SMEI Photometric Images and 3D Reconstructions

SMEI Photometric Images and 3D Tomographic Reconstructions

Introduction

This Solar Mass Ejection Imager (SMEI) data set and the 3D tomographic reconstructions from it are one of the primary UCSD end products in the analyses of these data. The SMEI sky maps, when sidereal and zodiacal-light background are removed, consist of sunlight Thomson-scattered from heliospheric electrons. Reaching this goal has included over twenty years of planning. The white-light photometers on the Helios spacecraft (Leinert et al., 1981) proved that this scattering from heliospheric electrons has a sufficient signal to be distinguished from other sources of variable inner-heliospheric brightness in deep space (see Richter et al, 1982). Further analysis of this data set proved that it is feasible to image heliospheric Coronal Mass Ejections (CMEs) and Corotating Structures (CIRs) (see Jackson and Leinert, 1985; Jackson, 1985; Jackson, 1991).

 

In the mid-1990’s as the SMEI concept began to gain momentum, it was realized that to become a truly viable instrument, the SMEI data would need to be analyzed photometrically – the brightness from heliospheric electrons (which move primarily radially outward from the Sun) decreases as they move outward such that structure distances from the Sun and from the spacecraft can be determined from this decrease. The photometric signal from a given structure needs to be sufficiently stable such that its surface brightness over time (that dims and evolves) presents the same calibrated measurement for the duration of its observation. From this was born the requirement for the SMEI program to provide photometric-image analyses, and the idea of employing computer assisted tomography (CAT) techniques to determine the distance to the structures being observed. Though details describing how this is done are found in many of the references herein, the basic concept is that the 3D reconstructions proceed by fitting a heliospheric model to the SMEI-observed brightness, using the Thomson-scattering parameters as given in Billings (1966). Since SMEI photometric data were not available to test prior to its launch, the CAT techniques were applied to the nearest available heliospheric data set, interplanetary scintillation (IPS) (see Jackson et al., 1998, Kojima et al., 1998). The set of scattering coefficients for the IPS data differ from those for Thomson-scattering brightness (see Jackson and Hick, 2004). This SMEI photometric imaging concept provides a good match to the U.S. Air Force space-weather forecast need: to understand the distance of CMEs and other heliospheric structures from the Earth as these move outward from the Sun.

Analyses

There are two parts to the SMEI sky-map analyses:

     

  1. Four-second-exposure individual data frame images are combined as completely as possible into photometric sidereal-coordinate sky maps. These preserve the resolution native to the original SMEI data frames, and optimally allow stellar images and other point sources to be removed. The originally recorded analog-to-digital surface-brightness units (ADUs) are retained through most of the subsequent analysis; a surface brightness of one S10 (the equivalent brightness of a 10th magnitude star spread over one square degree, see Cox (2000, page 330) is about 0.46 SMEI ADUs (Buffington et al. 2007). Because typically ten or more data frames combine to provide a surface-brightness measurement at any one sidereal location, this multiplicity allows removal of the most rapidly-varying sources of noise, such as high-energy-particle hits on individual data frames and the rapid passage of space debris near the SMEI orbit, by median filtering techniques. Details are found in articles by Jackson et al. (2004), Hick et al. (2005; 2007) and Buffington, et al. (2007). These processed sky maps have the zodiacal background removed by the subtraction of a multi-parameter zodiacal-light sky-map model (Buffington et al., 2006). This zodiacal-light model removes the bulk of the long-duration solar position-angle-dependent and annually-varying background. At this stage (and for several intermediate steps) the SMEI data are available for viewing and are presented to the public in the form of sky maps and sky-map differences for the entire SMEI data set to date (see http://smei.ucsd.edu/sky/index.html ). On a present-day single PC this entire process generally operates at approximately 10 times real time per SMEI camera.

     

  2. The individual sky maps thus processed are next analyzed to remove zodiacal light residual and aurora from successive 102-minute-orbit sky maps. How this is done is presented in publications by Jackson et al. (2004; 2006; 2008). The fully-processed sky maps of (1) have a polar mean minimum removed from them per camera, to further reduce zodiacal-light residue and some of the stray light present in individual maps from 600-orbit stretches of the data. For the present 3D reconstructions, data are then extracted from the sky maps for 600 consecutive orbits at a set of sidereal locations having approximate 5º centers binned to 1º × 1º and then culled to enhance the number of lines of sight towards the Sun. Following this, the aurora “noise” is removed as completely as possible using an iterative edge filter analysis technique that has been shown in spot checks to approximate an auroral-removal hand editing scheme devised in the preliminary analyses of these data. This same procedure adds a long-term temporal e-x/200-orbit base to each time series to further reduce zodiacal light trends from the line-of-sight time series. This auroral–removal process operates on a single PC at about 10 times real time to combine data from all three cameras and provide the time-series data edited for use in the 3D reconstruction analysis. This further step (the 3D reconstruction) is similar to that provided in the real-time analysis of IPS data on the Web site http://ips.ucsd.edu/ , but has been upgraded to provide higher-resolution analysis in keeping with the more numerous lines of sight present in the SMEI data (generally 200-700 lines of sight per SMEI orbit – or ~8000 lines of sight per day). This further step operates at these current resolutions again at about 10 times real time on a single PC. From these 3D reconstruction runs the 3D volumes (the “nv3 files”) are analyzed to provide the image data products of them using a set of IDL routines.

 
The location where the 3D reconstruction is most easily monitored is near Earth, and for several large events observed to date (see Jackson et al., 2007; 2008; Bisi et al., 2008), the 3D reconstruction analyses from SMEI brightness provide a reasonable approximation to the in situ densities observed by near-Earth spacecraft with the caveat that there are often differences noted (especially for the largest events) in various spacecraft proton and electron density values for the same event. Now that the STEREO spacecraft also measure in situ densities along the ecliptic, they give additional input and confirmation for the reconstructed densities. From the years 2007 to present, the STEREO spacecraft locations are indicated in the ecliptic cuts of the reconstructed densities.

These analyses are different from most of the other SMEI (and STEREO) analyses where in these other presentations orbit-to-orbit differences or subtractions from a local mean are used to enhance outward structure motion in the images. Here we are interested in the total electron content of the interplanetary medium and its variation. To obtain the total electron value, we have removed a long term base from the images, and modeled the ambient density. This plus the “averaging effect” of the current solar wind model leads to many of the differences present between the reconstructed sky maps in this study, and those from other studies.

 

Difficulties with the present-day reconstructions

 Finally, there are several problems in the 3D reconstructions currently being worked on:

  1. Present analyses use Solar Terrestrial Environment Laboratory (STElab) IPS velocities to help maintain the solar wind outward velocity for the 3D reconstructions. When available, these provide a general background velocity for the SMEI 3D reconstructions. How this operates is described in the recent papers by Jackson et al., (2006, 2008). These velocities can also be displayed separately as 3D solar wind reconstructions. When STELab velocities are not available during portions of the year when the Japanese radio telescopes are not operating, a constant background velocity is used for the reconstructions. The few numbers of velocity lines-of-sight from STELab (or the constant velocity assumption) allow only an average solar wind speed to be fit using the tomographic analysis. When CMEs have fast initiation speeds, these reconstructions do not take into account the rapid deceleration of these events as they merge into the background solar wind. Thus, fast-CME reconstructions sometimes show the event near the Sun beginning prior to its known initiation at the solar surface.
  2. The significant editing to remove artifacts from the SMEI data usually works well, but sometimes noise remains in the 3D reconstructions. This noise manifests itself as apparent outward flowing solar material near the edge of the camera 3 outage area, and also near the ecliptic to the West of the Sun. The latter is presumably caused by an inaccurate subtraction of the zodiacal cloud, which we are presently working to improve.
  3. The 3D reconstructions are run for each Carrington rotation period of 27.275 days as observed from Earth. The UCSD 3D reconstructions do not always blend precisely from the end of one Carrington rotation to the beginning of another in the analyses.

 

We continue to update our analyses as errors are discovered, and their effects mitigated by better processing, the inclusion of better data sets, or the inclusion of better modeling techniques, and as time permits.

 

Recent UCSD innovations:

2007/11/01 – The entire SMEI data set of image frames through to 2007 has been analyzed and made into preliminary sky maps.

2008/05/25 – The additional SMEI data set from the beginning of 2007 has been added to the sky map analysis through to 2008/05/01.

2008/06/24 – A newer high resolution 3D analysis has reconstructed CME sheath regions for a few large events as well as CME central cores in the PC reconstruction analyses. These 3D reconstruction analyses take approximately triple the computer time to provide and 100 times more computer disk space to store (~200 Mbytes/volume).

2008/07/01 – The entire SMEI data set has been processed through to 2008/05/01 using the preliminary low-resolution PC 3D reconstruction analyses; the IDL images provided from this data set are being produced to populate the UCSD Web site with completion expected in about one month at http://smei.ucsd.edu/smeidata.html .

2008/07/18 – A better technique to deal with SMEI camera 3 base subtractions has been determined. This innovation will undoubtedly result in SMEI camera 3 sky maps with far fewer radial streaks, and fewer errors that propagate into the resulting data analyses and 3D reconstructions.

2008/07/23 – One potential source of line of sight time series error has been identified; the lines of sight near the edge of the SMEI field of view can potentially be contaminated by bright star “bleed-in” from outside the field of view into the time series field. Steps are being taken to mitigate these effects in our analyses, and subsequent updates of the 3D reconstruction analyses will include the latest “bleed in” removal procedure.

2008/08/29 - The new higher-resolution 3D analysis described on 2008/06/24 has been used to analyze selected sets of SMEI data during periods of interest. These higher resolutions show the density enhancements behind shocks for the larger CME events (see Selected High Resolution Events , in particular 2003-10-31 => 2003-11-08)

2008/09/10 - A confidence level 3D matrix showing a composite number of line of sight crossings is now available on a regular basis for selected time periods. (see 2008-03-31 => 2008-04-07 WHI Period Extract) Generally 3D reconstructions require at least 3 perspective-view line of sight crossings present within a single resolution element to be tomographically reconstructed.

References:

Billings, D.E. (1966), A guide to the solar corona, Academic, New York, p. 150.

Bisi, M.M., B.V. Jackson, P.P. Hick, A. Buffington, D. Odstrcil, and J.M. Clover (2008), 3D Reconstructions of the Early-November 2004 CDAW Geomagnetic Storms: Analysis of STELab IPS speed and SMEI density data, J. Geophys. Res. – Space Physics: Special Edition - Geomagnetic Storms of Solar Cycle 23, in-press

Buffington, A., D.L. Band, B.V Jackson, P.P. Hick, and A.C. Smith (2006), A Search for Early Optical Emission at Gamma-Ray Burst Locations by the Solar Mass Ejection Imager (SMEI), Astrophys. J. 637, 880.

Buffington, A., B.V. Jackson, P. Hick, and S.D. Price (2006), An Empirical Description of Zodiacal Light as Measured by SMEI, EOS Trans. AGU 87(52), Fall Meet. Suppl., Abstract, SH32A-06.

Buffington, A., J.S. Morrill, P.P. Hick, R.A. Howard, B.V. Jackson, and D.F. Webb (2007), Analyses of the comparative responses of SMEI and LASCO,Proc. SPIE6689, 66890B, 1−6, doi: 10.1117/12.734658.

Cox, A.N. (2000), Allen’s Astrophysical Quantities, Fourth Edition, Springer, New York.

Hick, P.P., A. Buffington, and B.V. Jackson (2005), The SMEI real-time data pipeline: From raw CCD frames to photometrically accurate full-sky maps, Proc. of SPIE, 59011B, doi: 10.1117/12.617996.

Hick, P.P., A. Buffington, and B.V. Jackson (2007), A Procedure for Fitting Point Sources in SMEI White-Light Full-Sky Maps, in Proc. SPIE, 66890C, 1−8, doi: 10.1117/12.734808.

Jackson, B.V. (1985), Imaging of Coronal Mass Ejections by the Helios Spacecraft’, Solar Phys., 100, 563.

Jackson, B.V., 1991, ‘Helios spacecraft photometer observations of elongated corotating structures in the interplanetary medium’, J. Geophys. Res. 96, 11307.

Jackson, B.V., and C. Leinert (1985), Helios images of solar mass ejections, J. Geophys. Res. 90, 10,759.

Jackson, B.V., and P.P. Hick (2004), Three-dimensional tomography of interplanetary disturbances, in: Solar and Space Weather Radiophysics Current Status and Future Developments, D.G. Gary and C.U. Keller (eds.), ASSL 314, Kluwer, The Netherlands, 355.

Jackson, B.V., P.L. Hick, M. Kojima, and A. Yokobe (1998), Heliospheric tomography using interplanetary scintillation observations, 1. Combined Nagoya and Cambridge observations, J. Geophys. Res., 103, 12,049.

Jackson, B.V., A. Buffington, P.P. Hick, R.C. Altrock, S. Figueroa, P. Holladay, J.C. Johnston, S.W. Kahler, J. Mozer, S. Price, R.R. Radick, R. Sagalyn, D. Sinclair, G.M. Simnett, C.J. Eyles, M.P. Cooke, S. J. Tappin, T. Kuchar, D. Mizumo, D.F. Webb, P. Anderson, S.L. Keil, R. Gold, and N.R. Waltham (2004), The Solar Mass Ejection Imager (SMEI) mission, Solar Phys. 225, 177.

Jackson, B.V., A. Buffington, P.P. Hick, X. Wang, and D. Webb (2006), Preliminary three-dimensional analysis of the heliospheric response to the 28 October 2003 CME using SMEI white-light observations, J. Geophys. Res. 111, A4, A04S91.

Jackson, B.V., P.P. Hick, A. Buffington, M.M. Bisi, and E.A. Jensen (2007), SMEI Observations in the STEREO Era, Proc. SPIE 6689, 66890G, 1−14, doi: 10.1117/12.734870.

Jackson, B.V., M.M. Bisi, A. P.P. Hick, A. Buffington, J.M. Clover, and W. Sun (2008), Solar Mass Ejection Imager (SMEI) 3D reconstruction of the 27-28 May 2003 CME sequence, J. Geophys. Res. (in review).

Kojima, M., M. Tokumaru, H. Watanabe, A. Yokobe, K. Asai, B.V. Jackson, and P.L. Hick (1998), Heliospheric tomography using interplanetary scintillation observations, 2. Latitude and heliocentric distance dependence of solar wind structure at 0.1-1 AU, J. Geophys. Res., 103, 1981.

Leinert C., H. Link and N. Salm (1981), Calibration and in-flight performance of the zodiacal light experiment on Helios, J. Space Sci. Instr. Vol. 5, 257.

Richter, I., C. Leinert, and B. Planck, (1982), Search for short term variations of zodiacal light and optical detection of interplanetary plasma, Astron. Astrophys., 110, 115.