1. IASF Bologna 1 Terrestrial Gamma-Ray Flashes Martino Marisaldi (INAF-IASF Bologna) 2 nd International Conference Frontiers in Diagnostic Technologies Laboratori Nazionali di Frascati, 28-30 November 2011 M. Marisaldi FDT2 Frascati 29/11/2011

2. IASF Bologna M. Marisaldi 2 Outline Observational evidence Theoretical models Open questions Latest results and AGILE contributions Credit: Alan Stonebraker M. Marisaldi FDT2 Frascati 29/11/2011

3. IASF Bologna Serendipity at play M. Marisaldi FDT2 Frascati 29/11/2011 Vela satellites '70-'80 looking down to Earth... GRB BATSE onboard CGRO 1991 – 2000 looking up to space... TGF

4. IASF Bologna 1992: BATSE discovery of Terrestrial Gamma-ray Flashes(TGFs) M. Marisaldi FDT2 Frascati 29/11/2011 ~ 70 TGF detected on 9 life-Years typically 100 counts/TGF Fishman et al., Science (1994)Inan et al., GRL (1996) Energy > 1MeV, harder than GRBs very bright, ~1ms duration associated to lightning

5. IASF Bologna 2005: RHESSI observation of TGFs up to 20 MeV M. Marisaldi FDT2 Frascati 29/11/2011 Contiuous time-tagged event list No on-board trigger logic 10– 20 TGF per month Typically 20-30 counts/TGF ~800 TGFs reported in the 1 st RHESSI TGF catalog (Grefenstette et al., JGR 2009) clear association to global lightning activity One-to-one sferics correlations Smith et al., Science (2005)

6. IASF Bologna Progresses in TGF science M. Marisaldi FDT2 Frascati 29/11/2011 TGF discovery by BATSE RHESSI AGILE Association to lightning Cumulative spectrum Energy up to 20 MeV production altitude < 20km Fermi + ground lightning location networks Energy > 40 MeV up to ~ 100 MeV First localization in  -rays from space TGF & global lightning activity Duration ~250  s Discovery of e + /e - flashes

7. IASF Bologna Physical scenario: runaway electrons M. Marisaldi FDT2 Frascati 29/11/2011 Cold runaway: any electron goes relativistic Conventional breakdown: ionization > attachment Relativistic runaway regime: seed electrons get accelerated to relativistic energies Figure by V. Pasko, from “Sprites, Elves and intense lightning discharges”, ed. M. Fullekrug, Springer (2006)

8. IASF Bologna Physical scenario: relativistic feedback M. Marisaldi FDT2 Frascati 29/11/2011 Positron feedback (Dwyer 2008) limits the number of avalanche length before electric field quenching But feedback allows true discharge even with only a few avalanche lengths Feedback predicts correct TGF duration and approximate luminosity (courtesy D. Smith, plot by J. Dwyer)

9. IASF Bologna Relativistic Runaway Electron Avalanche (RREA) M. Marisaldi FDT2 Frascati 29/11/2011 Relativistic Runaway Electron Avalanche (RREA) with relativistic feedback (Dwyer 2008) Bremsstrahlung + Compton scattering RHESSI cumulative spectrum is compatible with a production altitude of 15-21 km (just above tropical thunderstorms) Still hint for individual spectral variability: differences in production altitudes or viewing angle? BATSE events seem produced at higher altitude, but discrepancy is reduced if dead- time effects are properly accounted for (Grefenstette et al., 2008; Ostgaard et al., 2008; Gjesteland et al., 2010) Dwyer and Smith, GRL (2005) Carlson, Lehtinen and Inan (2007)

10. IASF Bologna Open questions What is the TGF maximum energy? What is the TGF angular distribution? Which lightning type? Which phase of the lightning process? Where does the acceleration takes place? How common are TGFs? What is their intensity distribution? What is the climatology of TGFs? M. Marisaldi FDT2 Frascati 29/11/2011

11. IASF Bologna Operating TGF detectors RHESSI Fermi-GBM 12xNaI Fermi-GBM NaI GBM 2xBGO GBM BGO AGILE-MCAL AGILE GRID Data from: Smith et al. (2002), Meegan et al. (2009), Labanti et al. (2009), Tavani et al. (2009) Effective Area vs. Energy LAT ? M. Marisaldi FDT2 Frascati 29/11/2011

12. IASF Bologna 30 CsI(Tl) bars with Photodiode readout 1400 cm 2 geometrical area ~300 cm 2 effective area @ 1 MeV 330 keV – 100 MeV energy range 14% energy resolution FWHM @ 1.3 MeV 2  s timing accuracy in photon-by-photon mode Clever, fully-programmable trigger logic on time scales from 8s to 16ms, 1ms and 300  s The AGILE payload 40 cm Labanti et al., NIM A (2009): instrument paper Fuschino et al., NIM A (2008): trigger logic M. Marisaldi Marisaldi et al., A&A (2008): GRB detections Marisaldi et al., JGR (2010): TGF detections FDT2 Frascati 29/11/2011

13. IASF Bologna M. Marisaldi 13 Maximum energy & spectral shape 110 TGFs 1806 photons 142  E> 10 MeV 26  E> 20 MeV Results published in Tavani et al., Phys. Rev. Letters 106, 018501 (2011) RREA cutoff powerlaw model significant detection of  >40 MeV unexplained by standard RREA model: challenge for emission models 100MV potential compatible with strong IC discharges Broken powerlaw model  = -2.7 M. Marisaldi FDT2 Frascati 29/11/2011

14. IASF Bologna M. Marisaldi 14 Beaming angle: TGF localization from space in  -rays by AGILE GRID AGILE footprint GRID  projection GRID pointing Reverse event direction (TGF source)‏ M. Marisaldi FDT2 Frascati 29/11/2011

15. IASF Bologna TGF 12809-19 in details (2010 Oct. 16 20:44:55 UT)  -ray source GOES IR image ~5 minutes before TGF TRMM-LIS pass ~1 hour before TGF Credits: B. Carlson, Univ. Bergen M. Marisaldi FDT2 Frascati 29/11/2011

16. IASF Bologna M. Marisaldi 16 Geographical distribution Event clustering at < 400 km from AGILE footprint: half cone angle of emission ~30-40° Consistency with pervious detections based on RHESSI TGFs and sferics (Cummer 2005, Cohen 2010, Hazelton 2009, Gjesteland 2011) Results published in Marisaldi et al., Phys. Rev. Letters 105, 128501 (2010) Cohen et al., GRL 2010 M. Marisaldi FDT2 Frascati 29/11/2011

17. IASF Bologna M. Marisaldi 17 Do GRID photons come directly from the production region? Compton and pair production cross section in air become equivalent at ~25MeV: Compton interaction for low-energy GRID events cannot be ignored <3% probability to scatter above 40 km: the GRID photon tracks the source within the angular resolution 15-20 km 540 km AGILE good TGF production region Compton interaction 40 km AGILE bad TGF production region Compton interaction M. Marisaldi FDT2 Frascati 29/11/2011

18. IASF Bologna M. Marisaldi 18 Implications for beaming angle and electric field orientation at the source 50 MeV High energy photons track well the electric field orientation at the source A new tool to probe remotely the production site electric field Roussel-Doupre et al. Sp. Sci. Rev. 2009 M. Marisaldi FDT2 Frascati 29/11/2011

19. IASF Bologna Parent lightning / lightning phase M. Marisaldi FDT2 Frascati 29/11/2011 Williams et al., 2006 Lu et al., 2010 Observations by lightning location networks point toward initial phase of IC lightning at 10-14 km  -rays before or after the return stroke? Measurements from space still inconclusive. On ground experiments and triggered lightning observations points toward  -ray production during leader phase (before return stroke) Cummer et al., 2011

20. IASF Bologna Toward a climatology of TGFs M. Marisaldi FDT2 Frascati 29/11/2011 Splitt et al., 2010 Smith et al., 2010 TGFs follow lightning distribution and ITCZ movement BUT with significant differences: TGF tend to peak at later storm phase (while Ics don’t) and there is a geographical asymmetry. Peculiar subset of lightning? Hint: water/ice mixed phase plays a role Few climatology studies yet

21. IASF Bologna LIS/OTD global lightning distribution (10 Years of data) 0.5 x 0.5 deg per bin Flash rate [fl/km -2 /year] MCAL exposure [Seconds per bin] (Mar 2009-Feb 2010) Trigger logic really active 2.5 x 1.0 deg [lon x lat] Lightning distribution multiplied by the MCAL exposure to direct comparison with 12 Months of AGILE TGFs LIS-OTD high resolution full climatology available at http://thunder.msfc.nasa.gov/ M. Marisaldi 0.68 correlation coefficient for global case BUT… 2D Kolmogorov Smirnov probability = 0.002 for TGFs to be drawn from the same distribution of lightning! FDT2 Frascati 29/11/2011

22. IASF Bologna TGF / lightning statistical comparison M. Marisaldi Continental region 1D KS Longitude P 1D KS Latitude P 2D KS PTGF / flash ratio America0.340.450.131.5 10 -4 Africa0.170.140.036.0 10 -5 South East Asia0.950.780.877.5 10 -5 All0.002 7.8 10 -5 FDT2 Frascati 29/11/2011 Results published in Fuschino et al., Geophys. Res. Lett. 38, L14806 (2011)

23. IASF Bologna M. Marisaldi 23 Summary TGFs are the manifestation of the highest-energy natural particle accelerators on Earth Role of the RREA process widely accepted, but difficult to explain recent spectral results (E ~100 MeV) Acceleration sites correlated to IC lightning at 10-14 km altitude, possibly during leader formation Climatology studies show discrepancies with lightning distribution: possible sub-class of lightning? AGILE is giving an important contribution to TGF science: Energy spectrum up to ~100 MeV First TGF localization in  -rays Highest TGF detection rate surface density: TGF / lightning climatology study M. Marisaldi FDT2 Frascati 29/11/2011

24. IASF Bologna What next? M. Marisaldi FDT2 Frascati 29/11/2011 ASIM ESA mission >= 2014 TARANIS CNES mission >= 2015 But AGILE, RHESSI and Fermi still have a lot more to say!

25. IASF Bologna THANK YOU! Credit: Alan Stonebraker M. Marisaldi First AGILE TGF detections: Marisaldi et al., JGR 115, A00E13 (2010) TGF  -ray localization: Marisaldi et al., Phys Rev Lett 105, 128501 (2010) TGF high-energy spectrum: Tavani et al., Phys Rev Lett 106, 018501 (2011) AGILE TGFs & lightning activity: Fuschino et al., GRL 38, L14806(2011) FDT2 Frascati 29/11/2011

26. IASF Bologna M. Marisaldi 26 Extra slides M. Marisaldi FDT2 Frascati 29/11/2011

27. IASF Bologna M. Marisaldi 27 Imaging TGFs in gamma rays Search for GRID events in temporal coincidence with 119 MCAL TGFs detected between Jun. 2008 – Dec. 2009 12   significance 13 GRID events within 2 ms from TGFs T0! 4 6 3 M. Marisaldi FDT2 Frascati 29/11/2011

28. IASF Bologna M. Marisaldi 28 MCAL TGF detection rate > 250 class A TGFs + ~130 class B TGFs since June2008 34 TGFs Published in M. Marisaldi et al., J. Geoph. Res., 115, A00E13, 2010. 1ms trigger onset After entering Spinning mode ~10 TGF/month since Mar.'09 ~10 class A TGF/month ~ 5 class B TGF/month 24 months 1 st AGILE TGF catalog in preparation M. Marisaldi FDT2 Frascati 29/11/2011

29. IASF Bologna M. Marisaldi 29 The AGILE TGF sample Average properties A AND B class: Number of counts = 14 +/- 9 Duration = (0.8 +/- 0.4) ms Energy = (4.0 +/- 1.7) MeV M. Marisaldi FDT2 Frascati 29/11/2011

30. IASF Bologna Basic properties M. Marisaldi Class A (old) Class B FDT2 Frascati 29/11/2011

31. IASF Bologna High energy events M. Marisaldi FDT2 Frascati 29/11/2011

32. IASF Bologna M. Marisaldi Longitude and local time RHESSI AGILE class A AGILE class B FDT2 Frascati 29/11/2011

33. IASF Bologna MCAL high energy calibration Giuliani et al., ApJL 708 (2010) L84 GRB 090510 promptExtended emission M. Marisaldi In flight MCAL data GRID UL GRID data FDT2 Frascati 29/11/2011

34. IASF Bologna M. Marisaldi 34 Why AGILE is good for TGF science? MCAL energy range is extended up to 100 MeV: probing the high energy tail of the TGF spectrum Efficient trigger at ms and sub-ms time scale (the TGF time scale): not biased toward brightest events segmented independent detectors: low dead time and pile-up photon-by-photon data download for triggered events with 2  s time resolution <100  s absolute timing accuracy: mandatory for sferics correlation AGILE orbit at 2.5° inclination is optimal for mapping the equatorial region, where most of the events take place, with unprecedented exposure M. Marisaldi FDT2 Frascati 29/11/2011

35. IASF Bologna M. Marisaldi 35 MCAL Burst Trigger Logic Long (SW evaluated) time windows: 64ms, 256ms, 1.024s, 8.192s 4 spatial zones and 3 energy ranges Short (HW evaluated) time windows: sub-millisecond, 1ms, 16ms First trigger logic at ~1ms time scale Very flexible: more than 2000 parameters for full configuration; dedicated look-up tables to accept/reject triggers Current threshold settings: 16ms: >22 counts 1ms: >10 counts 293  s: > 8 counts M. Marisaldi FDT2 Frascati 29/11/2011

36. IASF Bologna M. Marisaldi 36 Trigger selection Key parameter: Hardness Ratio HR = (n. evt E>1.4 MeV) / (n. evt. E<1.4 MeV)‏ Selection criteria to reject known instrumental triggers: HR > 0.5 Known instrumental triggers All other triggers M. Marisaldi FDT2 Frascati 29/11/2011