Solar position dependence of rise-phase upflows

fl041.mariska01
Posted:  22-Feb-92
Updated: 05-Aug-92, 01-Nov-92, 11-Mar-93, 31-Jul-93, 21-Jan-94, 17-Sep-94, 13-Feb-95
Events specified: N/A

Collaborators: John Mariska (NRL, 11334::mariska), G. A. Doschek (NRL, 11334::doschek), any other interested parties.

Motivation: In the simple (and no doubt incorrect) loop model for a small flare, there should be a longitude dependence to the blueshifts observed early in the rise-phase of the flare. This, of course, assumes that the loop is always oriented in the same way and that the position on the Sun is producing the change in Doppler shift. In reality the situation is more complex. The flaring structures may have many possible orientations. If a large sample of similar flares, say M1 to M3 are observed, however, any underlying longitude dependence should emerge. Our preliminary evaluation of BCS flare observations shows that at any given position on the Sun a wide range of upflow velocities are observed for the same flare class. This study will build up a statistical picture of how the Doppler shifted components observed in the BCS early in flares depend on position on the Sun.

Required Observations/Analysis Techniques: All BCS-observed events in the M1 to M3 range that show reasonable rise-phase spectra will be considered for this study. For each channel, we will establish a standard method for determining the magnitude of the blue-shifted component. Additional parameters, such as the emission measure and the temperature, will, of course, also be calculated. SXT images early in the flare will provide some context for the BCS data. These images should, for example, show whether there is some common feature to the to soft X-ray plasma in flares that show large Doppler shifts. Then end result should be a "statistical" test of evaporation.

Update 13-Feb-95

I have now completed all the work I had planned on BB topic fl041: Solar position dependence of rise-phase upflows. The work has resulted in the following three papers. The first two have been published and the third is in press. You can now close out this BB topic. Here is the relevant information about the three papers:

Mariska, J.T., Doschek, G.A., and Bentley, R.D. 1993 Flare Plasma Dynamics Observed with the Yohkoh Bragg Crystal Spectrometer. I. Properties of the Ca XIX Resonance Line, ApJ, 419, 418

Abstract: Using data from the Bragg Crystal Spectrometer on the Yohkoh spacecraft we have computed measures of the total intensity, centroid position, and line width for the resonance line of Ca XIX during the rise phase and after maximum for 219 solar flares. The difference between the centroid positions early and late in each flare yields a measure of the line-of-sight velocity shift of the line centroid. We find a trend in the average value of the centroid shift with distance from Sun center suggesting radial mass motions with a characteristic velocity of 58 km s^-1. There is a correlation between the rise-phase line widths and the centroid shift. We find no correlation between the centroid shift and the peak intensity, rise time, and total flare duration; and no correlation between the line width and the distance from Sun center, the peak intensity, rise time, and total flare duration. These results do not conclusively support or refute the simple electron-beam-driven model or the thermal model for the rise phase of a solar flare.

Mariska, J.T. 1994, Flare Plasma Dynamics Observed with the Yohkoh Bragg Crystal Spectrometer. II. Properties of the Fe XXV, Ca XIX, and S XV Resonance Lines, ApJ, 434, 756

Abstract: Using data from the Bragg Crystal Spectrometer (BCS) on the Yohkoh spacecraft, we have computed measures of the total intensity, centroid position, and line widths for the resonance lines of Fe XXV, Ca XIX, and SXV during the rise phase and after maximum for 190 solar flares. The difference between centroid positions early and late in each flare yields a measure of the line-of-sight velocity shift of the line centroids. Roughly 25% of the flares show evidence for spatial shifts of the emitting plasma during the flare. For all three resonance lines in those flares that show only Doppler shifts, we find a trend in the average value of the centroid shift with distance from Sun center suggesting radial mass motions with characteristic velocities of 76, 63, and 58 km s^-1 in the resonance lines of Fe XXV, Ca XIX, and S XV, respectively. For each resonance line there is also a correlation between the rise-phase line widths and the centroid shift. For each resonance line we find no correlation between the centroid shift and the peak intensity, rise time, and total flare duration; and no correlation between the line width and the distance from Sun center, peak intensity, rise time, and total flare duration. Comparing data from the different BCS wavelength channels, we find that the centroid shift in each channel is correlated with the centroid shifts in the other channels. Similarly, the line width in each channel is correlated with the line width in the other channels.

Mariska, J.T. 1995, Flare Plasma Dynamics Observed with the Yohkoh Bragg Crystal Spectrometer. III. Spectral Signatures of Electron-Beam-Heated Atmospheres, ApJ, 444, in press (May 1)

Abstract: Using numerical simulations of an electron-beam-heated solar flare, we investigate the observational consequences of variations in the electron beam total energy flux and the low energy cutoff value for models with both low and high initial densities. To do this we use the evolution of the physical parameters of the simulated flares to synthesize the time evolution of the spectrum in the wavelength region surrounding the Ca XIX resonance line. These spectra are then summed over a 9 s time interval to simulate typical spectra from the Yohkoh Bragg Crystal Spectrometer and the first three moments are computed for comparison with observational results. This comparison shows that no single low or high initial density model satisfies the observed average behavior of the Ca XIX resonance line. Low initial density models produce too large a blueshift velocity, while high initial density models have lines that are too narrow. Comparison of these models with the Yohkoh data suggests that the key problem for models of the impulsive phase of a solar flare is producing significant amounts of stationary hot plasma early in the flare.

Update 17-Sep-94

I have completed a comparison of the results of my first survey with the predictions of a set of simulated model flare atmospheres. The results indicate that low initial density active region loops (5 to 10 dynes cm**-2) never yield a spectroscopic signature that matches the results of the survey. Higher density loops may produce profiles that agree with the measured centroid shifts, but do not agree with the width of the lines. A paper entitled "Flare Plasma Dynamics Observed with the Yohkoh Bragg Crystal Spectrometer. III. Spectral Signatures of Electron-Beam-Heated Atmospheres" is in preparation for submission to ApJ.

Paper II in the series has been accepted for publication in ApJ.

Update 21-Jan-94

The second paper dealing with a survey of BCS channels 2, 3,
and 4 has been submitted to ApJ.  Work has now shifted to using
the same moment analysis techniques on theoretical
spectra generated from the output of numerical simulations 

Mariska, J.T. 1994, , Flare Plasma Dynamics Observed with the Yohkoh Bragg Crystal Spectrometer. II. Properties of the Fe XXV, Ca XIX, and S XV resonance lines, ApJ, submitted.

Abstract: Using data from the Bragg Crystal Spectrometer (BCS) on the Yohkoh spacecraft, we have computed measures of the total intensity, centroid position, and line widths for the resonance lines of Fe XXV, Ca XIX, and S XV during the rise phase and after maximum for 190 solar flares. The difference between centroid positions early and late in each flare yields a measure of the line-of-sight velocity shift of the line centroids. Roughly 25% of the flares show evidence for spatial shifts of the emitting plasma during the flare. For all three resonance lines in those flares that show only Doppler shifts, we find a trend in the average value of the centroid shift with distance from Sun center suggesting radial mass motions with characteristic velocities of 76, 63, and 58 km/s in the resonance lines of Fe XXV, Ca XIX, and S XV, respectively. For each resonance line there is also a correlation between the rise-phase line widths and the centroid shift. For each resonance line we find no correlation between the centroid shift and the peak intensity, rise time, and total flare duration; and no correlation between the line width and the distance from Sun center, peak intensity, rise time, and total flare duration. Comparing data from the different BCS wavelength channels, we find that the centroid shift in each channel is correlated with the centroid shifts in the other channels. Similarly, the line width in each channel is correlated with the line width in the other channels.

Update 31-Jul-93

The first part of my survey has been accepted for publication in ApJ. I am now beginning to look at the results of my survey of 13 months of BCS data in the Fe XXV, Ca XIX, and S XV channels and expect to begin writing a second paper in the near future.

Information on the first paper:

Mariska, J.T., Doschek, G.A., and Bentley, R.D. 1993, Flare Plasma Dynamics Observed with the Yohkoh Bragg Crystal Spectrometer. I. Properties of the Ca XIX Resonance Line, ApJ, in press.

Abstract: Using data from the Bragg Crystal Spectrometer on the {\it Yohkoh\/} spacecraft we have computed measures of the total intensity, centroid position, and line width for the resonance line of Ca~XIX during the rise phase and after maximum for 219 solar flares. The difference between the centroid positions early and late in each flare yields a measure of the line-of-sight velocity shift of the line centroid. We find a trend in the average value of the centroid shift with distance from Sun center suggesting radial mass motions with a characteristic velocity of 58~km~s$^{-1}$. There is a correlation between the rise-phase line widths and the centroid shift. We find no correlation between the centroid shift and the peak intensity, rise time, and total flare duration; and no correlation between the line width and the distance from Sun center, the peak intensity, rise time, and total flare duration. These results do not conclusively support or refute the simple electron-beam-driven model or the thermal model for the rise phase of a solar flare.

Update 11-Mar-93

A paper entitled "Flare Plasma Dynamics Observed with the Yohkoh Bragg Crystal Spectrometer. I. Properties of the Ca XIX Resonance Line" is complete and ready for submission to ApJ. This paper summarizes the Ca XIX results for 219 flares observed from 1991 Nov 1 to 1992 Nov 30. From NOAA I have positions for 162 of the events. Many of the remainder have positions based on the data in the observing log for the associated SXT partial frames.

I am now extending the study to include Fe XXV and S XV. This is being done by using the Fe XXV light curve to select locations on the rise phase light curve and after maximum for computing the moments of the Fe XXV resonance line. The same times are then used to compute moments for the resonance lines of Ca XIX and S XV. This will allow comparisons of data from the various channels.

Update 01-Nov-92

I have now completed a survey of the behavior of the Ca XIX resonance line for 142 flares observed with the Yohkoh BCS between 1 October 1991 and 31 July 1992. For each flare I have computed the first three moments of the resonance line for a spectrum early in the rise phase and for a spectrum later in the flare. Comparison of the first moments (the centroid of the line) yields an estimate of the characteristic velocity of the Ca XIX emitting plasma early in the flare. The second moment provides an estimate of the nonthermal broadening of the resonance line early in the flare and late in the flare. For 122 of the flares I have positions from the SEL Event Listings. Using this data I have constructed plots of the characteristic velocity and the second moment as a function of the radius vector from Sun center as observed from the Earth. These plots show a clear tendency for the characteristic velocity to decrease near the limb. They appear to show a tendency for the second moment early in the flare to show an increase toward the limb. A plot of the characteristic velocity against the second moment shows a clear relationship between the two quantities. Plots of the moments against the peak count rate in the flare may also show some trends. I am currently beginning to write up the results and am continuing extending the data set using more recent flares. I am also investigating using the SXT partial frame locations associated with many of the flares to determine positions for those that have no positions in the SEL Event Listings.

Update 05-Aug-92

Update (5 August 1992): I have completed an initial survey of all final BCS data from 1 November 1991 to 31 May 1992 and found about 120 candidate flares for the study. Initial examination of the first part of the data set shows the following characteristics:

No significant shift 19 (4 at 80 degrees or greater)

Blue shift small 34 (1 at 80 degrees or greater) large 3

Blue wing 2 (+17 associated with other categories)

Red shift 11 (1 at 80 degrees or greater)

Total flares examined 69 (6 at 80 degrees or greater, 15 without positions)

These initial results were reported at the AAS Columbus meeting in June