THEMIS-Wind-ACE Solar Wind Field & Plasma Comparisons Joe King and Natalia Papitashvili April 6, 2011 Introduction This note addresses comparisons of THEMIS-B and -C (ARTEMIS P1 and P2) interplanetary magnetic field and plasma parameters with concurrent parameter values from the Wind and ACE spacecraft. We address, in sequence, the data, the tools developed to enable the comparisons, and the results of a preliminary comparisons. The bottom lines Wind and TH-C largely agree (to within ~0.1 nT) in magnetic field magnitude and components. These spacecraft agree with TH-B except that Bz(TH-B) is less negative than the others by ~0.2 nT. TH-B and TH-C flow speeds agree with each other to within 0.5% and are less than Wind flow speeds by ~2%. TH-B densities are less than TH-C and Wind densities by 20+ %, while TH-C densities are less than Wind densities by ~10%. ____________________ The Data ______ THEMIS-B and -C For THEMIS-B and -C, we started with CDAWeb-resident Level 2 data from the Fluxgate Magnetometer (FGM) and the Electrostatic Analyzer (ESA). We started with the ARTEMIS phase of the spacecraft lives (>~September, 2010), but then extended back in time to 1/1/2010 to get more spacecraft overlap. We selected THEMIS intervals when the spacecraft were well upstream of the average bow shock location (rather than identifying individual shock crossings from field and plasma time series data) and then examined data plots to ensure no non-solarwind THEMIS intervals had been included. ESA has both solar wind and magnetosphere operating modes that differ in the use of energy range and resolutions most appropriate to solar wind and magnetospheric plasmas. When we started this THEMIS-Wind-ACE comparative data set and interface, we included THEMIS-B and -C data only for times when the spacecraft was in the solar wind and ESA was in its solar wind mode. However, in September, 2011, the THEMIS/ARTEMIS science team judged that it would be better to keep ESA operating in its magnetosphere mode. We confirmed that the mode has very little impact on THEMIS-Wind agreement levels for solar wind ion flow speeds and densities, although the same is not true for solar wind flow direction or temperature. We then dropped the requirement for ESA to be operating in the solar wind mode for inclusion of THEMIS data in this THEMIS-Wind-ACE comparative data set. (2012/01/16) [Users should note that moments become unreliable in the presence of sufficiently high fluxes of solar energetic particles. Examples are seen at January 23-24 and 27-28, 2012. From Jim McFadded of the ESA team: SEP events will cause high background which will distort the moment calculation -- it's like adding an isotropic component that is not drifting -- density,pressure goes up and velocity goes down. You need a background subtraction routine to recover moments on the ground if you want to correct this.] (March, 2012) From ESA, we worked with ion density, average temperature, and GSE flow velocity Cartesian components flagged as good quality and as full mode. These are at 96s resolution for several hours most days and at 384s resolution for the remainder of each day. The plasma parameter values are as computed by the ESA team from distributions accumulated over 3 seconds. From FGM, we took the 3-sec GSE magnetic field average whose time tag was closest to that of a 96s or 384s plasma data point. We took THEMIS GSE spacecraft position data from SSCWeb, and we determined selenocentric spacecraft position data by subtracting SSCWeb's GSE position vector of the moon from the GSE spacecraft position vector. 96s/384s merged magnetic field - plasma - geocentric and selenocentric position data are available at http://omniweb.sci.gsfc.nasa.gov/ftpbrowser/themisb_mrg.html and http://omniweb.sci.gsfc.nasa.gov/ftpbrowser/themisc_mrg.html We next shifted (assigned new time tags to) each 96s/384s point to the Earth (X,Y,Z=0,0,0), following the technique used for Low Resolution OMNI and described in detail at http://omniweb.gsfc.nasa.gov/html/omni2_doc.html#shift. Briefly, we use spacecraft position vectors, the observed solar wind flow speeds, the Earth's ~30 km/s motion about the sun, and the assumptions of (helio)radial flow direction and variation phase front normals lying in the ecliptic plane and oriented half way between radial and an in-ecliptic normal to the Parker IMF spiral angle. For compatibility with Wind and ACE data, hourly averaged flow speeds, and flow direction angles computed from hourly averaged flow vector Cartesian components, were determined and carried forward. ______ Wind and ACE Hourly averaged Wind and ACE field and plasma parameters, built from Earth-shifted higher resolution data, have long been accessible and used in comparing Wind and ACE data to each other and to IMP8 and Geotail data. The shifting has been the same as for the THEMIS data. See http://omniweb.sci.gsfc.nasa.gov/ftpbrowser/mag_iwa.html and http://omniweb.sci.gsfc.nasa.gov/ftpbrowser/pla_iwa.html for IMP-Wind- ACE-Geotail comparisons for magnetic field and plasma parameters, respectively. ______ Spacecraft-merged data sets. Finally, we created magnetic field-specific and plasma- specific data sets having hourly averaged magnetic field or plasma parameters, built from Earth-shifted higher resolution parameters, from each of the four spacecraft in each hourly record. In addition to time tags and field or plasma data, the records of these data sets also contain Moon-Wind and Moon-ACE Impact Parameters. These are measures of the distance by which a plasma element seen by Wind or ACE would miss being seen by a near-moon spacecraft. They are useful as filters (explained below) during the near- moon phases (i.e., the ARTEMIS phases) of the THEMIS-B and -C lives. See http://omniweb.sci.gsfc.nasa.gov/ftpbrowser/impact_des.html for further discussion of Impact Parameters. _____________________ The Tools We have developed seven tools for display/analysis of the THEMIS, Wind and ACE interplanetary magnetic field and plasma data discussed above. These fall into three families that produce: intensity-time profiles with one multi-source color-coded panel for each physical parameter displayed (alternative: numeric listings for selected times and parameters); parameter value occurrence frequency distributions for a given parameter, source and time span, plus medians, means and standard deviations of the distributions; scatter plots and linear regression fits for any given parameter, time span and pair of sources. For the latter two families of tools, one can do filtering by the parameter being analyzed and/or by any other parameter(s) in the data records. "Filtering" means specifying a parameter value range outside of which data from a given record will not be included in an analysis. This functionality makes it possible, for example, to study dependency of some correlations on solar wind speed. Filtering by Impact Parameter is enabled fo the scatter plot/linear regression interfaces The seven tools are at: Magnetic field Intensity-time profiles: http://omniweb.sci.gsfc.nasa.gov/ftpbrowser/mag_thbcwa.html Occurrence frequency distributions and statistics: http://omniweb.sci.gsfc.nasa.gov/ftpbrowser/mag_thbcwa_d.html Scatter plots and linear regression fits: http://omniweb.sci.gsfc.nasa.gov/ftpbrowser/mag_thbcwa_s2.html Plasma Intensity-time profiles: http://omniweb.sci.gsfc.nasa.gov/ftpbrowser/pla_thbcwa.html Occurrence frequency distributions and statistics: http://omniweb.sci.gsfc.nasa.gov/ftpbrowser/pls_thbcwa_d.html Scatter plots and linear regression fits (LGF): http://omniweb.sci.gsfc.nasa.gov/ftpbrowser/pla_thbcwa_s2.html Scatter plots and LGF, logarithmic in N and T: http://omniweb.sci.gsfc.nasa.gov/ftpbrowser/pla_thbcwa_s3.html ____________________ The Comparisons We focus on THEMIS-Wind comparisons, as our ACE field and plasma data only run to 2010/095, and have no overlap with THEMIS-C. ______ Magnetic field comparisons. Between 1/1/2010 and 3/3/2011 there are 3250 hours common to Wind and THEMIS-B. For these hours, we have <|B|>(Wind) = 4.67 ± 2.07 <|B|>(TH-B) = 4.67 ± 2.08 (Wind) = 0.43 ± 2.61 (TH-B) = 0.40 ± 2.60 (Wind) =-0.36 ± 2.22 (TH-B) =-0.16 ± 2.24 Between 1/1/2010 and 3/3/2011 there are 2303 hours common to Wind and THEMIS-C. For these hours, we have <|B|>(Wind) = 4.67 ± 2.14 <|B|>(TH-C) = 4.73 ± 2.15 (Wind) = 0.61 ± 2.63 (TH-C) = 0.60 ± 2.68 (Wind) =-0.36 ± 2.25 (TH-C) =-0.42 ± 2.29 Between 1/1/2010 and 3/3/2011 there are 1769 hours common to THEMIS-B and THEMIS-C. For these hours, we have <|B|>(TH-B) = 4.71 ± 2.13 <|B|>(TH-C) = 4.74 ± 2.13 (TH-B) = 0.62 ± 2.60 (TH-C) = 0.63 ± 2.63 (TH-B) =-0.27 ± 2.29 (TH-C) =-0.43 ± 2.32 In terms of means, the three spacecraft agree to within 0.05 nT in |B| and Bx. (We assume By agreement level would be comparable to that for Bx.) Agreement for Bz is less good. (Wind) is 0.20 nT more negative than (TH-B) and 0.06 nT less negative than (TH-C) These two imply (TH-B) should be 0.26 nT less negative than ). Direct comparison shows (TH-B) to be less negative than (Wind) = 400.3 ± 88.9; median = 375 km/s (TH-B) = 392.4 ± 87.8; 369 (Wind) = 6.33 ± 4.76; 5.0 /cc (TH-B) = 4.83 ± 3.30; 3.9 Between 1/1/2010 and 3/3/2011 there are 2350 hours common to Wind and THEMIS-C with plasma V & N data. For these hours, we have (Wind) = 407.1 ± 95.0; median = 379 km/s (TH-C) = 397.7 ± 93.8; 372 (Wind) = 6.61 ± 5.84; 5.0 /cc (TH-C) = 5.96 ± 4.91; 4.5 Between 1/1/2010 and 3/3/2011 there are 1919 hours common to THEMIS-B and THEMIS-C with plasma V & N data. For these hours, we have (TH-B) = 401.8 ± 91.9; median = 376 km/s (TH-C) = 399.7 ± 91.6; 375 (TH-B) = 4.98 ± 3.82; 3.8 /cc (TH-C) = 6.03 ± 4.84; 4.5 Based on these means/medians, we see that speeds for TH-B and for TH-C agree to within ~2 km/s (or 0.5%); they are less than the Wind speed by ~8 km/s, or ~2%). The same conclusions follow from the linear regression fits for TH(B)-W and for TH(B)-TH(C), although the fit for TH(C)-W, i.e., V(TH-C) = -15.9 + 1.021*V(W), yields a 3% difference at V ~300 km/s and a <= 1% difference in the 500-700 km/s range. TH-C mean density exceeds TH-B density by 1.05 (~19%) and Wind density exceeds TH-C density by 0.65 (~10%). From these, we expect the Wind density to exceed the TH-B density by ~31%. Direct comparison shows Wind density exceeding TH-B densith by 1.50 (or ~27%), in reasonably good agreement with expectation. Approximately, and in summary, the TH-C density exceeds the TH-B density by ~20%, and, in turn, the Wind density exceeds the TH-C density by ~10%. Linear regressions in N also show that N(W) and N(TH-C) exceed N(TH-B) by ~22%, while N(W) exceeds N(TH-C) by ~10% across the 5-15 /cc range