1 / 33

In the terminology of predictive modeling, the performance of

S.M.P. McKenna-Lawlor 1 , M.Dryer 2,3 , C.D. Fry 2 , Z. Smith 3 , M. D. Kartalev 4 , W. Sun 5 , C. S. Deehr 5 , K. Kecskemety 6 , K. Kudela 7 , S. Barabash 8 , Y. Futaana 8 , R. Lundin 8 and R. Courtney 9

svea
Download Presentation

In the terminology of predictive modeling, the performance of

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. S.M.P. McKenna-Lawlor1, M.Dryer2,3, C.D. Fry2, Z. Smith3, M. D. Kartalev4, W. Sun5, C. S. Deehr5, K. Kecskemety6, K. Kudela7, S. Barabash8, Y. Futaana8, R. Lundin8 and R. Courtney9 1 Space Technology Ireland, National University of Ireland, Maynooth, Co. Kildare, Ireland (E-mail; stil@nuim.ie ). 2 Exploration Physics International, Inc., Huntsville, Alabama, 35806, USA. 3 NOAA Space Environment Center, Boulder, Colorado, 80305, USA. 4 Institute of Mechanics, Academy of Sciences, 1113 Sofia, Bulgaria. 5 Geophysical Institute, University of Alaska, Fairbanks, Alaska, 99775, USA. 6 KFKI Research Institute for Particle and Nuclear Physics, Budapest, Hungary. 7 Institute of Experimental Physics, Kosice, Slovakia. 8 Swedish Institute of Space Physics, Kiruna, Sweden 9 Space Operations Centre, Air Force Weather Agency, Omaha AFB, Omaha, Nebraska, USA. Comparison between the arrival times predicted using the HAFv.2 model of flare related particles/shocks associated with the disk passage of Active Region 0930 in December 2006 and the measured arrival times of these disturbances at Earth, Mars and Venus 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 1

  2. Active NOAA Region No. 0930 which transited the solar east limb on 5 December, 2006 (S06o, ~E90o), was associated during its disk passage in the minimum phase of Solar Cycle 23 with the production of significant solar flares, energetic particles and coronal mass ejections. The predicted arrivals at the Earth (1 AU), Mars and Venus of shocks generated during four of these events in December, 2007 were estimated using the Hakamada-Akasofu-Fry version 2 (HAFv.2) model and compared with in-situ observations recorded at each of the planets. Correspondences found between the predicted and observed arrival times of these particle/shock signatures at particular spacecraft are discussed in the context of developing capability to forecast the arrival of solar disturbances at different locations within the heliosphere. 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 2

  3. The HAFv.2 Model provides real-time operational forecasts of the disturbance-driven solar wind through employing a modified kinematic approach to simulate SW conditions. In this scenario, fluid parcels are emitted along radials from the rotating Sun. The spatial distribution of speed on the Sun-centered, spherical inner boundary is non-uniform. The speeds of the particles along a particular radial consequently vary as higher and lower-speed streams sweep past a particular radial as the Sun rotates. If magnetic flux conservation and a highly conducting solar wind plasma are each assumed, this leads to a frozen-in field condition which prohibits higher-speed streams from overtaking streams with slower speeds. Solar wind acceleration/deceleration is accounted for by introducing a set of parametric equations. Internal algorithms adjust fluid parcel positions to account for fast stream-slow stream interactions, compression of the plasma and evolution of the IMF . For details see Hakamada and Akasofu (1982) Note that the internal free parameters of the model were set following an early calibration with 1-D and 2-D MHD models/empirical studies and these are held constant (Fry et al., 2007). 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 3

  4. Potential Field Source Surface (PFSS) maps provide solar wind speed and radial magnetic field on the HAFv.2 inner boundary [See Wang and Sheely (1990); Arge and Pizzo (2000)]. Proxy parameters for significant disturbance drivers are solar optical, X-ray and radio events that are accompanied by a reported shock (or CME) with a speed of at least 400 km/s. INDIVIDUAL EVENT PARAMETERS Optical/X-ray Event start time (taken to be within 0.5h of the accompanying shock start) Disk location of the parent solar event Event duration (piston driving time of shock: determined from the GOES soft X-ray profile of the flare) Shock start (determined from metric Type II radio burst data) Initial speed (Vs) of the shock near the Sun (estimated from reported metric Type II speed, or plane of the sky CME speed) 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 4

  5. In the terminology of predictive modeling, the performance of a model in predicting shock arrivals is expressed in the following terms: HIT: Shock predicted and observed to arrive at a particular heliospheric location within ±24h of its observed detection time. MISS: Shock detected at a particular heliospheric location but predicted to arrive at a time more than 24h before or after this detection, or predicted not to arrive at all. FALSE Shock predicted to arrive, but not detected, within a ALARM window of 1-5 days (Earth) following a particular solar event. CORRECT Shock neither predicted nor detected at a particular heliospheric NULL location within a window of 1-5 days (Earth) following a particular solar event. 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 5

  6. Temporally close shocks can potentially interact. In the simplest case involving two interacting events, the predicted arrival of the second shock is either at the same time, or earlier, than that of the preceding event. To take account of such a pair of interacting shocks, the definitions presented previously are modified. As before each shock is assigned a hit (h), miss (m), correct null (cn) , false alarm (fa) and correct null (cn) classification. However, only one hit is recorded and the contributing event is assigned to the category ‘correct null’. For the present paper, the arrival times at Earth, Venus and Mars of flare related shocks identified exiting the Sun (using metric radio burst drift data) during the disk passage of Active Region No. 0950 were forecast in near-real time using the Hakamada-Akasofu-Fry Model, version-2/(HAFv.2). These predictions are compared with the measured arrivals at L1, Mars and Venus of shocks recorded in plasma and magnetic data aboard the ACE, SOHO, Mars Express, Venus Express and GOES spacecraft. The influence of interplanetary conditions in determining the outcome at individual planetary targets is discussed having regard to the prevailing geometry. East Limb Passage of Active Region No. 0950 Interacting Events 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 6

  7. (Top left) East limb passage of Active Region 0490 recorded aboard TRACE. (Top right) Full diskmagnetogram recorded by the Michelson Doppler Imager (MDI) aboard SOHO. (Bottom) MDI Potential Field Source Surface plot showing field lines out to 2.5 Rs. 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 7

  8. (Top right) An X9 flare at S07, E79 was recorded aboard GOES 12 on December 5 (10.34-12.15 UT), followed on December 6 (1842-1854 UT at S 04 E 64 by a further X 6.5 flare. (Bottom right) shows gradually rising protons recorded aboard GOES 11. GOES X-rays GOES protons 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 8

  9. Particle profile (protons) recorded by the EPHIN instrument aboard SOHO of a (relatively rare) increase in protons up to MeV energies associated with the east limb flares in Region 0490. 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 9

  10. The flare of December 5 (10.34-12.15 UT) was accompanied by a Type II metre wave burst with shock velocity 836 km/s. (Station SVI/San Vito, Italy). The later flare of December 6 (18.42-18.54 UT) was accompanied by a further Type II burst with shock velocity 2000 km/s (Rec. aboard STEREO/ WAVES, private communication ). 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 10

  11. Magnetic, energetic proton and solar wind data recorded aboard ACE during December, 2006 (Top panel) note the arrival of a shock in MAG data at 04.11 UT on December 7. The SWEPAM/SW Level 1 data (density, vel. temp) are unreliable from ~ 07.00 UT, December 7 until ~16.00 UT on December 8 and again on 13 December 13 ( ~ 13.40 - 18.00 UT) due to the prevailing high proton background (snowstorm- effect). Note that a second shock arrived on December 14 at 13.52 UTand a third shock on December 16 at 17.21 UT 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 11

  12. On December 13 an X3.4 flare occurred in the same active region (at S06, W23) See the GOES 11 and GOES 12 X-ray fluxes (left) and EIT picture of the flare location (right) 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 12

  13. A well defined proton enhancement (> 50 MeV protons)on December 13 indicates that the magnetic field was well connected on this day. 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 13

  14. An associated metric Type II burst was recorded at Learmonth (1534 km/s). LASCO reported a full disk, asymmetric halo event (seen here projected from behind the occulting disk). Also a partial halo (> 120o) was projected out of the ecliptic plane. 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 14

  15. A shock associated with the December 13 flare arrived at L1 at 13.36 UT on December 14. Solar wind speeds increased from 600 to approx. 950 km/s. The Bz component oscillated from + 15 nT to – 15 nT between 14.00 -18.00 UT (shock compression of the IMF in front of a magnetic cloud). From 18.00-22.00 UT the Bz component was uniformly north (maximum deflection ~ 15 nT). It then shifted (uniform deflection ~ 18 nT) and the field remained open until about 21.00 UT on December 15 as the magnetic cloud moved through. 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 15

  16. An X 1.5 flare ( 22.10-22.22 UT) occurred on December 14 (S06.W36). An associated metric Type II burst was recorded at Culgoora (1600 km/s). 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 16

  17. A proton event which was an order of magnitude weaker than in the case of the previous flare (max flux 13 pfu at E > 50 MeV ) occurred. Also an asymmetric full halo was reported by LASCO 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 17

  18. A weak shock and magnetic cloud reached L1 in association with the flare of December 14 at ~ 17.00 UT on December 16. Solar wind speeds increased from 550 km/s to ~ 750 km/s and BT increased to ~ 10 nT. The Bz deflections were mostly to the north throughout the enhanced period and there was, thus, a general lack of connectivity. 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 18

  19. Modeling of Shock Arrival at L1 (HAFv2 SW model) 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 19

  20. Modeling of Shock Arrival at L1 (HAFv2 SW model) continued 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 20

  21. The HAFv.2 model indicated that disturbances associated with the X9 and the X6.5 flares on December 5 and 6 interacted with each other to produce a composite shock that was predicted to arrive at L1 on December 7 at 08.00 UT. A shock was recorded in ACE data on December 8 at 04.11 UT, some 20 hours late but within the period of ± 24 hours considered to constitute a “hit” in making such predictions. On December 14, a shock predicted by HAFv.2 to arrive at 13.56 UT was detected at 14.00 UT in ACE data (hit). On December 16 a shock predicted to arrive at 12.00 UT was recorded in ACE data at 17.22 UT about 5.5 hours late (hit). HAFv.2 Predictions 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 21

  22. Locations of the inner planets relative to the Sun during December 2006 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 22

  23. The ASPERA 3 and ASPERA 4 instruments aboard Mars Express (MEX) and Venus Express (VEX) are each composed of four similar instruments NPI (Neutral Particle Imager) measures the integral energetic neutral atom (ENA) flux with no mass and energy resolution but with high angular resolution. NPD (Neutral Particle Detector)resolves mass (hydrogen and oxygen) and velocity (energy range 0.1 - 10 keV) of the ENA. EIS (Electron and Ion Spectrometer)determines the electron and ion distributions at energies up to 40 keV IMA (Mass resolving Ion Analyser)measures the main ion components (H+, H2+, He+, O+), molecular ions from 20 to 80 amu/q and up to 106 amu/q for dusty plasmas in the energy range from 100 eV to 40 keV/q. Measurements at Venus and Mars 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 23

  24. Aboard MEX, ASPERA-3 data are typically recorded close to the Bow Shock crossings in 3-4 hour intervals. On VEX, ASPERA-4 observations are typically made 60 min before and after the inbound and outbound bow shock crossings. Since VEX only observes at pericenter there is a data gap of about 20 hours between measurement sets. 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 24

  25. Ecliptic plane plots generated by HAFv.2 showing solar wind conditions at Earth, Mars and Venus. IMF pattern: toward field lines (blue); away field lines (red). 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 25

  26. The largest flare in the sequence (X9/FF 663) began on December 5 at ~ 10:34 UT. Neither ASPERA-3 or 4 were functioning at this time. However, ASPERA-3 detected an extremely high background level of ions and electrons at Mars from 14:00 UT when the next sequence of spacecraft operations was initiated.  This enhancement endured for at least 3 days (i.e. spanning the occurrence of the X6.5/FF 664 flare of December 06 at ~ 18.42 UT and was present until Mars entered an ‘away’ sector. 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 26

  27. On December 8 the eastern flank of interacting FF events 663 and 664 was predicted by HAFv.2 to arrive at Venus between 03.00-05.00 UT. Put in Ghee pictures for December 8 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 27

  28. There was a gap in the ASPERA-4 observations from December 07 at 10.00 to December 08 at 05.30 UT and the shock was not observed in the data recorded thereafter. However, following 05.30 UT time the background ions were found to have substantially increased in energy since the previous day 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 28

  29. The HAFv.2 model predicted that, on December 20 at 00.00 UT, the shock accompanying FF event 666 overtook that of event 665 just as the western flank of these interacting shocks reached Mars There was also a field reversal at this time 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 29

  30. The ASPERA-3 ion data show a signature of heating between December  19 (23:45UT) and December 20 (04:00UT), indicating the arrival during that interval at Mars of an interplanetary shock. The arrival time of this shock is in good conformity with the prediction of HAFv.2. By December 21 the solar wind had recovered to a cool beam. 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 30

  31. HAFv.2 has previously been shown to provide useful predictions of shock arrivals at Earth (e.g. Fry et al., 2003, McKenna-Lawlor et al., 2006) and, in the present case the model has provided predictive hits with regard to shock arrivals measured at L1 in association with the flares of December 5, 6, 13 and 14, 2006. Consideration of HAFv.2 predictions in relation to in situ measurements made at Mars by ASPERA-3, again suggest the usefulness of these predictions in the case of an event on December 20 when the Earth and Mars were located on opposite sides of the Sun. A prediction by HAFv.2 of the arrival of a shock at Venus required more continuous observations than were available in ASPERA-4 data to monitor its arrival. A hint of a possible in situ response to the predicted disturbance may be contained in the presence in ASPERA-4 data later on the day concerned of a significantly enhanced particle background. Conclusion 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 31

  32. A reasonably accurate description of pre-event heliospheric conditions made by utilizing the PFSS method and the HAFv.2 code is now to hand. A full 3D MHD global description of conditions at the Sun which will provide improved pre-event simulations is awaited. Definitive validation of predictive models at Venus and Mars requires for their implementation continuous observations at these planets. Conclusion continued 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 32

  33. Go raibh maiτh agaibh go léir 5th SECCHI Meeting LAL, Orsay, France. March 5-8, 2007 page 33

More Related