University of Kansas
X-Ray Emission in the Solar System
Temporal Variations of Geocoronal and Heliospheric X-Ray Emission Associated with the Solar Wind Interaction with Neutrals
by Cravens et al.
|
|
University of KansasX-Ray Emission in the Solar System |
Draft Temporal Variations of Geocoronal and Heliospheric X-Ray Emission Associated with the Solar Wind Interaction with Neutrals by Cravens et al. |
Image: Jovian soft X-rays from ROSAT; courtesy of J. H. Waite.
In this section we first directly compare ROSAT 1/4 keV channel LTE data for the 1990-1993 time period with measured solar wind proton fluxes. This comparison is limited because solar wind data were not available for many parts of this period. The solar wind data availability was better for the later 1996-1998 time period, but LTE data were not available for this period. Another limitation in the comparisons we will present in this section is that we did not select the LTE data for look direction, partly because this would reduce the number of available data points needed to track time variations. In addition, our simple model does not accurately determine the X-ray intensity for all directions. A useful future project would be to construct a more elaborate model and include the LTE look direction when comparing this data with the model. Note that in constructing the LTE part of the X-ray background signal, the steady (or almost steady) part of the background is removed [Snowden et al., 1994].
Time histories of the ROSAT LTE count rate and the measured solar wind proton flux are shown for a typical month in Figure 3. One gets the impression that the two data sets are positively correlated, and this is also true of most other months (not shown).
In order to obtain a more complete time history of the solar wind for this particular month, we "interpolated" the solar wind data gaps apparent in Figure 3 by using solar wind data from 27-day solar rotation periods either before or after the month shown, up to three periods. An obvious coronal mass ejection in one of the rotation periods was not included in this procedure, which assumes that major solar wind structures corotate with the Sun. A handful of short data gaps remained after this procedure was completed, and these were filled in with straight-line interpolations. Left out of this interpolation procedure is any evolution of solar structure (and the associated solar wind structure) within a solar rotation period and noncorotating events, such as coronal mass ejections, which occurred during a data gap.
We applied our model to this new solar wind data (and interpolated data) as described for Figure 3, and the results are shown in Figure 4. Figures 2 and 4 give a similar impression. The hydrogen component of the X-ray intensity has the greatest overall magnitude but exhibits little temporal variation, the geocoronal contribution has the smallest overall magnitude yet displays the greatest relative variation, and the helium component lies between these extremes.
Figure 5 compares the model results with the ROSAT LTE data separately for the total intensity, the helium component, and the geocoronal component. This is for the same time period as Figure 4. For this comparison, we subtracted the average monthly background for each of these components individually before plotting them. In each case, the ROSAT LTE data scale was separately adjusted to facilitate the comparison and to check which model component has a time variation most like the LTE data. In all three cases a reasonable correspondence exists between the model X-ray intensity and the LTE data, but the total intensity arguably shows the best agreement followed by the pure geocoronal component.
We also directly compared measured solar wind proton fluxes with ROSAT LTE data using daily averages for all days for which both data types were available (N = 121). The only selection criterion for the LTE data is that actual solar wind data (no interpolations from other times) should exist on the same day; we did not select for look direction. The resulting scatter plot is shown in Figure 6. A clear correlation is displayed in Figure 6 with a linear correlation coefficient of R = 0.71. The correlation is highly significant statistically with t = 9.6. The portion of the overall ROSAT variance explained by the linear regression (and hence by variations of the solar wind proton flux) is ~50% (i.e., R2). The correlation is surprisingly good given that (1) according to the SWCX theory the X-ray emission should be better correlated with the solar wind heavy ion flux than with the proton flux, as demonstrated by Neugebauer et al. [2000] for cometary X-rays; (2) no time delays associated with solar wind travel time and longitudinal structure were introduced into the data comparison (this is somewhat ameliorated by using daily averages and is relevant to the heliospheric helium component but not the geocoronal component); and (3) the ROSAT data were not sorted for look direction and the X-ray emission is expected to have spatial structure both in its heliospheric components [Robertson et al., 2001] and in its geocoronal component (due to the structure of the solar wind flow around the magnetopause).
The linear relation between solar wind flux and LTE count rate that one would expect from Figure 3 is also shown in Figure 6. By combining this linear relation with the relation between solar wind proton flux and total X-ray intensity implicit in Figure 2 (this also can be seen in Figure 5, top), we find that 1 ROSAT 1/4 keV channel count per second should correspond to a total soft X-ray intensity of ~5.7 keV cm-2 s-1 sr-1.
Next: 8. Discussion and Summary
 |
Tizby Hunt-Ward tizby@ku.edu |
