated nucleation (IMN); Yu (2010: JGR)

29 The fate of the nucleated aerosols Growth Nucleated aerosols (~2 nm) may grow by condensation and coagulation to CCN size (~100 nm) Limitations Re-evaporation Competition for condensable vapors => Slower growth Scavenging by coagulation

30 Summary Do GCR influence Earth s climate via clouds? Mechanisms involving charges on clouds some support from observations; poorly understood Mechanisms involving particle formation; some support from observational and modeling studies Aerosol nucleation important for CCN GCR stimulates aerosol nucleation But, globally, GCR variations do not seem to have a large influence on [CCN] clouds climate Growth of aerosols to CCN size still highly uncertain Other mechanisms?? E.g., involving ice clouds??

31 Thank you! Photo: Michael Gauss

32 Cosmic Rays Aerosol nucleation Sulfate aerosols, growing in size by coagulation, spreading horizontally H 2 O, SO 2 TROPICS SUBTROPICS STORM TRACKS

33 Solar Min minus Solar Max: Simulated Changes in Aerosol Number Conc. Solar signal weak in all cases! Snow-Kropla et al. (2011: ACP)

34 Simulated Forbush Decrease Events Cosmic Ray signals with a few days delay Quite strong signals also outside FD time window Snow-Kropla et al. (2011: ACP)

35 Aerosol nucleation vs [H 2 SO 4 ] 248 K 292 K 278 K GCR ionization enhances nucleation rate twofold at 292 K and more than tenfold at 278 K and 248 K Binary nucleation extremely slow in the PBL, compared to observations Unrealistically high [H 2 SO 4 ] applied Kirkby et al. (2011: Nature)

36 6 Largest FD events of Ion production rate calculated as a function of latitude, longitude and height 3-hourly ISCCP D1 cloud data Gridded data at 5 x 5 resolution No relation found between GCR signal and cloud amount Calogovic et al. (2010: Geophys.Res.Lett.)

37 Sensitivity Experiments Results robust to uncertainties in assumptions concerning background aerosols Cosmic Ray signal real, but small Snow-Kropla et al. (2011: ACP)

38 Simulated GCR Ionization Rate Kazil et al. (2012: GRL)

39 Snow-Kropla et al. (2011: ACP)

40 What are galactic cosmic rays? Galactic cosmic rays are high-energy charged particles that enter the solar system from far away in the galaxy. They are composed of protons, electrons, and fully ionized nuclei of light elements. Source: Wikipedia

41 Signals in Aerosols and Clouds Kazil et al. (2012: GRL)

42 Reflected SW radiation at TOA: Sensitivity to selected number of FD events a) b) c) As more and weaker events are added, signals become weaker, less significant

43 TOA Radiation (CERES data) vs. Cosmic Rays a) b) Signal a few days after strong FD events in mainly SW radiation, followed by an opposite signal a few days later in mainly LW radiation

44 Focus on areas of high cloud susceptibility No signals in cloud droplet size, cloud cover or cloud water path associated with FD events Kristjánsson et al. (2008: ACP)

45 Findings from CERN so far For typical boundary-layer NH 3 mixing ratios (< 1 pptv), ternary nucleation NH 3 -H 2 SO 4 -H 2 O, with or without ions, is unable to explain atmospheric observations The fraction of the freshly nucleated particles that grow to sufficient sizes to seed cloud droplets remains an open question experimentally The same goes for the role of organic vapours in the nucleation Kirkby et al. (2011: Nature)

46 Exploring Model Uncertainties Snow-Kropla et al. (2011: ACP)

47 Modeling of aerosol nucleation CCN We estimate that 45% of global low-level cloud CCN at 0.2% supersaturation are secondary aerosol derived from nucleation (ranging between 31 49% taking into account uncertainties in primary emissions and nucleation rates), with the remainder from primary emissions. The model suggests that 35% of CCN (0.2%) in global lowlevel clouds were created in the free and upper troposphere. In the marine boundary layer 55% of CCN (0.2%) are from nucleation, with 45% entrained from the free troposphere and 10% nucleated directly in the boundary layer. Merikanto et al. (2009: Atm.Chem.Phys.)

48 Particle Formation over Europe Monthly data Annual data No link found between: - variations in aerosol nucleation events - variations in GCR flux Kulmala et al. (2010: ACP)

49 Data Sources MODIS = MODerate resolution Imaging Spectro-radiometer Measures cloud and aerosol properties using 36 bands in VIS, NIR and IR Time resolution: Daily Spatial resolution: 1 km 1 degree CERES = Clouds and the Earth s Radiant Energy System Measures broadband SW and LW energy fluxes from µm Time resolution: Hourly Every 3 hours Daily Spatial resolution: 20 km 1 degree

50 AERONET data AERONET = AErosol RObotic NETwork Surface-based measurements using sun photometer instruments Measure aerosol properties Ångström Exponent is a Measure of Particle Size: 440 / AE log 440 / 340 log 340 Purely Gaseous Atmosphere: AE ~ 4 Small Particles: AE ~ 1 Large Particles: AE ~ 0

51 The GCR clear-air mechanism Carslaw et al. (2002: Science) With increasing nucleation rates, there is increased competition between the new particles for condensable material, which slows the growth rates of these new particles. Slower growth rates increase the probability of scavenging by coagulation Therefore, we would expect a smaller fractional change in CCN than the fractional change in the cosmic-ray flux Away from sources of precursor gases it may take on the order of a week or more for nucleated aerosols to grow to CCN sizes Snow-Kropla et al. (2011: ACP)

52 Aerosol Sources and Sinks Accumulation mode: Most CCN Seinfeld and Pandis (1998)

53 Aerosol Nucleation due to Cosmic Rays Simulated Ion- Mediated Nucleation Maxima: Upper tropical troposphere Lower mid-latitude troposphere Yu et al. (2010: JGR)

54 Global Cloud Cover ITCZ: Deep Convection => High Clouds (ice) Mid-latitude Storm tracks => Mid-level Clouds (mixedphase) Subtropical subsidence inversion => Low Clouds (liquid)

55 CCN at SolarMin SolarMax Standard assumption on the influence of ionization Extreme assumption on the influence of ionization Pierce & Adams (2009: Geophys.Res.Lett.)

56 The simulated signal in Cloud Condensation Nuclei (CCN) from changes in cosmic rays over a solar cycle is very weak The resulting radiative forcing is only ~0.01 W m -2 Pierce & Adams (2009: Geophys.Res.Lett.)

57 GLOMAP results vs. obs. Spracklen et al. (2010: Atmos.Chem.Phys.)

58 Validation of IMN nucleation scheme Observed IMN scheme Yu et al. (2010: JGR) Good agreement between model results and aircraft campaign observations

59 Validation of IMN nucleation scheme Yu et al. (2010: JGR) Snow-Kropla et al. (2011: ACP)

60 Solar Min minus Solar Max: Simulated Changes in Aerosol Properties Solar signal weak in all cases! Snow-Kropla et al. (2011: ACP)

61 Ionization Chamber as a proxy for the Earth s atmosphere Kirkby et al. (2011: Nature)

62 Pathways that need to be explored Near-Cloud Mechanism (G. Harrison) largely unexplored Favorable Conditions - Pristine Areas? Mechanisms involving Ice Clouds? Relevance for Paleoclimate?

63 Kirkby et al. (2011: Nature) 278 K [H 2 SO 4 ] = cm K [H 2 SO 4 ] = cm -3 - Nucleation rate roughly proportional to negative ion concentration -Binary nucleation extremely slow in the PBL, even with ion enhancement

64 248 K NH 3 < 35 pptv 278 K NH 3 < 35 pptv 292 K NH 3 < 35 pptv 292 K NH 3 = 230 pptv - A clear progression from almost binary nucleation at 248 K to pure ternary nucleation at 292 K; both contributing at 278 K - At 292 K clusters grew by a striking stepwise accretion of NH 3 molecules, each stabilizing a distinct additional number of acid molecules Kirkby et al. (2011: Nature)

65 292 K [H 2 SO 4 ] = cm -3 [H 2 SO 4 ] = cm K - Nucleation rate highly sensitive to small additions of ammonia up to ~100 pptv - Saturation at higher NH 3 mix. ratios [H 2 SO 4 ] = cm -3 Kirkby et al. (2011: Nature)

66 Approach Study recent Forbush Decrease events (time scale ~2 weeks) Search for signals of cosmic rays / solar activity in observational data (CERES) of TOA radiative fluxes Compare to signals of cloud and aerosol properties from MODIS data CERES and MODIS are instruments on the NASA Aqua and Terra satellites we use Terra data at 1 x 1 resolution

67 CERES data (annual average) Reflected SW Emitted LW SWtotal 200 LWtotal Signals dominated by latitude and clouds 0 Low clouds: Large SW effect High clouds: Large LW effect

68 Summary We have investigated possible links between cosmic rays, clouds and climate in Forbush Decrease events of ~2 week duration Signals for strong events only (< 1 such event per year) At +5-7 days, a signal is found in Ångström exponent at the surface, TOA SW fluxes, LWP and mid-level cloud amount At days an opposite signal is found in TOA LW fluxes, IWP and high clouds not understood! Applying a longer time window, signals of similar amplitude, unrelated to cosmic ray events, are found Our results indicate that the GCR cloud link is real, but weak Consistent with recent model studies

69 Cloud Properties vs. Cosmic Rays a) b) c) d) Signals a few days after strong FD events in several cloud parameters

70 Cloud Amount (blue) vs. TOA Radiation (red, green) Liquid Clouds SW signal PRECEDES cloud signal by 2 days Ice Clouds LW signal correlates well with cloud signal (negative correlation) Mid-Level Clouds SW signal correlates well with cloud signal Signal at +5 days correlates well with mid-level clouds Signal at +12 days correlates well with ice clouds

71 Cloud Properties vs. TOA Radiation a) b) Cloud Optical Depth ~ SW Cloud Droplet Size does not correlate with TOA SW or LW radiation! c) d) Liquid Water Path ~ SW Ice Water Path ~ LW, SW

72 Spatial Signatures - Chaotic The mean of (days 5 to 8) minus the mean of (days -15 to -1 plus days 10 to 20) a) b) % % c) d) W m -2 W m -2

73 Ångström Exponent

74 Why 5 episodes? How about 4, 6 or 7? Does not seem to make much difference

75 Removing the Jan05 event CALIQ, mean of strong FD jan05 removed all 5 FD SWtotal, mean of strong FD jan05 removed all 5 FD Days (0 at FD minimum) Days (0 at FD minimum)

76 Statistical procedure Arbitrary 5 * 36 (or 130) day periods, averaged Repeated 15 times Calculate 2 standard deviations around the mean

77 Residual from vertical partitioning of the cloud cover IR only IR LOW, IR MID, IR HIGH do not add up to ISCCP total cloud cover! IR+VIS+NIR

78 Simulated Cosmic Ray Ion-Pair Formation Rate Snow-Kropla et al. (2011: ACP)

79 Simulated Rate of Aerosol Nucleation Negative values in some areas due to more competition and larger coagulation sink Snow-Kropla et al. (2011: ACP)

80 ISCCP Low Clouds MODIS Liquid r = The high correlation disappears after 1998 Gray et al. (2010: Rev. Geophys.)