1. Title Real time GLE modeling and Relativistic Solar Cosmic Ray parameters deriving E.V. Vashenuyk, Yu.V. Balabin, B.B. Gvozdevsky Polar Geophysical Institute Apatity, Russia NMDB meeting Kiel, Germany, 3-8 December, 2008

2. Real time operating GLE alerting systems: 1. T Kuvabara, J.W.Bieber, P.Evenson, R.Pyle, et al., Real time cosmic ray monitoring system for space weather. Space weather, V.4, 2006, H.Mavromichalaki, G.Souvatzog,lou, C.Sarlanis, G.Mariatos, M. Gerontidou, A.Papaioannu, Plainaki, S.Tatsis, A.Belov, E.Eroshenko, and V.Yanke. The new Athens center on data processing from the neutron monitor network in real time. C.PAnn. Geophys. 23, 3103, 2005 Langer R,May be there are some others

3. The GLE onset detected.. Next we want to know the solar cosmic ray characteristics, to estimate the hazardous consequents. But deriving SCR parameters from the neutron network data is not an easy task. The problem becomes more complicated if even more it to solve for the limited time and at the limited number of stations. As we face in a case of the NMDB

4. OUTLINE Neutron monitor network as an instrument for relativistic solar cosmic ray studies and GLE modeling technique Results of relativistic solar cosmic ray events study with the GLE modeling Real time GLE modeling in frames of NMDB project. Approaches to the problem

5. Variations of cosmic ray intensity as recorded by the neutron monitors with solar activity. By black triangles are shown GLE occurrences.

6. PGI NMs Two neutron monitor stations of the Polar Geophysical Institute Apatity (67.55N 33.34E) Barentsburg (78.08N 14.12E) 4 INTERNET http://pgia.ru/cosmicray SERVER Neutron monitors computer and electronics racks

7. Effect of magnetosphere on cosmic rays 5 Asymptotic Cone of Acceptance is formed by trajectories of particles contributing into response of NM Barentsburg is a high-latitude station and accepts radiation from high latitudes of the selestial sphere and a subpolar station Apatity from equatorial latitudes

8. Anisotropy Solar cosmic rays anisotropy effect during the GLE on December 13, 2006 Apatity (10 s data) Barentsburg (1 min) 6

9. Set of NMs The worldwide network of neutron monitors as a multidirectional cosmic ray spectrometer 7

10. GLE modeling technique of deriving the characteristics of relativistic solar protons (RSP) from the neutron monitor network data consists of a few steps: 1. Definition of asymptotic viewing cones (taking into account not only vertical but also oblique incident on a detector particles) by the particle trajectory computations in a model magnetosphere (Tsyganenko 2002) 2. Calculation of the NM responses at variable primary solar proton flux parameters. 3. Application of a least square procedure for determining primary solar proton parameters (namely, energy spectrum, anisotropy axis direction, pitch-angle distribution) outside the magnetosphere by comparison of computed ground based detector responses with observations

11. 9 Method: 8 direct. Scheme of asymptotic cones calculations: ~20° Starting directions at a launching point Asymptotic directions at magnetopause To account the contribution of oblique incident particles we calculate 8 trajectories of particles launched at zenith angle 20 o and 8 azimuths Calculated asymptotic directions are then used in the following modeling of a NM response SCR GCR

12. Formula 8 The response function of a i-th neutron monitor to anisotropic flux of solar protons. (dN/N) i is percentage increase effect at a given neutron monitor i J(R) = JoR -  * is rigidity spectrum of RSP flux with changing slope  * =  +  ·(R-1) where  is increase per 1 GV (Cramp et al., 1997) S(R) is specific yield function (Debrunner et al., 1984), θ(R) is pitch angle (angle between the anisotropy axis given by  ;  parameters) F(θ(R )) ~ exp(-θ 2 /C) is pitch-angle distribution in a form of Gaussian (Shea&Smart, 1982)

13. Thus, 6 parameters of anisotropic solar proton flux outside magnetosphere  ; , Jo, , , C are to be determined by a solving of the nonlinear least square problem by comparison of computed responses with observations As example of such study we consider the last GLE on December 13, 2006.

14. GLE 70 9 Increase profiles at some NM stations: Oulu, Apatity, Moscow, Barentsburg, Fort Smith The asymptotic cones (1-20 GV), for the above NM stations and Th- Thule, McM-McMurdo, SA-SANAE, Ma-Mawson, No-Norilsk, Ti-Tixie, CS- Cape Shmidt, In-Inuvik, Pe-Pewanuk. The derived anisotropy axis and pitch angle grid lines for solar proton flux at 03.00 UT are shown. The cross is the IMF direction (ACE data). GLE 70 13.12.2006

15. Fitting Observed and modeled responses at a number neutron monitor stations at ─── increase profiles at neutron monitors ●●● modeling responses

16. 1 2 3 4 5 6 Dynamics of pitch angle distributions (PAD) derived from neutron monitors data 1 3 5 to Sun Numbers mark the moments of time PAD demonstrates an initial highly collimated beam of particles (prompt component) followed by a delayed quasi-isotropic population (delayed component)

17. Peak Peak caused by the prompt component of relativistic solar protons (Buetickofer &Flueckiger, 2006) Neutron monitors Hodoscope MEPHI observation of collimated particle flux: D. A. Timashkov, Yu.V. Balabin, V. V. Borog, K. G. Kompaniets, A. A. Petrukhin, E.V.Vashenyuk et al, 30icrc, paper 0298 Calculated magnetopause projection of the beam, registered by the Hodoscope MEPHI (blue spot). Asymptotic cones of NM stations (on the left) are shown.

18. Dynamics of energetic spectra of relativistic solar protons GLE 70

19. GLE 20.01.2005 Increase profiles as registered by a number of NM stations and EAS array “Carpet”(Baksan, North Caucasus) The spectrum derived in moment (1 ) when the prompt component was dominated is exponential in energy: J= 1.5  10 5 exp(-E/0.92), and spectrum of delayed component (2) has a power-law form: J = 7.5  10 4 E -4.9. (Vashenyuk et al.2006, 2007, Perez-Peraza et al., 2008)

20. Increase profiles at the McMurdo and Mawson neutron monitors (a), rigidity spectra derived at the moments 07:00 (1) and 08:00 (2) UT (b), SYF and spectra 1 and 2 (c); differential responses (d) of the McMurdo neutron monitor to the exponential spectrum at the moment 1 (blue shading) and to the power-law spectrum at the moment 2 (red shading). SYF- specific yield function Debrunner et al., 1984 Exponential spectrum of the prompt component was a cause of a giant increase effect at McMurdo neutron monitor and power law spectrum of delayed component produced rather moderate effect at this and other NM stations during the GLE 20.01.2005 a b c d

21. (a) Increase profiles at the Leeds and Ottawa neutron monitors; (b) energy spectra derived at the moments 04:00 (1) and 06:00 UT (2), ( c) SYF and spectra (1 and 2) and differential responses of the Leeds neutron monitor to the exponential spectrum (1,blue) and to the power- law spectrum (2,red). Two relativistic solar proton components in the GLE 23 February, 1956 By numbers are marked, respectively, the moments when the prompt component (1) or delayed one (2) were dominating. One can see comparable responses of both neutron monitors to the power-law spectrum at moment (2).

22. Results of modeling analysis of 22 major GLEs showing existence of two RSP components Spectrum of prompt component: J=J 0 exp(E/E 0 ), E (GeV); J 0, J 1 (m 2 s st GeV) -1 Spectrum of delayed component J=J 1 E - γ E.V. Vashenyuk, Yu.V. Balabin, L.I. Miroshnichenko J. Perez-Peraza, A. Gallegos-Cruz3, 30 icrc, Merida, Mx, paper 0588

23. Prompt and delayed components of relativistic solar protons (RSP) The modeling analysis of 22 large GLEs occurred in the period 1956-2006 on the data of the worldwide neutron monitors carried out by us revealed two distinct RSP populations (components): 1.Prompt Component (PC): the early collimated impulse-like intensity increase with exponential energy spectrum, 2.Delayed component (DC): the late quasi-isotropic gradual increase with a softer energy spectrum of the power law form. 3.The exponential spectrum may be an evidence of the acceleration by electric fields arising in the reconnecting current sheets in the corona. The possible source of delayed component particles can be stochastic acceleration at the MHD turbulence in expanding flare plasma. E.V. Vashenyuk, Yu.V. Balabin, L.I. Miroshnichenko J. Perez-Peraza, A. Gallegos- Cruz,J ASR, V.38 (3), 411; (2006); 30 icrc, Merida, Mexico, paper 0658 (2007)

24. GLE No Spectra of prompt and delayed solar proton components derived from neutron monitor data for a number of GLEs Prompt component Delayed component GLE No Points are direct solar proton data from spacecrafts and balloons Spectra of the prompt component as a rule have exponential dependence upon energy Spectra of the delayed component have close to the power law dependence upon energy

25. GLE No Spectra of prompt and delayed solar proton components derived from neutron monitor data for a number of GLEs Prompt component Delayed component GLE No Points are direct solar proton data from spacecrafts and balloons Spectra of the prompt component as a rule have exponential dependence upon energy Spectra of the delayed component have close to the power law dependence upon energy PC 20.01.05

26. Asymptotic cones calculated for particles launched under inclined directions. Focusing effect is seen for rigidities from 1 to 5-10 GV Focusing effect of geomagnetic field

27. Stations NStationLatLonRCRC 1 APATITY67.5533.330.65 2 ALMAATA-B43.1476.606.61 3 ATHENS37.5823.47 4 BERN46.957.45 5 BAKSAN43.2842.696.40 6 BARENTSBURG78.1214.120.00 7 CALGARY51.08-114.131.09 8 CAPE SCHMIDT68.92-179.470.55 9 CLIMAX39.37-106.18 10 EREVAN40.5044.17 11 FORT SMITH60.00-112.00 12 HERMANUS-34.4219.224.58 13 INUVIK68.35-133.720.16 14 IRKUTSK52.10104.003.56 15 JUNGFRAUJOCH46.5507.984.53 16 KERGUELEN-49.3570.251.10 17 KINGSTONE-42.99147.291.88 18 KIEL54.3010.102.32 19 LARC-62.20-58.96 20 LOMNICKY STIT49.1120.133.84 21 McMURDO-77.51166.430.00 22 MAGADAN60.10151.002.11 23 MOSCOW55.4737.322.39 24 MAWSON-67.6062.680.22 25 MEXICO19.33-99.18 26 NAIN56.60-61.70 27 NORILSK69.2688.050.57 28 NOVOSIBIRSK54.8083.002.78 29 NEWARK39.70-75.702.02 30 OLCH-33.49-70.72 31 OULU65.0225.500.77 32 POTCHEFSTROOM-26.6827.927.00 33 PEAWANUCK55.00-85.00 34 ROME41.9012.526.24 35 SANAE-71.67-02.850.86 36 TERRE ADELIE-66.65140.000.00 37 THULE76.50-68.700.00 38 TSUMEB-19.2017.609.21 39 TIXIE BAY71.60128.900.45 40 YAKUTSK62.02129.721.63 41 YANGBAJING30.1190.5314.10

28. Cut model, NMDB stations only 1.APTY 2.ATBA 3.ATHN 4.BKSN 5.BRBG 6.CAPS 7.IRKT 8.JUNG 9.KERG 10.KIEL 11.LARC 12.LMKS 13.MGDN 14.MOSC 15.NRLK 16.NVBK 17.OULU 18.TXBY 19.YKTK

29. Spectra derived at 3.55, 400 and 4.20 UT. Dashed curves: full model, solid curves- the cut one

31. Time UT Model 1 (full) ------------ 2(cut) gamma Delta gamma C I0I0 Residual error % 03:00 1-4.70.210.241.2 10 5 2.5 2 (cut)-4.30.25 1.6 10 5 3.9 03:55 1-6.7017.26.1 10 5 1.5 2(cut)-6.30.079.13.6 10 5 4.7 04:00 1-6.9018.412.5 10 5 1.7 2(cut)-6.708.24.0 10 5 5.3 04:05 1-7.0021.811.0 10 5 1.9 2(cut)-6.7011.13.8 10 5 3.8 05:00 1-7.6023.75.4 10 5 3.9 2(cut)-6.708.02.1 10 5 4.5 Results of GLE 70 modeling by the full and cut models